key: cord- -toevn u authors: venkatesan, sudhir; carias, cristina; biggerstaff, matthew; campbell, angela p; nguyen-van-tam, jonathan s; kahn, emily; myles, puja r; meltzer, martin i title: antiviral treatment for outpatient use during an influenza pandemic: a decision tree model of outcomes averted and cost-effectiveness date: - - journal: j public health (oxf) doi: . /pubmed/fdy sha: doc_id: cord_uid: toevn u background: many countries have acquired antiviral stockpiles for pandemic influenza mitigation and a significant part of the stockpile may be focussed towards community-based treatment. methods: we developed a spreadsheet-based, decision tree model to assess outcomes averted and cost-effectiveness of antiviral treatment for outpatient use from the perspective of the healthcare payer in the uk. we defined five pandemic scenarios—one based on the a(h n ) pandemic and four hypothetical scenarios varying in measures of transmissibility and severity. results: community-based antiviral treatment was estimated to avert – % of hospitalizations in an overall population of . million. higher proportions of averted outcomes were seen in patients with high-risk conditions, when compared to non-high-risk patients. we found that antiviral treatment was cost-saving across pandemic scenarios for high-risk population groups, and cost-saving for the overall population in higher severity influenza pandemics. antiviral effectiveness had the greatest influence on both the number of hospitalizations averted and on cost-effectiveness. conclusions: this analysis shows that across pandemic scenarios, antiviral treatment can be cost-saving for population groups at high risk of influenza-related complications. influenza pandemics are rare, unpredictable events with potentially serious consequences. they are considered to be important public health emergencies by the world health organization, and a number of countries, with many having specific pandemic preparedness plans. [ ] [ ] [ ] neuraminidase inhibitors (nai) often feature prominently in pandemic influenza preparedness plans and several high-income countries have acquired nai stockpiles because pandemic specific vaccines may not be widely available for up to months. clinical trials show nai effectiveness in modestly reducing duration of symptomatic illness in patients with uncomplicated seasonal influenza. [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] however, these trials were under-powered to assess nai impact on secondary outcomes such as hospitalizations. [ ] [ ] [ ] two meta-analyses of the extant clinical trial data, examining outcomes based on the intention-to-treat-influenza infected (itti) approach, found that early nai treatment (≤ h of symptom onset) was associated with a risk reduction of and % for hospital admission in otherwise healthy patients with influenza. other meta-analyses of trial data that evaluated all outpatients with influenza-like-illness (ili) using the intention-to-treat (itt) approach did not find a reduction in hospitalizations in those treated with nais. , if a future pandemic is severe, hospital capacity may be exhausted and therefore reserved for the severely ill who are most likely to benefit. countries may decide to focus a significant part of their pandemic response plan towards community treatment aimed at averting hospitalizations. policy makers considering nai stockpiling for a future pandemic of unknown severity will have to consider both number of hospitalizations averted and the cost-effectiveness of such an intervention. nai treatment for pandemic influenza has generally been estimated to be cost-effective for higher-income countries. [ ] [ ] [ ] however, a review identified that previous health economic evaluations often neglected pandemic uncertainty by only evaluating singular, fixed pandemic scenarios. moreover, few models have incorporated the increased risks of adverse pandemic influenzarelated outcomes for patients with at-risk conditions. we present a spreadsheet-based decision tree model that evaluates the impact of community-based nai treatment in terms of the averted influenza-related hospitalizations and associated costeffectiveness in a range of pandemic scenarios. we built a decision tree model ( fig. ) to calculate the impact of community-based nai treatment for five pandemic scenarios. the first scenario is based on the uk's a(h n ) pdm experience, with a clinical attack rate (car) of % and a case hospitalization risk (chr) of . and . % among non-high-risk and high-risk patients, respectively (table ) . the other four scenarios were based on hypothetical pandemics that varied the car ( and %) and the chr ( . - . % for non-high-risk patients; - % for high-risk patients) ( table ). the hypothetical scenarios are based on a risk assessment framework developed by the cdc. , a standardized risk space was defined based on previous influenza pandemics, and hypothetical pandemic scenarios were identified from this risk space to allow easy comparisons to future economic evaluations. the chrs for the high-risk groups in these four hypothetical pandemics were assumed to be five times the chr for the non-high-risk group of patients based on estimates from the a(h n ) pandemic. we also assumed that the percentage of patients seeking outpatient/ambulatory care would increase with the chr of the pandemic, ranging from % among non-high-risk patients in a -type pandemic to~ % among high-risk patients when the chr is % (table ) . we estimated the number of deaths averted through averting hospitalizations by multiplying the number of hospitalizations averted with an inhospital mortality risk that was constant across the scenarios. we did not differentiate between oseltamivir and zanamivir in the definition of nais in our model; however, we based our cost and treatment effectiveness estimates on data specific for oseltamivir. we focus on community-based treatment and do not consider nai prophylaxis. we used nai effectiveness estimates from an individual participant data (ipd) meta-analysis of clinical trials data on otherwise healthy patients with seasonal influenza based on itti analysis (relative risk: . , % confidence interval: . - . ) since nais are not active against non-influenza respiratory infections. to account for nai prescriptions to patients with non-influenza ili, we assumed a 'wastage factor' of %, i.e. patients with non-influenza ili would be prescribed % of the number of regimens that are prescribed to patients with influenza. we assumed that all patients would start nai treatment ≤ h of symptom onset in our main model and then performed a sensitivity analysis varying the promptness of care-seeking within h of symptom onset from to % (percentage of all care-seeking patients who do so ≤ h of symptom onset). based on estimates from , we also assumed that % of patients would be compliant with the prescribed regimen. unit cost data for our model were obtained from secondary sources including the british national formulary and uk-based reports on the cost of health and social care (table ) . briefly, we used a weighted average cost of physician-based consultation of £ . . this cost was calculated as a weighted average cost of either a conventional primary care consultation or a phone-based consultation with the national pandemic flu service (npfs). the weighting of the costs was done using the proportion of assessments routed through each consultation service in . we used a cost of £ for an nai prescription, which included the cost of delivery. costs of hospitalizations ranged from £ for non-high-risk patients to £ for highrisk patients (table ). all costs were inflated to the british pound sterling (£) using the hospital and community health services (hchs) index. the overall population of . million was based on the uk population. we performed the analyses from the perspective of the healthcare payer, the uk national health service (nhs). given that we did not undertake a full costutility analysis, we chose to measure our outcomes in natural units (deaths and hospitalizations) rather than in standardized units (qalys). we considered a time horizon of less than one year (one pandemic event), therefore a discounting rate would not apply. in each pandemic scenario, we compared the number of outcomes averted (hospitalizations and deaths) and total costs associated with nai treatment compared to no nai treatment. we assessed cost-effectiveness of communitybased nai treatment by estimating the cost per averted hospitalization. our primary analysis was performed using the middle values of our input parameters using formulas provided in appendix . to account for uncertainty in parameter estimates, we performed sensitivity analyses by probabilistically varying input parameters along pre-defined probability distributions (table ) and using monte carlo simulations ( iterations using latin hypercube sampling) to calculate mean output values and % confidence intervals for different combinations of input parameters. the sensitivity analyses were performed using the software @risk version . (palisade corporation). further, we also performed two-way sensitivity analysis to assess the impact of varying nai effectiveness and patient compliance on the outcome (hospitalizations averted). in a -like pandemic scenario, we estimated that in our base-case model (no nai treatment) there would be hospitalizations in the overall population. we estimated that . million regimens of nais would be dispensed for outpatient treatment. nai treatment would have averted ( %) hospitalizations in a population of . million ( hospitalizations averted/million population) at a cost of £ per hospitalization averted ( table ). the cost to avert one hospitalization was £ in high-risk populations and £ in the non-high-risk population ( table ). in the % car-severity scenario (chr: non-high-risk = . %; high-risk = . %), we estimated that hospitalizations would occur. the . million regimens of nais would be dispensed, averting ( . %) hospitalizations at a cost per averted hospitalization of £ in the overall population and £ in the non-high-risk population. nai treatment was seen to be cost-saving in the high-risk population. in the % car-severity scenario (chr: non-high-risk = %; high-risk = %), we estimated that over . million hospitalizations would occur. the . million nai regimens would be dispensed, averting ( . %) hospitalizations in the total population at a cost per averted hospitalization of £ in the non-high-risk population. nai treatment was seen to be cost-saving in the overall population and in the high-risk population. in the % car-severity scenario, (chr: non-highrisk = . %; high-risk = . %), we estimated that over hospitalizations would occur. the . million nai regimens would be dispensed, averting ( . %) hospitalizations at a cost per averted hospitalization of £ in the overall population and £ in the non-high-risk population. nai treatment was seen to be cost-saving in the high-risk population. in the fourth pandemic scenario, (chr: non-high-risk = %; high-risk = %), we estimated that over . million hospitalizations would occur. the . million nai regimens would be dispensed, averting ( . %) hospitalizations in the overall population at a cost per averted hospitalization of £ in the non-high-risk population. nai treatment was seen to be cost-saving in the overall population and in the high-risk population. we found that varying the proportion of care-seeking patients who do so within h of symptom onset, while keeping all other variables constant, lowered the percentage of averted hospitalizations in the overall population from . % (assuming %) to . % (assuming %) in the -like pandemic scenario ( table , supplemental table s ). our sensitivity analyses revealed that using just the middle values of input parameters in a simple multiplicative model without probability distributions was likely to overestimate the number of hospitalizations averted and underestimate the cost per averted hospitalization. for the -like pandemic scenario, multiplying the middle values of input parameters (table ) overestimated the overall number of averted hospitalizations by % and underestimated the overall cost peraverted hospitalization by % when compared to the mean estimated from the monte carlo simulation (supplemental table s ). similar differences in estimates were observed in the other scenarios as well. the sensitivity analyses, based on a -like pandemic scenario, indicated that nai effectiveness had the greatest impact on both the total number of hospitalizations averted, as well as on the cost per hospitalization averted (see fig. for scenario). when the nai effectiveness was varied from to %, the resulting overall proportion of averted hospitalizations ranged between and %, at a cost per averted hospitalization of £ -£ . the percentage of care-seeking patients who were prescribed nai, the proportion of nai prescriptions to non-influenza patients, and nai treatment compliance were in the top three influential parameters for one or both outcomes (fig. ) . in our two-way sensitivity analysis we varied the treatment compliance level along with nai effectiveness beyond the % confidence intervals of our input parameter (from % effectiveness to % effectiveness). increased compliance levels were consistently associated with an increased number of averted hospitalizations across nai effectiveness estimates (fig. ) . the impact of prescribing nais to non-influenza ili patients had a considerable effect on the cost per averted hospitalization. for the -like pandemic scenario, this ranged from £ per averted hospitalization (wastage factor = %) to £ per averted hospitalization (wastage factor = %). main finding of this study we found that community-based nai treatment would avert a significant proportion of hospitalizations and deaths, particularly in high-risk patients, across the pandemic scenarios we explored in this analysis. however, a substantial number of hospitalizations and deaths would continue to occur even with community-based nai treatment. the proportion of hospitalizations averted by nais could be an important consideration while planning for conditions when hospital capacity could be exceeded. community-based nai treatment was seen to be cost-saving for the overall population in a pandemic with a high car and high severity, and costsaving for patients at high risk of complications from influenza across all the pandemic influenza scenarios tested. the value of nai treatment for population groups not at high risk and for milder pandemic scenarios will have to be determined by careful review under country-specific willingnessto-pay thresholds and the desire to reduce the number of hospitalizations and potential hospital capacity issues. what is already known on this topic nai treatment for pandemic influenza has generally been shown to be cost-effective, when compared to no nai treatment. [ ] [ ] [ ] previous studies have found that nai effectiveness is, by far, the most influential factor affecting the numbers of outcomes averted and the associated cost-effectiveness. , results from our sensitivity analysis support this finding. a study based in the united states that used a similar model showed slightly lower proportions of hospitalizations averted due to nai treatment when compared to ours, but the difference could be because of the lower level of treatment effectiveness assumed in the us study. the us study further found that while nai treatment averted many hospitalizations, large numbers of hospitalizations would remain, which is similar to what we have found. we found that variations in nai prescription rate, treatment compliance and healthcare-seeking behaviour (to include the choice to seek care and the promptness in care-seeking) impacted considerably on the outcomes, suggesting that even with a drug of fixed effectiveness, factors relating to healthcare-seeking and healthcare delivery could significantly influence the total number of hospitalizations and deaths averted. these data indicate that a successful pandemic stockpiling strategy must be linked to operational procedures which optimize timely access to antivirals, widespread treatment implementation, and high levels of compliance in targeted groups. one recognized limitation of some previous economic analyses of nai treatment has been that entire populations have been modelled homogenously without accounting for the increase in the likelihood of influenza-related care-seeking and complications in patients with underlying at-risk conditions. , in our model, we vary the propensity to seek care and chr by patients' at-risk status. the significance of this is that countries with limited resources could consider obtaining smaller antiviral stockpiles to target at-risk population groups and avert a higher number of hospitalizations and deaths for each antiviral course dispensed than if they adopted a treat-all approach. the car was an important factor in determining the number of nai regimens that would be needed for communitybased treatment. our model showed that a highly transmissible, but low severity pandemic would require a larger nai stockpile than a pandemic with lower transmissibility and higher severity. however, across all pandemic scenarios, the number of nai regimens dispensed for outpatient treatment was well below the uk's published national nai stockpile size of almost million courses of the drug. we have adopted a simple and transparent approach to model building in which we account for important epidemiological factors, population healthcare-seeking behaviour and service utilization rates in a range of pandemic scenarios. our analyses are uk-focussed, but the spreadsheet tool is easily adaptable to represent other healthcare systems. while the epidemiological parameters are unlikely to change drastically by country, input parameters relating to healthcare utilization and costs will need to be replaced with country-specific ones. we provide the simple version of the spreadsheet tool (without the sensitivity analysis) in appendix . we used updated nai effectiveness estimates from seasonal influenza data, although observational data from the a(h n ) pandemic in a high-severity (high risk of hospitalization) population suggest similar estimates of nai effectiveness (≤ h from symptom onset). we assumed nai effectiveness is the same in patients with and without atrisk conditions. while there is some evidence to suggest that the level of effectiveness against hospitalization is similar for both groups, there is also evidence that suggests a reduction in nai effectiveness in patients with at-risk conditions. this study is subject to limitations. we used a decision tree model (not a transmission dynamic model) and assumed no effect of nai treatment on transmission. there is evidence to suggest that nai treatment, at a population level, is likely to have minimal impact on influenza transmission. however, decision tree models are known to be limited, especially in their ability to describe the change in influenza attack rates in different risk groups over the course of a pandemic. a comparison of static and dynamic models of nai treatment for pandemic influenza concluded nai treatment was seen to be costeffective with both modelling paradigms; although the associated cost-effectiveness ratios were seen to differ. due to a lack of evidence specific to hospitalization, we did not consider benefits of nai treatment > h of symptom onset. nai treatment has, however, been shown be beneficial even when started beyond h from symptom onset. the use of nais may be associated with additional costs to the healthcare system due to possible adverse effects of nais but we have not considered these costs in our model since most side effects are known to be minor. finally, we have assumed that the multiplier for high-risk patients remains constant between severity scenarios resulting in a chr as high as %. chrs of %, even for high-risk patients, may be unlikely. our analyses show that nai treatment in outpatients can be cost-saving, particularly for population groups at high risk of influenza-related complications. model-based estimates like these of the potential hospitalizations, deaths and costs associated with different pandemic scenarios can help countries consider different treatment options and inform stockpiling decisions while developing pandemic preparedness plans. nai stockpiling decisions are also influenced by other costs to the healthcare system related to storage and maintenance of the nai stockpile. currently, the shelf-life for the mg hard capsules of oseltamivir phosphate that comprise most of the nai stockpile is estimated to be years if stored as per instructions. however, influenza pandemics cannot be predicted, and nai stockpiles could remain unused at the end of their shelf-life, or they may be rendered ineffective or less relevant by the development of antiviral drug resistance or newer, more effective influenza antiviral therapies. additionally, evidence suggests that in-hospital nai treatment may also be associated with protective effects , and nai treatment has been shown to be cost-effective if the benefits of nai usage are confined only to those treated in hospital. if a pandemic treatment policy was pursued which combined community use of nais to prevent hospital admission and nai treatment of hospitalized patients to reduce mortality, then cost-effectiveness and stockpile strategies across both scenarios would need to be considered. future research in optimizing nai distribution to risk groups during a pandemic will further inform the costeffectiveness of stockpiling. supplementary data are available at the journal of public health online. dh pandemic influenza preparedness team. uk influenza pandemic preparedness strategy pandemic influenza preparedness and response: a who guidance document department of health and human services selecting viruses for the seasonal influenza vaccine efficacy and safety of the neuraminidase inhibitor zanamivir in the treatment of influenza a and b virus infections randomized, placebocontrolled studies of inhaled zanamivir in the treatment of influenza a and b: pooled efficacy analysis efficacy and safety of oseltamivir in treatment of acute influenza: a randomised controlled trial efficacy and safety of the oral neuraminidase inhibitor oseltamivir in treating acute influenza: a randomized controlled trial zanamivir for treatment of symptomatic influenza a and b infection in children five to twelve years of age: a randomized controlled trial efficacy of oseltamivir treatment started within days of symptom onset to reduce influenza illness duration and virus shedding in an urban setting in bangladesh: a randomised placebo-controlled trial intravenous peramivir for treatment of influenza a and b virus infection in high-risk patients single dose peramivir for the treatment of acute seasonal influenza: integrated analysis of efficacy and safety from two placebo-controlled trials evidence and policy for influenza control antivirals for influenza: where now for clinical practice and pandemic preparedness? neuraminidase inhibitors for preventing and treating influenza in healthy adults and children impact of oseltamivir treatment on influenza-related lower respiratory tract complications and hospitalizations oseltamivir treatment for influenza in adults: a meta-analysis of randomised controlled trials effectiveness of oseltamivir in adults: a meta-analysis of published and unpublished clinical trials oseltamivir for influenza in adults and children: systematic review of clinical study reports and summary of regulatory comments pandemic influenza: guidance for primary care trusts and primary care professionals on the provision of healthcare in a community setting in england cost-effectiveness of antiviral stockpiling and near-patient testing for potential influenza pandemic strategies for antiviral stockpiling for future influenza pandemics: a global epidemic-economic perspective cost-benefit of stockpiling drugs for influenza pandemic cost-effectiveness analysis of pandemic influenza preparedness: what's missing? novel framework for assessing epidemiologic effects of influenza epidemics and pandemics standardizing scenarios to assess the need to respond to an influenza pandemic pandemic (h n ) in england; an overview of initial epidemiological findings and implications for the second wave use of neuraminidase inhibitors in influenza estimating the united states demand for influenza antivirals and the effect on severe influenza disease during a potential pandemic access to the nhs by telephone and internet during an influenza pandemic: an observational study department of health. the national pandemic flu service: an evaluation uk population total health outcomes in economic evaluation: the qaly and utilities dynamic versus static models in cost-effectiveness analyses of anti-viral drug therapy to mitigate an influenza pandemic public accounts committee. stockpiling tamiflu and the management of the stockpile in impact of outpatient neuraminidase inhibitor treatment in patients infected with influenza a(h n )pdm at high risk of hospitalization: an individual participant data metaanalysis effectiveness of neuraminidase inhibitors in preventing hospitalization during the h n influenza pandemic in british columbia the early transmission dynamics of h n pdm influenza in the united kingdom electronic medicines compendium (emc). tamiflu , and mg hard capsules effectiveness of neuraminidase inhibitors in reducing mortality in patients admitted to hospital with influenza a h n pdm virus infection: a metaanalysis of individual participant data impact of neuraminidase inhibitors on influenza a(h n )pdm -related pneumonia: an individual participant data meta-analysis evidence synthesis and decision modelling to support complex decisions: stockpiling neuraminidase inhibitors for pandemic influenza usage scientific summary of pandemic influenza & its mitigation: scientific evidence base review using an online survey of healthcare-seeking behaviour to estimate the magnitude and severity of the h n v influenza epidemic in england influenza-like illness, the time to seek healthcare, and influenza antiviral receipt during the - influenza season-united states seasonal influenza vaccine uptake amongst gp patient groups in england: provisional monthly data for influenza pandemic: an independent review of the uk response to the influenza pandemic predictors of clinical outcome in a national hospitalised cohort across both waves of the influenza a/h n pandemic - in the uk unit costs of health and social care department of health. nhs reference costs vaccination against pandemic influenza a/h n v in england: a real-time economic evaluation we would like to thank anita patel from cdc, atlanta, for reviewing this manuscript and offering helpful comments. key: cord- -d saaiu authors: eijsink, job f. h.; al khayat, mohamed n. m. t.; boersma, cornelis; ter horst, peter g. j.; wilschut, jan c.; postma, maarten j. title: cost-effectiveness of hepatitis c virus screening, and subsequent monitoring or treatment among pregnant women in the netherlands date: - - journal: eur j health econ doi: . /s - - - sha: doc_id: cord_uid: d saaiu background: the prevalence of diagnosed chronic hepatitis c virus (hcv) infection among pregnant women in the netherlands is . %, yet many cases remain undiagnosed. hcv screening and treatment of pregnant hcv carriers could reduce the burden of disease and limit vertical transmission from mother to child. we assessed the impact of hcv screening and subsequent treatment with new direct-acting antivirals (daas) among pregnant women in the netherlands. methods: an hcv natural history markov transition state model was developed, to evaluate the public-health and economic impact of hcv screening and treatment. besides all , pregnant women in the netherlands (cohort ), we modelled further cohorts: all , first-time pregnant women (cohort ), , pregnant migrant women (cohort ) and , first-time pregnant migrant women (cohort ). each cohort was analyzed in various scenarios: i no intervention, i.e., the current practice, ii screen-and-treat, i.e., the most extensive approach involving treatment of all individuals found hcv-positive, and iii screen-and-treat/monitor, i.e., a strategy involving treatment of symptomatic (f –f ) patients and follow-up of asymptomatic (f ) hcv carriers with subsequent treatment only at progression. results: for all cohorts, comparison between scenarios (ii) and (i) resulted in icers between € , and € , per qaly gained and year budget impacts varying between € , , and € , , . for all cohorts, comparison between scenarios (iii) and (i) resulted in icers between € , and € , per qaly gained and budget impacts varying between € , , and € , , . for all cohorts, the icers (scenario iii versus ii) involved in delayed treatment of asymptomatic (f ) hcv carriers varied between € , and € , , well above the willingness-to-pay (wtp) threshold of € , per qaly gained and even above a threshold of € , per qaly gained. conclusion: universal screening for hcv among all pregnant women in the netherlands is cost-effective. however, it would be reasonable to consider smaller risk groups in view of the budget impact of the intervention. electronic supplementary material: the online version of this article ( . /s - - - ) contains supplementary material, which is available to authorized users. hepatitis c is a serious disease caused by infection with hepatitis c virus (hcv). worldwide - million people are chronically infected with hcv [ , ] . exposure to the virus results in % of cases in a chronic infection [ ] . approximately % of chronically infected patients develop serious hcv-related liver disease after onset of the infection [ ] . currently, hepatitis c affects % of pregnant women globally [ ] . hcv may be transmitted vertically, mostly perinatally, from mother-to-child [ ] [ ] [ ] . with the development of new drug therapies which are highly effective and well tolerated, there is a potential for these drugs to be used by pregnant patients with hepatitis c [ ] . hcv screening of pregnant women potentially contributes to the goal of the world health organization (who) to achieve % diagnosis of hcv and % treatment by worldwide through scaling-up screening strategies and prevention of hcv transmission [ ] . two major developments have contributed to the demand for hcv screening of specific risk groups. the first and most important development is the improved hcv treatment with direct-acting antivirals (daas) [ ] . more than % of chronically infected hcv patients are cured through daa treatment compared to only % with previous treatments [ ] [ ] [ ] . the second development is the increase of hepatocellular cancer (hcc) incidence, hcv infection being the leading cause of hcc in western countries [ ] . screening and daa treatment of risk groups could prevent reinfection, new infections, hcc and vertical transmission from mother-to-child. the health council in the netherlands has recommended to investigate the cost-effectiveness of screening of pregnant women for hcv with subsequent daa treatment [ ] [ ] [ ] . prevalence of diagnosed chronic hcv infection among women in the dutch population is . % ( % confidence interval (ci): . - . %), which is similar to the prevalence in the general population in europe [ ] . firstgeneration non-western migrants are more likely to be hcvpositive ( . - . %) than western women ( . - . %) [ ] . notably, these immigrants represent . % of the total dutch female population [ ] . in industrialized countries, hcv is the most common cause of chronic liver disease among children and perinatal transmission is the leading cause of infection [ ] . the current best estimate of vertical transmission risk is between . and . % [ ] . treatment with daas during pregnancy is not yet recommended, and lactation during treatment is contra-indicated, because of a lack of information on potential toxicity [ ] . however, it is conceivable that in the near future daa treatment of hcv-infected women during pregnancy becomes available, not only to limit disease progression in the patient, but also to prevent vertical transmission of the virus to the child. the aim of this study is to estimate the public-health and economic impact of hcv screening and treatment among pregnant women from a public-health perspective [ ] . in particular, we estimated the health gains, cost-effectiveness and the budget impact of implementing such a programme. the results of our study can be used to reach a rational decision as to whether hcv screening and potential treatment of pregnant women should be implemented in the netherlands and elsewhere [ ] . a screening model linked to hcv-disease states within a markov model was used to evaluate the cost-effectiveness (ce) of hcv screening of pregnant women, with initial treatment during pregnancy, compared to current practice (no screening and no intervention) from a health-care payer perspective in the netherlands. our ce analysis includes health benefits for pregnant women and their children, and the corresponding budget impact. the costs and effects of hcv screening and various modalities of subsequent treatment versus current practice were calculated for four cohorts of pregnant women and were expressed in terms of incremental cost-effectiveness ratio (icer), as further elaborated below. we used a deterministic, hcv natural history, closed-cohort markov model, as presented in fig. . the model includes annual cycles and a life-time horizon of years, representing the approximate period from the age at which a woman can become pregnant until her death. hcv carriers were classified in metavir scores f -f . f is a (fully) healthy, but hcv-infected, state. f -f represent mild to severe stages of liver fibrosis. f represents liver cirrhosis. in the model, patients with metavir score f may develop hepatocellular cancer (hcc), decompensated cirrhosis (dcc) and, subsequently, patients with dcc can progress to liver transplantation (lt). lt-patients move to the follow-up state (post-lt). post-lt patients are described as patients during the first months after their liver transplantation. after year, they move to the follow-up state post-lt + until their death. without screening, hcv-infected patients generally develop symptoms in a late stage of infection [ , ] . implementation of screening will result in detection of increased numbers of asymptomatic patients [ ] and, later on, fewer patients with fibrosis or cirrhosis relative to the current situation without screening. in this study, we assumed that testing a cohort comprising of all pregnant women is a 'one-time' screening for each women (independent of the number of pregnancies), rather than having repeat testing in their potential subsequent pregnancies. in the model, we used a conservative sustained virologic response (svr) of % for patients with metavir scores of f , f , f and f , and % for f patients [ ] . we only included treatment regimens for weeks, independent on the metavir scores and in accordance with the dutch hcv-guidelines [ , ] . it was assumed that if patients were not cured, they proceed to the next lower health state. vertical transmissions are included in the model as potentially prevented hcv infections, after screening of the mothers and subsequent daa treatment. the probabilities to move from one health state to another are given in table s of the appendix. the first hcv screening step represents a serologic antibody test to determine the presence of a current or past hcv infection. the second test is a reverse-transcription polymerase-chain reaction (rt-pcr) viral rna test to confirm the serologic test, and to determine whether the hcv infection had been cleared spontaneously. the rt-pcr test has a sensitivity between . % and . % and a specificity between . and . % [ ] . the third test concerned a fibroscan examination, which is a quantitative analysis technique to support the diagnostics of liver fibrosis in patients and to determine the metavir score (f -f ). individuals are screened first for anti-hcv antibodies and, if found positive, are subsequently screened for hcv rna. outpatient visit consultation costs were included for each test. for individuals who are rna-positive, we incorporated fibroscan costs for disease staging. the annual costs for daa treatment were assumed at their list price levels in the netherlands [ ] . weighted average treatment costs for daa were estimated at € , based on actual use of daa medication (sofosbuvir, ledipasvir/ sofosbuvir, grazoprevir/elbasvir, velpatasvir/sofosbuvir, daclatasvir, ombitasvir/paritaprevir/ritonavir) for a -week treatment period in [ ] . we did not consider other treatments for hcv infection, such as protease inhibitors, ribavirin or peg-interferons. for the budget-impact analysis, we included total medical costs in the first years, costs of hcv treatment, screening costs and follow-up costs with possible hcv-related diseases. the pregnant women included for hcv screening in this study are between and years of age, with an average hcc hepatocellular cancer, dcc decompensated cirhossis, lt liver transplantation. lrd: liver-related death. *in case of treatment failure, patients will be in the same metavir state after the treatment age of [ , ] . we excluded women with recurrent hcv infection, women with hiv infection and injecting drug users [ ] . in this study, we considered four different cohorts of pregnant women. the characteristics of the four cohorts were obtained from statistics netherlands (cbs); we took the average size of the years to . details of the cohorts, including size, hcv prevalence and vertical transmission estimates [ ] , are as follows: quality of life depends on the state of health and the age of the pregnant woman. in the model, all hcv health states were assigned a particular utility, ranging from to . utility reflects death and utility reflects full health without any complaints. the utility of hcv-positive, but asymptomatic, patients (f ) was reduced with . [ , ] , because of reasons of anxiety and worries and among these individuals. utilities after successful treatment were assumed to increase by . [ ] . the utilities are presented in s of the appendix. quality-adjusted life years (qalys) were calculated as the product of remaining life years of the patient in a particular health state after the intervention (screening and monitoring or treatment) and the quality of life after the intervention [ ] . we investigated three scenarios with different comparisons between the scenarios. scenario i, the no-intervention scenario, reflects the current practice of absence of screening. scenario ii, the screen-and-treat scenario, reflects the most extensive approach with daa treatment of all individuals found hcv-positive after screening. finally, scenario iii, the screen-and-treat/monitor scenario, reflects the approach in which, after screening, the f patients are not treated but actively monitored (and, if indicated, treated later on). we specifically considered this third scenario to avoid delayed overtreatment. indeed, % of asymptomatic hcv-infected individuals spontaneously clear the virus and, in addition, approximately % of chronically infected patients will never develop hcv-related liver disease [ ] . obviously, one does not know a priori which patients will develop chronic infection and symptoms of disease. therefore, we chose to periodically monitor these patients. we assumed that just monitoring asymptomatic hcv carriers instead of treatment would contribute to lower treatment costs and result in higher patient value. three comparisons between the different scenarios were performed: • scenario ii versus scenario i, reflecting screening and treatment of all hcv-positive patients versus the current practice of no intervention. • scenario iii versus scenario i, reflecting treatment of symptomatic (f -f ) patients and monitoring of asymptomatic (f ) hcv carriers versus the current practice. • scenario iii versus scenario ii, focusing specifically on the additional costs and health gains due to immediate treatment of all f hcv carriers versus just monitoring these asymptomatic individuals until some of them progress to disease. as avoidance of mother-to-child transmission of hcv is one of the most important reasons for hcv screening and treatment of pregnant women, vertical transmissions are explicitly taken into account in the model. specifically, we included the effects of vertical transmission on the healthcare costs, treatment costs and qalys for the (unborn) children. we express the cost-effectiveness of the different scenarios described above in terms of incremental cost-effectiveness ratio (icer), using the following formula: in which c represents the costs and e the quality-adjusted life years (qalys); subscript represents the case where the intervention has been applied and subscript represents the case where the intervention has not been applied. therefore, the icer represents the costs per quality-adjusted life year (qaly) gained. in the netherlands, icers are considered against an informal willingness-to-pay (wtp) threshold of € , per qaly gained [ ] . notably, we also considered a wtp-threshold of € , per qaly gained, reflecting the burden of disease [ ] . the budget-impact analysis gives a perspective on total future hcv-related costs. for the budget-impact analysis, we included direct medical costs, costs of hcv treatment and costs of screening, in the first years, years and years of implementation of screening according to the budget impact guidelines [ ] . the total costs were discounted with an annual rate of %, the qalys were discounted with . %, according to dutch guidelines [ ] . price levels in the year were applied. a one-way sensitivity analysis was performed to estimate the effect of variation in specific parameters on the icer and to determine which parameter has the most pronounced effect on the icer. the parameters were varied between minus % and plus % of the base-case parameter value. the prevalence was varied in the range of the % ci of the hcv prevalence of . % ( . - . %). a probabilistic sensitivity analysis (psa) was performed to assess the uncertainty around the different input parameters and the effect on the cer. here, input parameters are considered as random quantities based on the underlying parameter distributions. for every simulation ( in total), the parameters were sampled from the parameter space of % ci. if the % ci was unknown for a specific parameter, we varied the parameter between minus % and plus %, following a triangle distribution. all variables and ranges are represented in table s of the appendix. we first determined the health benefits involved in implementation of hcv screening and daa treatment among pregnant women in the netherlands. in all four cohorts, we found significant reductions in liver disease after - decades, specifically a reduction of % in dcc, of % in hcc, of % in liver transplantation (lt) and of % in liver-related death (lrd). we also found significant reductions in vertical hcv transmissions. since each cohort consists of a different number of pregnant women with a specific hcv prevalence, the absolute number of avoided vertical transmission varied between the different cohorts. specifically, in the cohort of all pregnant women, we found avoided cases of vertical transmission, in the cohort of first-time pregnant women avoided cases, in the cohort of pregnant migrants avoided cases, and in the cohort of first-time pregnant migrants avoided cases. we subsequently determined the cost-effectiveness and budget impact of hcv screening and treatment among the four cohorts of pregnant women following the different scenarios and comparisons. table presents an overview of the results. for each of the cohorts, the table shows the values of the icer for comparisons between the two respective intervention scenarios and the scenario; no intervention (current practice), table also presents the icers for scenario screen-and-treat versus screen-and-treat/monitor and in table the total years, years and years budget impact (bi) of the different interventions. below, we further elaborate on the results obtained for each of the cohorts. the (table ) . limiting the intervention to the cohort of first-time pregnant migrants further improved cost-effectiveness results, with the most favorable outcomes for the screen-and-treat/ monitor scenario. specifically, comparison in this group between the screen-and-treat and no intervention scenarios yielded qalys gained at incremental costs of € , , , resulting in an icer of € , per qaly gained. the total bi over years of this screening scenario we conducted an additional comparison ( as indicated above, hcv screening and treatment of pregnant women prevents significant numbers of vertical transmission cases. yet, the effects of vertical transmission on the icers of the screen-and-treat and the screen-and-treat/ monitor scenarios remain limited. this is primarily due to the relatively low rate of vertical transmission of . - . % [ ] . with inclusion of vertical transmission in the markov model, the icers for the four different cohorts range between € and € , , and without inclusion of vertical transmission in the model, the icers range between € and € . we performed both a one-way sensitivity analysis and a probabilistic sensitivity analysis (psa), to assess the effect of parameter uncertainty on the cost-effectiveness outcomes. the effect of the cohort size on the icer outcomes was found to be minimal. here, we present the result on the univariate sensitivity analysis for the screen-and-treat/monitor versus no intervention scenario in the cohort of first-time pregnant migrants. this is the scenario with the most favorable cost-effectiveness. the one-way sensitivity-analysis for this scenario in this fig. one-way sensitivity analysis for the comparison between the screen-and-treat/monitor and no intervention scenarios among first-time pregnant migrants. the diagram shows the change in the icer when each parameter is increased or reduced with % cohort shows that the cost-effectiveness outcome is most sensitive to variation in the prevalence of hcv (fig. ). for the screen-and-treat versus no intervention scenario in the same cohort, the cost-effectiveness outcome was most sensitive to variation in medication price (fig. ). for the screen-and-treat versus screen-and-treat/monitor scenario, monitoring disutility is most sensitive to variation (fig. ) . the results of the ceac are presented in figs. and . these results indicate that, among the four cohorts investigated, the icers for both the screen-and-treat versus no intervention and screen-and-treat/monitor versus no intervention scenarios remain well below the informal dutch wtp-threshold of € , per qaly gained. overall, the results of the psa showed limited variation around the mean cost-effectiveness estimate upon varying the model inputs independently, underlining the robustness of the model. finally, fig. shows the respective cost-effectiveness acceptability curve based on varying the wtp-threshold. these results indicate that among the four cohorts investigated, the icers for screen-and-treat versus screenand-treat/monitor is not below the informal dutch wtpthreshold of € , per qaly gained. our study demonstrates that, after screening of pregnant women, identification of hcv patients at early metavir stages and implementation of daa treatment would prevent one out of three liver-related diseases caused by hcv on the long term. in addition, depending on the specific screening/ treatment strategy, the size and the hcv prevalence of the cohorts, hcv screening and treatment results in prevention of - vertical transmissions in the netherlands. our present study demonstrates that hcv screening of pregnant women and subsequent immediate treatment of all hcv-positive individuals with daas is a cost-effective intervention in the netherlands. this applies not only to the cohorts of non-western migrant women in the netherlands with a relatively high hcv prevalence, but also to the cohorts of all pregnant dutch women in which on average the hcv prevalence is lower. indeed, in all four different cohorts studied, the icers of the screen-and-treat versus no intervention scenario were similar, varying between € and € , per qaly gained, and thus remained well below the wtp-threshold of € , per qaly gained. still considerably lower icers were obtained for the screen-and-treat/monitor scenario in which only the symptomatic f - patients are treated and the asymptomatic f hcv carriers are just monitored until some of them progress fig. one-way sensitivity analysis for the comparison between the screen-and-treat and no intervention scenarios among first-time pregnant migrants. the diagram shows the change in the icer when each parameter is increased or reduced with % fig. one-way sensitivity analysis for the comparison between the screen-and-treat and screen-and-treat/monitor scenarios among first-time pregnant migrants. the diagram shows the change in the icer when each parameter is increased or reduced with % cost-effectiveness acceptability curve (ceac) for the comparison between the screen-and-treat and screen-and-treat/monitor scenarios among the four cohorts of pregnant women to disease. indeed, for this scenario the icers among the different cohorts varied between only € and € per qaly gained, the most cost-effective result being obtained for the cohort of first-time pregnant migrant women. while, as indicated above, the icer of hcv screening and treatment (or monitoring of f and treatment of f - patients), remained below the dutch wtp-threshold of € , , the budget impact of these interventions was substantially different between the four cohorts. clearly, the budget impact is directly proportional to the size of the cohort, and thus was much higher for the cohorts of all pregnant dutch women, as opposed to the migrant women. for the screen-and-treat scenario, the budget impact varied between € , , and € , , in the migrant cohort and all pregnant women, respectively. also, the extent of treatment strongly affects the budget impact. for example, in the cohort of all pregnant women, the budget impact of the screen-and-treat/monitor scenario was, with € , , , much lower than the € , , of the screen-and-treat scenario. likewise, in the cohort of migrant women, the budget impact varied substantially between these two scenarios, ranging from € , , and € , , . the above results illustrate that implementation of a strategy of active monitoring of f patients, rather than immediate treatment of these asymptomatic individuals, represents an effective way of reducing the costs of hcv screening and treatment. the reason is that approximately % of hcvinfected individuals spontaneously clear the virus, while furthermore % of those who do become chronic hcv carriers, will never develop hcv-related liver disease [ ] . clearly, postponing treatment of f patients saves potentially unnecessary costs. accordingly, restriction of treatment to f - patients represents the most cost-effective scenario and thus contributes to optimization of value for hcv patients. this is also illustrated by the comparison of our scenarios ii and iii, resulting in an icer above € , per qaly gained in all cohorts studied, which directly demonstrates that treatment of f patients is not cost-effective. a % daa discount, in the comparison of screen-and-treat versus screen-and-treat/monitor all hcv-infected pregnant women, would be cost-effective at a threshold of € , per qaly gained, in different cohorts of pregnant women. % discount, is comparable with the discount rate from biologicals versus biosmilars in the netherlands, therefore in the future screen-and-treat could also be a cost-effective scenario compared to screen and monitoring [ ] . while just monitoring of asymptomatic chronic hcv carriers does reduce costs, it does not prevent spread of the virus through vertical transmission from mother-to-child. monitoring of f carriers does not prevent hcv infection in subsequent pregnancies either; our model did not take further transmission of hcv and spreading of infection into account in untreated women. in this respect, our model can be considered to reflect a conservative estimate of cost-effectiveness. inclusion of transmission effects beyond the child would further enhance the cost-effectiveness profile. however, these effects do not outweigh the benefits of restricting treatment to f - patients. we therefore conclude that monitoring of f hcv-positive patients instead of immediate treatment prevents significant costs and thus results in the most favorable cost-effectiveness with a substantially lower budget impact [ ] . in this study, we focused on screening of pregnant women and subsequent treatment of hcv-positive individuals with daas. however, currently, hcv treatment with daas is contraindicated for pregnant women, because of a lack of studies regarding direct teratogenic effects and pharmacological effects later in life of the offspring. consequently, under the present circumstances, hcv-positive mothers can only be treated after childbirth and thus only children from subsequent pregnancies would be protected. according to bernstein et al., universal hcv screening and treatment with daas during pregnancy is on the horizon [ ] . clearly, these interventions should be urgently evaluated for safety and implemented if appropriate [ ] . several studies regarding daa treatment of hcv infection during pregnancy are ongoing. for example, the results of a phase i study in magee women's hospital in pittsburgh are expected to be presented in [ ] . in the future, we anticipate a development for hcv screening and treatment similar to that in the case of hiv/aids, where hiv-positive women are treated with combination antiretroviral therapy (cart) to prevent mother-to-child transmission of the virus [ , ] . perinatal transmission is the primary hcv transmission route among children responsible for - % of cases. many children often remain untested and potentially hcv undiagnosed. therefore, next to the direct benefit of treatment for the women in curing their infection and preventing serious liver-related diseases, benefits for the child exist in avoiding hcv with possible extrahepatic effects of hcv infection in childhood and significant reductions in both physical and psychosocial health as well as in cognitive functions. the outcome of our study that hcv screening and treatment of pregnant women in the netherlands is a cost-effective intervention against the informal dutch wtp-threshold of € , per qaly gained, is in apparent disagreement with the findings of urbanus et al. [ ] in . these authors estimated that only if costs per treatment were to decline to € (a reduction in price of € , ), screening of all pregnant women would be cost-effective. however, the results of urbanus et al. [ ] were obtained before the introduction of the highly effective daas in . now, it appears that screening and daa treatment, of hcv-positive individuals would be a cost-effective intervention. nonetheless, as discussed above, screening of the entire population of pregnant women is not necessarily preferred, because of the large budget impact of the intervention and the low hcv prevalence in the total dutch population. kracht et al. have proposed "micro-elimination" of hcv by screening and treatment of various pre-defined hcv risk groups [ ] . these authors concluded, in agreement with our results, that hcv screening of risk groups is the most pragmatic and efficient approach. our study could be helpful with decisions on the implementation of hcv screening programmes in europe. the estimated fraction of hcv cases that remain undiagnosed in the general or proxy populations in europe ranges between % in denmark to . % in greece [ ] . razavi et al. estimated the overall proportion of undiagnosed hcv cases in the eu at % [ ] . daa treatment of hcv in pregnancy is not (yet) in clinical guidelines, our model is hypothetical currently in that respect. the main difference between, for example, the assld/idsa-guidelines and our model is that we assumed that pregnant women are treated with daas after hcv diagnosis during pregnancy. a simplified treatment algorithm [ ] , for treatment-naive patients without cirrhosis, possibly would reduce the costs in the model, which could also improve the performance of treatment, corresponding with favorable to cost-effectiveness [ ] . this study reflects a single cohort model in the netherlands, with effects on children for that specific cohort. our current analysis does not include future pregnancies in the very same cohorts. in the future, the total amounts of screened and treated pregnant women will be lower and preferably result from the standard prenatal screening for infectious diseases, which means higher numbers to screen to identify patients, but also less patients to be treated with relatively expensive treatments. our study demonstrates that screening and monitoring or treatment of smaller subgroups of pregnant women is highly cost-effective approach and has a comparatively low budget impact in the netherlands. on the other hand, in other countries with a higher hcv prevalence, screening of all pregnant women could be a more cost-effective option [ , ] . our study indicates that universal hcv screening of pregnant women in the netherlands is cost-effective, independent of the specific cohort involved. however, the budget impact is substantially different between subgroups, and is largely determined by the cohort size and by the extent of treatment of hcv-positive individuals. screening and subsequent monitoring of f patients and treatment of f -f patients with the daas appeared to be the most cost-effective approach. hcv screening and treatment of pregnant women results in a substantial reduction of hcv-related liver diseases and deaths. it also prevents vertical transmission of the virus from mother to child. from a public-health and health-economic perspective, it would be reasonable to consider smaller risk groups of first-time pregnant or/and non-western pregnant women for an active hcv screening programme in the netherlands, and possibly elsewhere. open access this article is licensed under a creative commons attribution . international license, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the creative commons licence, and indicate if changes were made. the images or other third party material in this article are included in the article's creative commons licence, unless indicated otherwise in a credit line to the material. if material is not included in the article's creative commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. to view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/ . /. the estimated future disease burden of hepatitis c virus in the netherlands with different treatment paradigms natural history of chronic hepatitis c the effects of female sex, viral genotype, and il b genotype on spontaneous clearance of acute hepatitis c virus infection hepatitis c virus treatment: is it possible to cure all hepatitis c virus patients? effects of mode of delivery and infant feeding on the risk of mother-to-child transmission of hepatitis c virus: european paediatric hepatitis c virus network the management of hcv infected pregnant women and their children european paediatric hcv network clinical course and management of acute and chronic viral hepatitis during pregnancy european convention on human rights hcv vertical transmission in pregnancy: new horizons in the era of daas hepatitis elimination by : progress and challenges systematic review: current concepts and challenges for the direct-acting antiviral era in hepatitis c cirrhosis peginterferon alfa- b plus ribavirin compared with interferonalfa- b plus ribavirin for initial treatment of chronic hepatitis c: a randomised trial peginterferon alfa- a plus ribavirin for chronic hepatitis c virus infection sterfte aan chronische hepatitis b en c in nederland hepatitis c virus and hepatocellular carcinoma: a narrative review screening van risicogroepen op hepatitis b en c (screening of risk groups for hepatitis b and c )| advies | gezondheidsraad hepatitis c in the general population of various ethnic origins living in the netherlands: should non-western migrants be screened? is adding hcv screening to the antenatal national screening program in amsterdam, the netherlands, cost-effective? a review of hepatitis c virus (hcv) vertical transmission: risks of transmission to infants born to mothers with and without hcv viraemia or human immunodeficiency virus infection the prevalence of hepatitis b and c in an antenatal population of various ethnic origins perinatal transmission of hepatitis c virus infection cost-effectiveness of universal hepatitis c virus screening of pregnant women in the united states hepatitis c in pregnancy: screening, treatment, and management hepatitis c virus prevalence in the netherlands: migrants account for most infections health strategy on hcv in the netherlands longterm treatment outcomes of patients infected with hepatitis c virus: a systematic review and meta-analysis of the survival benefit of achieving a sustained virological response hcv richtsnoer weeks treatment daas systematic review of the accuracy of antibody tests used to screen asymptomatic adults for hepatitis c infection. c open ank?infot ype=g&label = -totaa l&tabel =b_ -basis &geg=vs_ gebr&item=j ap ank?zoekt erm=sofos buvir % fvel patas vir% fvox ilapr evir&toedi ening svorm =table tten en capsules the use of highly active antiretroviral therapy (haart) in patients with advanced hiv infection: impact on medical, palliative care, and quality of life outcomes guideline for conducting economic evaluations in healthcare utiliteitsmeting bij de klinische besluitvorming systematic review: health-state utilities in liver disease: a systematic review natural history of hepatitis c the nice cost-effectiveness threshold: what it is and what that means quantification of the potential impact of cost-effectiveness thresholds on dutch drug expenditures using retrospective analysis principles of good practice for budget impact analysis: report of the ispor task force on good research practices -budget impact analysis. value heal richtlijn voor het uitvoeren van economische evaluaties in de gezondheidzorg (protocol for the execution of economic evaluation in healthcare cost effectiveness of screening for hepatitis c virus in asymptomatic, average-risk adults identifying key benefits in european off-patent biologics and biosimilar markets: it is not only about price! hepatitis c virus prevalence and level of intervention required to achieve the who targets for elimination in the european union by : a modelling study hepatitis c in pregnancy in the era of direct-acting antiviral treatment: potential benefits of universal screening and antepartum therapy considering direct-acting antivirals to cure hepatitis c virus during pregnancy: is this the last treatment frontier? ther study of hepatitis c treatment during pregnancy -clinical-trials.gov n.d treatment of hepatitis c during pregnancy-weighing the risks and benefits in contrast to hiv strategies for achieving viral hepatitis c micro-elimination in the netherlands hepatitis b and c testing strategies in healthcare and community settings in the eu/ eea: a systematic review a simplified algorithm for the management of hepatitis c infection liver diseases in pregnancy: liver transplantation in pregnancy key: cord- -diqt g authors: wang, ying; yuan, yufeng; lin, likai; tan, xiaodong; tan, yibin title: determining the ideal prevention strategy for multidrug-resistance organisms in resource-limited countries: a cost-effectiveness analysis study date: - - journal: nan doi: . /s sha: doc_id: cord_uid: diqt g the aim of this study was to determine the most cost-effective strategy for the prevention and control of multidrug-resistant organisms (mdros) in intensive care units (icus) in areas with limited health resources. the study was conducted in icus of four hospitals. the total cost for the prevention of mdros and the secondary attack rate (sar) of mdros for each strategy were collected retrospectively from subjects from january to december . the average cost-effectiveness ratio (cer), incremental cost-effectiveness ratio (icer) and cost-effectiveness acceptability curve were calculated. hand hygiene (hh) had the lowest total cost ( . rmb) and sar of mdros ( . %) while single-room isolation showed the highest cost ( . rmb) and contact isolation had the highest sar of mdros ( . %). the average cost per unit infection prevention was . rmb, with the hh strategy followed by the environment disinfection strategy (cer = . ). hh had the highest iterative cost effect under willingness to pay less than rmb. due to the low cost for repeatability and obvious effectiveness, we conclude that hh is the optimal strategy for mdros infections in icus in developing countries. the cost-effectiveness of the four prevention strategies provides some reference for developing countries but multiple strategies remain to be examined. the prevention and control of multidrug-resistant organisms (mdros) is one of the most urgent public health concerns worldwide, especially in countries with limited health resources [ ] [ ] [ ] . mdros lead to a substantial economic burden due to unnecessary longer hospital stays, higher risk of readmissions and additional disease costs [ , ] . apart from the economic factors, the human disease burden is high, since mortality due to healthcare-associated infection (hai) has been suggested to be % higher in patients with mdros compared with those with antibiotic-susceptible infections [ ] . patients in the intensive care unit (icu) are regarded as those at the highest risk of infection with mdros due to invasive procedures, the use of immunosuppressive agents as well as a number of drugs (including antibiotics) and their underlying diseases [ , ] . in china, a recent study showed that the detection rate of multidrug-resistant acinetobacter baumannii (mdr-ab) in the icu was . %, compared with . % in non-surgical departments [ ] , while in german icus, the prevalence of carbapenem-resistant organisms (cro) was higher than in the general wards [ ] . direct contact with infected patients, carriers, medical equipment and the contaminated environment is believed to be the main route of transmission for mdros in the icu [ ] . hence, guidelines published by the world health organization (who) in proposed prevention and control measures for mdros based on multimodal infection prevention and control (ipc) strategies [ ] . unfortunately, many hospitals in low and middle-income countries lack the necessary infrastructures, medical equipment and experienced professionals, leading to substantial challenges and dilemmas for preventing mdros effectively [ ] . a national health resource survey in china showed an unpromising status between the healthcare demands and poor support because the ratios of patient beds to doctors and nurses were : and : . , respectively and each icu only had on average two single rooms available for isolation [ ] . the who has identified the obvious gaps in understanding the cost-effectiveness and practicability in isolating patients with mdros [ ] , in particular, how to maximise the effectiveness of prevention and control of mdros in hospitals with limited resources especially in low-and middle-income countries and how to optimise the use of prevention and control resources. health economic analysis is a useful tool to evaluate the effect of various health policies and is being increasingly used in the analysis of the prevention and control measures of mdros [ ] [ ] [ ] . previous studies have mainly focused on a single measure evaluation in developed countries [ ] [ ] [ ] [ ] , but did not consider the most cost-effective isolation measures with regard to limited health resources. hospitals with low-and mid-level budgets and resources need to select and focus on an optimal strategy based on practical cost-effectiveness, rather than implementing a whole host of measures against mdros. given the importance of the scientific and rationale prevention strategies for mdros worldwide, the limited isolation resources in developing countries and the lack of related cost-effectiveness analysis researches, this study aimed to determine the ideal prevention strategy for mdros in icus in areas with limited health resources, based on a decision tree model. this retrospective study was conducted in icus of four hospitals (three in each hospital) (zhongnan hospital of wuhan university, central hospital of wuhan, union hospital affiliated to tongji medical college of huazhong university of science and technology and tongji hospital affiliated to tongji medical college of huazhong university of science and technology) and covered the time period from january to december . the four hospitals were large general tertiary grade a centres with , , and beds. their icus are national largeand medium-sized and cover three key specialities in each hospital; icu bed numbers were , , and , respectively, and the icu wards were comparable in the number of beds and the condition of the admitted patients. the study was approved by the ethics committee of zhongnan hospital of wuhan university. the authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the helsinki declaration of , as revised in . the inclusion criteria of patients were: ( ) complete basic information; and ( ) patient consent to be included in the database and agreement on personal data collection. the four hospitals represented the four core infection prevention and control strategies, namely hand hygiene (hh), contact isolation, single room isolation and environmental surface cleaning and disinfection [ ] performed according to each hospital's own guidelines. we hypothesised that under different preventions, the number of patients infected with mdros would vary in the icus over the observation period. at the same time, the cost of implementing prevention measures will vary according to the number of patients infected with mdros in the icu and the costs needed for prevention will increase with an increasing number of patients infected with such organisms. therefore, under these assumptions, the secondary attack rate (sar) of mdros infection rates in each of the four hospitals (each applying one of the four strategies) were taken as the index of a prevention effect. the prevention cost of the four strategies multiplied by the number of infections was regarded as the accounting total cost. based on patients' outcomes under the different prevention strategies and associated costs, an analytic decision model was built to evaluate the cost-effectiveness of different prevention strategies. this economic evaluation is in accordance with the who costeffectiveness analysis guideline [ ] . there were two main data inputs in the decision model: the definition and calculation of the four prevention measures' cost and effectiveness. since there are few comprehensive prospective studies on the dynamic transmission of mdros in icus, especially in developing countries, we sought to retrospectively collect actual data from the study hospitals over year. based on the model construction, patients could be admitted to the icu as being either infected, colonised, or uncolonised with mdros. each prevention strategy had a focused key implementation and the cost per patient infected with mdros was calculated accordingly ( table ) . because of the limited information and published literature, we derived the detailed costs from three channels: the official price bureau, hospitals' material supply system and direct observation as supported by screenshots or paper trail. in addition, two other kinds of costs during mdros prevention were included, based on a previously published study [ ] namely nursing time ( . yuan/infected patient) and medical waste disposal (safe handling and transportation of medical waste bags once a day and transportation of sharps boxes every days at . yuan/infected patient). the total cost of each strategy was calculated by multiplying the cost per unit infected patient by the number of infected patients in the ward during the study period. we postulated that under the different prevention strategies implemented in each hospital, the patients' infection status and the rate of new infections in the wards would be different. the effectiveness of this evaluation was regarded as the sar of mdros in the wards. the reverse calculation was used as the standard in the statistical analysis. sar is an effective index widely used in epidemiology and infection for evaluating the prevention and control measures of infectious diseases [ ] . a retrospective cohort study used sar to assess the effect of different prevention measures on middle-east respiratory syndrome [ ] . similarly, sar was used as the major index to assess the effectiveness of antiviral prophylaxis during the hin outbreak [ ] . the calculation and estimation of the number of cases were derived from the retrospective data and based on the following formula [ ] : sar = number of secondary patients infected with mdros number of susceptible contact patients × k the key indicators in this study were: ( ) cost-effective frontier (cef): in the cost-effect diagram, the cost for all strategies are connected by line segments to form a cost effect boundary for cases of multiple strategies and absence of inferior strategies; ( ) average cost-effectiveness ratio (cer): average cost per unit infection prevented; ( ) incremental cost-effectiveness ratio (icer): represents the cost of unit effectiveness (Δc/Δe). the effectiveness of strategies was based on sar and icer referred to the cost to prevent one new secondary infected patient in the ward; ( ) a tornado analysis was performed to determine the value that would change the choice of the optimal decision within the change the statistical significance of the basic patients' characteristics among the four strategies was determined using χ tests and anova with tukey's post hoc test, using the stata software (statacorp lp, college station, tx, usa). p values < . denoted statistical significance. for the cost-effectiveness analysis, the relevant parameters were entered into excel (microsoft, redmond, wa, usa) and analysed using treeage . (treeage software, inc., williamstown, ma, usa). a total of patients from the icus were included in the study. the numbers of patients in each strategy were with hh, with single-room isolation, with contact isolation and with surface cleaning and disinfection. there were significant differences among the four groups regarding age, icu stay, surgery, type of sample for bacterial detection, numbers of routine or abnormal tests and use/duration of airway ventilation, central and urinary catheters (all p < . ; table ). the sar of mdros was highest with the contact isolation strategy ( . %) and the lowest with hh strategy ( . %), single-room isolation ( . %) and environmental disinfection ( . %) falling in between (table ) . table lists the multidrug-resistant microorganisms recovered from patients and shows that the most frequent were carbapenem-resistant enterobacteriaceae, followed by carbapenem-resistant acinetobacter baumannii and methicillinresistant staphylococcus aureus. among the four major prevention measures, only hh and contact isolation were visible on the cost-effectiveness curve, while isolation in single-rooms and environmental disinfection fell beyond the cost-effectiveness curve (fig. ) . hh had the lowest total cost ( . rmb) and single-room isolation had the highest cost ( . rmb). the average cost per unit infection prevention was . rmb, with the hh strategy followed by environmental disinfection (cer = . ). incremental costs for each additional unit of infection prevented were and - . for the hh and environment disinfection strategies, respectively ( table ). the tornado analysis showed that the number of new patients infected with mdros under the hh strategy was the only factor that had the largest impact on the net benefit of the overall strategy. the cost-effectiveness acceptance curves for the four strategies were compared when the wtp for measures to prevent the transmission of mdros was set at rmb. as shown in figure , hh had the highest iterative cost effect under willingness to pay less than rmb meaning that it can be repeated at a very low cost. when the wtp was increased to rmb, single room isolation became the best strategy. the cost-effectiveness analysis of four prevention strategies against mdros in icus was studied in china, a country with limited health resources. hh proved to be the optimal infection prevention strategy due to its excellent performance and lower costs for each additional unit of sar prevented, followed by contact isolation. the net benefit of the overall infection prevention strategy was mostly due to the effectiveness of the hh strategy which due to its low cost (< rmb) for repetitive operation, would be more likely to be adopted by health care providers at the national or local level, for the prevention of mdros. while hh practice can readily be adopted by all healthcare staff, other strategies, in particular, environmental cleaning often requires dedicated staff members and specialist equipment. likewise, single-room isolation requires that wards are designed to include such rooms, which is not the case in most hospitals in china since most wards only have one or two single rooms [ ] . as expected, contact isolation proved the most expensive and impractical infection control strategy for developing countries because of the need for expendables and single-use supplies. it follows that the ideal setting, like in most hospitals in developed countries, is the simultaneous use of all four modalities [ ] , ying wang et al. but the key finding from the present study is that hh alone was the most efficient of all approaches for preventing mdros due to the direct reduction of sar of mdros, combined with the lowest cost. nevertheless, human factors will ultimately determine the efficacy of these methods and the transmission of mdros in hospitals [ ] . the who- campaign has been shown to be effective in improving hh [ , ] , but in low-and middle-income countries, poor hh compliance of < % remains typical [ , ] . the benefit of hh alone is supported by a number of studies, as shown by a meta-analysis of the relation between hh and the incidence rate reduction of nosocomial infections in icus [ ] . moreover, a recent study by luangasanatip et al. [ ] showed that hh interventions are cost-effective in preventing mrsa in icus in middle-income countries and a study from vietnam reported a saving of $ per prevented hai through improved hh [ ] . individualised bundles of infection control measures, including hh, were also identified as a recommended strategy (icer = $ . ) in the latter study. due to its simplicity, attainability and economic benefit, our study underlines the effectiveness of hh for preventing mdros transmission in hospitals and additionally can be extended to areas with limited health resources [ ] . it is noteworthy that the effectiveness of hh depends on the compliance of healthcare workers. a study of > hh opportunities monitored by unique observers showed that an improvement of hh compliance from % to % could decrease the hai rate and result in savings of about $ million [ ] . hence, for hh, the compliance-dependent effect is particularly important when the cost fluctuation is small. for hh, the monetary investment is small, but the investment in time is more important, which is a primary factor playing against the strategy [ ] . insisting on better practice, using surveillance and alarms and providing feedback to staff could all contribute to improving the hh rates [ , ] . contamination of the near-patient environment by mdros was found to be responsible for patient-to-patient transmission of these organisms [ ] [ ] [ ] [ ] [ ] . nevertheless, as for hh, compliance with the best principles of surface and environmental cleaning was found to be relatively low, with an average of % [ ] [ ] [ ] . in addition, the significant material and human resources are needed for optimal environment control and have been shown to be less cost-efficient than hh [ ] . single-room isolation is one of the oldest methods for infection control, but it is time-consuming, may impede proper care and is inefficient if a total contact isolation strategy is used since the health care staff may spread infections [ ] . the contact isolation strategy, especially for msra, remains controversial. spence et al. [ ] concluded that contact isolation was costly and unnecessary for patients colonised with mrsa, but an evaluation of independent studies found there was some evidence supporting the practice [ ] . considering the low compliance level of hh in some developing countries [ , ] , single-room isolation or other sequential isolation measures could prove easier to implement than routine hh. some previous studies have explored the cost-effectiveness of prevention measures for hais, including mdros, using a mathematical model; a systematic review documented that most of the cost-effectiveness analyses of control measures for mrsa remained at the level of individual measures, with very few studies comparing multiple measures [ ] . rattanaumpawan et al. [ ] compared a combination of measures with traditional infection control care but did not analyse the individual components of the combination. moreover, up to , most of the economic evaluations of prevention measures against mrsa were carried out in developed countries such as the usa, germany and commonwealth countries [ , , , , , ] , with important gaps in knowledge regarding developing countries. the latter gaps combined with other shortages in health resources together contribute to hinder the prevention and control of mdro transmission in such countries. effective strategies are necessary to slow down the epidemics of mdros, which potentially will have disastrous impacts on public health in the future [ , ] . the present study has some limitations. first, while it investigated the transmission data of patients infected with mdros and a decision-making model was developed, we recognise that the development of infectious diseases is an event-dependent process. second, costs were calculated according to each patient with the total cost over during the study (expenditure per patient × number of infected patients in the ward), it did not take into account the patient's hospitalisation time and other factors. hence, for future studies, we intend to establish a complete markov model, including a time model based on the complete data of the outcomes of patients infected with mdros in icus. based on this, we will include the incremental cost that relies on the transition period of patients to perform in-depth cost-benefit analysis [ ] . nevertheless, we suggest that our study was innovative in determining the cost-benefit analysis of the different measures against mdros in icus and constitutes an exemplar for developing countries. third, each hospital relied on a different strategy for preventing mdros and thus some bias could have resulted from differences between hospitals that were not taken into account in the study. finally, multivariable analysis of the use of devices was not performed owing to the wide variability of the number of devices and timing of use among the patients. such shortcomings need to be addressed in further studies to determine the optimal mdro prevention strategy in different settings. in summary, to our knowledge, this is the first study to evaluate the optimal strategy among hh, single-room isolation, contact isolation and environmental sanitation in developing countries for the prevention of mdros. due to the low cost for repeatability and obvious effectiveness, hh represents the optimal strategy to reduce the incidence of mdros infection in icus in limited health resources and our findings may be relevant to low-and middle-income areas. further research is needed on how to maximise the effectiveness of hh and the optimum combination of different prevention measures. cer, cost-effective frontier; incrcost, incremental cost; increff, incremental effectiveness; icer, incremental cost-effectiveness ratio; nmb, net monetary benefits; abs. dominated, absolute dominated (i.e., when there are multiple strategies to choose from, there is always a strategy that is better than an absolute disadvantaged strategy); undominated (i.e., in the cost-benefit analysis, an undominated strategy indicates that when there are multiple strategies to choose from, the dominant strategy is better than the others). a average cost-effectiveness ratio. b incremental cost-effectiveness ratio. ying wang et al. a view on antimicrobial resistance in developing countries and responsible risk factors antibacterial resistance worldwide: causes, challenges and responses hospital infections by antimicrobial-resistant organisms in developing countries economic burden associated with multi-resistant gram-negative organisms compared with that for methicillin-resistant staphylococcus aureus in a university teaching hospital epidemiology and economic impact of health care-associated infections and cost-effectiveness of infection control measures at a thai university hospital impact on morbidity, mortality, and length of stay of hospital-acquired infections by resistant microorganisms an integrative review of infection prevention and control programs for multidrug-resistant organisms in acute care hospitals: a socio-ecological perspective multi-drug resistant organism infections in a medical icu: association to clinical features and impact upon outcome a multi-center study on the risk factors of infection caused by multi-drug resistant acinetobacter baumannii prevalence of carbapenem-resistant organisms and other gram-negative mdro in german icus: first results from the national nosocomial infection surveillance system (kiss) systematic literature analysis and review of targeted preventive measures to limit healthcare-associated infections by meticillinresistant staphylococcus aureus who guidelines for the prevention and control of carbapenem-resistant enterobacteriaceae, acinetobacter baumannii and pseudomonas aeruginosa in health care facilities. geneva: world health organization status of the prevention of multidrug-resistant organisms in international settings: a survey of the society for healthcare epidemiology of america research network the professional status of infectious disease physicians in china: a nationwide cross-sectional survey extended models for nosocomial infection: parameter estimation and model selection cost-effectiveness of a national initiative to improve hand hygiene compliance using the outcome of healthcare associated staphylococcus aureus bacteraemia cost-effectiveness of a model infection control program for preventing multi-drug-resistant organism infections in critically ill surgical patients cost-effectiveness of strategies to prevent methicillin-resistant staphylococcus aureus transmission and infection in an intensive care unit challenges and strategies for prevention of multidrug-resistant organism transmission in nursing homes evaluating infection prevention strategies in outpatient dialysis units using agent-based modeling efficacy of infection control interventions in reducing the spread of multidrug-resistant organisms in the hospital setting making choices in health: who guide to cost-effectiveness analysis. geneva: world health organization a literature review of nursing turnover costs secondary attack rate outcome measurement of extensive implementation of antimicrobial stewardship in patients receiving intravenous antibiotics in a japanese university hospital assessment of secondary attack rate and effectiveness of antiviral prophylaxis among household contacts in an influenza a (h n )v outbreak in kobe environmental cleaning and disinfection of patient areas guidelines on hand hygiene in health care (first global patient safety challenge -clean care is safer care). geneva: world health organization interventions to improve hand hygiene compliance in patient care successful implementation of the world health organization hand hygiene improvement strategy in a referral hospital in mali effectiveness of a catheter-associated bloodstream infection bundle in a thai tertiary care center: a -year study comparative efficacy of interventions to promote hand hygiene in hospital: systematic review and network meta-analysis cost-effectiveness of interventions to improve hand hygiene in healthcare workers in middle-income hospital settings: a model-based analysis cost-effectiveness of a hand hygiene program on health care-associated infections in intensive care patients at a tertiary care hospital in vietnam health care-associated infections: a meta-analysis of costs and financial impact on the us health care system factors impacting hand hygiene compliance among new interns: findings from a mandatory patient safety course factors affecting hand hygiene adherence at a private hospital in turkey contamination, disinfection, and cross-colonization: are hospital surfaces reservoirs for nosocomial infection? transfer of multidrug-resistant bacteria to healthcare workers' gloves and gowns after patient contact increases with environmental contamination risk of acquiring antibiotic-resistant bacteria from prior room occupants risk of acquiring multidrug-resistant gram-negative bacilli from prior room occupants in the intensive care unit risk of hand or glove contamination after contact with patients colonized with vancomycin-resistant enterococcus or the colonized patients' environment systematic review of studies on compliance with hand hygiene guidelines in hospital care universal glove and gown use and acquisition of antibiotic-resistant bacteria in the icu: a randomized trial interventions to reduce colonisation and transmission of antimicrobial-resistant bacteria in intensive care units: an interrupted time series study and cluster randomised trial preventing the transmission of multidrug-resistant organisms: modeling the relative importance of hand hygiene and environmental cleaning interventions is patient isolation the single most important measure to prevent the spread of multidrug-resistant pathogens? contact precautions for methicillin-resistant staphylococcus aureus colonization: costly and unnecessary? isolation measures in the hospital management of methicillin resistant staphylococcus aureus (mrsa): systematic review of the literature cost-benefit of infection control interventions targeting methicillin-resistant staphylococcus aureus in hospitals: systematic review burden and management of multidrug-resistant organisms in palliative care a cost-effectiveness modelling study of strategies to reduce risk of infection following primary hip replacement based on a systematic review conflict of interest. none. key: cord- -jik j authors: agrawal, anshu title: sustainability of airlines in india with covid- : challenges ahead and possible way-outs date: - - journal: j revenue pricing manag doi: . /s - - -z sha: doc_id: cord_uid: jik j coronavirus outbreak has been highly disruptive for aviation sector, threatening the survival and sustainability of airlines. apart from massive losses attributed to suspended operations, industry foresee a grim recession ahead. restrictive movements, weak tourism, curtailed income, compressed commercial activities and fear psychosis are expected to compress the passenger demand from to %, endangering the commercial viability of airlines operation. fragile to withstand the cyclic momentary shocks of oil price fluctuation, demand flux, declining currency, airlines in india warrants for robust structural changes in their operating strategies, business model, revenue and pricing strategies to survive the long-lasting consequences of covid- . paper attempts to analyze impact of lockdown and covid crisis on airlines in india and possible challenges ahead. study also suggests the possible way-out for mitigating the expected losses. the world, at present is combating with pandemic covid- . emerged from wuhan (china) in december , within few months it has taken countries across the globe into its clutches. with reported cases on january , the infected cases have crossed . million as on july . india is no different; with cases reported as on february, the number has surpassed , , whilst three extended phases of lockdown. the magnitude of virus contagion spread in the absence of any antidote developed so far has left the countries across the world with quarantine as the only remedy, despite of its drastic consequences on the economy. aviation sectors is perhaps worst hit with covid impact. the preventive restrictive movements have drastically dented the airlines and allied services with huge losses. airlines passengers' services in india remain suspended for sixty day ( march to may ), bringing massive loss to the industry. according to dgca, six days suspended operations of march leads to % decline in passenger traffic (from . million reported in march to . million in march ). as per crisil infrastructure advisory report, the expected revenue loss to the indian aviation sector due to lockdown amounts to billion; inter se, airlines account for % losses, followed by allied servicesground handling, etc. capa india estimates the industry staggering losses of to billion in april-june quarter, assuming operations to remain suspended till june . this imply loss of . to billion for per day of extended lockdown. in addition to the above losses, the industry foresees grim recession ahead. restrictive movements and destinations, truncated consumable income, decline in tourism, and fear psychosis are expected to significantly curtail the passenger traffic for the current fiscal or perhaps longer. india is the third largest domestic civil aviation market in the world (ibef, report). however, thin profit margins, high operating cost, inflated taxes and cut-throat price war make it one of the toughest aviation market (saranga and nagpal ) . the cost structure of airlines in india is believed to be highly bloated with atf taxes, landing and parking charges, which are perhaps highest in india. the industry is exposed to high operating leverage. the airlines operating cost structure consists of nearly to % of fuel cost, % lease rental, nearly % for other operating expenses (including general administrative, operating expenses such as flight equipment, maintenance, overhaul, user charges including landing, airport charges and air navigation charges (dgca report). other than fuel cost, maintenance of aircraft, selling & distribution cost, and parking & landing charges, rest other expenses are fixed and are to be honored irrespective of flight operations. burden of fixed charges-lease rental, interest charges, and crew salaries keep the airlines on their toes for managing cashflows. the high operating cost and cut-throat competition compel the airlines to struggle with low margins. the airlines demand in india is highly price elastic (wang et al. ). entry of lowfrill competitors has changed the airlines price dynamics of pricing the services that were earlier based on additional frills (saranga and nagpal ) . any hike in the expenses, prima-facie, is a pinch on the airlines margin as ugly fare wars restrict to surpass the uncertain hike in costs on ticket prices. commercials of airlines revolve around available seat kilometers (ask)-capacity, revenue per kilometer (rpk)-income earned, passenger load factor (plf)-capacity utilized, break-even load factor (belf)-operating cost per ask over operating revenue per rpk. higher the distance flown, more is the opportunity for the airlines to spread the operating fixed cost over longer distance and thereby reducing their adjusted operating cost. in this backdrop, improving plf by offering lucrative offers is prevalent trend in the industry. as provided, the airlines in india performed at decent capacity of nearly % in the month of january, . in the month of february, irrespective of low demand, the spicejet, go air, indigo, air asia, vistara have managed high plf by providing attractive offers (fig. ) . higher plf, however, does not implies profitability. it only represents the successful selling of available seats. operating viability requires the plf to exceed belf. the irony is that despite of heavy demand, the airlines strive hard for making break-even due to tough competition. in the dilemma of managig operating cash flows, the cash stripped airlines with overmounted fixed operating costs emphasize on selling more seats, ignoring the break-even. as provided in the fig. , only five airlines operate above belf during fy - , with the safety margins from . to . %, whilst rest all were in red. the covid outbreak has added financial woes of the sector. with the dwindling demand anticipations, capacity utilization certainly will be a major challenge ahead for airlines sustainability. hitherto combating for break-even, low passenger traffic possibly restrain the airlines from recovering their variable expenses, thereby obstructing the commercially viability of their operations. present study attempts to analyze the financial impact of covid outbreak on airlines and challenges ahead. possible suggestion for sustainable operations of airlines are suggested. findings are expected fig. plf of airlines in india during january and february . source dgca https ://www.ibef.org/indus try/india n-aviat ion.aspx#:~:text=india 's% avi ation % ind ustry % is% exp ected ,aviat ion% nav igati on% ser vices % by% https ://www.dnain dia.com/busin ess/repor t-the-cost-of-flyin g-high- to contribute in the restructuring of the airlines for operating viability and sustainability. airline industry has been one of the fastest growing industry globally in terms of demand as well as capacity (lee ) . over the past century, commercial aviation has been observed as integral part of economic prosperity, stimulating trade, cultivating tourism development. its relative affordability in recent years has inculcated it in people's lifestyles (o'connell ) . aviation sectors economic contribution (direct, indirect, induced and tourism concomitant) in global gdp is estimated as usd . trillion (atag ). notwithstanding the growth in demand as well as capacity, the sectors has always been financially challenging struggling with thin margins (o'connell ), vulnerable to fuel prices, foreign exchange, interest rates and high competition (merkert and swidan ; stamolampros and korfiatis ) . the industry has been exposed to dynamic external environment, regulations, technology, customers preference, intense competition, labor cost, fuel prices and security measures and so forth (riwo-abudho et al. ) . airline industry performance is contingent to macro-predictability, micro-uncertainty and macro environmental factors (mhlanga ) . airlines industry has always been exposed to exogenous events. terrorist attack of / has put the industry into depression making number of airlines bankrupt. those who rescued from the effect have been grabbed with the oil crisis of (yang ) . the entry of low cost carriers (lcc) in triggered turnaround changes in the industry in terms of pricing strategies and well as competition level (belobaba ) . the lccs pricing and revenue management strategies threaten the commercial viability of traditional model, compelling the changes in conventional airline revenue management practices (michaels and fletcher ). the paper examines how they differ in their approach, how airlines are responding and what constitutes an effective response in the changed airline business world. this includes consideration of all the marketing levers (product, price, promotion and distribution) in an integrated way, as well as developments needed in the core revenue management systems themselves (michaels and fletcher ) . online bookings, access to airline tickets on internet has made price competitiveness as an important parameter of airline's success (ratliff and vinod ) . india airlines market despite of being the fastest growing market (mahtani and garg ) , has been one of the toughest aviation markets in the world, due to high fuel prices, overcapacity and intense price competition (saranga and nagpal ) . notwithstanding the extensive infrastructural development supported by government, airlines in india often combat financial distress with the changing dynamics of internal and external environment (mahtani and garg rapid transformation with the liberalization of indian aviation sector (singh ) . india began to relax controls on its airline industry in , allowing willing entrants to add system's capacity. however, financial performance of the airlines remains challenging owing to inappropriate policies, restricted capacity allocation on profitability basis (hooper ) . liberalization of air travel services and the advent of low-frill airlines have changed the panorama of indian civil aviation in terms of demand as well as supply (ohri ; srinidhi ) . reformation of regulatory policies resulted in three-fold increase in the number of scheduled airlines and a five-fold increase in the number of aircraft operated (o'connell and williams ) . the increased interconnectivity within the global airline markets has altered the dynamics of external environment and internal operations (riwo-abudho et al. ; singh ) . success and survival in this milieu warrants for coherent strategies adapting with market flavor (pathak ) . entry of the lccs in india in , with first 'no-frills' airlines-air deccan has changed the dynamics of indian domestic aviation market (sakariya et al. ). low-cost carrier (lcc) by enhancing affordability of air travel has stimulated the demand for air travel in india (krämer et al. ; wang et al. ). undoubtedly, low-frill operation has proved to be a successful business model in the industry (alamdari and fagan ) . budget airlines and small chartered airlines witnessed more efficient in the system (dhanda and sharma ; jain and natarajan ; saranga and nagpal ) and dominated the indian airline market (deeppa and ganapathi ; wang et al. ) . the lcc in india have managed to achieve significant operational efficiencies with the rigid cost structure, heavy taxes, high landing and parking charges, undesirable regulatory factors (saranga and nagpal ) . india's low cost carriers show better scale efficiency vis-à-vis their full service competitors (sakthidharan and sivaraman ) . low cost airlines have been witnessed advantageous in utilizing their capacity compare to the full service airlines which strives hard to attain break-even capacity (thirunavukkarasu ) . however, the inexorable rise of lcc has made the industry more volatile (doganis ) . intense competition and enhanced capacity have made cost effectiveness as the daring need for survival and sustainability. financial performance of airlines is vulnerable to both internal conditions of the company and as well the external environment. operating factors, namely, operating revenue per air kilometers, capacity, cost structure, load factor dictate the operational output of the airlines and their commercial stability. from the external environment, atf prices largely affect airlines profitability in india. also, annual inflation and gdp growth rate in the country has a major influence on the sustainability of the airlines in india (mahtani and garg ) . with uncontrollable cost behavior, tight margins and cut-throat market, survival and subsistence of airlines largely depends on its ability to maximize their customer base (singh ) . fierce competition compel the airlines to optimizes their revenues (josephi ; krämer et al. ). in the backdrop of covid pandemic outbreak, the globally airline industry has been adversely affected. airlines in india which have been observed vulnerable to withstand the cyclic economic disruption (of fuel prices, inflation, devaluation of currency and demand shock), certainly be entering into a tough time with extremely low demand and ever mounting losses. present study attempts to analyze the financial impact of covid pandemic on airlines in india and possible impact of their financial strengths and weakness. further study suggests possible way-outs of sustaining operating viability. the indian aviation industry is characterized by high fixed costs of nearly to %. these costs include lease rental, employees cost, interest charges. per day of suspended operations has hit the industry at the rate of - crore loss per day. table exhibits fixed-cost information pertaining of four key airlines of india for last three years (fy to ). the costs mentioned signify the charges that are to be met irrespective of the business operations. the increasing pattern of expenses over years, prima-facie, signify the expanded operations' size over years. ceteris paribus, no significant change in the operations size and cost for the fy - , per day loss of suspended operations for interglobe aviation accounts for crores, followed . crores for spicejet, . crores for go airlines and . crores for air asia (based on the - estimates). in capital intensive industries, such as airlines, liquidity plays an important role in boosting profits (merkert and swidan ) . perhaps the cash rich airlines are in better position to negotiate with the suppliers-oil companies, lessor, bankers, employees for favorable deals and heavy discounts. airlines in india suffers from weak liquidity. cash burn rate of airlines in india during the years to is provided in table . the cash burn rate indicates the number of days for which a company can sustain its operations with the available cash reserves. the data contained in table , suggests few days of cash back-up available to most of the airlines, excluding interglobe aviation which is exhibiting consistent pattern of satisfactory cushion of more than a quarter. the aggregated cash reserves of interglobe aviation as on december were reported to be . crores . assuming, . crores of daily fixed cost (refer table ), the reserves of crores possibly have been wiped out amid seventy days of lockdown. remaining cash balance of . crores suggest the probability of days of survival, based on estimated burn rate of . per day. however, for other airlines resuming operations with insufficient operating cash seems to be a challenge. in the backdrop of tight liquidity, thin margins and high burn rate, the airlines have always been fragile to withstand the normal demand shocks, oil price fluctuation, depreciating currency, etc. industry has vouched the devastating impact of these events ranging deep losses to airlines bankruptcy. table exhibits onwards financial performance of airlines in india in terms of profits margins, rate of returns, assets turnover ratio and interest coverage ratios. as provided, the profit margins of the airlines are highly thin and unsatisfactory to insulate the firms from sudden shocks. median net profit margin − . , prima-face, corroborate that net profits of all the airlines in india are occasionally positive. there appears only three airlines, interglobe, go air, and blue dart (cargo airline) with positive net profit margin in all the five years. in terms of magnitude, the net profit margin . to % and ebit margin of to % does not seems satisfactory to justify the corpus invested and the risk involved there in. oil price hike of has plunged the sector into deep losses. interglobe aviation that appears to be best performer of the industry has experienced deep shrinkage in its net profit margin of from to . % (table ). unable to take the hit, loss running jet airways blown out of the race with its operations meeting grinding halt in april . previously also, industry has a history of several starts and may failures; east west airlines and damania airways in s, kingfisher airlines in are classic instances of airlines financial failure. table exhibit the altman z-score of select four airlines. altman z-score model (altman ) was developed by edward altman in . it gauges the likelihood of bankruptcy of business concern within two years, using multiple corporate income and balance sheet values. z-scores are used to predict corporate defaults and an easy-to-calculate control measure for the financial distress status of companies. the z-score is calculated using liquidity, profitability, leverage and turnover parameters. (altman ) . here x working capital/total asset, x retained earnings/total asset, x ebit/total x market capitalization/ book value of debt, x total sales/total assets. score below . signifies high probability of bankruptcy; . to . is considered as grey zone and score of above . is considered as safe zone. this model was applicable for manufacturing sector. for predicting the bankruptcy of service sector firms in emerging market modified atman score was proposed (altman ). as per the model, z − score = . + . x + . x + . x + . x , here x working capital/total assets, x retained earnings/ total assets, x ebit/total assets, x market capitalization/ book value of debt. score above . is considered safe zone, . to . as moderate risk and score below . indicates high risk of bankruptcy. table exhibits the altman z-score of airlines in india computed using traditional altman model and modified altman model for emerging market. in the backdrop of unavailability of market capitalization information of all the airlines, the enterprise value minus book value of debt is considered as value of equity. the findings of both the models lend credence to the sustainability of indigo aviation and spicejet. nevertheless, the decline in the scores is very likely, due to deteriorated finances amid lockdown and grim prospect of passenger demand ahead. covid- pandemic has proven highly disruptive. it has wreaked havoc with the global economy, economically, socially and financially (laing ; wren-lewis ). the aftermath of the disasters is perhaps more threatening, endangering the survival and sustainability of various businesses. airline industry is worst hit sector, which is expected to lose usd . billion in , the highest loss the sector has ever witnessed (iata). owing to the restricted movements and destinations, the industry expects severe decline in its passenger load (thams et al. ) , perhaps, a significant parameter of airlines profitability (baltagi et al. ; clark and vincent ; sibdari et al. ) . as provided in fig. , month of march has witnessed sharp decline in plf of airlines all across the globe. other than the loss amid suspended operation, the future prospect of the industry seems more dreadful for sustainable operations of airlines. in the backdrop of aggressive multiplication in covid cases, the likelihood of normal passenger traffic seems distant. restricted movements, fear psychosis, declined tourism, reduced commercial activities, curbed disposable income is expected to have significant impact on passenger airlines demand. tourism sector is considered as significant driver/ stimulator of airlines business (bieger and wittmer ) . an important aspect of international traffic to and from india pertains to trend in foreign tourist arrivals in india. the months from april to july are generally observed as peak season for the airlines, with the maximum passenger load factor (plf). in the fy - , yoy growth in plf is positive only in the month of april & july. as per the dgca report, % of international passenger traffic during fy was attributed to tourism sector. in view of expected decline in tourism amid covid pandemic, the airline business foresees a major disruption ahead. according to icao united aviation study, depending upon the duration and intensity of outbreak, control measures and economic and psychological impact, the global pink cells portray risky zone and green cells represent safe zone as per altman z-score airlines industry may witness decline of to % seats offered, reduced passenger traffic from to million and gross operating revenue loss of approximately usd to million for the year . as per the report, the estimated decline is the worst ever observed before during any of the crisis, economic or otherwise (fig. ) . airlines in india are vulnerable to high operating leverage (sakthidharan and sivaraman ) . operating leverage signifies an ability of a firm to use its fixed operating expenses to magnify the impact of change in its sales on its operating profit. degree of operating leverage (dol) is calculated as total contribution /total ebit. high the degree of operating leverage, higher will be the magnifying impact of increased operations/sales on ebit (chen et al. ; garcía-feijóo and jorgensen ; mandelker and rhee ) . for instance, times of dol implies that if sales increase by % than ebit will increase by × , i.e., times. it is worth mentioning, that use of fixed operating cost signifies the risk in operations; the risk of repaying the fixed charges in case income fall short of expectations (gahlon ; mcdaniel ) . performance of high levered firms significantly reduced compared to their competitors in industry downturns due to enhanced cost of financial distress (gonzález ). in the backdrop of severe downturn expected in the industry, the highly levered airlines in india are likely to suffer heavy losses. table exhibits the degree of operating leverage of four airlines in india and the consequences on the ebitda of the airlines, with the different expectations of possible decline in sales amid covid impact. the rationale of including select airlines for analysis is the unavailability of the data for the year . as provided, air asia (india) is in losses; go air, spicejet and intergloble are reflecting alarming degree of operating risk. high the dol, higher will the expected losses. with . times of dol, ebitda of interglobe aviation is expected to decline by . times with % dip in its revenue, i.e., from . lac crores of ebitda to negative- . lac crores. social distancing practices initiated by regulatory authorities and airlines to prevent infection outbreak will be financial hit on airlines pocket (iata economics ). declined plf coupled with cost of social distancing is expected to threaten the commercial viability of airlines operations. measures such as leave empty seats between passengers in the aircraft will reduce the seating capacity by to %. in india, dgca laid down social distancing norms and sanitization norms for airlines to be followed during passengers handling, sanitizing aircrafts, checkpoints and baggage, ppe kits, medical team, etc. this cost will further dig the profitability of the airlines. the dgca advisory of blocking middle seat, will compressed the seat offering capacity of airlines to %. plf is an important driver of airline financial performance. based on a sample of airlines, on average, airlines break even at a load factor of %. notwithstanding the high plf of to %, airlines are witnessed struggling for breakeven. as per iata analysis, out of the sample of airlines across globe, only airlines will manage break-even below %. in the present scenario, where airlines plf is expected to decline by to %, the financially feasibility of airlines operations seems scary. additionally, the cost of implementing other social distancing and sanitization norms will further enhance the airlines' costing. airlines perhaps find it difficult to cover the variable cost of their operations. post-lockdown world will be not be the business as usual. the airline industry combat with covid- and its after effects seems taxing and perhaps long drawn-out. the sustainability and survival of airlines warrants for turnaround changes in their strategies and business model to strengthen their financial stamina. overcapacity, intense competition and high operating cost are the major factors affecting airlines performance. to overcome the present challenge of covid crisis, optimal utilization of resources, cooperation rather than competition, and cost optimization seem to be the possible way-outs for sustaining with commercially viable take-off on rough terrain. air cargo business despite of being a least preferred choice of airlines compared to passenger business, has an important role to play in the airline's profitability. threatening subsistence with the growing challenges of the industry warrants for major structural changes in the present business model. accommodating the cargo business in the existing business model perhaps be an effective steps towards the improved performance (reis and silva ) . the globalization of the supply chain has resulted in competitive pressure on the air cargo industry. with independent and improved supply chain strategies, airlines can positioned themselves in the global supply chain market (hong et al. ) . high degree of cargo business is evident to improve the operational efficiency of combination as well as cargo airline (hong et al., ) . airlines with a high share of cargo business in their overall operations are significantly more efficient than airlines ( hong and zhang ) . however, challenges for handling cargo makes it less attractive to airlines compared to passenger business. combination airlines use the belly space of passenger aircrafts to substantiate the cargo. these airlines often experience the problem of freight orders exceeding the airline's fixed capacity, particularly for hot selling routes (feng et al. ) . in present scenario, where a severe decline in passenger traffic as well as restricted destinations is expected amid infection paranoia, cargo business perhaps can be used a rescue boat to safeguard the airlines from expected the crash landing. it is a saying in management accounting, that in short-term if profits can't be maximized, focus should be on minimizing the losses. for optimum capacity utilization, cargo-cum-passenger model can be an effective way-out. at present nearly % of freight business in india is done through belly cargo. only blue dart is fully dedicated airlines for freight cargo business (fig. ) . in view of restricted passenger movements, from january onwards airlines across the globe have started engaging passenger aircrafts entirely for cargo (fig. ) . indigo followed by spicejet have also joined the race. figure portray the average seat capacity of scheduled airlines in indian during fy - . as provided, most of the airlines have the average capacity of to . with the expected - decline in passenger traffic in current fiscal, accommodating cargo load for unutilized seats can mitigate the revenue losses of the airlines. amid low passenger traffic, dedicating small aircrafts (with less seating capacity) for passenger business relatively will be more financially viable for the airlines. big aircrafts can be temporarily converted in cargo planes for carrying supplies. depending upon the cargo load, large capacity planes can be fully dedicated or utilized as passenger cum cargo planes. in view of dgca advisory to leave middle seat vacant, some temporary arrangements for accommodating cargo in provided space can be worked out. for instance, vacant seat can be used for carrying passenger's luggage and the side carriers can be utilized for lesser weights parcels. also, the space used for accommodating passengers' check-in luggage can be utilized for cargo business. the passenger cabin can be restructured in such a manner that its front and back seats can be used for passenger traffic and middle space can be utilized for cargo services. airline industry is known for ugly competition and fare wars that perhaps has been the prime reason for their meagre profit margin (eng and vichitsarawong ) . in the backdrop of trimmed passenger traffic expected for upcoming months, pooling of resources perhaps can be useful step in this direction. airlines industry needs to adapt cooperation model instead of competition. alliance in the airline industry is a widely used strategy to stimulate competition (cobeña et al. ) . alliances are useful rescue for the firms with vulnerable strategic positions either because of competition or when they are attempting pioneering technical strategies (eisenhardt and schoonhoven ) . they enhance value by facilitating optimal utilization of pooled resources (das and teng ) . the alliance, perhaps, can be better way-out for balancing these demand and supply fluctuations. alliance for aircraft sharing can possibly assist the airlines in optimizing their aircraft capacities and mitigating their operating losses. the covid economic impact on aviation is extreme and perhaps uncertain. higher the reduction in plf, more difficult will be the attainment of break-even for the airlines. the lockdown of two months with zero revenue and spiraled fig. cargo business using passenger flights during january to april . source icao https ://www.hindu stant imes.com/busin ess-news/india -s-domes ticair-traffi c-to-fall-to- -mn-this-fisca l-repor t/story -ynobz buttd z g d sn fm.html using passenger planes for carrying cargo in belly space fixed charges, particularly, loan instalments and lease rental, perhaps has drained out the liquidity of airlines. with the trimmed air traffic estimated in the coming months, there seems meagre probability of recovering the past losses. with the reduced plf the recovery of variable cost of operating a flight will be challenging, threating the operation viability of airlines. figure portray projection done by icao regarding commercials of aviation sector. as provided, in all the situations the operating losses are confirmed, with the only difference in the magnitude of losses from high to low. in the given situation, bailout package, particularly, waivers of interest charges pertaining to lockdown period, reduced landing and parking charges, atf taxes, seems essential for the stability of the sector. the cost waivers by reducing operating cost of airlines will enhance the airlines probability of attaining break-even. in fact, in view of dipped consumable income, the reduced cost possibly be a relief for passengers in terms of affordable flying. air travelers rate assurance (singh, ) and financial conditions of airlines significantly affect the quality of air travel. product quality decreases when airlines are in financial distress (phillips and sertsios ) . given the deteriorating finances and demand crunch ahead, airlines service quality and safe operations may be compromised. further, the reasonable ticket cap as a safeguard to airlines as well as passengers' interest may be implemented. present paper attempts to analyze the vulnerability of airlines in india to withstand covid- after effects. lockdown of two months has been drastic for the fragile airlines business distressed with thin margins, liquidity crisis, over mounting fixed cost and debt. zero revenue, albeit spiraling fixed expenses has been a drain on the cash reserves of airlines dragging them towards insolvency. above all, the sector is viewing grim recession ahead. in this backdrop, the operation viability of airlines seems conditional on the recovery of variable expenses. sustainability of airlines warrants of turnaround changes in their revenue strategies and operating models. focus on minimizing losses rather than profit maximization possibly can help the airlines to combat current situation. impact of the adherence to the original low-cost model on the profitability of low-cost airlines predicting financial distress of companies: revisiting the z-score and zeta® models altman z-score the measurement of firm-specific indexes of technical change did lccs save airline revenue management? air transport and tourism: perspectives and challenges for destinations, airlines and governments operating leverage, profitability, and capital structure possible scenarios of covid impact on airlines commercials. source icao sustainability of airlines in india with covid- : challenges ahead and possible way-outs capacity-contingent pricing and competition in the airline industry diversity in airline alliance portfolio configuration a resource-based theory of strategic alliances customers' loyalty towards low cost airlines in india a study on analysis of operating efficiency of air india and indigo airlines the airline business. the airline business: routledge resource-based view of strategic alliance formation: strategic and social effects in entrepreneurial firms competition and profits in the airline industry: the case of airasia and malaysia airlines tying mechanism for airlines' air cargo capacity allocation operating leverage as a determinant of systematic risk leverage and corporate performance: international evidence estimation viability of dedicated freighter aircraft of combination carriers: a data envelopment and principal component analysis an efficiency study of airlines and air cargo/passenger divisions: a dea approach liberalisation of the airline industry in india iata economics chart of the week-social distancing would make most airlines financially unviable a dea study of airlines in india oliver's yard, city road london ec y sp are airline passengers ready for personalized dynamic pricing? a study of german consumers covid- ): implications for the mining industry effects of operational performance on financial performance an analysis of key factors of financial distress in airline companies in india using fuzzy ahp framework the impact of the degrees of operating and financial leverage on systematic risk of common stock operating leverage and operating risk flying with(out) a safety net: financial hedging in the airline industry impacts of the macro environment on airline performances in southern africa: management perspectives competing in an lcc world the routledge companion to air transport management transformation of india's domestic airlines: a case study of indian airlines, jet airways, air sahara and air deccan discussion paper: airport privatization in india survival lessons from a dying kingfisher how do firm financial conditions affect product quality and pricing? future of revenue management: airline pricing and revenue management: a future outlook assessing the air cargo business models of combination airlines kingfisher's acquisition of air deccan: altering india's lcc scenario? the case centre impact of operating cost components on airline efficiency in india: a dea approach drivers of operational efficiency and its impact on market performance in the indian airline industry on the impact of jet fuel cost on airlines' capacity choice: evidence from the u.s. domestic markets competitive service quality benchmarking in airline industry using demand model for air passenger traffic on international sectors airline service quality and economic factors: an ardl approach on us airlines tourism & hospitality: an initial assessment of economic impacts and operational challenges for the tourism & hospitality industry due to covid- an analysis on domestic airlines capacity performance in india key determinants of airline pricing and air travel demand in china and india: policy, ownership, and lcc competition airlines' futures key: cord- -fj j authors: morin, b. r.; kinzig, a. p.; levin, s. a.; perrings, c. a. title: economic incentives in the socially optimal management of infectious disease: when [formula: see text] is not enough date: - - journal: ecohealth doi: . /s - - - sha: doc_id: cord_uid: fj j does society benefit from encouraging or discouraging private infectious disease-risk mitigation? private individuals routinely mitigate infectious disease risks through the adoption of a range of precautions, from vaccination to changes in their contact with others. such precautions have epidemiological consequences. private disease-risk mitigation generally reduces both peak prevalence of symptomatic infection and the number of people who fall ill. at the same time, however, it can prolong an epidemic. a reduction in prevalence is socially beneficial. prolongation of an epidemic is not. we find that for a large class of infectious diseases, private risk mitigation is socially suboptimal—either too low or too high. the social optimum requires either more or less private mitigation. since private mitigation effort depends on the cost of mitigation and the cost of illness, interventions that change either of these costs may be used to alter mitigation decisions. we model the potential for instruments that affect the cost of illness to yield net social benefits. we find that where a disease is not very infectious or the duration of illness is short, it may be socially optimal to promote private mitigation effort by increasing the cost of illness. by contrast, where a disease is highly infectious or long lasting, it may be optimal to discourage private mitigation by reducing the cost of disease. society would prefer a shorter, more intense, epidemic to a longer, less intense epidemic. there is, however, a region in parameter space where the relationship is more complicated. for moderately infectious diseases with medium infectious periods, the social optimum depends on interactions between prevalence and duration. basic reproduction numbers are not sufficient to predict the social optimum. during an epidemic, there are several options available to private individuals interested in managing infectious disease risk. these include self-quarantine, vaccination, voluntary travel restrictions, contact reduction, prophylaxis, and preferential mixing (philipson ; geoffard and philipson ; philipson ; taylor and ampt ; brownstein et al. ; ferguson et al. ; germann et al. ; hunter ; rao et al. ; curtis et al. ; fenichel et al. ) . the choice of which measures to use frequently depends on costs-the cost of illness and the cost of risk mitigation. the cost of illness is determined by the likelihood that people will become infected if they make contact with an infectious individual, together with the cost of treatment and loss of earnings if they do become infected. the cost of risk mitigation is the resources sacrificed to reduce risk by some amount. diseases that are perceived to be highly infectious and life-threatening warrant greater private risk mitigation effort than diseases that are perceived to be only moderately infectious and with few symptoms (sanchez ; leroy et al. ; team ). diseases that are novel, and therefore uncertain in their effects, warrant greater private risk mitigation effort than diseases that are more familiar (taylor et al. ; presanis et al. ; rubin et al. ). it follows that private risk mitigation efforts might be expected to evolve over the course of an epidemic as information improves. a number of studies have, for example, tracked changes in the efforts made by private individuals to reduce disease risk during the a/h n epidemic bayham et al. ; springborn et al. ) . the problem addressed in this paper is that disease risk mitigation by one individual affects the well-being of both that individual and others. people vaccinating themselves against seasonal influenza, for example, reduce the likelihood that they will become infected, but also reduce the likelihood that they will become infectious. vaccination, in such cases, is said to be an impure public good. it confers benefits on the individual, but also on those with whom the individual interacts (sandler et al. ) . on the other hand, people avoiding infection now by sequestering themselves may simply delay the moment at which they become infected and infectious and hence may unwittingly prolong an epidemic . private diseaserisk mitigation, in this case, is an impure public bad. in providing benefits to the individual, it imposes costs on society. because people cannot capture any wider benefits they confer, and are not confronted by any wider costs they impose, they have little incentive to take those benefits and costs into account when making their decisions. the public good (bad) is therefore under-provided (over-provided). the general response of public health authorities to the under-provision of private disease risk mitigation is the adoption of social distancing measures that include quarantines, travel restrictions, and school closures. these are not, however, the only options available. since private disease risk mitigation efforts reflect the costs of illness and illness avoidance, interventions that change those costs are also among the options available to public health authorities . in this paper, we consider policy instruments that close the gap between the socially and privately optimal levels of disease risk mitigation. such policy instruments align the socially and privately efficient solutions by aligning the social and private cost of diseaserisk mitigation. we assume private disease-risk mitigation strategies similar to morin et al. ( ) and consider a control that operates on the cost of illness. this has parallels with the health belief model originally developed in the s, which supposes that private health-risk mitigation depends on beliefs about the likelihood and consequences of contracting disease, the efficacy and costs of the proposed action, and the decision makers' capacity to implement the action (rosenstock ) . we note that risk is used here in the economic sense. it is the product of the probability that an event will occur and the cost of the event if it does occur. the two most common approaches to the management of disease risk are mitigation and adaptation. mitigation implies an action that reduces the probability that the individual will fall ill. adaptation implies an action that has no effect on the probability that the individual will fall ill, but reduces the cost of illness. the probability that an infection will occur, cbp si (/(t)), depends on three things: contact volume c, the probability that contact results in infection b, and the probability that contacts are with infectious individuals p si (/(t)), a function of mitigation effort /(t). we treat the first two of these parametrically, although we acknowledge that they may be affected by, respectively, strategies to quarantine infectious individuals or to reduce activity levels , and strategies such as prophylaxis or vaccination (weycker et al. ) . we focus instead on the probability that contact will be made with infectious individuals. it was shown in fenichel et al. ( ) that despite travel restrictions during the h n swine flu outbreak, individuals did not cancel travel plans (reduce contact volume) so much as alter the timing of their travel. in previous work, we have shown the equivalence of private strategies targeting c or p si ) and have identified the social costs and benefits of private strategies that alter disease trajectories (morin et al. ) . in this paper, we study the impact of changes in the private cost of illness on mitigation effort, and show how diseases with the same r can trigger very different interventions, depending on which parameter drives the basic reproduction number. the paper is structured as follows. we first describe a model of preferential mixing based on observable disease states. we then detail two economic decision models, one private and one public, and show how they are coupled. since the models are not tractable enough to yield analyt-ical solutions, we solve both numerically in order to show the epidemiological effects of public taxes or subsidies on private cost of illness. this is followed by a discussion. our modeling approach builds on existing affinity-based mixing compartment models where compartments represent different disease states (busenberg and castillo-chavez ; blythe et al. ; castillo-chavez et al. ; fenichel et al. ; morin et al. ) . we suppose that individuals mix preferentially, conditional on their own disease state and the (observable) disease states of others. the only information available to individuals is the expression of symptoms in either themselves or others. as in , the resulting mixing strategy depends on the relative costs of illness and illness avoidance. this framework has been shown to provide the most mathematically general solution to the problem of who mixes with whom under the assumptions of symmetric contacts (blythe et al. ) . in what follows, individuals are defined only by their health state, although they could just as easily be grouped according to various shared attributes such as economic status, cultural or ethnic identity, geographical location, age, or disease awareness. we suppose that all individuals who do not show symptoms-because they have either not become ill, are asymptomatically ill or have recovered from illness-are treated equally by others. the use of the affinity framework allows three factors to determine the volume of contact between groups of like individuals: ( ) the size of each group, ( ) the nominal activity level or disease-free contact rate of each group and ) the relative affinity/disaffinity between groups. in what follows, we treat the affinity/disaffinity between groups as endogenous to the epidemiological system. susceptible individuals, or at least individuals who believe themselves to be susceptible, choose the people with whom they mix in order to alter the probability that they will encounter infectious individuals and subsequently become ill themselves. in the most general case, people who believe themselves to be susceptible at some time are taken to include all those who have been free of symptoms up to that time. this includes those who are actually susceptible, those who are asymptomatically infectious and those who are recovered but have never had symptoms. we hold the nominal level of activity (the contact rate) constant throughout the course of the epidemic and take it to be equal for all individuals. this makes it possible to consider only the effect of changes in mixing preferences. see fenichel et al. ( ) for a treatment that selects the volume of contacts, and for an analysis of the conditions under which choice of contact rates and avoidance effort are equivalent strategies. the main dif-ference between strategies is that while choice of contact rates allows complete isolation, choice of avoidance effort does not. to illustrate the approach, we first focus on a susceptible-infectious-recovered (sir) model-but note that we will be reporting results for a range of other models including susceptible-exposed or latent-infectious-recovered (seir), one-path and two-path susceptible-asymptomatical infectious-infectious-recovered (one-path sair, two-path sair). in the sir case, only susceptible individuals are free of symptoms. the disease dynamics are summarized in three differential equations: as is standard with the sir model, we let c be the nominal contact volume of all individuals. p si (/(t)) is the conditional probability that a contact made by a susceptible individual, committing /(t) effort to avoiding infection at time t, is with an infectious individual, and c is the rate at which an individual recovers and becomes immune. i t ð Þ and r t ð Þ are, respectively, the numbers of infectious and recovered (immune) individuals. the conditional probability that an individual in the ith disease state encounters an individual in the jth disease state is given by the elements of a time-dependent mixing matrix, p(t) = (p ij (t)), that is taken to satisfy three axioms economic incentives in the socially optimal management (busenberg and castillo-chavez ; blythe et al. ; castillo-chavez et al. ): . p ij ; for all i; j fs; e; a; i; rg . p j fs;e;a;i;rg p ij ¼ ; for all i fs; e; a; i; rg; . iðtÞp ij ¼ jðtÞp ji ; for all i; j s; e; a; i; r f g : these three axioms imply that, collectively, susceptible individuals have the same expectation of encountering infected individuals as infected individuals have of encountering susceptible individuals. it has been shown that the unique solution to these mixing axioms is given by where mðtÞ ¼ sðtÞm s ðtÞ þ iðtÞm i ðtÞ þ rðtÞm r ðtÞ; and uðtÞ ¼ ð/ ij ðtÞÞ is a symmetric affinity matrix, in this case . the main difference between the approach here and previous use of affinity models is that we treat the elements / ij (t) of the affinity matrix are private choice variables. they may be interpreted as the effort that the representative individual in disease state i makes to avoid individuals in disease state j, if / ij (t) < , or to associate with individuals in disease state j, if / ij (t) > . if the representative individual in every disease state i makes no effort to avoid individuals in disease state j, and vice versa, then / ij (t) = . we then have classic proportionate mixing. we take zero elements in the affinity matrix to be evidence of 'avoidance-neutrality.' that is, they show the representative individual to be neutral about a pairing event with someone from another disease class. by contrast, negative (positive) elements reflect the desire of the representative individual in one disease state to avoid (seek out) individuals in other disease state. avoidance, / ij (t) < , can result from individuals in both states wishing to avoid one another; individuals in one state wishing to avoid individuals in other states who may be neutral to the pairing; or individuals in one state wishing to avoid individuals in other states more strongly than those individuals favor the pairing. similarly, engagement results from individuals in both states favoring the pairing; individuals in one state seeking out individuals in other states who may be neutral to the pairing; or individuals in one state wishing to engage with individuals in other states more strongly than those individuals wish to avoid the pairing. this is a similar measure to that used in models of assortative mating (karlin ) and selective mixing (hyman and li ) and is a form of a contact kernel (gurarie and ovaskainen ) . the elements of the affinity matrix, / ij (t), describe what the representative individual in each health state wants. what they actually get depends both on the preferences of others in the population and on the relative size of all health classes. more particularly, the elements of the mixing matrix p = (p ij ) depend both on the proportion of the population in each disease state and on the affinity matrix. they describe the conditional probabilities that an individual of disease state i contacts someone in disease state j. in what follows, we focus on individuals who believe themselves to be susceptible (who have been symptom free up to that point) and assume that they maximize the net present value of the contacts they make, taking into account the cost of illness and illness avoidance, by choosing the effort to commit to preferential mixing: the elements of u t ð Þ. formally, the decision problem for individuals who believe themselves to be susceptible, collectively labeled x, is to choose the level of mitigation effort, / xi (t), in order to maximize the difference between the benefit of not being symptomatic, b, and the cost of mitigation effort, c(/ xi (t)), given the weight they place on future well-being (the discount rate d) and their planning horizon, t. if susceptible individuals are averse to mixing with symp-tomatic (infectious or otherwise) individuals in the sir model, and if all others are neutral, u t ð Þ has the structure: in which represents neutrality and / si (t) < represents the effort susceptible individuals make to avoid mixing with infectious individuals at time t ). this defines we may then write the mixing matrix of conditional probabilities as each figure shows the prevalence with public health authority intervention, i pha (t), and with only the private mitigation effort, i priv (t), for various b. the present value difference (pvd) in these curves is shown along with the intervention level of the public health authority. positive pvd, corresponding to an early reduction in incidence that is not offset by the future ''fatter'' tail, is analogous to a tax on the cost of illness. negative pvd was always found to include an early increase infection and is analogous to a subsidy on the cost of illness. the magnitude of intervention is directly related to the magnitude of the pvd. as shown in morin et al. ( ) , we note that mitigation effort is restricted to the range: with proportionate mixing resulting from applying effort. p si (t) = i(t) is the maximum probability of contact and private quarantine of infectious individuals at the left endpoint, p si (t) = is the minimum probability of contact. from an economic perspective, private quarantine implies that the marginal cost of illness is greater than the marginal cost of illness avoidance for all levels of illness avoidance effort. we assume that a forward-looking representative individual, who believes himself/herself to be susceptible, seeks to mitigate disease risks by avoiding those who are obviously (symptomatically) infectious. we suppose that individuals may belong to any one of the following epidemiological states at a given time: (s)usceptible to the disease, (e)xposed or latently infected being asymptomatic and noninfectious, (a)symptomatically infectious, (i)nfectious with symptoms, or (r)ecovered and immune to the disease. individuals who believe themselves to be susceptible may include those in states s, e, and a. we further assume the motivation for selective mixing is the desire to avoid the costs of illness only. we do not allow individuals to behave altruistically. we also assume that individuals who know themselves to be infected have no incentive to avoid others. only susceptible people (or those who believe themselves to be susceptible) react to disease risk. this includes all individuals in health classes x ¼ s; e; a; or r a (recovered from asymptomatic infection). because all individuals who react to disease risk consider themselves to be equally susceptible to the disease, their mixing decisions are both identical and symmetric [i.e., all / xi (t) = / ix (t) are equal to one another and all other entries in u t ð Þ are ]. formally, these individuals screen contacts by choosing the elements / ij (t) of the matrix uðtÞ, i, j [ {s, e, a, i, r a } so as to maximize the difference between the expected benefits of contact and the expected cost of illness and illness avoidance, given their current health state. the benefits of contact range from the satisfaction to be had from purely social engagement to the financial gains to be had from market transactions with others. for simplicity, we assume that the benefits of a contact are financial gains and that they are the same for individuals in all health states, b i ¼ b; i s; e; a; i; r f g . the cost of illness generally includes both forgone earnings, lost wages, and the cost of healthcare. for simplicity, we take the cost of illness to be the cost of treatment. the cost of illness avoidance is simply the cost of the effort made to avoid contact with people who are ill: the cost of choosing / ii t ð Þ; i s; e; a; r f g . the net benefits of contacts with others by an individual in the ith health state at time t thus comprise the difference between the benefit of contacts made in that health state, b i , and the cost of disease and disease risk mitigation, table ). all individuals within a particular disease class are assumed to behave in the same way. we define u i ð/ ij ðtÞÞ ¼ b i À c i ð/ ij ðtÞ; iðtÞÞ to be the net benefits of contact to the representative individual in health state i at time t. we define e v j t þ s; / ij t þ s ð Þ to be the individual's expected value function from time t + s forward, where the probability that the individual will be in health state j in the future depends on the disease recovered from symptomatic infection, immune no b r susceptible, exposed, asymptomatically infectious, and recovered (from asymptomatic infection) individuals all choose to mitigate risk as if they were susceptible and thus carry the cost of mitigation. dynamics and their mixing strategy while in health state i. the decision problem for the representative susceptible individual in health state i may be expressed via the hamilton-jacobi-bellman equation: where s is a short interval of time. this is subject to the dynamics of the disease: eq. ( ), table the h-j-b equation identifies the problem solved by the representative individual in state i: to maximize the expected net value of current and future contacts by choosing the extent to which they mix with individuals in other health classes. following fenichel et al. ( ) , we assume that individuals form their expectations adaptively. the value function v i (t, h(t)) is defined recursively as the sum of the current net benefits of contact in health state i given the information available at time t plus the discounted stream of expected net benefits over the remaining horizon. this expectation is conditional on the effects of disease risk mitigation decisions on the probability of transitioning between health states. more particularly, we assume individuals observe the state of an epidemic at time t and make a forecast for the epidemiological trajectory over the interval s. their mixing strategies are then adapted over time as they make new observations on the state of the epidemic. we assume that individuals make the simplest forecast-that all disease classes are constant over the interval s, but that they adapt to new data as it emerges. the representative individual will increase effort to avoid infection up to the point where the marginal cost of illness avoidance is just offset by the marginal benefits it yields-the avoided cost of illness. efforts to avoid infection will be increasing in the cost of illness and decreasing in the cost of illness avoidance. in models without risk mitigation, disease dynamics may be completely characterized from initial conditions. with risk mitigation, the evolution of the epidemic reflects feedback between the cost of disease and disease avoidance on the one hand, and averting behavior on the other [see (fenichel and horan ; horan et al. ) for further discussion]. to solve the problem, we take a discrete time counterpart to eq. ( ) and solve numerically using techniques similar to those in fenichel et al. ( ) . specifically, we solve the adaptive expectation problem using a method we call cast-recast. at each time, the individual solves the h-j-b equation using backwards induction from their time horizon ( days) to the present, while supposing that there is no change to the state variables over the time horizon. having determined the optimal mitigation effort, / ij (t), the individual commits that effort until the next time step (day). that is, their mitigation effort is held constant for time step. the ordinary differential equations describing the disease dynamics are advanced, and the process is repeated. note that the private forecast used has little effect on the optimal outcomes so long as the forecast period, the time interval s, is short relative to the disease dynamics. if the epidemic evolves rapidly relative to the period over which the individual commits to a fixed level of risk mitigation, then the assumption that the disease states are constant may induce errors. in previous work using this method morin et al. morin et al. , morin et al. , , we found a smooth response-the decisions made by individuals were much the same from day to day. since the epidemic evolves on a timescale of weeks, this gives us confidence in the cast-recast method for the private problem. in the discussion, we outline when the results from such a method may favor more rapid transmission of epidemiological state variables, and when individual behavior choices need to be more ''agile'' in order to match timescales with the disease spread. the choice of / ij (t) maximizes the private net benefits of contact for the individual over the course of an epidemic, given the private cost of illness and illness avoidance. however, as was shown in morin et al. ( ) this may well be publically suboptimal, depending on the social rate of discount or the social cost of illness. the social cost of illness is the sum of the costs borne by all infected and symptomatic individuals, together with the cost of disease avoidance by all others. in a real system, it would also include the infection risk borne by healthcare workers, but we do not address that here. of the many intervention options open to public health authorities-quarantine, social distancing measures such as school closures, vaccination campaigns and so on-we focus on instruments that change risk mitigation by changing the private cost of illness. in doing this, we follow a literature that integrates epidemiology and economics to explore ways in which economic behavior affects disease spread . this literature has concentrated on the economic causes and epidemiological consequences of peoples' contact decisions hammer , ; barrett and hoel ; funk et al. ; fenichel et al. ; funk et al. ; springborn et al. ). by treating the economic factors behind contact and mixing decisions as central elements in disease transmission, the approach opens up a new set of disease management options. in what follows, we suppose that the public health authority is able to use an economic policy instrument to alter the private cost of illness. we consider the instrument, r d , d sir, seir, one path sair, two path sair f g see tables and , which may be interpreted as a disease-specific tax or a subsidy on the cost of illness, c i i; r d ð Þ := c i þ r d ð Þ. if r d > (illness is 'taxed'), the private cost of illness is increased. enforced, uncompensated sick 'leave' would be an example of this. we expect this to increase disease-risk mitigation effort and hence the illness avoidance costs carried by individuals. if r d < (illness is 'subsidized'), the private cost of illness is reduced. subsidized health insurance schemes would be an example of this. we expect this to reduce disease-risk mitigation effort, and with it the illness avoidance costs carried by reactive individuals. by changing the privately optimal level of disease-risk mitigation, it is possible to change overall disease dynamics. there are many ways in which interventions change the private cost of illness in real-world conditions, ranging from direct subsidies or taxes on drugs and treatment, through health insurance costs and coverage, to statutory obligations on sick leave. we suppose that r d can be applied in a way that proportionately reduces or increases the relative private cost of illness. this would be consistent with, for example, mandatory insurance cover for a specific proportion of potentially forgone earnings. our baseline case assumes that the policy instrument is revenue neutral. if r d < (illness is subsidized), the cost is met by a levy on all income from contacts. if r d > (illness is taxed), the revenue is returned as a tax benefit on all income from contacts. since taxes and subsidies both potentially impose an efficiency cost in the form of a deadweight loss of consumer and producer surplus, we include a proxy for this in the optimization problem. more particularly, we include a term that specifies any deadweight loss as a proportion of the cost of taxes or subsidies. the public health authority's problem for an sir disease thus takes the form: subject to the disease dynamics described by the relevant compartmental epidemiological model and to the private decision problem described in eq. ( ). that is, the public health authority selects r d so as to maximize the net benefits of risky contacts to society-where society is the sum of all individuals in all health classes. the final term in the public health authority's problem is our proxy for the deadweight loss associated with taxes or subsidies on the cost of illness. a [ [ , ] is the proportion of the cost of the intervention that is recovered. to solve these two problems, we maximize the integral in eq. ( ) over the entire epidemic by solving the complete private problem for each ''guess'' of r d . this was implemented using matlab's fminbnd function. we investigated the effect of a subsidy/tax on the cost of illness on private disease risk mitigation in the key epidemiological classes across four compartmental models: sir, seir, and two sair models, a one-path progression and a two-path progression. within the one-path model, susceptible individuals are first asymptomatically infectious and then progress to symptomatically infectious and then immune. in the two-path model, a susceptible individual becomes either asymptomatically or symptomatically infectious and then recovers. individuals who recover from asymptomatic infection (r a ) are expected to behave as if they are susceptible. individuals who recover from symptomatic infection (r i ) are not. because we have assumed no heterogeneity aside from health status, the use of a disease-specific instrument, r d , has the potential to be efficient. if the population were heterogeneous, however, we would expect to need more targeted interventions. we did not consider models with reentry to the susceptible class, e.g., sis, sirs, and other cyclical models. this is for two reasons. first, each of these models is capable of endemic levels of infection. this, combined with the fact that people may experience reinfection, would require individuals to form expectations (possess memory) with respect to the impact of different avoidance strategies. second, numerical simulations of these models reveal very broad oscillations that confound comparison with single outbreak models. there are no entries (births) or removals (deaths) from the system (see table for example diseases listed for each model). each model considered here is therefore a so-called single outbreak model; the population is not only kept at a fixed number, but it is also closed to the introduction of new individuals. we modeled the dynamics of the epidemic types in table using ordinary differential equations (see table ). this has two main implications for disease dynamics: (a) once nonzero, the state variables will never again be zero in finite time, and (b) in an infinitely small amount of time ''mass'' will move into each compartment as long as the transition rates are nonzero. these affect the interpretation to be given of the point at which an epidemic is ''over.'' it is feasible that extreme risk mitigation early in the course of the epidemic could wipe out the infection within a population. however, within the differential equation framework, as soon as mitigation weakens, coupled with the fact that there is a nonzero infectious population with potentially a very large susceptible population, the infection will again spread, potentially causing additional peaks. the instantaneous transfer of individuals from one compartment to another also serves to induce reaction timing that may not conform to data. in recognition of this, we supposed that the differential equations represent an expectation of outcomes over a population divided between three health classes. in previous work (morin et al. , , we demonstrated that private disease-risk mitigation reduces peak symptomatic infection levels and the total number of people who experience symptomatic infection while prolonging the epidemic. we seek to understand how public health interventions aimed at minimizing the cost of disease and disease avoidance affect individual risk mitigation decisions. our measure of mitigation effort is a relative one. it is the proportion of maximum mitigation effort undertaken, where maximum effort is defined as that which results in p si (t) = . this enables us to compare effort across diseases. the background against which public health authority intervention occurs is that private mitigation effort is increasing in infectiousness (b), infectious duration ( = c ), and the share of the public cost of intervention recovered by the public health authority (a). in other words, private mitigation effort is increasing in the cost of illness. we find that in all cases peak effort occurs early in the epidemic-within the first days for our parameters-and that action is taken sooner, the greater the severity of the disease. our results on the socially optimal response this induces for different diseases follow. these results are driven by the assumptions we make about the cost of illness (equal to the benefit gained from being healthy) relative to the cost of mitigation. specifically, we assume that the cost of mitigation is low relative to the cost of illness. to explore the sensitivity of the optimal intervention strategy of the public health authority to the severity of disease, we varied b and c over intervals so that b=c : ; : ½ ; and a in steps of . from to , with baseline values of a = , b ¼ : , and c = / . while the socially optimal public health authority intervention decision can be viewed as a function of the disease's basic reproduction number r = b/c, the two components of r (infectiousness and duration) have rather different effects on the public health authority intervention. nor is the optimal social response monotonic in either case. r (b) tests the sensitivity of health interventions to the impact of infectiousness on r and tests the sensitivity of health interventions to the impact of disease duration on r . for r (b) [ [ . , . ] , the optimal public health authority intervention involves an increase in the private cost of illness-a 'tax' on illness that will stimulate higher levels of private disease risk mitigation. as infectiousness falls, the optimal public health authority incentive to mitigate first rises and then falls, the turning point being determined by parameters describing both the cost of illness and the cost of illness avoidance. the optimal tax in this case reaches a maximum of % for r (b) = . . for diseases where infectiousness is either very low (r (b) < . ) or very high (i.e., r (b) > . ), the optimal public health authority intervention involves a reduction in the cost of illness-a 'subsidy' on illness that lowers private diseaserisk mitigation effort (fig. ) . the duration of illness has a slightly different effect. note that small r (c) values are indicative of diseases of short duration. for r (c) = . , for instance, the illness is symptomatic for less than days. we found that for diseases of moderate to short duration, r (c) < . , it would generally be optimal for the public health authority to stimulate an increase in private risk mitigation by raising the private cost of illness-by taxing illness. on the other hand, for diseases of longer duration, r (c) > . , we found that it would always be optimal for the public health authority to reduce private risk mitigation by lowering the private cost of illness. given the assumed cost of illness and mitigation effort, private individuals will overreact to the risk of diseases of long duration and high infectiousness and will underreact to diseases of short duration and low infectiousness. to see the trade-off between infectiousness and duration, we considered the range of (b, c) that generates r [ . , . ]. figures and show the parameter combinations that leave the public health authority indifferent between intervening or not over this range of values, or that favor intervention to increase or decrease privately optimal disease risk mitigation. while infectiousness and duration trade off against each other, the relation is not linear. where private individuals take excessive precautions, pha interventions will generally discourage private risk mitigation for highly infectious diseases of long duration. this changes, however, if the infectiousness of diseases of long duration is very low or the duration of highly infectious diseases is very short. symmetrically, pha interventions will generally encourage private risk mitigation for less infectious diseases of short duration. once again, though, this changes if the infectiousness of diseases of short duration is very high or the duration of less infectious diseases is very long. the socially optimal intervention will also dependintuitively-on the degree to which the susceptible and infected individuals carry the cost of public health authority interventions. to capture this, we included a cost recovery parameter, a, and explored the sensitivity of the optimal intervention strategy of the pha to variation in a (fig. ) . the proportion of the cost of a subsidy, on the cost of illness, recovered through taxation (or the proportion of tax revenues returned to the wider population) is given by a. specifically, we varied a in steps of . from to with a baseline value of a = . for the baseline values for infectiousness and duration, b ¼ : and c = / , we found that the greater the value of a, the less one needs to subsidize the cost of illness in order to align private and social optima. the reason is that the cost of illness is affected both by the instrument itself and by cost recovery. as the rate of cost recovery increases, the cost of illness rises along with the privately optimal level of risk mitigation. indeed, for a ! . the public health authority switches from discouraging to encouraging private risk mitigationfrom a subsidy to a tax. the inclusion of a class of individuals who have unknowingly contracted the illness but are not a danger to others (they are asymptomatic and noninfectious) changes the economics of the problem in important ways. on the one hand, it unnecessarily increases the aggregate cost of mitigation since exposed individuals continue to mitigate, even though they there is no need. on the other hand, by driving up p se and reducing p si , exposed individuals reduce infection rates below those that would occur if only susceptible individuals mitigated risk. that is, the risk mitigation undertaken by exposed individuals confers an external benefit on society. the result is that public health authority intervention in the seir case increases private risk mitigation for all but the most severe diseases (i.e., for all r . ) (fig. ) . we also note that the basic reproduction number for the seir model is identical to that for the sir model and does not involve the latent period. over reasonable latent periods from to days, we found an increase in the 'tax' rate applied by the public health authority, but no switching or nonlinearities in behavioral response. the big differences between the sir and seir cases are that the introduction of a latency period slows the spread of disease, while the mitigation by exposed individuals reduces the rate of new infections. interestingly, when we decomposed the marginal benefit of public health authority intervention by the infectiousness and duration of disease, we found the marginal benefit of intervention to be increasing in the duration of disease, but decreasing in infectiousness (fig. ) . as b increases and causes r (b) to pass . , we found the marginal benefits of public health authority intervention to be monotonically decreasing. this is because @ / si @r d @b < : however, as duration of disease increases, causing r c ð Þ to increase, we found the marginal benefits of pha intervention to rise monotonically after approximately . . private disease-risk mitigation in the sir case reduces peak prevalence but prolongs the epidemic at higher levels than would have occurred in the absence of mitigation. in the seir case, by contrast, private disease risk mitigation reduces peak prevalence and prolongs the epidemic, but at lower levels than would have occurred in the absence of mitigation. moreover, the longer the duration of the disease the greater the social net benefits this offers. the one-path sair model introduces an asymptomatically infectious stage between susceptible and symptomatically infectious. these individuals now undertake risk mitigation because they do not know they're infected. unlike the seir case, however, this is potentially harmful to the population. while @p si @/ si < , it is also true that @p sa @/ si [ . we may assume that the two classes have different levels of infectiousness. the net effect of private disease risk mitigation is therefore given by the combined impact of two things: the marginal reduction in the infection rate due to the avoidance of infected people and the marginal increase in infections due to susceptible-asymptomatic contact. the marginal effect of risk mitigation through mixing is @p si @/ si b i and the marginal effect of susceptible-asymptomatic contact is @p sa private risk mitigation reduces prevalence regardless of the actions of asymptomatically infec- it has the opposite effect. whether the public health authority seeks to increase or decrease private disease risk mitigation effort depends on the net effect. given our baseline parameters, if b a < À : b i þ : then the public health authority will seek to increase private mitigation, and if b a [ À : b i þ : , it will seek to reduce private mitigation. the strength of intervention increases with the distance of (b i , b a ) from the line b a ¼ À : b i þ : (which is roughly analogous with the condition that the basic reproduction number is ). in the two-path sair model, asymptomatically infectious individuals can spread disease due to their (redundant) risk mitigation actions. as in the seir model, those who've recovered from asymptomatic infection undertake mitigation that does not benefit them, but may offer a benefit to susceptible individuals by reducing their chance of making infectious contacts. in some sense, the r a class becomes like a vaccinated class of potentially highly connected individuals. they therefore dilute the contact pool for susceptible individuals. we find the interesting effect that whether to tax or subsidize is sensitive only to the recruitment into the i class. if the process favors the generation of symptomatic infection, then the pha intervenes with a tax on illness. if the process favors the generation of asymptomatic infection, then the pha intervenes with a subsidy. a process favorably generates asymptomatic infection when b a p sa p aa þ b i p si À p ii ð Þ[b a p sa À p aa ð Þ þ b i p si p ii where p xx is the probability that an infected person of type-x creates another individual of type-x when they spread infection. therefore, adjustments will be made to the cost of where @/ si @r d cancels out from both sides. if p aa ; p ii < , then the marginal condition is never met and no mitigation will occur. likewise, if both p aa ; p ii [ then mitigation will occur indefinitely until p si = . if one is less than / with the other greater, then the level of mitigation is unclear because it now depends on the relative strengths of infection and the sensitivities of pair probabilities to mitigation. when private disease-risk mitigation has general epidemiological consequences, the optimal level of mitigation should be determined by reference to the costs and benefits to society at large. because disease-risk mitigation changes the characteristics of an epidemic, it changes the social costs and benefits of disease. if private mitigation reduces prevalence, for example, it confers benefits on society. if it increases the duration of an epidemic, on the other hand, it imposes costs. since the private and social calculus of the costs and benefits of mitigation are different-individuals calculate only the costs and benefits to themselves, society calculates the costs and benefits to all-we expect the privately and socially optimal level of disease risk mitigation to differ. we have argued that individuals base their mitigation decisions on the expected net benefits of particular actions. very similar arguments have long been made in the medical literature. the health belief model, for example, was developed to explain why people undertook (or abstained from) private disease-risk mitigation (champion and skinner ) . the model, originally developed in the s, assumed that individuals engaged in health-risk mitigation to reduce the threat of illness based on their beliefs about four things: the likelihood of contracting disease; the consequences of falling ill (symptoms, loss of work wages, loss of personal interactions); the efficacy and costs of the proposed action; and their capacity to undertake the action adequately. it also assumed that individuals responded to incentives-'cues to action' that signaled susceptibility, severity, costs and the like. the perceived benefits and costs of alternative behaviors determined which behaviors would be undertaken (rosenstock ) . in the context of this paper, the perceived susceptibility of an individual at time t , conditional on a mitigation effort /(t), is given by r t t exp Àbp si / s ð Þ ð Þ ð Þ ds, and the expected cost of illness is the daily forgone income (adjusted by public health authority intervention) multiplied by the expected duration of illness, c ill ( + r d )/c. while we only allow a single risk mitigation choice, affinity-based mitigation, we allow it to vary from ''doing nothing'' (/(t) = ) where x(t) is the total population of susceptible individuals). as in the framework of the health belief model, we expect that mitigation effort applied will be increased up to the point at which its cost is just offset by the benefits, in terms of the reduced probability of illness. applications of the health belief model to disease-risk mitigation in the swine flu outbreak found that the framework adequately segmented the respondent populations into those who vaccinated and those who did not (janz and becker ) . they found that perceived susceptibility, benefits, and barriers were strongly correlated with vaccine-seeking behavior and that these results were strengthened by the fact that, even if it was unsuccessful, the vaccine alleviated the symptoms of illness (aho ; cummings et al. ; larson et al. ; rundall and wheeler ; janz and becker ) . the central proposition of this paper is that since disease-risk mitigation is a function of the private cost of disease, it can be managed through changes in the private cost of disease. whether private disease-risk mitigation is above or below the socially optimal level of mitigation, there exist interventions that have the potential to align private choices with the interests of society. our numerical results on the difference between the privately and socially optimal mitigation reflect the specific assumptions made about the relative cost of disease and disease avoidance. these were selected to illustrate the range of potential responses and the sensitivity of responses to disease characteristics. given our cost assumptions, we find that for sir diseases individuals confronted with highly infectious illnesses of long duration overreact to the associated risks. they mitigate more than is socially optimal. conversely, individuals confronted with short illnesses of low infectiousness underreact. for seir diseases, on the other hand, the privately optimal level of disease-risk mitigation is less than the socially optimal level for almost all parameter values. the gap between privately and socially optimal investments in risk mitigation is partly due to the fact that private individuals base their risk mitigation decisions on the observed health state of others-whether or not others are symptom-free. this is a source of error that is greater in some cases (seir, two-path sair) than others (sir, onepath sair). it is also due to the epidemiological effects of actions that influence when a susceptible person becomes infected in the course of an epidemic. one of the main external effects of private disease-risk mitigation is the prolongation of epidemics and with it the cost of diseaserisk mitigation. we therefore considered interventions that ran in both directions: subsidies on the cost of illness (e.g., health insurance) that discourage mitigation and taxes (e.g., uncompensated mandatory days off or other such penalties) that encourage mitigation. it can, by the way, be shown that subsidies or taxes on the cost of mitigation rather than the cost of illness have similar effects. for any adjustment to the cost of illness, there exists an adjustment to the cost of mitigation that produces the same private response. it does not follow, however, that the two instruments are equally efficient. adjusting the cost of illness alters 'severity' in the language of the health belief model while adjusting the cost of mitigation alters 'barriers.' which is the more cost-effective in any real application would depend on the effort needed to achieve the desired epidemiological effect and the relative cost of effort in each case. within the sir model, each individual has perfect information on their infection state and all relevant information on others (infectious or not). the other three models each introduce error due to the existence of asymptomatic classes. they also introduce changes in disease dynamics that affect the socially optimal level of risk mitigation. whether it is socially optimal to intervene in ways that increase or decrease private disease risk mitigation depends on the effect of mitigation on the time profile of infections. the point has already been made that mitigation reduces peak prevalence but also prolongs epidemics. when the additional cost of the longer (and potentially fatter) tail of an epidemic outweighs the benefits offered by a reduction in peak prevalence, the pha will intervene to reduce private risk mitigation-and vice versa (as shown in fig. ) .the advantages of instruments that operate on the cost of illness are that they can be put in place in advance of an outbreak and that they operate automatically. diseasespecific insurance or treatment costs, for example, can be established in advance but only influence private decisions when an outbreak occurs. we have not considered particular instruments in this paper and note only that the effect of each depends on how it would impact the relative costs and benefits of mitigation actions. there is a range of cost-based instruments that could be applied to diseases whose properties are well understood. the instruments that should be applied to specific diseases are those expected to cost effectively close the gap between socially and privately optimal disease risk mitigation. while we consider only policy instruments determined in advance of an epidemic-i.e., that correspond to the known properties of known diseases-we acknowledge that this would not be reasonable in the case of novel or emerging infectious diseases. there is less scope for the use of incentives of this type in such cases, but it would be worth considering how the choice of r d might evolve during the course of an epidemic in future work. finally, it is worth noting that any intervention of this kind in a real-epidemiological-economic environmental would have distributional consequences. it would make some people better off and some people worse off. we have abstracted from any distributional consequences in this paper by treating all individuals in each disease class as homogeneous, and by assuming that the payoff to contact is the same for everybody. implementation of a policy instrument of this kind in a real system would, however, need to take account of the distributional goals of the society concerned. infectious diseases are a comparatively small part of the burden of disease in highincome countries, but still the largest part of the disease burden in low-income countries. this is precisely where distributional issues are of greatest concern and where individuals are least able to bear the cost of disease-risk avoidance. participation of senior citizens in the swine flu inoculation program: an analysis of health belief model variables in preventive health behavior measured voluntary avoidance behaviour during the a/h n epidemic toward a unified theory of sexual mixing and pair formation empirical evidence for the effect of airline travel on inter-regional influenza spread in the united states interaction, pair formation and force of infection terms in sexually transmitted diseases. mathematical and statistical approaches to aids epidemiology pair formation in structured populations. differential equations with applications in biology the health belief model psychosocial determinants of immunization behavior in a swine influenza campaign disgust as an adaptive system for disease avoidance behaviour jointly-determined ecological thresholds and economics trade-offs in wildlife disease management adaptive human behavior in epidemiological models management of infectious wildlife diseases: bridging conventional and bioeconomic approaches skip the trip: air travelers' behavioral responses to pandemic influenza strategies for mitigating an influenza pandemic modelling the influence of human behaviour on the spread of infectious diseases: a review the spread of awareness and its impact on epidemic outbreaks disease eradication: private versus public vaccination mitigation strategies for pandemic influenza in the united states the economical control of infectious diseases infectious diseases, public policy, and the marriage of economics and epidemiology towards a general formalization of encounter rates in ecology public-private'health law: multiple directions in public health behavior changes in sis std models with selective mixing the health belief model: a decade later models of multifactorial inheritence: i, multivariate formualtions and basic convergence results the relationship of health beliefs and a postcard reminder to influenza vaccination fruit bats as reservoirs of ebola virus sir dynamics with economically driven contact rates. natural resource modeling disease risk mitigation: the equivalence of two selective mixing strategies on aggregate contact patterns and resulting epidemic spread the social benefits of private infectious disease-risk mitigation merging economics and epidemiology to improve the prediction and management of infectious disease private vaccination and public health: an empirical examination for us measles economic epidemiology and infectious diseases the severity of pandemic h n influenza in the united states adherence to self-quarantine recommendations during an outbreak of norovirus infection the health belief model and preventive health behavior public perceptions, anxiety, and behaviour change in relation to the swine flu outbreak: cross sectional telephone survey factors associated with utilization of the swine flu vaccination program among senior citizens in tompkins county ebola viruses a conceptual framework for understanding global and transnational public goods for health optimal random exploration for trade-related non-indigenous accounting for behavioral responses during a flu epidemic using home television viewing risk factors for human disease emergence travelling smarter down under: policies for voluntary travel behaviour change in australia ebola virus disease in west africa-the first months of the epidemic and forward projections population-wide benefits of routine vaccination of children against influenza this study was made possible by grant # r gm - from the national institute of general medical sciences (nigms) at the national institutes of health and contract hshqdc- -c- from the science and technology directorate, department of homeland security. the contents of this manuscript are solely the responsibility of the authors and do not necessarily represent the official views of dhs or nigms. the funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. it was also funded by nsf grant as part of the joint nsf-nih-usda ecology and evolution of infectious diseases program. economic incentives in the socially optimal management key: cord- -ed gawoj authors: barron, sarah p.; kennedy, marcus p. title: single-use (disposable) flexible bronchoscopes: the future of bronchoscopy? date: - - journal: adv ther doi: . /s - - - sha: doc_id: cord_uid: ed gawoj the coronavirus disease (covid- ) pandemic has highlighted the importance of reducing occupational exposure to severe acute respiratory syndrome coronavirus (sars-cov- ). the reprocessing procedure for reusable flexible bronchoscopes (rfbs) involves multiple episodes of handling of equipment that has been used during an aerosol-generating procedure and thus is a potential source of transmission. single-use flexible bronchoscopes (sufbs) eliminate this source. additionally, rfbs pose a risk of nosocomial infection transmission between patients with the identification of human proteins, deoxyribonucleic acid (dna) and pathogenic organisms on fully reprocessed bronchoscopes despite full adherence to the guidelines. bronchoscopy units have been hugely impacted by the pandemic with restructuring of pre- and post-operative areas, altered patient protocols and the reassessment of air exchange and cleaning procedures. sufbs can be incorporated into these protocols as a means of improving occupational safety. most studies on the efficacy of sufbs have occurred in an anaesthetic setting so it remains to be seen whether they will perform to an acceptable standard in complex respiratory procedures such as transbronchial biopsies and cryotherapy. here, we outline their potential uses in a respiratory setting, both during and after the current pandemic. the development of reusable flexible bronchoscopes (rfbs) in was a ground-breaking development in diagnostic and therapeutic bronchoscopy. common indications include diagnostic washings, endobronchial biopsy and brushings and transbronchial needle aspiration (tbna). many therapeutic procedures are now possible with both rigid and flexible bronchoscopy including foreign body removal and tumour debulking while recent advances include asthma and chronic obstructive pulmonary disease (copd) therapies. by comparison, in the intensive care unit (icu) their uses include the confirmation of endotracheal tube positioning in difficult airways as well as diagnostic sampling. bronchoscopy poses challenges from the perspective of infection prevention with a risk of transmission to both the patient and the personnel involved [ ] [ ] [ ] [ ] [ ] . patient infections can arise exogenously as a result of contaminated equipment [ ] and, while the majority of outbreaks of pseudo and actual infection have been linked to breaches in bronchoscope reprocessing guidelines, a recent study demonstrated that even with complete adherence to protocol, contamination and microbial growth persisted on fully reprocessed rfbs [ ] . the main risk for the personnel involved is the transmission of acute respiratory infection (ari) via aerosols generated during the procedure [ , ] . currently, the risk of transmission of severe acute respiratory syndrome coronavirus (sars-cov- ) to both patients and healthcare personnel (hcp) is of huge concern [ ] and there is evidence of transmission of the virus in healthcare settings [ ] . bronchoscopy should be avoided in people with confirmed or suspected coronavirus disease [ , , ] ; however, if essential, several organisations recommend avoiding rfbs to reduce the risk of viral transmission [ ] . until now, disposable or single-use flexible bronchoscopes (sufbs) have primarily been used by anaesthetists in an icu or peri-operative setting where they perform to an acceptable level in comparison to rfbs [ , ] combined with the distinct advantage of a reduced risk of infection owing to their sterility [ ] . several studies have assessed their cost compared to rfbs [ , ] and a recent review that incorporated the cost of treating the exogenous infections that might be caused by rfbs found that sufbs were significantly more cost-effective [ ] . in this review, the risk of infection with standard rfbs will be outlined as will the advantages of sufbs, with comment on their cost profile compared to rfbs and attempt to suggest a rationale for their use during the covid- pandemic and in a respiratory setting. this article is based on previously conducted studies and does not contain any studies with human participants or animals performed by any of the authors. standard review article methodology was used. search terms including 'reusable bronchoscope', 'single-use bronchoscope', 'disposable bronchoscope and covid- pandemic' were placed in pubmed, google and embase search engines and the resulting english language papers that were available were read. additionally, the references of all these papers were read and any citations deemed appropriate were also sought and read and their references were reviewed. bronchoscopy is a semi-critical procedure (spaulding classification)-there is a moderate risk of infection as the bronchoscope is in contact with mucous membranes but does not enter sterile tissues or the vasculature. devices in this category warrant high level disinfection (hld) [ ] . when a bronchoscope is used for a procedure that breaches the mucosa, it is recommended that the accessory that breaches the mucosa is either single-use or undergoes sterilisation. hld involves the elimination of all bacteria, viruses and fungi with the exception of some bacterial spores which are only removed with sterilisation [ ] . reprocessing aims to stop the transmission of exogenous infection to the patient. outbreaks of bacterial infection associated with rfbs have primarily occurred in the setting of breaches in the reprocessing protocols whilst pseudo infection (cultural evidence of transmission of organisms without evidence of patient infection) has also occurred [ ] . the transmission of viral respiratory pathogens via rfbs has not been reported to date. the major recommendations from the various guidelines [ , , ] regarding the appropriate reprocessing of rfbs are the same. mechanical cleaning is performed as soon as the procedure is finished with leak testing to assess the integrity of the scope coupled with brushing (ideally with single-use brushes) and flushing. the rfb then undergoes hld-previously a manual process [ ] ; however, use of an automated endoscope reprocessor (aer) is now preferred [ ] . bronchoscopes must be stored in a hanging position in a cabinet with appropriate aeration and with adequate space between them to prevent cross-contamination [ ] . staff should receive training and wear personal protective equipment (ppe) while reprocessing the bronchoscopes [ ] . if ppe is not used, staff are at an increased risk of infection and may recontaminate fully processed scopes [ ] . in some institutions, rfbs are cleaned and then sterilised [ ] ; however, the chemicals usedeither ethylene oxide or hydrogen peroxideare expensive and interfere with the mechanical properties of flexible bronchoscopes [ ] . certain infectious agents are unusually resistant to standard methods of disinfection, sterilisation and uv radiation e.g. the prions that cause the transmissible spongiform encephalopathies (tses). if a patient undergoing bronchoscopy has a suspected diagnosis of a tse, the rfb should be incinerated after use; or if there is an expectation of repeat bronchoscopy in the same patient, the rfb should be set apart for use in that patient only [ , ] . a recent study over three different clinical sites inspected rfbs and measured levels of protein, adenosine triphosphate (atp) and infectious organisms both before and after manual cleaning and hld [ ] . at all sites, the patient-ready bronchoscopes had visible defects ( %) and harboured antimicrobial growth ( %). at two of the sites, the reprocessing was inadequate as a result of multiple episodes of non-compliance with the guidelines e.g. disabling of the cycles of the aer, ungloved handling of bronchoscopes and dirty storage cabinets. however, even at the third site where the reprocessing procedures met national guidelines, there was still an unacceptably high level of bio-burden on reprocessed bronchoscopes leading to the conclusion that a movement towards sterilisation of rfbs might be warranted [ ] , though this has its own disadvantages as outlined earlier. in , ecri highlighted the recontamination of flexible endoscopes due to mishandling or improper storage as one of the top ten health technology hazards. they referred particularly to the recontamination of disinfected endoscopes caused by failure of staff to change their gloves between inserting and removing the endoscope from the aer [ ] . assuming that there is no risk of exogenous infection with a sufb, a systematic review in used avoidance of this risk as an effect measure while trying to elucidate the true cost of rfbs when cross-contamination and infection are taken into consideration. the studies eligible for inclusion in the analysis involved flexible bronchoscopy performed in both icu and respiratory units. the results revealed an overall . % infection risk to the patient which considerably decreased the cost-effectiveness of rfbs compared to sufbs [ ] . sufbs are not designed to withstand the standard reprocessing of rfbs. it has been shown that following basic cleaning of a sufb, there was significant microbial colonisation of the devices at h including high-risk pathogens for causing pneumonia. this study confirms that sufbs are only appropriate for single use as opposed to single patient use [ ] . a recent morbidity and mortality report from the usa showed that hcp represented % of the population infected with sars-cov- and the majority ( %) of this group only had exposure to an infected person within a healthcare setting [ ] . occupational status as hcp was only specified in % of all submissions so it is likely that their actual infection rates were significantly underestimated. in those hcp on whom outcomes were available, - % required hospitalisation and . - . % died from the infection, highlighting the significant risks associated with the occupation [ ] . transmission of the virus occurs via droplets and fomites. droplets are respiratory aerosols that are larger than lm in diameter while fomites are inanimate objects that can transmit disease if they are contaminated with an infectious agent. bronchoscopy is considered an aerosol-generating procedure (agp) [ ] . thus, there is a risk of viral transmission both from the aerosols generated during the bronchoscopy and again during staff reprocessing of contaminated bronchoscopes. as a result, bronchoscopy is relatively contraindicated in patients who are suspected of or confirmed with sars-cov- infection in a bid to reduce disease transmission and protect hcp [ , ] . despite this, a balance must be achieved to ensure that patients who need urgent therapeutic and diagnostic interventions are not neglected. in situations where bronchoscopy is essential, for instance a suspicion of malignancy in a patient fit for cancer therapy, any day-case bronchoscopy should be delayed by days following confirmed or suspected diagnosis of covid- and if proceeding thereafter, only essential personnel should be present with all staff wearing ppe and avoidance of high flow nasal oxygen [ ] . the american association of bronchology and interventional pulmonology (aabip) has gone as far as to recommend avoiding the use of rfbs in this situation and using sufbs instead [ ] with some institutes having already adopted sufbs as a result [ ] . exhaled air from a human patient simulator (hps) receiving - l/min of oxygen via nasal cannula in a negative pressure room can travel as far as cm. coughing can cause air dispersal to cm which is reduced to cm with donning of a surgical mask by the hps. thus, it is advised that all patients with covid- wear surgical masks to reduce transmission. additionally, when using a nasal approach during bronchoscopy, the patient's mouth should be covered by a mask and if they require non-invasive ventilation (niv), it should be administered through a hole in the patient's mask [ ] . the construction of a surgical tent with disposable drapes may improve the safety of hcp during the procedure [ ] , although any barrier method needs research to evaluate whether it might reduce covid- transmission whilst increasing the transmission of other hospitalacquired infections. similar principles apply in unavoidable surgical procedures in sars-cov- -infected patients that require a general anaestheticsingle-use equipment is recommended where possible e.g. a sufb could be used to ensure the correct positioning of an endotracheal tube [ ] . a systematic review completed in compared the risk of transmission of acute respiratory infections (aris) to hcp involved in agps. the ten studies that met the inclusion criteria all pertained to transmission of severe acute respiratory syndrome coronavirus (sars-cov) during the outbreak of / and the agps included tracheal intubation, bronchoscopy and niv. the data was only slightly significant for tracheal intubation increasing the risk of transmission of sars-cov and there was insufficient data to come to any conclusion about bronchoscopy. unfortunately, the included studies were categorised as providing very low-quality evidence; however, the review did emphasise the importance of using appropriate ppe during agps [ ] . another study in singapore general hospital during the sars-cov outbreak noted the high risk of transmission to anaesthetists particularly during intubation and bronchoscopy. following the implementation of stringent infection control measures including single-use devices where possible and appropriate use of ppe, they successfully reduced transmission to hcp even when patients presented asymptomatically or with atypical infections [ ] . therefore, evidence supports the introduction of sufbs to decrease viral transmission both from and to staff and patients. furthermore, eliminating the requirement for reprocessing can help to counteract a reduction in staff numbers due to local outbreaks [ ] . several companies produce single-use bronchoscopes with some of them currently on fourthgeneration devices that have improved image quality and degrees of angulation (fig. ) . the range of devices includes a selection of channel diameters (table ) . each company has produced a small portable reusable screen that is easy to clean and from which videos or images can easily be saved or downloaded. as sufbs do not require any reprocessing after use, the risk of transmission of infectious particles to hcp is reduced by minimising exposure to fomites or aerosols. the portable screens are also easy to clean and a less complex circuit allows for easier tracing of any potential contaminants [ ] . the ability of trainees to learn is limited while current recommendations advise that only essential personnel are allowed into the room during bronchoscopy [ , , , ] ; however, the video function on sufbs allows you to easily record and store images which can be used to demonstrate clinical anatomy and pathology [ ] from a remote location where social distancing can easily be observed. they are also useful as general teaching aids such as simulation training with mannequins which has been shown to reduce the amount of subsequent damage to rfbs [ , ] . sufbs clearly have advantages in centres performing bench, cadaveric or large animal research, reducing cost of equipment, cleaning costs and storage. sufbs may also be of use in training and research in the veterinary field where bronchoscopy is performed for a variety of indications [ , ] . the other major advantage for sufbs is the option for parallel as opposed to linear use in the respiratory suite which can decrease delays between procedures and increase the number of bronchoscopies that can be performed. their immediate availability and the possibility of out-of-hours use is also a distinct advantage in the anaesthetic setting for the unanticipated difficult airway [ , ] . in the immunocompromised patient and in rare cases of prion contamination due to tse, they offer a safer alternative to rfbs. sufbs have been shown to be acceptable compared to rfbs in an anaesthetic setting [ , ] and for performing bronchoalveolar lavage (bal) in healthy volunteers for research purposes [ ] (table ) , though some users comment that image technology and handling is not yet equivalent to rfbs [ ] . the studies of the efficacy of sufbs to date have compared them to rfbs in an anaesthetic setting only, with no studies analysing their efficacy in a clinical pulmonology setting. in anaesthetics where bronchoscopes are often used in emergency situations such as outer diameter (mm) . , . , . , . , . . (agile) (fig. ) [ ] , it is essential that they are fit for purpose. the demonstration that sufbs are adequate in performing bal for research [ ] suggests that they are also likely to be acceptable for diagnostic purposes in a clinical respiratory setting. an increased yield with bal using sufbs [ ] has the potential to reduce postprocedural side effects and, if this finding is reproducible, could make them the preferred choice over rfbs. during the covid- pandemic the cost of healthcare services and management of resources will ultimately affect patient outcomes. one might anticipate that sufbs will be more expensive than rfbs; however, in addition to the initial cost of the rfb, a significant amount of resources are required for the appropriate training of personnel, provision of designated cleaning areas, ppe, maintenance of the aer as well as supply of disinfectants, enzyme reagents and detergents [ ] . the limited number of studies in this area (none conducted solely in a bronchoscopy unit) have demonstrated that sufbs can be as cost-effective as rfbs in a variety of situations (table ) . local factors such as initial purchase price, service agreements and reprocessing protocols influence the cost-effectiveness of rfbs-repair costs can vary significantly between different hospitals which will alter the cost per use [ , ] . additionally, different procedures require higher maintenance or are associated with greater damage to bronchoscopes e.g. more write-offs when performing percutaneous tracheostomy compared to bal [ ] . there is a grs global rating scale (a validated score for benchmarking operators who perform clinical bronchoscopy), bal bronchoalveolar lavage, rml right middle lobe procedure number at which rfbs become more cost-effective than sufbs [ , ] with mathematical modelling tools available to assess the cost-effective number of rfbs or sufbs that should be purchased whilst allowing for locally variable factors. when device demand is incorporated into these algorithms, there is a suggestion that units performing a smaller number of interventions could manage solely with sufbs whilst rfbs become more economical with increased demand [ ] and it is likely to be cost-effective to have a subset of sufbs available for emergency use [ , ] . to date, no studies have been published looking at the cost- effectiveness of sufbs in a respiratory setting, and the costs associated with rfb maintenance and repair in an anaesthesia department may not be comparable. prior to the recommendation to introduce sufbs into bronchoscopy units, it would be important to optimise the current costs associated with maintenance and repairs of rfbs. it has been shown that as much as . % of the cost of bronchoscope repair can be attributed to preventable damage e.g. unsheathing a biopsy needle within a working channel [ ] , and the introduction of educational programmes which focus on the cost of rfb repairs as well as emphasising safety regulations and procedure could drop the repair cost as much as % per procedure [ ] . quality improvement campaigns are similarly useful in reducing the incidence of scope damage and decreasing episodes of rfb unavailability [ ] . purchasing an insurance policy for an rfb is another way of reducing the cost of repairs [ ] . in developing countries, adherence to bronchoscopy guidelines may impose prohibitively expensive costs on the development of highquality flexible bronchoscopy units e.g. because of the recommendation for aers to reprocess rfbs [ ] . depending on the anticipated number of procedures sufbs may be a solution to this problem. during the covid- pandemic, sufbs have the potential to create a safer working environment in situations where agps such as bronchoscopy or intubation are unavoidable. prior to introduction of sufbs under normal circumstances, it would be necessary to assess their use and cost-effectiveness in a respiratory setting. in the interim, there are many strategies that can be employed to improve the cost-effectiveness of rfbs. it is likely that in the future, mathematical modelling tools will be used to guide procurement decisions for single-use and reusable bronchoscopes depending on local maintenance agreements, the number of procedures performed and sufbs would be purchased to make up the capacity shortage until the demand reaches a level that makes further purchase of reusable devices more cost-effective. funding. no funding or sponsorship was received for this study or publication of this article. the rapid service fee was funded by the authors. authorship. all named authors meet the international committee of medical journal editors (icmje) criteria for authorship for this article, take responsibility for the integrity of the work as a whole, and have given their approval for this version to be published. disclosures. sarah p barron and marcus p kennedy have nothing to disclose. compliance with ethics guidelines. this article is based on previously conducted studies and does not contain any studies with human participants or animals performed by any of the authors. data availability. data sharing is not applicable to this article as no datasets were generated or analysed during the current study. commons attribution-non-commercial . international license, which permits any non-commercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the creative commons licence, and indicate if changes were made. the images or other third party material in this article are included in the article's creative commons licence, unless indicated otherwise in a credit line to the material. if material is not included in the article's creative commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. to view a copy of this licence, visit http:// creativecommons.org/licenses/by-nc/ . /. transmission of infection by flexible gastrointestinal endoscopy and bronchoscopy protecting healthcare workers from sars-cov- infection: practical indications summarizing societal guidelines regarding bronchoscopy during the covid- pandemic american college of chest physicians and american association for bronchoscopy consensus statement: prevention of flexible bronchoscopy-associated infection effectiveness of reprocessing for flexible bronchoscopes and endobronchial ultrasound bronchoscopes aerosol generating procedures and risk of transmission of acute respiratory infections to healthcare workers: a systematic review covid- : protecting health-care workers characteristics of health care personnel with covid- -united states bronchoscopy services during the covid- pandemic irish thoracic society statement on bronchoscopy and sars covid- . statement. ir thorac soc american association for bronchology and interventional pulmonology (aabip) statement on the use of bronchoscopy and respiratory specimen collection in patients with suspected or confirmed covid- infection randomised controlled trial comparing the ambu Ò ascope tm with a conventional fibreoptic bronchoscope in orotracheal intubation of anaesthetised adult patients the disposable ambu ascope vs. a conventional flexible videoscope for awake intubation-a randomised study single-use flexible bronchoscopes compared with reusable bronchoscopes: positive organizational impact but a costly solution cost comparison of reusable and single-use fibrescopes in a large english teaching hospital a systematic review and cost effectiveness analysis of reusable vs. single-use flexible bronchoscopes british thoracic society guideline for diagnostic flexible bronchoscopy in adults health service executive standards and recommended practices for endoscope reprocessing units. health technical memorandum management and decontamination of flexible endoscopes. health technical memorandum transmissible spongiform encephalopathy agents: safe working and the prevention of infection top health technology hazards. executive brief. pennsylvania: ecri institute, health devices contamination of single-use bronchoscopes in critically ill patients with the current covid- pandemic: should we use singleuse flexible bronchoscopes instead of conventional bronchoscopes? pediatric laryngoscopy and bronchoscopy during the covid- pandemic: a four-center collaborative protocol to improve safety with perioperative management strategies and creation of a surgical tent with disposable drapes surgery in covid- patients: operational directives how severe acute respiratory syndrome (sars) affected the department of anaesthesia at singapore general hospital single use bronchoscopes: applications in covid- pandemic reducing maintenance and repair costs in an interventional pulmonology program implementation of an anaesthesia quality improvement programme to reduce fibreoptic bronchoscope repair incidents transtracheal wash and bronchoalveolar lavage randomized clinical trial to evaluate the pathogenicity of bibersteinia trehalosi in respiratory disease among calves difficult airway society guidelines for management of unanticipated difficult intubation in adults single use and conventional bronchoscopes for broncho alveolar lavage (bal) in research: a comparative study (nct ) infections and damaged flexible bronchoscopes-time for a change cost analysis of (ambu Ò ascope tm ) and reusable bronchoscopes in the icu decision support for the capacity management of bronchoscopy devices: optimizing the cost-efficient mix of reusable and single-use devices through mathematical modeling comparative study on environmental impacts of reusable and single-use bronchoscopes flexible bronchoscope damage and repair costs in a bronchoscopy teaching unit implementing flexible bronchoscopy in least developed countries according to international guidelines is feasible and sustainable: example from phnom-penh, cambodia key: cord- -ddve mga authors: li, tianyang; han, zhongyi; wei, benzheng; zheng, yuanjie; hong, yanfei; cong, jinyu title: robust screening of covid- from chest x-ray via discriminative cost-sensitive learning date: - - journal: nan doi: nan sha: doc_id: cord_uid: ddve mga this paper addresses the new problem of automated screening of coronavirus disease (covid- ) based on chest x-rays, which is urgently demanded toward fast stopping the pandemic. however, robust and accurate screening of covid- from chest x-rays is still a globally recognized challenge because of two bottlenecks: ) imaging features of covid- share some similarities with other pneumonia on chest x-rays, and ) the misdiagnosis rate of covid- is very high, and the misdiagnosis cost is expensive. while a few pioneering works have made much progress, they underestimate both crucial bottlenecks. in this paper, we report our solution, discriminative cost-sensitive learning (dcsl), which should be the choice if the clinical needs the assisted screening of covid- from chest x-rays. dcsl combines both advantages from fine-grained classification and cost-sensitive learning. firstly, dcsl develops a conditional center loss that learns deep discriminative representation. secondly, dcsl establishes score-level cost-sensitive learning that can adaptively enlarge the cost of misclassifying covid- examples into other classes. dcsl is so flexible that it can apply in any deep neural network. we collected a large-scale multi-class dataset comprised of , chest x-ray examples: examples from confirmed covid- cases, , examples with confirmed bacterial or viral pneumonia cases, and , examples of healthy people. extensive experiments on the three-class classification show that our algorithm remarkably outperforms state-of-the-art algorithms. it achieves an accuracy of . %, a precision of %, a sensitivity of . %, and an f -score of . %. these results endow our algorithm as an efficient tool for the fast large-scale screening of covid- . as covid- continues to affect our world, automated screening of coronavirus disease (covid- ) is urgently needed to realize large-scale screening to combat it. since the outbreak of covid- in december to date, more than , , people have been infected around the world. and more than , deaths from the virus have been recorded according to the world health organization. fast and large-scale screening is necessary to cut off the source of infection. however, the rapidly growing amount of covid- cases makes global medical resources unbearable. automated screening systems can correspondingly assist in speeding up screening and would reduce the workload of radiologists. therefore, it is urgent to realize the automated screening of covid- that helps to accelerate the large-scale screening and alleviate the global shortage of medical resources. nowadays, medical imaging examinations, such as chest ct, x-ray, play an essential role in the diagnosis process of covid- . clinically, although nucleic acid detection is the gold standard, the availability, stability, and reproducibility of nucleic acid detection kits are problematic [ ] . for example, the quantity of nucleic acid kits is limited in many countries or regions, resulting in the slower screening of new coronary pneumonia [ ] . many patients with new coronary pneumonia cannot be tested in time and thus cannot be admitted to the hospital, which accelerates the widespread of the novel virus. on the contrary, medical imaging examinations can help clinical to carry out disease detection conveniently and quickly, and thus make patients be treated timely. accordingly, medical imaging examinations with symptom observation are widely used for the early diagnosis of covid- worldwide [ ] . x-rays have unique advantages of light, quick, and availability in the screening of covid- . on the one hand, regular x-ray machines can be accessed in most primary hospitals where ct scanners are insufficient. most ambulatory care facilities have deployed x-ray units as basic diagnostic imaging. most importantly, the easily accessible x-ray examination is beneficial for fast large-scale screening. on the other hand, the radiation dose of x-ray is a few hundredths of the chest ct [ , ] . one statistic from the san francisco bay area shows that hospitals add one more case of cancer for every to , additional routine chest ct examinations [ ] . this statistic indicates that a single chest ct examination increases the lifetime risk of cancer by . % - . %. also, the x-ray examination is more economical than ct. therefore, this paper addresses the novel task of automated screening of covid- from chest x-rays. however, robust and accurate screening has two crucial bottlenecks. firstly, the imaging features of covid- cases share some similarities with other pneumonia cases on chest x-rays. even radiologists cannot distinguish them based on chest ct accurately without other inspection methods [ ] , let alone based on chest x-rays. the second bottleneck is that the misdiagnosis rate of covid- is very high. high misdiagnosis rate of covid- has a prohibitive cost that is not only delaying the timely treatment of patients but also causing the widespread of the virus with high social costs. misdiagnosis cost sensitivity should be fully considered in the automated screening of covid- . therefore, these limitations impedes the accurate screening of covid- patients from the susceptible population. while a few pioneering works have made much progress, they neglect both crucial bottlenecks. all of them adopted common machine learning classifiers. for example, both hassanien et al. [ ] and sethy et al. [ ] used support vector machines (svm) to screen covid- based on extracted features. farooq et al. [ ] , narin et al. [ ] , wang et al. [ ] and etc adopted popular deep neural networks with a minor modification of network architecture. more interestingly, zhang et al. [ ] viewed this screening task as an anomaly detection problem and proposed to use existing anomaly detection techniques. however, in fact, the screening of covid- is a fine-grained cost-sensitive classification problem, as mentioned before. only from this perspective can we design a satisfactory solution. therefore, in this study, we attempt to design an efficient solution to combat the bottlenecks. in this paper, we propose an innovative discriminative cost-sensitive learning (dcsl) for the robust screening of covid- from chest x-rays. dcsl combines both advances from fine-grained classification and cost-sensitive learning techniques. to overcome the subtle difference bottleneck, we propose a conditional center loss function to learn the discriminative representation between fine-grained classes. by combing with vanilla loss function, the conditional center loss can discover a weighted center for each class and efficiently enlarge the inter-class manifold distance as well as enhancing the intra-class compactness in deep representation space. to combat the cost sensitivity, we propose scorelevel cost-sensitive learning. it introduces a score-level cost matrix to reshape the classifier confidences by modifying the classifier output, such that the covid- examples have the maximum score, and the other classes have a substantially lower score. based on the domain knowledge that the costs between misclassifying covid- into different classes or misclassifying other classes into covid- are not equal, we define new score-level costs to encourage the correct classification of covid- class. we combine both advances into a deep neural network with end-to-end optimization, successfully achieving fine-grained cost-sensitive screening of covid- . a series of experiments show that our algo-rithm remarkably outperforms previous methods. the contributions of this work include: • for the first time, we formulate the task of screening of covid- from chest x-ray as a fine-grained costsensitive classification problem. accordingly, we propose a practical solution, discriminative cost-sensitive learning, that achieves much high screening accuracy. • we propose a new conditional center loss that considers the class-conditional information when learning the center points per class. the conditional center loss successfully overcomes the bottleneck of feature similarities. • we propose a new score-level cost-sensitive learning that introduces a domain knowledge-based cost matrix to enlarge the cost of misclassifying covid- examples into other classes. it greatly reduces the misdiagnosis rate. the remainder of this paper is organized as follows. section presents the related works in terms of artificial intelligence assisted analysis of covid- and involved methodologies. section gives in detail the proposed discriminative cost-sensitive learning. section presents detailed descriptions of collected datasets, experiment settings, and exhaustive results. section concludes this work comprehensively. this section shows related works about the automated screening of covid- and involved algorithms of our work. to take part in the global fight against covid- , many studies have designed ai-empowered technologies for improving the clinical diagnosis efficiency. shi et al. [ ] comprehensively summarized lots of emerging works, including automated screening [ , , , , , , , , ] , patient severity assessment [ ] , infection quantification [ ] , and infection area segmentation [ , ] . among them, automated screening received the most attention, involving chest x-ray based and chest ct based works. since d ct scans have spatial complexity, existing ct based works have proposed to design three types of solutions, including patchbased methods [ , , , ] , slice-based methods [ , , , ] , and d ct-based method [ ] . as we mentioned before, ct has some disadvantages of expensive, high radiation, and inaccessibility, thus lots of screening works are based on d chest x-rays. we comprehensively review chest x-rays based methods as follows. hassanien et al. [ ] used a multi-level threshold segmentation algorithm to crop lung areas and adopted svm to classify covid- and normal cases based on chest xrays. ozturk et al. [ ] ensembled several feature extraction algorithms and used a stacked autoencoder with principal component analysis to make decisions. they showed that handcrafted features based classifiers perform better than deep models on small data. several studies applied popular deep learning techniques for the screening of covid- . hemdan et al. [ ] validated multiple popular deep models and demonstrated their effectiveness on this new task. farooq et al. [ ] utilized the residual networks (resnet) to validate the screening performance of covid- . narin et al. [ ] tested an inception architecture (inceptionv ) on the screening task. they showed that pre-trained models are useful. the vgg and vgg networks are adopted by hall et al. [ ] and apostolopoulos et al. [ ] , respectively. khalifa et al. [ ] used generative adversarial networks and a finetuned deep transfer learning model, achieving promising and effective performance. their results also confirm that chest x-rays based screening of covid- has great research significance. moreover, several studies designed specialized solutions for the screening of covid- according to the characteristics of the task. afshar et al. [ ] adopted a capsule network for handling small data. abbas et al. [ ] designed a decompose, transfer, and compose (detrac) network based on class decomposition for enhancing low variance classifiers and facilitating more flexibility to their decision boundaries. wang et al. [ ] proposed a new deep network called covid-net, which consists of stacked residual blocks for achieving easily training and deepening the architectures. to improve diagnostic performance, ghosha et al. [ ] used bayesian convolutional neural networks to estimate uncertainty. more interestingly, zhang et al. [ ] viewed this screening task as an anomaly detection problem and proposed to use existing anomaly detection techniques. specifically, they used a hybrid loss that combines a binary cross-entropy loss and a deviation loss to assign anomaly scores to covid- examples. while a few pioneering works have made great progress, they neglect both cost sensitivity and fine-grained bottlenecks. to the best of our knowledge, it is the first time that we insightfully view the screening of covid- from chest x-rays as a fine-grained cost-sensitive classification problem. the goal of fine-grained classification is to classify data belonging to multiple subordinate categories, e.g., covid- , common pneumonia, normal chest x-rays. the facing problem of fine-grained classification is that these subordinate categories naturally exist small inter-class variations and large intra-class variations. the common solutions of this problem can be organized into three main paradigms, including ) localization-classification networks based methods [ , , ] , ) external information-based methods [ , , ] , and ) end-to-end feature coding-based methods [ , ] . localization-classification networks based methods first learn part-based detectors or segmentation model to localize salient parts for improving the final recognition accuracy. however, this type of paradigm needs additional part annotations. external information-based methods leverage exter-nal information, i.e., web data [ ] , multi-modality [ ] , or human-computer interactions [ ] . different from the previous two paradigms, end-to-end feature coding-based methods learn a more discriminative feature representation directly [ ] . among them, several works specifically designed extra useful loss functions for learning discriminative fine-grained representations. for instance, the contrastive loss is designed for dealing with the relationship of paired example points effectively [ ] . triplet loss correspondingly constructs loss functions for example triplet [ ] . however, contrastive loss and triplet loss are required that the number of training pairs or triplets dramatically grows, with slow convergence and instability. center loss is proposed to solve this issue by minimizing the deep feature distances of intra-class only [ ] . this loss function learns a center for each class and pulls the deep features of the same class to their centers efficiently. however, center loss easily suffers from the issue of class imbalance. therefore, in this paper, we propose to learn a new conditional center loss with joint balance optimization for the robust screening of covid- . cost-sensitive learning is a learning method that considers the cost of misclassification, and its purpose is to minimize the total cost [ ] . in the classical machine learning setting, the costs of classification errors of different classes are equal. unfortunately, the costs are not equal in many real-world tasks. for example, in covid- screening, the cost of erroneously diagnosing a covid- patient to be healthy may be much higher than that of mistakenly diagnosing a covid- patient to be common pneumonia. costsensitive learning is proposed to handle this problem and has attracted much attention from the machine learning and data mining communities [ ] . existing works on misclassification costs can be categorized into two classes, including example-dependent cost [ , , ] and class-dependent cost [ , , ] . example-dependent cost-based methods consider the misclassification cost of each example and are required example-level annotations, which are impractical in real-world tasks. therefore, most methods are focused on class-dependent costs. cost-sensitive learning is also suitable to handle the problem of class imbalance [ ] . in this study, we introduce a score-level cost-sensitive learning approach based on an expert-provided cost matrix to improve the screening accuracy of covid- from chest x-rays. in this section, we first introduce the necessary notations and the objective for the task of screening of covid- from chest x-rays (see section . ). we then present the newly-proposed discriminative cost-sensitive learning (dcsl) framework, which consists of a conditional center loss (see section . ) and a score-level cost-sensitive learning approach (see section . ). we finally combine the two novel modules to construct the dcsl framework for the fine-grained cost-sensitive classification problem (see section . ). figure : overview of the proposed discriminative cost-sensitive learning (dcsl) framework. based on a data pool of clinical x-rays, a comprehensive analysis is conducted to obtain the class-conditional information (class balance weight). in the training period, we randomly draw a batch of data to optimizer softmax cross-entropy and conditional center loss with the class-conditional information. specifically, an input image is firstly processed by a backbone network, which mainly includes convolutional layers (conv) and fully-connected layers (fc). after that, deep features are obtained to minimize conditional center loss. score-level costs are applied to the outputs from the final output layer to get the final cost-sensitive prediction. the task of covid- screening is under the familiar supervised learning setting where the learner receives a sample of labeled training examples {( , )} = drawn from a joint distribution defined on  × , where  is the example set of d chest x-ray images, and  is the label set of patient conditions, such as covid- , common pneumonia, others.  is { , } in binary classification and { , … , } in multi-class classification. specifically, is any chest x-ray image of one patient, and is the label of this patient. we denote bŷ the empirical distribution. we denote by  ∶  ×  → ℝ any loss function defined over pairs of labels, such as - loss, cross-entropy loss, etc. for binary classification, we denote by for any distribution on  × and any labeling function ℎ , we denote  (ℎ ) = e ( , )∼ (ℎ ( ), ) the expected risk. our objective is to select a hypothesis out of a hypothesis set  with a small expected risk  (ℎ ) on the target distribution. this section presents the newly-proposed conditional center loss. to better understand the role of conditional center loss, we first consider the softmax based cross-entropy loss (softmax loss) that is presented as follows. where ∈ r (to reduce abuse notations) denotes the th deep feature of the x-ray image , belonging to the class . ∈ r denotes the th column of the weights ∈ r × in the last fully-connected layer, and ∈ r is the bias term. and are the size of mini-batch and the number of classes, respectively. intuitively, the softmax loss first computes the probability of correct classification and then takes the logarithm of this probability. since the logarithm value of probability is negative, a minus sign is added in front of it. while the softmax loss is good at common object recognition tasks, it cannot learn enough discriminative features in processing fine-grained classification tasks in which significant intraclass variations exist. according to our in-depth analysis of the characteristics of the screening of covid- from chest x-rays, we, with keen insight, view it as a fine-grained classification problem. as we discussed before, the center loss has advantages of flexibility, easy-to-implement, and stability. therefore, we initially leverage the center loss to learn more discriminative features. the goal of center loss is to directly improve the intra-class compactness, which is conducive to the discriminative feature learning. center loss is widely embedded between the fully connected layers of deep neural networks for decreasing the intra-class variations in the representation space (the dimension is two). it is commonly appeared with softmax loss and used for face recognition and fine-grained classification. the center loss function is formulated as follows. where the denotes the th class center of deep features. the center is updated according to the mini-batch data. and it is computed by averaging the features of the corresponding classes. a scalar is used to control the learning rate of the centers. the update equation of is represented as follows. where is an indicator function in which = if the condition is satisfied, and = if not. denotes the iteration of training. the final joint loss function is given by where is used for balancing the joint loss function. while the joint loss function has achieved great success in practice, however, it quickly losses efficiency in the class imbalance situation. in other words, the center loss does not work in the screening task of covid- according to our observations. after an in-depth analysis, we found that the problem is that the learned center points are unrepresentative. to handle this problem, we propose a conditional center loss that considers the class-conditional information when updating the center points and optimizing the center loss. we denote by the weight of th class, and is computed by the ratio between the number of th class's training examples and the total training examples. the update equation of center points is reformulated by meanwhile, we found that embedding the clas-conditional information into the center loss can significantly improve the total screening accuracy. the reason is that it makes the center loss learn more balance center points and thus enhances the intra-class compactness to obtain discriminative deep features. accordingly, the conditional center loss with softmax loss is represented as follows. the conditional center loss can effectively handle the problem that imaging features of covid- share some similarities with other pneumonia on chest x-rays by enlarging their feature distance in high-level representation space. the goal of cost-sensitive learning is to classify examples from essential classes such as covid- correctly. we propose a score-level cost-sensitive learning module that can efficiently learn robust feature representations for both the critical and common classes. it thus can enhance the accuracy of covid- without unduly sacrificing the precision of the overall accuracy. generally speaking, we introduce a handcrafted cost matrix whose design is based on clinical expert experience and then incorporate it after the output layer of deep neural networks. in this manner, we can directly modify the learning process to incorporate classdependent costs during training and testing, without affecting the training and testing time of the original network. we will show that the proposed algorithm can efficiently work for the screening of covid- and can be inserted into any deep neural networks. in this section, we first present the traditional cost-sensitive learning, which usually adds the cost matrix into the loss functions. we then detail the score-level cost-sensitive learning with an advanced learning strategy. formally, we denote by ′ the cost matrix whose diagonal ′ , represents the benefit for a correct prediction. we also denote by , the misclassification cost of classifying an example belonging to a class into a different class . the expected risk defined on the target distribution is given by where ( | ) is the posterior probability over all possible classes given an example . the goal of a classifier is to minimize the expected risk ( | ), however, which cannot be reached in practice. thus, we use its empirical distribution of̂ to minimize the empirical risk as follows. where (⋅) denotes a neural netowrk and ( ) is the neural netowrk output. ∈ r denotes the one-hot of the label . for neural networks, loss function  can be a cross-entropy loss function with softmax, which is penalized by the cost matrix ′ as follows. where denote the deep feature from the output of the penultimate layer. the entries of a handcrafted cost matrix usually have the form of inserting such a cost matrix into loss function can increase the corresponding loss value of an important class. however, such a manner would make the training process of neural networks unstable and can lead to non-convergence [ ] . therefore, we propose an alternative cost-sensitive learning. to make the learning process more stable and convergence, we propose a new score-level cost matrix to modify the output of the last layer of a convolutional neural network (cnn). as shown in figure , the location of the score-level cost matrix is after the output layer, and before the loss layer with softmax. we introduce the new score-level costs to encourage the correct classification of essential classes. therefore, the cnn output ( ) is modified using the cost matrix as follows. where ( ) ∈ r × and ∈ r × , such that ∈ r × . during the training process, the output weights are modified by the score-level cost matrix to reshape the classifier confidences such that the desired class has the maximum score, and the other classes have a considerably low score. note that the score-level costs perturb the classifier confidences. such perturbation allows the classifier to give more attention to the desired classes. in practice, all cost values in are positive, which enables a smooth training process. when using the score-level cost matrix, the cross-entropy loss function obtain outputs = ( + ) : compute the joint loss by  ( , , , , ) : update ⇐ − ∇ update ⇐ − ∇ : end while with softmax is finally can be revised by combining the conditional center loss and the score-level cost-sensitive learning, we propose the discriminative costsensitive learning (dcsl) framework. as shown in figure , given a chest x-ray image , dcsl first uses a backbone network parameterized by to extract deep features which is used for finding the center points and optimizing the conditional center loss. dcsl then uses an output layer parameterized with and to obtain the output vector ( ), where is a scoring function and is consisted of the backbone and the output layer. dcsl finally applies the score-level cost matrix on the output ( ) to obtain the new output , which is inputted into the joint loss layer when training and into the softmax layer when testing. the joint loss is revised by the workflow of learning and optimization of dcsl is shown in algorithm . we evaluate our algorithm on a newly-collected dataset against state-of-the-art algorithms. the code and dataset will be publicly available. to evaluate the performance of our method on screening of covid- , we collected a multi-class multi-center chest x-ray dataset. this dataset includes , examples with image-level labels. specifically, we collected the dataset from three different sources. the first source is from a github collection of chest x-rays of patients diagnosed with covid- . the second source is from a kaggle dataset , which is thoroughly collected from the websites of ) radiological society of north america (rsna), ) radiopaedia, and ) italian society of medical and interventional radiology (sirm). the third source is from a collection of x-ray images of bacterial and viral pneumonia [ ] . the collected data consists of chest x-rays with confirmed covid- , , chest x-rays with confirmed bacterial and viral pneumonia, and , examples of healthy condition. we selected out low-quality images in the dataset to prevent unnecessary classification errors. representative chest x-ray images of different classes are illustrated in figure , which shows the subtle differences between different classes as well as proving the necessity of fine-grained classification. in our experiments, we conduct a three-class classification task for better verifying the proposed scsl algorithm in the screening task. the first class is healthy x-ray images, the second class is confirmed covid- x-ray images, and the third class is other confirmed pneumonia x-ray images, which include both bacterial and viral pneumonia. we employ standard five-fold cross-validation on the dataset for performance evaluation and comparison. the dataset is divided into five groups. among them, four groups are used for training the deep neural networks, and the last group is used for testing the performance. this procedure is repeated five times until the indices of all of the subjects are obtained. the evaluation metrics include accuracy, precision, sensitivity, and f score. in order to verify the effectiveness of our proposed algorithm, we compare our designed dcsl algorithm with state-of-the-art methods: covid-net [ ] , vgg [ ] , incep-tionv [ ] , resnet [ ] . covid-net is a newly-proposed deep convolutional neural network tailored for the detection of covid- from chest x-rays. vgg , inceptionv , and resnet are popular convolutional neural networks which have made great success in various tasks. we implement our algorithm in keras. we use the common vgg network as the backbone [ ] . the original vgg network has layers. there are convolutional layers with a small filter with a size of × for extracting deep features. five max-pooling layers with × kernel are deployed after each block of the convolutional layers. we set the output shape of the features of the last convolutional layer to be × × and flatten them. the original fully connected layers of vgg are removed and replaced by two trainable fully-connected layers. the channel numbers of the two fully-connected layers are and three, respectively. since the collected dataset is too small to obtain promising results through training the deep network from scratch, we use a transfer learning strategy, i.e., the parameters of the convolutional layers are initialized from the pre-trained model on imagenet [ ] . also, all the compared algorithms are implemented according to their open-source codes with pretraining. the chest x-ray images were resized into × × . we also use an augmentation strategy to expand the dataset: each random-selected example is rotated by degrees clockwise or counterclockwise. the is set as . . is set to one without loss of generality. adam optimizer is used with an initial learning rate of e- . we set the training epoch to . the score-level cost matrix is designed according to the clinical expert experience as following. first of all, the cost of misclassifying covid- is higher than misclassifying other classes. among them, the cost of misclassifying covid- into healthy patients is higher than the cost of misclassifying covid- into other pneumonia patients. second, the cost of misclassifying other pneumonia is smaller than the cost of misclassifying covid- . specifically, the cost of misclassifying other pneumonia into covid- is higher than the cost of misclassifying other pneumonia into healthy patients. third, the cost of misclassification of healthy people is smaller than in the previous two situations. among them, the cost of misclassifying a healthy person into covid- is greater than the cost of misclassifying a healthy person into other pneumonia. accordingly, the final score-level cost matrix is designed as illustrated in figure . the proposed discriminative cost-sensitive learning algorithm (dcsl) achieves the highest results on the screen-ing of covid- from chest x-rays. table reports the results of our algorithm and compared algorithms. our algorithm obtains a classification accuracy of . %, a precision of . %, a sensitivity of . %, and an f -score of . %. our algorithm remarkably outperforms covid-net [ ] , which achieves state-of-the-art results before our work. also, our algorithm significantly outperforms all the compared algorithms on all the metrics. as shown in figure , even both the complex lung structures and indiscernible infection areas lead to unusual difficulties; our algorithm still obtains accurate performance, which demonstrates its robust strengths. figure displays the confusion matrixes of our algorithm and covid-net. owing to our score-level cost-sensitive learning, we achieve % accuracy in the class of covid- . such a result demonstrates the effectiveness of incorporating the score-level matrix after the output layer of deep neural networks to modify the learning process. figure presents the sensitivities of each class, where our algorithm achieves % sensitivity of covid- , which is much higher than compared methods. these results once verify the advantages of score-level cost-sensitive learning. both figure and figure show that our algorithm also achieves the highest accuracy in other classes, which demonstrates the critical role of conditional center loss that can improve the intra-class compactness evenly. we further perform statistical analysis to ensure that the experimental results have statistical significance. a paired t-test between the covid-net and our algorithm is at a % significance level with a p-value of . . this analysis result clearly shows that the improvement of our method is noticeable. the p-values of the vgg , inceptionv , and resnet models are less than . , which proves that popular classifiers are not suitable for the task of screening covid- from chest x-rays. these analyses verify that our insight that viewing the screening of covid- from chest x-rays as a fine-grained cost-sensitive classification task is correct. this section further gives in-depth ablation studies to demonstrate the effect of conditional center loss (ccl) and score-level cost-sensitive learning (slcsl), respectively. we construct four ablation models based on the backbone network vgg- . the first model is only using the cross-entropy loss with softmax called as softmax loss. the second model combines center loss and cross-entropy loss with softmax called as softmax+center loss. similarly, the third model combines conditional center loss and softmax loss called as softmax loss + ccl. the final model is our algorithm dcsl that combines score-level cost-sensitive learning, conditional center loss, and softmax loss. generally speaking, our final model dcsl achieves the best performance than the other ablation models, as shown in figure . these confusion matrixes strongly prove that our algorithm can accurately screen covid- from chest xrays without any missing case. both figure and figure demonstrate the convergence and stability of dcsl in the training and validation period. these excellent results show that our algorithm successfully achieves accurate and robust screening of covid- from chest x-rays. these extensive results once verify the correctness of our insight that this task is a fine-grained cost-sensitive classification problem. (a) softmax loss (b) softmax+center loss (c) dcsl figure : accuracy curves in the training and validation period. they show that dcsl has higher accuracy with stable convergence. table reports the results of out ablation study on different loss functions. our conditional center loss (softmax loss + ccl) remarkably outperforms the center loss and softmax loss. these results demonstrate the importance of considering the class-conditional information when updating the center points and optimizing the center loss. can contribute to improving the intra-class compactness. moreover, figure shows that the conditional center loss has fewer mistakes and achieves a balance performance on the three classes. also, both figure and figure show that the conditional center loss has excellent stability and fast convergence. in summary, the conditional center loss has a significant impact on the performance of our proposed architecture. when the conditional center loss is not used, the result of classification is obviously decreased, and the learned deep features contain significant intra-class variations. another goal of this work is to enhance the sensitivity of covid- without decreasing the overall classification accuracy. although all the results have verified the advantages of score-level cost-sensitive learning, we should dissect its strengths. figure shows that using score-level costsensitive learning achieves zeros mistake of the covid- class. moreover, figure demonstrates that dcsl makes the % sensitivity in covid- without decreasing the overall classification accuracy. experimental results show that dcsl can significantly improve the sensitivity and precision of covid- . to conclude, cost-sensitive learning plays a crucial role in the screening of covid- . during the global outbreak of covid- , the cost of misclassifying covid- patients into other types of pneumonia or even healthy people are much higher than the cost of misclassifying other classes. the proposed score-level cost-sensitive learning has significantly improved the sensitivity of covid- , proving our hypothesis that cost-sensitive learning is very suitable for the in this paper, we reported a new attempt for the finegrained cost-sensitive screening of covid- from chest x-rays. we proposed a novel discriminative cost-sensitive learning (dcsl) that includes a conditional center loss function and a score-level cost-sensitive learning module. to the best of our knowledge, this is the first method that formulates this novel application as a fine-grained cost-sensitive classification problem. extensive results have demonstrated that dcsl can achieve reliable and accurate results. indepth analyses have revealed the effectiveness and potential of dcsl as a clinical tool to relieve radiologists from laborious workloads, such that contribute to the quickly large-scale screening of covid- . classification of covid- in chest x-ray images using detrac deep convolutional neural network covid-caps: a capsule networkbased framework for identification of covid- cases from x-ray images covid- : automatic detection from x-ray images utilizing transfer learning with convolutional neural networks support vector machines with example dependent costs classification and regression trees deep learning-based model for detecting novel coronavirus pneumonia on high-resolution computed tomography: a prospective study imagenet: a large-scale hierarchical image database leveraging the wisdom of the crowd for fine-grained recognition metacost: a general method for making classifiers cost-sensitive maximum-entropy fine grained classification covid-resnet: a deep learning framework for screening of covid from radiographs radiation dose in computed tomography pv-lvnet: direct left ventricle multitype indices estimation from d echocardiograms of paired apical views with deep neural networks estimating uncertainty and interpretability in deep learning for coronavirus (covid- ) detection rapid ai development cycle for the coronavirus (covid- ) pandemic: initial results for automated detection & patient monitoring using deep learning ct image analysis coronavirus detection and analysis on chest ct with deep learning finding covid- from chest x-rays using deep learning on a small dataset automatic x-ray covid- lung image classification system based on multi-level thresholding and support vector machine deep residual learning for image recognition covidx-net: a framework of deep learning classifiers to diagnose covid- in x-ray images serial quantitative chest ct assessment of covid- : deep-learning approach radiation doses and risks in chest computed tomography examinations development and evaluation of an ai system for covid- diagnosis ai-assisted ct imaging analysis for covid- screening: building and deploying a medical ai system in four weeks a rapid advice guideline for the diagnosis and treatment of novel coronavirus ( -ncov) infected pneumonia (standard version) identifying medical diagnoses and treatable diseases by image-based deep learning detection of coronavirus (covid- ) associated pneumonia based on generative adversarial networks and a fine-tuned deep transfer learning model using chest x-ray dataset cost-sensitive learning of deep feature representations from imbalanced data deep lac: deep localization, alignment and classification for fine-grained recognition bilinear cnn models for fine-grained visual recognition cost-sensitive learning and the class imbalance problem learning when data sets are imbalanced and when costs are unequal and unknown automatic detection of coronavirus disease (covid- ) using x-ray images and deep convolutional neural networks classification of coronavirus images using shrunken features. medrxiv facenet: a unified embedding for face recognition and clustering detection of coronavirus disease (covid- ) based on deep features lung infection quantification of covid- in ct images with deep learning review of artificial intelligence techniques in imaging data acquisition, segmentation and diagnosis for covid- large-scale screening of covid- from community acquired pneumonia using infection size-aware classification very deep convolutional networks for large-scale image recognition deep learning enables accurate diagnosis of novel coronavirus (covid- ) with ct images do ct scans cause cancer? learning from web data using adversarial discriminative neural networks for fine-grained classification deep learning face representation by joint identification-verification rethinking the inception architecture for computer vision covid-net: a tailored deep convolutional neural network design for detection of covid- cases from chest radiography images a deep learning algorithm using ct images to screen for corona virus disease mask-cnn: localizing parts and selecting descriptors for fine-grained bird species categorization a discriminative feature learning approach for deep face recognition fine-grained image classification by visual-semantic embedding deep learning system to screen coronavirus disease pneumonia learning and making decisions when costs and probabilities are both unknown a simple method for costsensitive learning covid- screening on chest x-ray images using deep learning based anomaly detection part-based r-cnns for fine-grained category detection deep learning-based detection for covid- from chest ct using weak label on multi-class cost-sensitive learning key: cord- -vtids ns authors: laxminarayan, ramanan title: trans-boundary commons in infectious diseases date: - - journal: nan doi: . /oxrep/grv sha: doc_id: cord_uid: vtids ns emerging threats to global health, including drug-resistant pathogens, emerging pandemics, and outbreaks, represent global trans-boundary commons problems where the actions of individual countries have consequences for other countries. here, we review what economic analysis can offer in countering these problems through the design of interventions that modify the behaviour of institutions and nations in the direction of greatest global good. the past century has been marked by significant improvements in life expectancy, due to greater child survival and reductions in infectious disease. the greatest victories in global health have come through globally coordinated actions-the eradication of small pox in , the global polio eradication initiative (still ongoing), and the sharp reductions in malaria through the global malaria eradication program (gmep) in the s. just the first two of these initiatives resulted in roughly . m deaths averted each year (unicef, ; ehreth, ) , and the gmep was responsible for eliminating malaria in countries (kouznetsov, ) . largely as a consequence of these efforts and of improvements in wellbeing that have translated into better ability to prevent and treat infectious diseases, these conditions have diminished in importance as a source of ill health across much of the world. according to the global burden of disease estimates, the percentage of disability-adjusted life years (dalys) due to prominent infectious diseases (comprised of the following four cause groups: hiv/aids and tuberculosis; diarrhoea/lower respiratory infections/other infectious diseases; neglected tropical diseases and malaria; and other communicable diseases) decreased from . per cent in to . per cent in , while the percentage of deaths due to prominent infectious diseases decreased from to . per cent (ihme ) . nevertheless, infectious diseases continue to be a significant source of ill health globally and a number of the world's emerging global health threats involve infectious diseases that can easily cross boundaries. the emergence of a new infectious disease-ebola being the most recent example-poses a significant risk to other countries, no matter where it arises. the risk is not uniform: countries that are connected by geography or population movement with the country where the disease emerges, and those with weak health systems are particularly vulnerable. but there are counter-examples as well. new delhi metallo-β-lactamase (ndm) enzymes that cause drug resistance in bacteria, which were first reported in from one patient hospitalized in sweden, are now reported globally (nordmann et al., ) . multiple factors including human population growth, land-use changes, and infectious diseases originating from wildlife (also known as zoonoses) are accelerating the frequency with which infectious diseases emerge (jones et al., ) . even if the direct health toll from these emerging infections does not approach the levels that were observed during the global flu pandemic, when nearly m people died, these infections can nevertheless do serious damage to economies, health, and health systems by virtue of their speed of attack. ebola has killed roughly , people in the last year, mostly focused in the west african countries of guinea, liberia, and sierra leone, and resulted in a per cent loss of gdp in these countries (world bank, ) . a recent study projected that after - months of disruptions, the accumulation of a large connected cluster of children unvaccinated for measles across guinea, liberia, and sierra leone resulted in between , and , additional child deaths due to measles alone (takahashi et al., ) . the deaths of healthcare personnel may have ripple effects down the road and could even discourage people seeking to train to be tomorrow's healthcare workers. the overall damage to health systems due to the large numbers of health system professionals lost to ebola will only become apparent in coming years. drug resistance is now a global problem and threatens public health in nations regardless of economic status (laxminarayan et al., ) . antibiotic-resistant gonorrhoea emerged in vietnam in (holmes et al., ) , before spreading to the philippines and finally to the united states (rasnake et al., ) . ndm enzymes are now in nearly every country, as discussed earlier. in this paper, we discuss the global health threats that involve 'commons' problems. with such problems, the actions undertaken in one country have consequences for other countries, but these are 'externalities' that are not taken into consideration by decisionmakers. for instance, a country may not report a disease outbreak for fear that it would discourage tourism, but the failure to report the outbreak could put other countries at risk. other examples of country-level actions with global consequences include inadequate vaccination coverage; slow progress on disease elimination; failure to report and contain pandemic flu, antibiotic resistance, and counterfeit drugs; and climate-related health threats. we provide some examples and case studies of such negative externalities across borders. then, we discuss the need for international cooperation for tackling these global health threats. the remainder of the paper is organized as follows. section ii describes trans-boundary externalities in tackling infectious diseases. section iii deals with incentives for surveillance and reporting of disease outbreaks. section iv addresses incentives for disease elimination and eradication. section v addresses incentives and financing mechanisms for controlling drug-resistant pathogens. section vi concludes the paper. early examples of international medical cooperation in the modern age were based on the idea that because infectious diseases do not respect national boundaries, meaningful control necessarily transcends national programmes. the first international sanitary conference was convened in paris in to discuss the quarantine of ships to contain plague, yellow fever, and cholera; it predated the first geneva conventions on treatment of war casualties by years (stern and markel, ) . more recently, campaigns to eliminate smallpox and eradicate malaria have been built on the idea that infectious disease control depends not just on national priorities but also on the priorities of one's neighbours and trading partners. an understanding of transnational disease transmission was deeply rooted in the gmep, which was launched in . funding from the top contributors to the special account for malaria by member countries during - accounted for per cent of overall contributions over this period (table ) . of these contributors, only saudi arabia had any significant malaria. malariacontrol investments in the current era are also likely to be largely externally funded, but contributions are not likely to continue indefinitely. therefore, the gains made from control have to be sufficiently large not just in the focal country but also in neighbouring countries so that malaria control will continue to be a priority for national planners even after the donors have exited. malaria control benefits the country in which it occurs, of course, but in the longer term, its neighbours benefit as well because they face fewer cases of imported malaria. the spatial coordination problems introduced by trans-boundary malaria are also relevant for the problem of regional elimination within large countries, especially those with frequent in-country movement, such as india. in contrast, china has managed to eliminate malaria from most of the interior of the country, but imported malaria remains a problem on its southern border. the extent of the 'external' benefit (to a neighbour) depends on malaria prevalence in that neighbour and the frequency and direction of overland migration. if malaria is common, then the benefit of fewer imported cases is minimal. however, the benefits can be large if the neighbour has eliminated malaria but still has to deal with cases imported from the focal country. barrett describes four equilibria in interactions between two countries that share an infectious disease (barrett, ) . in the first equilibrium, neither country engages in control, irrespective of what the other country does. in the second, each country eliminates the disease, irrespective of what its neighbour decides to do. in the third, each country eliminates the disease only if the other can be relied upon to do so. in the fourth, one country does not eliminate the disease, irrespective of what the other does. when countries are not identical in either epidemiological conditions or economic prosperity, it may be in the interest of some countries to eliminate malaria but for others not to, even if all others have eliminated malaria. yet elimination may be the optimal outcome for the two countries as whole. this is the case in which richer adjacent countries have financed elimination in poorer countries, as we observe in the lubombo spatial development initiative (lsdi). lubombo spatial development initiative lsdi offers a recent example of trans-boundary control of infectious disease (sharp et al., ) . malaria control was seen as an essential element of economic development in the lubombo region of eastern swaziland, southern mozambique (maputo), and north-eastern kwazulu natal province in south africa. malaria prevalence in these three regions was closely intertwined because of the frequent migration of people (sharp and le sueur, ) . most malaria cases in swaziland and kwazulu natal were imported from mozambique: for instance, nearly per cent of the malaria cases in kwazulu were in the district adjoining mozambique. between november and february , indoor residual spraying with bendiocarb insecticide was carried out twice a year in mozambique. spraying started in zone ( figure ) and proceeded incrementally, eventually covering seven districts and a population of roughly , people. in swaziland, where there were no other changes in malaria control efforts over the same time period, new malaria cases declined by per cent (table ). malaria cases declined by per cent in mpumalanga province, probably because during this period, indoor residual spraying and artemisinin-combination treatment were introduced on the south africa side of the border. nevertheless, the sharp decline in malaria in swaziland and south africa was attributable at least in part to efforts in mozambique, which were largely paid for by south africa and, to a lesser extent, by the global fund to fight aids, tuberculosis and malaria. west african river blindness programme coordinated financing, specifically with reference to multi-lateral financing to more than one country, is essential to permit a coordinated approach to disease control. however, such coordination has rarely been accomplished outside of global disease eradication programmes. there are a few examples of regionally coordinated financing such as against river blindness. the onchocerciasis control programme (ocp), which was launched in , covered major portions of seven western african countries (burkina faso, benin, ghana, côte d'ivoire, mali, niger, and togo). because the initial set of countries did not cover the limits of the breeding sites of the main vector, the savannah blackfly, the programme was expanded in to also include guinea, guinea-bissau, senegal, and sierra leone. a rare example of a transnational disease control effort launched by the world bank (kim and benton, ) , ocp relied on regionally coordinated larvicide spraying along the niger river to control black fly populations, and, at its peak, the programme covered m people in countries. this coordinated funding was in recognition of the fact that controlling black fly populations in a single country would be infeasible and required the cooperation of all seven countries on the niger river. through the mectizan donation programme, which was initiated in , onchocerciasis was eliminated as a public health problem in west africa. over the period - , the programme prevented , cases of blindness, and brought about m hectares of arable land-enough to feed an additional m people a year-back into productive use. the earliest efforts in global cooperation in the context of sanitary conventions, which required countries to report cholera outbreaks, subsequently led to the establishment of the pan american health organization, a pre-cursor to the world health organization (who) in the twentieth century. despite the benefits of warnings and reports on infectious disease outbreaks, there are few incentives for countries to report disease outbreaks that occur within their borders. current international health regulations, which were first enacted in and most recently revised in , require countries to report disease outbreaks. however, as there are no penalties for non-reporting, reporting depends on the goodwill of nations (baker and fidler, ) . this may not be entirely true since 'the consequences of non-compliance may include a tarnished international image, increased morbidity/mortality of affected populations, unilateral travel and trade restrictions, economic and social disruption and public outrage' (who, ) . specifically, if countries do not report promptly, other countries may take actions to moderate their trade and travel relations with the target country for fear that a future outbreak may also not be reported. we have discussed this in detail below as ex ante sanctions that precede an actual future outbreak. from a practical standpoint, countries face conflicting incentives as to whether or not to report an outbreak. on the one hand, reporting brings the near certainty of trade sanctions that can impose large costs. for example, when peru reported an outbreak of cholera in , its south american neighbours imposed an immediate ban on peruvian food products. the $ m cost of these sanctions and the additional $ m lost from reduced tourist activity far exceeded the domestic health and productivity costs of the epidemic (panisset, ) . on the other hand, countries may report an outbreak in the belief that the information will be reported anyway through the media or informal channels. furthermore, reporting an outbreak may result in international assistance for containing the outbreak. for instance, in the same peruvian outbreak, foreign aid in the form of rehydration salts, saline solution, and antibiotics, while unable to prevent an epidemic, helped to significantly reduce the death rate (brooke, ; suárez and bradford, ) . the appearance of new infections is determined by a number of factors, but generally is mediated by large growing populations that have poor nutrition and lack access to medical care (woolhouse and gowtage-sequeria, ) . however, despite the regular appearance of novel infections (woolhouse et al., ) , few infections are able to spread effectively within a population. over the last century, although more than diseases are believed to have emerged, only five novel diseases have swept across the globe-three were novel strains of influenza, another was hiv/aids, and more recently we saw the spread of sars, which emerged in china in november and spread around the world infecting more than , people in countries and killing approximately before it was contained (zhong et al., ) . further delays in reporting sars by china could have resulted in catastrophic consequences worldwide if the pathogen had been more virulent (heymann and rodier, ) . there is evidence that countries respond to external incentives on whether or not they report infectious disease outbreaks. an outbreak of meningococcal meningitis during the hajj resulted in more vaccination requirements for travellers coming to saudi arabia (laxminarayan et al., ) . these requirements, which were introduced in , were associated with reduced reporting of meningitis outbreaks among countries in sub-saharan africa, especially among countries with relatively few cases reported between and . the announcement of a programme in to assist countries with immediate vaccines conditional on their reporting of outbreaks was associated with an increase in reporting among countries that had previously not reported meningitis outbreaks (laxminarayan et al., ) . incentives for surveillance and reporting lie at the heart of an effective strategy to respond to avian influenza . mathematical models have suggested that it may be possible to contain an emerging pandemic of avian influenza if detection and reporting of cases suggestive of increased human transmission occurs within approximately weeks of the initial case (ferguson et al., ; longini et al., ) . while the who is responsible for coordinating the global response to human cases of avian influenza, decisions on establishing surveillance networks and reporting of outbreaks are the province of national governments. incentives to report an outbreak once it has been detected are only one part of the story, since an outbreak must first be detected. incentives to invest in surveillance to detect an outbreak are likely to be endogenous, and depend on whether or not a country wishes to report an outbreak (malani and laxminarayan, ) . these incentives are driven in part by the 'private' value of early detection to the individual country, but also by the likely consequences of the availability of this information to the rest of the world, either through the act of formal reporting or by informal channels, such as news reports or rumours. the greater the anticipated sanctions, the less likely a country will be to invest in surveillance. conversely, the higher the perceived benefit of international assistance in preventing or ameliorating the cost of an outbreak, the greater the likely investment in surveillance. current international mechanisms to encourage better reporting of disease have, by and large, ignored the economic dilemma and strategic behaviour of countries with emergent outbreaks. investments in surveillance also depend on the likelihood that the detected outbreak will produce a significant epidemic. the more a country believes a disease will arise and spread, the more significant the incentive to invest in surveillance. however, this investment can be tempered by the likelihood of false positives-the detection of a disease when none exists (malani and laxminarayan, ) . thus, a trade-off exists between investing in increased surveillance and investing in more accurate surveillance. a government's decision to report an outbreak can be modelled as a signalling game in which a country has private but imperfect evidence of an outbreak (malani and laxminarayan, ). an important conclusion is that not all kinds of sanctions may discourage reporting. what does this mean? let us divide sanctions into two kinds. ex ante sanctions are imposed in the form of reduced trade and travel contact with countries that are perceived to be poor at reporting disease outbreaks promptly. it is for this reason that west africa is not a favoured tourist destination-even in the absence of ebola, one is never quite sure if the system is able to detect and report this and other diseases. in contrast, an ex post sanction is imposed following a disease outbreak. ex post sanctions discourage detection and reporting since they kick in only after an outbreak has been announced. however, ex ante sanctions do not deter reporting and if anything they encourage reporting so that countries can signal that they are on top of their disease surveillance programmes. furthermore, ex ante sanctions based on fears of an undetected outbreak can reduce reliance on ex post sanctions as ways of controlling outbreaks. second, improving the quality of surveillance networks to detect outbreaks may not promote the disclosure of an outbreak because the forgone trade from reporting truthfully is that much greater. in sum, obtaining accurate information about potential epidemics is as much about incentives for reporting as it is about the capability and accuracy of surveillance networks. solving trans-boundary disease problems requires coordinated financing solutions, as has been evident with global eradication programmes. eradication of a disease means that it is no longer prevalent in any country in the world and requires elimination in every country. elimination, however, requires only the absence of the disease from a single country. global small pox eradication was largely paid for by the united states, even though countries like india stood to gain from the reduction in the number of deaths but were unable to achieve elimination on their own. however, the united states continues to recoup its roughly $ m investment in small pox eradication every days through not having to vaccinate its citizens against the disease. the optimal coverage with a vaccination programme of a disease that can be eradicated is given by p c = /( −r ) where r is the reproductive number of the disease-the number of secondary infections generated by a single infected patient entering a completely susceptible population. note that this critical rate of vaccination coverage depends only on the reproductive number (an epidemiological variable) and not on the costs of vaccination averted or any other economic variables. eradication may not be optimal in the case of all diseases, however. for diseases like measles, where the pathogen can be easily engineered through artificial methods and re-introduced into the population, there is no option of stopping vaccination. indeed, the current cohort of immunized individuals represents a valuable stock that is not easily replaceable in the short term. the optimal level of vaccination coverage of a disease for which vaccination must continue even after the disease has been eliminated can be computed as below. total costs to society include the costs of the vaccination campaign (vaccination costs), which we assume to increase exponentially with coverage, and costs of infection (infection costs) that we use as an index of the severity of the disease. the assumption that costs are increasing exponentially with coverage is consistent with the idea that reaching the most difficult to access and geographically remote populations involves increasing marginal costs. the total infection costs are proportional to the total number of the infective individuals in the population. because there is little evidence for increasing or decreasing marginal costs of infection within a single population (the change in total costs that arises from having one additional infection in the population), we assume constant marginal cost and model the costs of infection as a linear function of the infected. the total cost of the vaccination plus infection is then, with per capita burden c i . the cost of coverage is c(p) = ae xp , where a is the cost of vaccinating the first child (the cost of setting up the programme), and x captures the increase in costs with the increasing coverage p. when there is no immigration, we can calculate the economic optimum by minimizing eq. to find the level of coverage that minimizes total costs, which is independent of transmission. if the economic optimum p i is above the critical elimination threshold, p c = − /  , the optimal strategy is to eliminate the infection locally: (details in appendix in klepac et al. ( ) ). local elimination can be optimal also in the case of very severe diseases. in fact, for large enough per capita burden c i , i.e. the economic optimum p i is always above p c , and optimal vaccination coverage p* is reduced to the critical elimination threshold determined by  (eq. ). the optimal level of vaccination coverage for a disease that cannot be eradicated is a function of only economic parameters. indeed, epidemiological parameters play no role at all. local elimination is optimal only for low  values that result in a critical elimination threshold p c that is smaller than p i . moreover, adding immigration of infection to a single population precludes elimination by local vaccination alone. drug resistance is a global commons problem and covers the full range of infectious disease-causing pathogens from viruses, bacteria, fungi, and parasites through to disease vectors including mosquitoes, blackflies, and sandflies. resistance can arise in any single country and move globally. in this section, we focus on bacterial resistance and parasite resistance in the context of malaria. the global burden of resistance is poorly quantified but is likely to be concentrated in three major categories: increasing costs of resistant infections, increasing costs of antibiotics, and inability to perform procedures that rely on effective antibiotics to prevent infection. a primary burden of resistance is that resistant infections are more expensive to treat, and patients infected with resistant strains of bacteria are more likely to require longer hospitalization and face higher treatment costs than patients infected with drug susceptible strains (holmberg et al., ; the genesis report, ). an estimated , people die each year in europe from antibiotic-resistant bacteria (ecdc/emea joint technical report, ). in the united states in , an estimated , invasive methicillin-resistant staphylococcus aureus, or mrsa, infections required hospitalization and were associated with , deaths (klevens et al., ) . these estimates are useful for indicating the order of magnitude, but are imprecise because resistant infections are more common in individuals on long courses of antibiotic treatment: it is difficult to ascertain whether resistance is the cause of death or a correlate of long antibiotic treatment, hospitalization, and underlying sickness. in low-and middle-income countries, where the ability to pay for second-line drugs is limited, worse health outcomes are common, particularly in newborn children. even with effective antibiotics, neonatal infections are the major cause of neonatal deaths, which in turn account for more than a third of the global burden of child mortality (zaidi et al., ) . over half of neonates with extended spectrum beta-lactamase (esbl) sepsis are likely to die (versus a quarter of neonates with non-esbl infections), and a half of neonates with mrsa die (versus per cent of neonates with methicillinsensitive staphylococcus aureus) (kayange et al., ) . at these rates of mortality, one can estimate roughly , neonatal deaths attributable to gram-negative organisms and s. aureus, and , neonatal deaths attributable to esbl resistance and mrsa in india alone. a further cost of resistance is that associated with the cost of introducing new, expensive, antimicrobials to replace old ineffective ones (office of technology assessment, ) . this represents forgone resources that society could deploy elsewhere (reed et al., ) . according to one estimate, between and , increases in drug resistance raised the cost of treating ear infections by about per cent in the united states ($ m) (howard and rask, ) . resistance can also render broader health system functions such as surgeries, transplantations, and chemotherapy ineffective (laxminarayan et al., ) . a recent study estimated that, without effective antibiotics, - per cent of patients undergoing total hip replacements would have a postoperative infection, with a case-fatality rate of roughly per cent (smith and coast, ) . this category of burden affects both low-and middle-income as well as highincome countries and is likely to be the predominant way in which resistance drives up health care costs. take the case of drugs to treat malaria. the use of antimalarials places selection pressure on parasites to evolve resistance to these drugs. moreover, resistance is bound to arise when these drugs are misused, and could have adverse consequences for all malaria-endemic countries. efforts to manage resistance across national borders would have to rely on international agreements and regulations (walker et al., ) or on tax or subsidy instruments (arrow et al., ) . in the absence of such agreements and regulation, countries are unable to commit themselves to an optimal use of antibiotics, which would be in all countries' interests. at the macroeconomic level, a too intensive use of antibiotics in the health sector results in excessive levels of resistance both for that country and to the rest of the world (cornes et al., ) . a supranational authority would have to consider both the externality benefits of antibiotic use, in terms of reducing infections, and the costs, in terms of resistance (rudholm, ) . whether antibiotic consumption should be taxed or subsidized to reach the first-best outcome then depends on the relative magnitude of the externalities. in practice, the consequences of antibiotic use in sectors such as to make livestock grow faster involve little by way of positive externalities but impose resistance costs on other sectors and should therefore be taxed. a new class of antimalarial drugs, called artemisinins, requires a different way of thinking about optimal subsidies to manage resistance. when chloroquine, a oncepowerful antimalarial drug, became obsolete, the public health world was left with the challenge of optimally deploying the last remaining effective drug class, artemisinins. the who has recommended that artemisinins be used in combination with a partner drug that is unrelated to artemisinin's mechanism of action and genetic bases of resistance, so that a single mutation cannot encode resistance to both components (who, ) . artemisinin combination treatments (acts), if used instead of monotherapies of either artemisinin or the partner drug on its own, should slow the emergence of antimalarial resistance. however, the who guidelines are routinely flouted because monotherapies are much less expensive than acts. in response to this problem, an institute of medicine report (arrow et al., ) recommended establishing an international fund to buy acts at producer cost and resell them at a small fraction of that cost. on economic efficiency grounds there is a second-best case for subsidizing acts, because the ideal policy-taxing monotherapies and other antimalarials according to the marginal external cost from the elevated risk of the evolution of resistance-is infeasible, given their widespread use in the informal sector. the efficiency argument is further strengthened by the positive externality, to the extent that effective treatment of one individual reduces the risk of infection transmission to other individuals. laxminarayan et al. ( ) show that it is possible to determine the optimal subsidy in a dynamic diseasemodelling framework. bioeconomic analysis has been helpful for determining whether the social benefit from the subsidy, in terms of delayed resistance and saved lives, exceeds the social cost of resistance because of increased use of acts (laxminarayan et al., ) . it was also instrumental in turning an idea into the affordable medicines facility for malaria (amfm), a global financing system launched in early . amfm was formally evaluated in . in the six pilots where the programme was implemented to a substantial degree, amfm met or exceeded benchmarks for availability, price, and market share of quality-assured acts. in private, for-profit pharmacies, the quality-assured act market share at baseline ranged from to per cent (tougher et al., ) . a drawback of this evaluation was that it did not attempt to measure the impact on malaria prevalence or artemisinin resistance, both of which would have been difficult to ascribe to the intervention in the timeframe of the evaluation. nevertheless, the global fund to fight aids, tuberculosis and malaria made a political decision to discontinue amfm based on political objections raised by some country delegations (arrow et al., ) . one way to improve the efficiency of amfm resources was possibly to target children, though it would avert significantly fewer deaths. however, the benefits of a child-targeted subsidy (i.e. deaths averted) are eroded as leakage increases (i.e. older individuals taking young child-targeted doses), with few of the benefits (i.e. reductions in overall prevalence) of a universal subsidy (klein et al., ) . although potentially more cost-effective, a child-targeted subsidy must contain measures to reduce the possibility of leakage. most global health problems are 'commons problems'. therefore, it is often essential to have cooperative financing mechanisms for global health interventions, whether to eradicate disease, encourage appropriate levels of disease surveillance and reporting, or to reduce the likelihood of drug resistance. innovative financing that takes into account cross-country spillovers can play a critical role in arriving at globally optimal outcomes. for instance, in the case of the amfm subsidy, a high-level financing mechanism that lowers the cost of quality acts to all countries, including those that were at highest risk of using monotherapies, both enabled access to effective treatment and also reduced the threat of resistance. no bilateral financing solution could have achieved the same impact because of potential leakage to other countries, as discussed earlier. a global mechanism that is able to provide resources that incentivize surveillance and reporting of disease outbreaks can successfully counter the disincentives faced by countries for prompt reporting. again, bilateral assistance that simply focuses on subsidizing surveillance but does not pay attention to the lack of incentives for reporting cannot solve the problem. the three exemplars of trans-boundary problems that we have discussed can be applied to other global health problems with a public goods nature. saving lives, buying time: economics of malaria drugs in an age of resistance, board on global health the affordable medicines facility-malaria: killing it slowly global public health surveillance under new international health regulations the smallpox eradication game peru's neighbors halt food imports', the new york times drugs and pests: intertemporal production externalities the bacterial challenge: time to react the global value of vaccination strategies for containing an emerging influenza pandemic in southeast asia the recent history of malaria control and eradication global surveillance, national surveillance, and sars health and economic impacts of antimicrobial resistance studies of venereal disease. i. probenecidprocaine penicillin g combination and tetracycline hydrochloride in the treatment of "penicillinresistant the impact of resistance on antibiotic demand in patients with ear infections gbd cause patterns global trends in emerging infectious diseases predictors of positive blood culture and deaths among neonates with suspected neonatal sepsis in a tertiary hospital cost-benefit analysis of the onchocerciasis control program (ocp)', world bank technical paper cost-effectiveness analysis of childtargeted subsidies for artemisinin combination therapies using a bioeconomic malaria model invasive methicillin-resistant staphylococcus aureus infections in the united states malaria control by application of indoor spraying of residual insecticides in tropical africa and its impact on community health will a global subsidy of new antimalarials delay the emergence of resistance and save lives? extending the cure: policy responses to the growing threat of antibiotic resistance should new antimalarial drugs be subsidized? antibiotic resistance-the need for global solutions containing pandemic influenza at the source incentives for surveillance and reporting of infectious disease outbreaks operational strategies to achieve and maintain malaria elimination global spread of carbapenemase-producing enterobacteriaceae impact of antibiotic-resistant bacteria: a report to the us congress international health statecraft: foreign policy and public health in peru's cholera epidemic history of us military contributions to the study of sexually transmitted diseases socioeconomic issues related to antibiotic use economic implications of antibiotic resistance in a global economy malaria in south africa--the past, the present and selected implications for the future seven years of regional malaria control collaboration the true cost of antimicrobial resistance the economic impact of the cholera epidemic in peru: an application of the cost of illness methodology', water and sanitation for health project reduced vaccination and the risk of measles and other childhood infections post-ebola the real war on drugs: bacteria are winning effect of the affordable medicines facility-malaria (amfm) on the availability, price, and market share of quality-assured artemisinin-based combination therapies in seven countries: a before-and-after analysis of outlet survey data vaccines bring diseases under control', unicef, the progress of nations frequently asked questions about the international health regulations host range and emerging and reemerging pathogens temporal trends in the discovery of human viruses the economic impact of ebola on sub-saharan africa: updated estimates for hospital-acquired neonatal infections in developing countries epidemiology and cause of severe acute respiratory syndrome (sars) in guangdong, people's republic of china key: cord- -ts llerc authors: wang, qiang; shi, naiyang; huang, jinxin; cui, tingting; yang, liuqing; ai, jing; ji, hong; xu, ke; ahmad, tauseef; bao, changjun; jin, hui title: effectiveness and cost-effectiveness of public health measures to control covid- : a modelling study date: - - journal: nan doi: . / . . . sha: doc_id: cord_uid: ts llerc background the severe acute respiratory syndrome coronavirus (sars-cov- ) was first reported in china, which caused a respiratory disease known as coronavirus disease (covid- ). since its discovery, the virus has spread to over countries and claimed more than deaths. this study aimed to assess the effectiveness and cost-effectiveness of various response public health measures. method the stochastic agent-based model was used to simulate the process of covid- outbreak in scenario i (imported one case) and ii (imported four cases) with a series of public health measures, involving the personal protection, isolation-and-quarantine, gathering restriction, and community containment. the virtual community was constructed following the susceptible-latent-infectious-recovered framework. the epidemiological and economic parameters derived from the previous literature and field investigation. the main outcomes included avoided infectors, cost-effectiveness ratios (cers), and incremental cost-effectiveness ratios (icers). the sensitivity analyses were undertaken to assess uncertainty. findings in scenario i and ii, the isolation-and-quarantine averted and humans infected respectively at the cost of us$ and us$ , both with negative value of icers. the joint strategy of personal protection and isolation-and-quarantine could avert one more case than single isolation-and-quarantine with additional cost of us$ and us$ respectively. the effectiveness of isolation-and-quarantine decreased as lowering quarantine probability and increasing delay-time. especially in scenario ii, when the quarantine probability was less than %, the number of infections raised sharply; when the quarantine delay-time reached six days, more than a quarter of individuals would be infected in the community. the strategy including community containment could protect more lives and was cost-effective, when the number of imported cases was no less than , or the delay-time of quarantine was more than five days, or the quarantine probability was below %, based on current assumptions. interpretation the isolation-and-quarantine was the most cost-effective intervention. however, personal protection and isolation-and-quarantine was the optimal strategy averting more infectors than single isolation-and-quarantine. certain restrictions should be considered, such as more initial imported cases, longer quarantine delay-time and lower quarantine probability. as of march , , about cases of coronavirus disease have been identified in china. the global number of reported cases of covid- has surpassed and the confirmed cases of covid- have been reported in more than countries. as date, the th century has witnessed several large-scale outbreaks of infectious diseases caused by coronaviruses. the cases infected with covid- were significantly higher than ones infected with severe acute respiratory syndrome (sars) and middle east respiratory syndrome (mers). the statistic china, restrictions on gathering referred to the restriction of crowd-gathering activities, especially catering and entertainment. the enforcement of community containment was a restriction on the movement of people throughout the community, minimizing human contact. the incubation period and serial interval came from the estimation of chinese center for disease control and prevention (cdc) and guangdong provincial cdc in the field work, , and were considered fitting to the gamma distribution in the model. the parameter of distance transmission probability has been reported in previous study. the protective effectiveness of personal physical interventions derived from the cluster randomized controlled trial. in our study, we converted odds ratio (or) of handwashing and mask-wearing into the relative risk (rr), and calculated the ( -rr)/rr as the personal protection effectiveness. in the model, we set the probability and delay-time for isolation and quarantine. the isolation delay-time meant that the time of dealing with patients lagged behind the time of infection onset, and the quarantine delay-time meant that the time of handling close contact lagged behind the time of exposing. initially, we assumed that the index case (initial imported case) would be % isolated with no time delay (infecting others and isolation were carried out within the same day and infecting others preceded isolation). the quarantine probability was % and delay-time was two days. in the sensitivity analysis, the probability of quarantine of close contacts was set from % to % and the delay-time was from zero day to six days. the economic data derived from the field work and previous literature (table ). the cost of personal protection included masks and handwashing (water and soap). the price of the mask was us$ · each and we assumed that two masks were used per person per day. given the soap using, the cost of handwashing per person per day was calculated as the formula provided in the previous study: . cc-by-nc-nd . international license it is made available under a perpetuity. is the author/funder, who has granted medrxiv a license to display the preprint in (which was not certified by peer review) preprint the copyright holder for this this version posted march , . . https://doi.org/ . / where the costpp = cost of hand washing, f= times of hand washing per day, and we set to six, v= volume of hand washing per time, and we set to c.c/ml, cwater = water cost per liter, and was us$ · , csoap = cost of soap, and was us$ · , t = the number of days soap available, and we set to . we assumed the day of personal protection was equal to the time from the day first case occurred to the last case recovered in the area plus days. the cost of cases included the direct medical cost and indirect cost. we searched the cost of sars patients to estimate the covid- cases. in guangzhou, china, the average hospitalization cost per patient was us$ , and the average hospital stay was days. the average hospitalization cost achieved us$ in canada, which was higher than that in china. we estimated the average medical cost of us$ for covid- patient. referring to human capital approach in disease burden, we estimate that the indirect cost of infected patient using per capita disposable income (pcdi)/ · * (hospitalization days added rest days). the average rest days were estimated to seven days. we assumed that the cost of isolation would be included in the cost of hospitalization. the cost of quarantine of close contacts included direct and indirect parts. the cost of quarantine (accommodation and surveillance daily) per day was us$ for each close contact. similar to human capital approach in disease burden, is the author/funder, who has granted medrxiv a license to display the preprint in (which was not certified by peer review) preprint main health benefits of our study were avoided infections conducting measures versus no-interventions. the cost-effectiveness ratios (cers) and incremental costeffectiveness ratios (icers) were calculated as the main cost-effectiveness outcomes. we calculated the cers for interventions through cost divided by humans protected (uninfected) . the icers were calculated as the difference in the total costs between the intervention cohorts and non-intervention cohorts, divided by the difference in the total avoided infection. positive icers showed the incremental costs required for avoiding infected person. negative icers indicated that intervention results in fewer costs while avoiding infected people than no intervention. the strategy was considered to be cost-effective if icers were lower than three times of per capita gdp. in , the per capita gdp in china was us$ . we did not discount the cost because of the short one-and-two-way sensitivity analyses were performed to explore impact of the parameters in the range to test the robustness of the findings, including the epidemiological characteristics, interventions implement, and economic parameters. . cc-by-nc-nd . international license it is made available under a perpetuity. is the author/funder, who has granted medrxiv a license to display the preprint in (which was not certified by peer review) preprint the copyright holder for this this version posted march , . . introduction of one case, each strategy could avoid the number of infectors and be costeffective compared with no intervention (table ) . the isolation-and-quarantine was the most cost-effective intervention, avoiding cases and saving us$ (icers < ). the most protective single strategy was community containment, which avoided one more case than the isolation-and-quarantine at the additional us$ . among the joint strategies, there was the lowest ratio of cost-effectiveness for the program a (cers= us$/ per human protected). the program a could avert one more infector comparing to single isolation-and-quarantine. in scenario ii (table ) , compared with no intervention, personal protection or gathering restriction was not cost-effectiveness (icers > three times of per capita gdp). the isolation-and-quarantine was still the most cost-effective, avoiding cases and saving us$ (icers< ). compared with the isolation-and-quarantine, community containment could avoid one more case with the additional us$ . among the joint strategies, there was the lowest ratio of cost effectiveness for the program a (cers= us$/ per human saved). similarly, the program a versus single isolation-and-quarantine could avert one more infector. the number of infectors depended on transmission constant in scenario i (appendix ). varying the transmission constant from the · to two, the isolation-and-quarantine was the most cost-effective single intervention, and program a was the most cost-effective joint intervention. . cc-by-nc-nd . international license it is made available under a perpetuity. is the author/funder, who has granted medrxiv a license to display the preprint in (which was not certified by peer review) preprint the copyright holder for this this version posted march , . . https://doi.org/ . / the number of imported cases was a key parameter influencing the effectiveness and cost-effectiveness analysis. there were not significantly differences in effectiveness between the program a and c, when the imported cases were set to ten or (figure a and appendix table ). when the imported cases were no less than , the program c including community containment could effectively decrease the infectors than program a including isolation-and-quarantine, but the former was not cost-effective. the cers of interventions increased significantly as the increase of imported cases (figure a). the threshold analysis showed that program c became cost-effective (icers< three times of per capita gdp) comparing to program a when initial cases increased to imported cases (appendix table ). the isolation delay-time did not contribute to the spread of infections in scenario i (figure b). the increase of isolation delay time, however, caused a significant increase in the number of infections in scenario ii. when the isolation delay of four index cases reached four days, there were more than humans being infected, which was three times as the one without isolation delay. the cers of interventions increased as the increase of the isolation delay-day (figure b). the program a dominated the program c in scenario i and ii within the sensitivity analysis of isolation delay-time (appendix table ) . the effectiveness of isolation-and-quarantine was sensitive to the low quarantine probability. when the tracing probability of close contact was reduced to %, the number of people infected increased significantly, especially in the scenario ii ( figure c ). in scenario i and ii, the effectiveness of outbreak controlling was close between program a and c when the probability of tracing above % (appendix table and table ). the cers decreased as the increase of quarantine probability, and was most unstable when the quarantine probability was % (figure c). in scenario i, the program c was not cost-effective comparing to program a. the icers of program c was close to three times of per capita gdp when the quarantine probability was % in scenario ii. the is the author/funder, who has granted medrxiv a license to display the preprint in (which was not certified by peer review) preprint the copyright holder for this this version posted march , . . https://doi.org/ . / threshold analysis showed that program c became cost-effective (icers< three times of per capita gdp) comparing to program a when quarantine probability was below % (appendix table ). varying the quarantine delay time from zero day to four days, it had little influence on averting infected cases (figure d). when the tracing delay-time of close contacts was extended to six days, the number of people infected increased significantly (appendix table and table ). in scenario ii, when quarantine delay-time reached six days, there were likely more than humans being infected, accounting for a quarter in the space. the cers of interventions was unstable when the quarantine delay-time was no less than five days (figure d). comparing with program a, the program c was costeffective when the delay-time more than five days in scenario i and four days in scenario ii respectively (icers< three times of per capita gdp). varying the cost of patient from us$ to us$ , the cers of interventions increased and icers of interventions comparing to the non-intervention decreased (appendix table and table ) . the most cost-effective strategy was isolation-andquarantine in scenario i and ii. in scenario i, the effectiveness of outbreak controlling was not sensitive to the transmission constant and quarantine probability (appendix table ). when the transmission constant was set to two, the outbreak could be controlled by the % probability quarantine. however, as the transmission constant increased in scenario ii, the control of outbreak required higher quarantine probability. when the quarantine probability was % and transmission constant was two, it was likely about a quarter of people would be infected in scenario ii (appendix table ). the program a dominated the program c in the scenario i and ii in general. when the transmission constant was above one and the quarantine probability was below than %, the . cc-by-nc-nd . international license it is made available under a perpetuity. is the author/funder, who has granted medrxiv a license to display the preprint in (which was not certified by peer review) preprint the copyright holder for this this version posted march , . . https://doi.org/ . / probability could accelerate the outbreak of covid- . the effectiveness and cost-effectiveness of interventions were sensitive to the initial imported cases. the increase of imported cases could lead to the increase of risk of covid- infection, even conducting the strict interventions. we suggested that the infectors avoided by isolation-and-quarantine and community containment were not significantly when the imported the cases below . when the imported cases reached , community containment could avoid more cases significantly. the strategy including community containment was cost-effective when imported cases reached , the · % of the community population ( humans). the current article found that the initial number of cases had an effect on the effectiveness of interventions. the choice of optimal strategy depended on the setting parameter of interventions. we compared the strategy of personal protection and isolation-and-quarantine (program a) with strategy of personal protection and community containment (program c). generally, program a was cost-effective versus program c. however, the program c was cost-effective at the % probability and more than two quarantine delay-days, or % probability and no less than five quarantine delay-days in the sporadic outbreak area. the program c would dominate the program a at the % quarantine probability or quarantine delay-time was more than three days in the cluster area. the effectiveness of isolation and contact tracing was associated with the extent of transmission before symptom onset. the proportion of asymptomatic infection would contribute to the outbreak of covid- , which was consistent with our findings. in our study, the community containment would be more efficient and cost-effective when the quarantine delay-time was more than latent period. we suggested that increase of is the author/funder, who has granted medrxiv a license to display the preprint in (which was not certified by peer review) preprint the copyright holder for this this version posted march , . . https://doi.org/ . / there were some limitations in the study. first, covid- was recently emerged disease first reported in wuhan, china, therefore the availability of epidemiological data is insufficient. we set the study parameters referring to the existing published epidemiological studies and adopted the gamma distribution to some of the parameters, which could improve the precision of estimate. second, the cost of societal interventions was difficult to estimate. in our study, human capital approach was borrowed which might more conservatively estimate the cost. the cost of the disease would also increase, if according to the actual situation in wuhan, china. third, our model simulated a local area with humans, which may result in limited extrapolation ability. finally, the simplification of the model will have some biases compared with the real situation, because the flow of people will be affected by many factors. in the sporadic and cluster outbreak area, the isolation-and-quarantine was the most cost-effective intervention. the personal protection and isolation-and-quarantine was the optimal joint strategy averting more cases than single isolation-and-quarantine. rapid and effective isolation and quarantine could control the outbreak of covid- . the strategy including community containment could be more effective and costeffective when low probability and long delay of implements of interventions or much imported cases. . cc-by-nc-nd . international license it is made available under a perpetuity. is the author/funder, who has granted medrxiv a license to display the preprint in (which was not certified by peer review) preprint the copyright holder for this this version posted march , . is the author/funder, who has granted medrxiv a license to display the preprint in (which was not certified by peer review) preprint the copyright holder for this this version posted march , . . https://doi.org/ . / is the author/funder, who has granted medrxiv a license to display the preprint in (which was not certified by peer review) preprint the copyright holder for this this version posted march , . . https://doi.org/ . / is the author/funder, who has granted medrxiv a license to display the preprint in (which was not certified by peer review) preprint the copyright holder for this this version posted march , . . https://doi.org/ . / is the author/funder, who has granted medrxiv a license to display the preprint in (which was not certified by peer review) preprint the copyright holder for this this version posted march , . . https://doi.org/ . / is the author/funder, who has granted medrxiv a license to display the preprint in (which was not certified by peer review) preprint the copyright holder for this this version posted march , . . https://doi.org/ . / economic evaluation of the routine childhood immunization program in the united states key: cord- -d l sbeb authors: oberoi, sumit; kansra, pooja title: economic menace of diabetes in india: a systematic review date: - - journal: int j diabetes dev ctries doi: . /s - - -z sha: doc_id: cord_uid: d l sbeb aim: diabetes mellitus is recognised as a major chronic pandemic disease that does not consider any ethnic and monetary background. there is a dearth of literature on the cost of diabetes in the indian context. therefore, the present study aims to capture the evidence from the literature on the cost of diabetes mellitus in india. methods: an extensive literature was reviewed from academia, ncbi, pubmed, proquest, ebsco, springer, jstor, scopus and google scholar. the eligibility criterion is based on ‘picos’ procedure, and only those studies which are available in the english language, published between and february , indexed in abdc, ebsco, proquest, scopus and peer-reviewed journals are included. results: a total of thirty-two studies were included in the present study. the result indicates that the median direct cost of diabetes was estimated to be ₹ , /- p.a. for the north zone, ₹ , /- p.a. for the south zone, ₹ , /- p.a. for the north-east zone and ₹ /- p.a. for the west zone. similarly, the median indirect cost of diabetes was ₹ , /- p.a. for the north zone, ₹ /- p.a. for the south zone, ₹ , /- p.a. for the north-east and ₹ /- p.a. for the west zone. conclusion: the present study highlighted that diabetes poses a high economic burden on individuals/households. the study directed the need to arrange awareness campaign regarding diabetes and associated risk factors in order to minimise the burden of diabetes. electronic supplementary material: the online version of this article ( . /s - - -z) contains supplementary material, which is available to authorized users. 'diabetes is a metabolic disease characterised by hyperglycemia resulting from defects in insulin secretion, insulin action or both' [ ] . with rising pervasiveness globally, diabetes is conceded as a major chronic pandemic disease which does not consider any ethnic background and monetary levels both in developing and developed economies and has also been designated with the status of 'public health priority' in the majority of the countries [ , ] . individuals with diabetes are more susceptible to develop any of the associated complications, viz. macrovascular or microvascular. as a consequence, people experience frequent and exhaustive confrontation with the health care systems [ ] . the treatment cost for diabetes and its associated complications exert an enormous economic burden both at the household and national levels [ ] [ ] [ ] [ ] [ ] . in a developing nation like india, the majority of diabetes patients experience a substantial cost burden from out-ofpocket (oop). also, the dearth of insurance schemes and policies escalate the cost of diabetes care [ ] . instantaneous urbanisation and socio-economic transitions, viz. rural to urban migration, low exercise regimen, lifestyle disorder, etc., have resulted in an escalation of diabetes prevalence in india over the last couple of decades [ ] [ ] [ ] [ ] [ ] . according to the international diabetes federation [ ] , 'india is the epicentre of diabetes mellitus and it was found that in india had the second-largest populace of million diabetic patients, after china. and the figure is expected to be just double million by '. considering that fact, the epidemiologic transition of diabetes has a colossal economic burden [ ] . the estimated country-level health care expenditure on diabetes mellitus in india after amending purchasing power difference was billion us dollars in , pushing india in fourth place globally after the usa, china and germany. looking at the economic burden, in india, diabetes alone exhausts to % share of an average indian household earning [ ] [ ] [ ] . chronic nature and the rising epidemic of diabetes have everlasting consequences on the nation's economy and health status [ ] . therefore, managing diabetes and its comorbidities is a massive challenge in india due to several issues and stumbling blocks, viz. dearth of awareness regarding diabetes, its risk factors, prevention strategies, health care systems, poverty-stricken economy, non-adherence to medicines, etc. altogether, these issues and problems remarkably contribute to the economic menace of diabetes in india [ ] [ ] [ ] [ ] [ ] . after a perspicuous representation of the economic menace of diabetes in india, policymakers and health experts should provide healthier prospects to enhance the quality of life of millions [ ] . thus, the present study aims at capturing the evidence from the literature on the cost of diabetes mellitus in india, reviewing the materials and methods used to estimate the costs and, lastly, exploring future research area. for the accomplishment of the objective, the paper has been divided into five sections. the 'introduction' section of the study discusses diabetes and its economic burden. the 'materials and methods' section deals with materials and methods applied for data extraction and quality assessment. the 'results' section of the present study reports the results of the study. the 'discussion' section concludes the discussion along with policy implications and limitations. a comprehensive literature review was carried out by following the 'preferred reporting items for systematic reviews and meta-analysis (prisma) guidelines' [ ] . the article suggests a minimum set of guidelines and procedures of writing items to enhance the quality of the systematic review. a search was performed between february and march for the accumulation and review of studies published up to january . [ ] . later, articles were identified to be duplicate and removed immediately. of the total articles, limited studies managed to clear the eligibility criterion based upon the significant elements of the 'patient intervention comparison outcome study (picos)' procedure [ ] . title, abstract and keywords of the remaining studies were assessed to determine their relevance. those articles which have been included (a) were available in english language; (b) were published between and february ; (c) were indexed under abdc, ebsco, proquest and scopus; (d) were under journals that are to be peer-reviewed in nature; (e) highlighted unprecedented research outcomes on costs; and (f) were comprising at least one or more demographic zones. thus, the screening procedure facilitated the selection of articles. majority of research publications were excluded on the grounds if they (a) did not provide the detailed analysis of how costs were estimated; (b) were conference articles or posters; (c) only presented the costs of diabetes prevention; and (d) were published in non-peer-reviewed journals. the exploration includes those articles which highlight the cost burden of diabetes in india. whilst performing the analysis, two interdependent excel spreadsheets were developed for data to be summarised. in the very first spreadsheet, a predefined category was used, viz. publication title/year, study type, location, diabetes type, methodology and findings. relevant information is drawn out and presented in table , highlighting the study characteristics of the included articles. the second excel spreadsheet focuses its attention on the list of technical criteria applied to assess the quality of the articles incorporated in the review process. copious checklist has been put forward for the quality assessment of the included studies and majority of them emphasise on the economic assessment, viz. cost analysis, cost-benefit analysis (cba), health care utility analysis, etc. [ , ] . therefore, the quality indicators developed for the present study were grounded on the criterions suggested by prior literature [ ] [ ] [ ] [ ] . a symbol of (√) yes, (×) no and (±) moderately available was assigned to individual quality indicator. each symbol was allocated with a score of , which leads to a maximum attainable score of for each study reviewed. hence, a complete detailed analysis of the parameters utilised is presented in table . the characteristics of the included thirty-two studies are presented in table were included. the cost of diabetes was estimated from various locations such as the south zone (n = ), followed by the north zone (n = ), the north-east zone (n = ) and the west zone (n = ). a large proportion of studies ( %) were defined under india as a whole. whilst conducting review studies, it is imperative to initially define the type, study interest, sample size, data source and outlook of the study. the included studies majorly focus on type diabetes (n = ), followed by both type and type studies (n = ), studies were identified under type diabetes and only study was acknowledged under gestational/foot ulcer category, whilst the remaining studies did not define any diabetes type (table ). of the total studies, % of studies focus on general costs and the remaining studies emphasise on foot ulcers and others. whilst discussing the cost interests, the complications associated with diabetes were estimated by merely studies and the remaining studies ( %) estimated the diabetes cost without any complications. defining sample size is the utmost priority of the study, studies ( %) of the total studies have properly identified the sample size to be ≤ respondents, only studies specified the population size to be > respondents and studies ( %) did not define or provide the sample size. under the source of the cost data section, studies ( %) retrieved data on cost from the patients themselves; for studies ( %), source of cost data was obtained from medical institutes; and the remaining studies ( %) acquired the data on cost from publications. studies on the economic burden of illness could be done through several perspectives, viz. household, patient, societal and governmental. in the particular study, the patient's perspective was most commonly the research question of the study was mentioned? epidemiological definition such as type of diabetes ( and ) studied was provided? complications associated with diabetes were clearly stated? the location of the study respondent was clearly defined? the sample size of the study was adequate? acknowledged by studies ( %), studies considered societal perspective, followed by government perspective for studies and lastly, household perspective was adopted by studies as highlighted in table . the quality of the included studies is broadly presented in table . for all studies, research questions and findings were discussed and explained in a very well-defined manner. the presentation of the results was completely in synchronisation with the aim and conclusions derived from the reviewed articles. it was found that % ( ) of the studies have comprehensively defined the epidemiological definition such as type of diabetes (type and type ). limitations experienced by the majority of studies that hampered the quality of the reviewed articles were the absence of a broad definition of diabetes and a lack of adequate sample size. a major proportion of studies ( %) did not extensively define diabetes and studies ( %) moderately considered the sample size. for most of the reviewed articles, the sampling technique for data collection was addressed and only study did not define the sampling technique. however, % ( ) of studies lucidly defined the tools and technique employed in the reviewed articles and the remaining studies moderately describe the tools and technique. a majority of studies ( %) have properly classified the cost of diabetes and the remaining studies defined moderately. hence, based on quality index scores, the majority of the studies (n = ) scored ' yes' on a -point scale. interestingly, studies attained a marginally higher score of ' yes' of the total studies as presented in table . the economic burden of diabetes mellitus has led to numerous studies on the cost of illness. the cost exerted by diabetes can be categorised into three groups: direct cost, indirect cost and intangible cost [ , ] . direct cost includes both direct health care costs (diagnosis, treatment, care and prevention) and direct non-health care costs (transport, housekeeping, social service and legal cost) [ , ] . indirect cost includes cost for absenteeism, loss of productivity and disability [ , ] . lastly, intangible costs embrace cost for social isolation and dependence, low socio-economic status, mental health and behavioral disorder and loss of quality of life [ , , ] . all twenty-one reviewed studies put forward data and statistics to evaluate per capita cost of individual/household at zone level and the remaining eleven studies highlighted the cost of diabetes at the national level (table ) . to have a clear insight on cost, the reviewed articles have been categorised into four different zones, viz. north zone, west zone, south zone and north-east zone. were lucidly defined? . cost of diabetes was properly classified? the estimated annual direct cost was ₹ /individual and indirect cost was ₹ , including productivity and income loss through illness. ---- the mean total cost of diabetes in india accounts to ₹ /p.a. the mean direct cost of diabetes was ₹ /-and indirect cost, viz. hospitalisation, was /-p.a. (some regional differences in patterns of expenditure exist, with patients in the west of india likely to spend % more on laboratory fees, check-ups and medicines than any other region.) ---- the mean annual direct cost of treatment was ₹ /and % of amount is spent on drugs and medicines. the mean annual indirect cost of treatment was ₹ /of which . % was wage loss. ---- the total cost for diabetes management was ₹ /p.a. of which ₹ /was direct cost for the treatment of diabetes and ₹ /was spent on indirect cost. the total cost for treatment of diabetes with comorbidities was ₹ /p.a. the direct cost with complications was ₹ /p.a. and indirect cost amounts to be ₹ /p.a. ---- the mean direct cost of diabetes for consultation, lab investigation, medicines etc. was ₹ /monthly, whereas indirect cost for outpatient care was ₹ /monthly and indirect cost for inpatient care was ₹ per month. ---- the total average yearly direct cost was observed to be ₹ /-. however, the mean direct cost for all patients with diabetes was ₹ /p.a. individuals with three or more comorbidities encountered % more cost of care, amounting to ₹ , /annually. [ ] katam et al. the average total direct cost per patient annually was amounted to be ₹ , /-. the highest portion of direct cost was spent on insulin and glucose test strips ( %). ---- [ ] khongrangjem et al. the total median cost of illness per month was ₹ /-. total cost was made up of ₹ /direct cost and ₹ /indirect cost. ---- [ ] kumar et al. the total mean evaluation of annual direct spending on ambulatory diabetes care was ₹ /-. ---- [ ] kumar and mukherjee the total direct expenditure incurred on diabetes was ₹ , /p.a. and total indirect expenditure was ₹ , /p.a. ---- kumpatla et al. the total direct cost estimates without any complication were observed to be ₹ /-. the total cost of expenditure with complication was ₹ , /-. (cost for patients with foot complication was ₹ , /-, also average cost for renal patients under the north zone, studies were included to calculate both direct and indirect costs of diabetes at the individual/household level (fig. ) . the median direct cost of diabetes is estimated to be ₹ , /per annum, ranging from ₹ /to ₹ , , /- [ , , , , , [ ] [ ] [ ] . the most commonly measured costing items under direct cost were expenditure on medicines ( studies), diagnostic expenses ( studies), transportation cost ( study), hospitalisation ( studies) and consultation fee ( studies). the median indirect cost of diabetes for the north zone was evaluated to be ₹ , /per annum, ranging from ₹ / -to ₹ , /- [ , , , ] . for all indirect cost studies, costing items, viz. wage loss and leisure time forgone, were used majorly. south zone includes studies, majorly from karnataka state ( studies), followed by tamil nadu ( studies) and andhra pradesh ( study). the median direct cost was assessed to be ₹ , /per annum (fig. ) , ranging from ₹ /to ₹ , /per annum [ , - , , , , , , ] . direct costing items, viz. medicine cost ( studies), consultation fees ( studies) and hospitalisation ( studies), were used in the reviewed article. the median indirect cost of diabetes was ₹ /per annum, ranging from ₹ /to ₹ /per annum with major cost items such as monitoring cost ( study), absenteeism ( studies) and impairment ( study) [ - , , ] . under the north-east and west zone, only one-one study was observed, to evaluate the direct and indirect cost of author publication year was ₹ , /followed by , /-for cardiovascular disease.) [ ] ramachandran the average inpatient and outpatient cost of diabetes is ₹ /p.a. and ₹ /p.a. ---- [ ] ramachandran et al. the total median direct expenditure on health care was ₹ /p.a. ---- the mean cost per hospitalizations was ₹ /p.a. for diabetes. ---- rayappa et al. the direct annual cost (incl. hospital, test, monitoring etc.) was ₹ , /and indirect annual cost was ₹ /-. the total direct cost (incl. drugs, tests, consultation, hospital, surgery, transport) was ₹ /half yearly. the total direct cost (drugs and medicine) for diabetes patients was ₹ p.m. ---- [ ] thakur et al. the mean annual direct expenditure for diabetes care was ₹ and indirect cost was ₹ . ---- [ ] tharkar et al. the total direct cost for hospitalisation was ₹ , p.a. the total direct cost for hospitalisation with comorbidities was ₹ , /p.a. [ ] tharkar et al. the median annual direct cost associated with diabetes care was ₹ , and indirect cost was ₹ , respectively. ---- diabetes at the individual/household level [ , ] . the median direct cost of diabetes for north-east was evaluated to be ₹ , /per annum and ₹ /per annum was observed for the west zone (fig. ) . commonly estimated costing items were surgical procedures, expenditure on drugs/medicines, clinical fees, etc. the median indirect cost estimated for the north-east zone was ₹ , /per annum and ₹ /per annum was analysed for the west zone. indirect costing items identified for both reviewed studies were loss of wage, spendings on health class, travelling expenditure and spendings on diet control. lastly, studies were incorporated to estimate the cost of diabetes for india as a whole at the individual/ household level [ , , - , , , , - ] . the median direct cost of diabetes for india as a whole was ₹ /per annum, ranging from ₹ /to ₹ , /per annum. also, the median indirect cost of diabetes at the individual/ household level was estimated to be ₹ /per annum, ranging from ₹ /to ₹ , /annually ( figs. and ). diabetes mellitus is associated with a large number of serious and chronic complications, which act as a major cause of hospitalisation, morbidity and premature mortality in diabetic patients [ , , , ] . diabetes mellitus is commonly associated with chronic complications both macrovascular and microvascular origin [ , ] . microvascular complications of diabetes mellitus include retinopathy, autonomic neuropathy, peripheral neuropathy and nephropathy [ , ] . the macrovascular complication of diabetes mellitus broadly includes coronary and peripheral arterial disease [ , ] . of the total reviewed studies, only studies estimated the cost of complications associated with diabetes (table ) . a couple of studies on diabetes assessed the cost of illness to be . times higher for individuals with complications as exhibited in table [ , ] . a similar study by sachidananda et al. [ ] concluded that the cost of diabetes is . times higher for complicated non-hospitalised patients and . times higher for complicated hospitalised patients. kapur [ ] inferred that individuals with three or more comorbidities encounter % more cost of care, amounting to ₹ , /annually. according to cavanagh et al. [ ] , india is the most expensive country for a patient with a complex diabetic foot ulcer, where . months of income was required to pay for treatment. three reviewed studies incorporated in the study estimated the cost of individual/household with both macrovascular and microvascular complications [ , , ] kansra [ ] , as suggested by moher et al. [ ] the cost of illness prompted by renal (kidney) complication [ , ] . lastly, eshwari et al. [ ] estimated the total cost for the treatment of diabetes with comorbidities was ₹ /annually. direct cost with complications was ₹ /per annum and indirect cost amounts to be ₹ /annually. rising menace of diabetes has been a major concern for india. with a frightening increase in population with diabetes, india is soon going to be crowned as 'diabetes capital' of the world. a swift cultural and social alteration, viz. rising age, diet modification, rapid urbanisation, lack of regular exercise regimen, obesity and a sedentary lifestyle, will result in the continuous incidence of diabetes in india. the primary objective of this article is to detect and capture the evidence from published literature on the per capita cost at the individual/household level for both direct and indirect costs of diabetes in india which are available and published since . of the total records, studies were identified to meet the inclusion criterion. therefore, the findings of the present study suggest that per annum median direct and indirect cost of diabetes at the individual/household level is very colossal in india. a large proportion of health care cost is confronted by the patients themselves, which affects the fulfilment of health care because of financial restraints [ ] . the proportion of public health expenditure by the indian government is the lowest in the world. as a consequence, out-of-pocket (oop) spending constitutes to be % of the total health expenditure. hence, financing and delivering health care facilities in india is majorly catered by the private sector for more than % of diseases in both rural and urban areas [ ] . direct cost items (expenditure on medicines, diagnostic expenses, transportation cost, hospitalisation and consultation fee) and indirect cost items (loss of wage, spendings on health class and travelling expenditure) were most commonly reported costing items in the present study [ , , , , , ] . most of the reviewed studies on the cost of diabetes highlighted expenditure on drugs/medicine as the foremost costing item which accounts for a significant share of all direct costs. the finding of the present study is consistent with yesudian et al. [ ] , 'cost on drugs constitutes % of the total direct costs'. the majority of the reviewed articles included in the study justify that the primary cause for such abnormal costs of medicines is the common practice adopted by physicians to prescribe brandnamed medicines, rather than generic medicines. in context to the quality of tools and techniques incorporated by the included studies, a large number of articles ( %) witnessed to acknowledge the standards of tools and techniques. similarly, the classification of the cost of diabetes was also determined by the majority of reviewed articles ( articles) . but the absence of a comprehensive definition of diabetes and a small size of individuals/ households produce dubiousness about the standards or quality of the study. hence, the limitations experienced by the majority of reviewed articles hampered the quality of the present study. thus, it is beneficial to develop and suggest standard procedures and framework to conduct a comprehensive and exhaustive study on the cost of diabetes. the present study holds few limitations. primarily the exclusion of the relevant articles presented as conference papers and those studies published under nonpeer-reviewed journals. with the omission of the above literature, some biasness might have been introduced into the review process. furthermore, the major limitation of the present study is the non-availability of published articles under the central and east zone of india. also, the studies published under the north-east zone and west zone were only one. lastly, the heterogeneity in material and methodology used in cost estimation are not analogous. as a consequence, conducting a metaanalysis is not feasible. the above discussion highlighted a huge economic burden of diabetes in india and variations were recorded in the different zones. it was observed that the cost of drugs/medicines accounts for a major burden of the cost of diabetes. the study suggested few policy interventions to cope with the high economic burden of diabetes. there is a dire need in the country to arrange awareness programmes on diabetes and associated risk factors. the menace of diabetes can be controlled by devising new health care policies, introducing new generic medicines and taxing alcohol/tobacco. diabetes is a lifestyle disease so along with the above measures, a change in dietary habits, physical activity, beliefs and behavior can reduce its economic burden. conflict of interest the authors declare that they have no conflict of interest. ethical approval the study is a review-based study, so it does not contain any studies with animals. the present study only reviews those studies which contain individual's performance. informed consent for the present study, it is not necessary to obtain any consent. diagnosis and classification of diabetes mellitus the costs of treating long term diabetic complications in a developing country: a study from india diabetes mellitus and its complications in india who multinational study group. mortality and causes of death in the who multinational study of vascular disease in diabetes cost of treating diabetic foot ulcers in five different countries dispensing patterns and cost of glucoselowering therapies in diabetes mellitus patients at a tertiary care hospital in southern india health-care cost of diabetes in south india: a cost of illness study assessment of cost of illness for diabetic patients in south indian tertiary care hospital annual cost incurred for the management of type diabetes mellitus-a communitybased study from coastal karnataka trends in prevalence of diabetes in asian countries time use and physical activity: a shift away from movement across the globe difference in prevalence of diabetes, obesity, metabolic syndrome and associated cardiovascular risk factors in a rural area of tamil nadu and an urban area of delhi economic vulnerability to health shocks and coping strategies: evidence from andhra pradesh, india. health policy plan prevalence of risk factors for non-communicable diseases in rural and urban tamil nadu international diabetes federation: idf diabetes atlas idf diabetes atlas: global estimates of diabetes prevalence for and projections for the economic burden of diabetes in india: a review of the literature awareness about diabetes among adolescents of mangalore economic burden of diabetes among women: a comparison of outpatient and inpatient care problems associated with diabetes care in india current scenario of diabetes in india the socioeconomics of diabetes from a developing country: a population based cost of illness study socio-economic inequalities in the financing of cardiovascular and diabetes inpatient treatment in india cost of diabetic care in india: an inequitable picture preferred reporting items for systematic reviews and meta-analyses: the prisma statement spider: a comparison study of specificity and sensitivity in three search tools for qualitative systematic reviews methodological considerations in cost of prostate cancer studies: a systematic review cost of illness for bipolar disorder: a systematic review of the economic burden cost-of-illness methodology: a guide to current practices and procedures cost of illness studies: what is good about them? cost-of-illness studies cost of diabetes mellitus in africa: a systematic review of existing literature economic cost of diabetes care: the bangalore urban district diabetes study increasing expenditure on health care incurred by diabetic subjects in a developing country: a study from india cost of ambulatory care of diabetes mellitus: a study from north india socio-economic burden of diabetes in india excess cost burden of diabetes in southern india: a clinicbased, comparative cost-of-illness study economic analysis of diabetes care burden of healthcare utilization and out-of-pocket costs among individuals with ncds in an indian setting expenditure on health care incurred by diabetic subjects in a developing country-a study from southern india economic burden of diabetes in urban indians health care cost incurred by patients of diabetes mellitus in a tertiary care hospital setting in coastal karnataka district global policy: aspects of diabetes in india direct cost of ambulatory care of type diabetes in the middle and high income group populace of delhi: the dedicom survey costs incurred by families having type diabetes in a developing country-a study from southern india expenditure pattern on diabetes care: a community based longitudinal study in resettlement colony of east delhi cost of illness (coi) of type-ii diabetes mellitus in shillong, meghalaya drug prescription patterns and cost analysis of diabetes therapy in india: audit of an endocrine practice economic impact of type- diabetes mellitus on households in hisar district of haryana state high direct costs of medical care in patients with type diabetes attending a referral clinic in a government-funded hospital in northern india the economic burden of diabetes in india: results from a national survey cost of medical care among type diabetic patients with a co-morbid conditionhypertension in india direct costs associated with chronic kidney disease among type diabetic patients in india factors influencing medical tourism in india: a critical review costs of diabetes the economic burden of depression in sweden from economic costs of diabetes in the us in a prospective study of outcomes, healthcare resource utilization, and costs associated with postoperative nosocomial infections the prevalence of multi-morbidity in primary care and its effect on health care utilization and cost the burden of epilepsy for the patient: the intangible costs economics of psychosocial factors in patients with cardiovascular disease the economic burden of diabetes in india: a review of the literature publisher's note springer nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations key: cord- -r ljahss authors: gudmundsson, sveinn vidar; merkert, rico; redondi, renato title: cost structure effects of horizontal airline mergers and acquisitions date: - - journal: transp policy (oxf) doi: . /j.tranpol. . . sha: doc_id: cord_uid: r ljahss the purpose of this research is to test the ex-post cost structure effects in horizontal mergers and acquisitions (m&a). our proposed methodology quantifies cost structure effects empirically to inform competition policy around m&as in the airline industry. the results show that horizontal m&as involving unprofitable firms significantly reduce variable costs and increase fixed costs ex-post. m&as involving only profitable firms show no significant impact on the cost structure. we offer support that the ex-post cost structure effects of airline m&as depend on the incentives to improve efficiency, reflected in the ex-ante performance of the merging firms. we further argue that market behavior may not just depend on market structure but cost structures too, all of which should be accounted for in antitrust decision making and regulation around airline m&as. mergers and acquisitions (m&as) constitute one of the key competitive strategies for firms, having global worth in excess of $ . trillion in (statista, ) . despite extensive literature documenting meager efficiency effects of m&as (bauer and matzler, ; datta, ; gates and very, ; ingham et al., ; lubatkin, ; seth, ; vaara et al., ) , managers readily embrace this strategy, policy makers such as antitrust authorities keep approving such transactions without having full information (for the airline context see for example, nannes, ) and academics are increasingly focusing on explaining their variance rather than overall performance (keil et al., ) . from a competition policy perspective, airline m&as are often approved by antitrust authorities and regulators (i.e. in the us) despite evidence suggesting that the proposed transaction would reduce the level and intensity of competition in this market. this happens because the antitrust authorities focus not only on consumer welfare but also employment and hence jobs by allowing a failing airline to merge (or be acquired by) a financially stronger airline (merkert and morrell, ) . we argue that ensuring healthy levels of competition will work best if the resulting (merged) entity is cost competitive and this paper aims to show that under certain conditions merging two airline firms can strengthen the cost position of the merged entity, assuming there are remaining players in the market ex-post. it would hence be justifiable for the regulator in certain circumstances, as highlighted in this research, to approve the transaction despite the reduced level of competition. williamson ( ) argued that mergers can affect variable costs (vc) and improve the firm's relative competitiveness despite an increase in fixed costs (fc) . we therefore predict that changes in vc and fc following m&as occur but the direction of these changes affect the competitiveness of the firm differently. our argument is that the direction of the cost changes depends on conditions surrounding ex-ante profitability of the firms and therefore management incentives to improve efficiency. efficiency improvement is not only useful for airline management to compensate for costs outside of their control ) but may also help or hinder airlines to compete, having implications on competition policy in the relevant jurisdiction (schnell, ) . there is a large body of extant literature on market power considerations in airline policy (e.g. oliveira and oliveira, ; manuela jr et al., ) demonstrating that an increase in market power is a potential downside of m&a transactions. in this present paper the focus is on the internal cost effects and we do not try to address revenue growth, market power or private benefit effects that may occur in parallel and which are important to competition policy. ample research already exists on these aspects due to the data being readily available, whereas cost structure effects have been largely neglected due to a lack of detailed cost data (e.g. gayle and le, ; jeziorski, ) . in our paper, the cost structure is seen as fixed with vc varying before/ex-ante and after/ex-post a specific m&a transaction. this is important as extant m&as studies have generally focused on total costs, or total factor productivity (tfp) but not explored specific cost types. most of this research found negligible or even negative gains (caves, ; gudmundsson et al., ; king et al., ; scherer, ; tichy, ) , while only few revealed positive gains (capron, ; chatterjee, ) , including studies using tfp analysis that showed positive gains following m&a in the airline industry (yan et al., ) . m&as scholars have as a consequence of these diverse results called for research using new approaches and detail industry data (meeks, ; moatti et al., ; tichy, ) . thus, the key contribution of our research is to illustrate how horizontal m&as affect the cost structure of the firm which has implications on strategies of airlines and competition policies of regulators. our specific contribution is to isolate the effects on vc and fc and therefore offer insights under what conditions the merged firm can strengthen its cost position in an industry (porter, ; powell, ; röller et al., ; williamson, ) which is of particular interests to situations where the merger is aimed to protect a financially ailing carrier. in the airline context vc tend to be volume-related costs as fuel, landing fees, catering, crew expenses (travel, hotel and per diem), crew salaries (if paid per flight hour), maintenance (most airlines consider maintenance costs as vc with the exception of annual inspections, scheduled overhauls and avionics updates). fixed costs fc tend to be time-related, such as salaries or rents, and are referred to as overhead costs. if an airline stops operating a specific flight, such as during the covid- crisis, it would continue to incur time related costs like rental, salaries, interest expenses, depreciation, and insurance expenses. our key contribution to the m&as literature is therefore filling a research gap pertaining to the varied effects of m&as on cost types. in what follows we provide the literature background for our study and the predicted relationship between the cost structure effects of m&as and efficiency improvement incentives of merging firms. then we describe the cost disaggregation measurement approach, followed by the results. we conclude the paper by discussing the academic and practical implications of our findings. in part we base our arguments in this paper on williamson ( ) , deliberating the competition effects of changes in the cost structure following m&as, and sitkin and pablo ( ) as well as lander and kooning ( ) when we argue that merger incentives and the conditions surrounding the initiation of the m&a process has a significant effect on the cost structure effects. as pointed out before, the extant literature is meager on studies focusing on vc following mergers, principally due to lack of data, but also suggesting a complex set of conditions that are necessary a priori, hindering all-encompassing generalizations. for example, scholars have pointed out that m&as are guided by external industry conditions (gort, ; white, ) that enable or hinder welfare enhancing effects of m&as, while mergers that generate efficiencies are likely to induce entry to an industry (werden and froeb, ) . in the background of this literature, we develop our cost structure arguments taking into account different types of m&as guided by specific conditions surrounding target selection and timing that affect the ex-ante performance disparity of the merging firms and therefore efficiency improvement incentives ex-post. for example, the initial delta -northwest merger discussions took place in the background of soaring oil prices and crashing post-bankruptcy share-prices. hence, understanding merger conditions and their merger impact have been suggested in order to preevaluate mergers and potential ex-post efficiency gains (xuejie et al., ) . the classic arguments for m&as are efficiency enhancements and market power (e.g. survival, competitive advantage, market-share, diversification, etc.). for example, managers may seek mergers to realize efficiency related synergies to prevent bankruptcy (shrieves and stevens, ) or to preempt competitors from acquiring strategically important firms in the industry (fridolfsson and stennek, ) . having these arguments in mind and our focus on the cost structure, the extant literature suggests that m&as cost effects are shaped by several conditions that are not necessarily exclusive. the first condition is relatedness and complementarity (yu et al., ; wang and zajac, ) . thomas ( ) found that integration costs are lower if both firms run similar operations, because they can better optimize and join together complementary resources for value creation, findings that were later supported by the work of wang and zajac ( ) . drawing on this work, we assume that horizontal mergers in the same industry involving both relatedness and complementarity should generate positive value in the form of lower costs, although in combination with the other conditions that follow here. the second condition is a performance deficit that generates stronger incentive to improve management processes (bruton et al., ) . in support of this condition we draw on leibenstein's ( ) work on value creation from superior management processes in m&as and later research that came to similar conclusions (berger and humphrey, ; deyoung, ) . these value creation effects have been found to exist when m&as involve unprofitable target firms (hotchkiss and mooradian, ) . if the acquiring firm is unprofitable it, may seek a profitable target to obtain superior management processes or some tangible assets (arikan, ) . in the case of two unprofitable firms the incentive to act and impact costs should be no different from the other two scenarios we just described. therefore, in all three cases a potent incentive should exist to improve efficiency of the ex-post merged firm. the third condition is adverse industry conditions. early research considered economic disturbance or shocks as an influencing factor both in the timing and intensity of m&as (gort, ; jensen, ) . under these circumstances, observant managers may assume that if two competitors merge, profits of their firms will decline because of outsider disadvantage (trautwein, ) . they therefore attempt to preempt their competitors from acquiring or merging with strategically important firms. in fact, in the context of our focus on the cost structure, fridolfsson and stennek ( ) argued that preemption exists, especially if m&as are likely to enhance the ability of the firm to compete through lower prices (shifting focus to vc reduction) and by inflicting negative externalities on merger outsiders. m&as vary in terms of the incentive to seek efficiency improvement ex-post. for example, firms in competitive industries with many players (e.g. airlines) have higher efficiency improvement incentive (hay and liu, ) than firms in less competitive industries with fewer players (e.g. aircraft manufacturing). in the same vain, unprofitable firms have higher efficiency improvement incentive than profitable firms. leibenstein ( ) in his work on x-efficiency, used the term incentive efficiency, to reflect the motivation to act, and we use the term here to underline the presence of an incentive to act when there is a performance gap. thus, in the present research we draw a parallel between ex-ante profitability of merging firms and the incentive to achieve specific cost effects (cost reduction) ex-post. managers of profitable firms have lower incentive to secure efficiency improvements ex-post compared to those of unprofitable firms. hence, in the former case m&as are largely motivated by other reasons than efficiency, such as market power or private benefit. to clarify our assumptions, we draw up in table the ex-post incentive matrix influencing efficiency improvement actions in m&as coming from the ex-ante profitability of the firms. as is clear from the table we assume that potent incentives for efficiency improvement exist in all ex-ante profitability combinations, except when two profitable firms are involved. according to our framework, such m&as are less likely to involve potent cost efficiency improvement incentives and affect the cost competitiveness of the ex-post firm. scholars have argued that external shocks to an industry create opportunities for better performing firms to add management value by acquiring and restructuring poorer performing firms (andrade et al., ; gort, ) . such mergers may preempt competitors from seizing firms with superior combination potential (fridolfsson and stennek, ) . however, this research departs from the view that mergers between large firms are predominantly a preemptive tactic against take-overs (gorton et al., ) and exclusive of efficiency motives. along these arguments we attempt to reveal evidence supporting our prediction that horizontal m&as involving unprofitable firm(s) have stronger management incentive for efficiency improvement ex-post than if both firms are profitable, which is important for competition policy and relevant to regulators approving mergers. in other words, we anchor our arguments in the efficiency domain by building upon the assumption that efficiency outcomes of m&as are not only affected by the ex-ante profitability of the firms themselves but also the variance in the average profitability of the industry itself, i.e. opportunities to acquire or merge with unprofitable and undervalued firms during industry downturns. incentive asymmetries are common in m&as (parvinen and tikkanen, ) . for example, executives of better performing firms may be motivated to pass on superior management processes to poorer performing firms and create value ex-post (akhavein et al., ; leibenstein, ; berger and humphrey, ; deyoung, ) , while executives of unprofitable firms, facing loss of confidence, may be motivated to consider bold strategic actions like becoming a target to restore confidence and bring in external resources (hitt et al., ; morrow et al., ; weitzel and jonsson, ) . according to the work of graves ( ) and napier ( ) employees of unprofitable firms are more prone to accept changes that bring about efficiency improvements, than employees of profitable firms. these contributions emphasize that unprofitable firms develop employee dissatisfaction that can be positively reduced following m&as through rationalization (reduced slack), more work to do (surge in activity to tackle inefficiency), increased scope and variety of work (graves, ) , and a feeling of increased job security (napier, ) . further elaboration on this theme highlights the relationship between employee satisfaction, efficiency motivation ex-post, and potential productivity increases to bring costs down (hotchkiss and mooradian, ; judge et al., ) . assuming that the ex-ante performance asymmetry of firms boosts incentives for efficiency and therefore the motivation to reduce costs ex-post brings us to the competitive implications in an industry. airlines were generally considered as fixed costs operations before the rise in fuel prices starting in the s. however, in today's context airlines face volatilities on both the demand and supply side, pushing vc to the fore. as mentioned earlier, williamson ( ) argued that reduction in vc can greatly improve competitive advantage of the firm if cost reductions are passed directly on to customers through lower prices, a point accentuated in conceptual studies (roller, stenek and verboven, ) independent of specific industries and case studies (ryerson and kim, ) on m&as in the airline industry that have shown that vc like fuel consumption are affected by fleet and network rationalization following mergers. harmonizing schedules of two merging airlines and rationalizing frequencies helps improve load-factors and increase average aircraft sizes (through density and scope economies), reducing vc per output unit. following m&as, airlines can also return aircraft to lessors, and sell or park inefficient aircraft to bring about reduction in vc (maintenance and fuel costs per asm). for example, in , based on information from airsafe (www.airsafe.com), the average fleet age of northwest airlines was . years and delta air lines . years. following the merger the combined fleet age was around years, but not evenly distributed across fleet sizes. for example, for narrow body aircraft the average age of the fleet went from years in to years in , and for wide body aircraft it went from . years to . years over the same period (airinsight, ) . having established the strategic importance of a reduction in vc, explanations must be sought why total costs may, on average, not be affected? one explanation rooted in the dynamics of costs is derived from chandler ( ) , arguing that when firms reduce average vc, fc may rise from acquiring more efficient technologies. for example, a firm aiming at reducing variable input costs, such as fuel consumption, can acquire more fuel efficient vehicles and realize cost reduction per output unit, but may incur higher fc associated with acquiring new vehicles. for example, a firm j o u r n a l p r e -p r o o f that decides to lease equipment would incur a reverse trade-off, namely an increase in vc and reduction in fixed costs. delta air lines, in the s, was a case in point, by choosing to operate and acquire older aircraft, resulting in lower fixed costs but higher vc (fuel and maintenance). this model allows quicker adjustment in supply to match demand, and therefore avoid the burden of higher fixed costs. however, a large increase in fuel prices would, using this strategy, trigger disproportionate increase in vc compared to a carrier with a younger fleet. reverting to fc, case-based evidence shows, that despite improvement in operations, the management may attempt to influence employee buy-in of m&as causing increases in employee related fixed cost. several studies have provided support of this analogy pointing out that merging firms tolerate duplication of activities and may harmonize payscales upwards to engage staff and achieve smoother integration ex-post (gayle and le, ; huck et al., ; prechel et al., ) . underlining that a firm, being acquired, commanding valuable resources (often a profitable firm) is more likely to be in a strong bargaining position causing upward shift in employee related fc. in addition, m&as do not just make firms larger, but also more complicated, moving from a simple to a complex organizational form (prechel et al., ) accelerating decreasing returns to size (merkert and morrell, ) . another important consideration is the role of the preemption motive on the cost structure following m&as. firms preempting competitors from merging with strategically important firms, often pay acquirer premium affecting the profitability of the ex-post firm (molnar, ; slusky and caves, ) . preemption holds that despite a loss in profitability due to a rise in fc, it is still a better position for the firm than if a rival merged with the target firm. hence, it follows that the ex-post merged firm has an incentive to offset such premium. this premise is supported by acquiring firms being more prone to pay acquisition premiums when expectations of post-merger synergies are higher (agarwal and kwan, ; slusky and caves, ) . if an efficiency improvement in vc is realized ex-post, it forges cost competitiveness of the merged firm in an industry. hence, we predict that managers accept, ex-post, an increase in fc (merger costs, acquirer premium, etc.) if anticipating reduction in vc. in contrast, if both firms are profitable either management team is less adept to seek efficiency improvements and if trying, employee tensions may rise as diverse but "good" management practices and processes compete for retention ex-post. supporting this, past research has shown that m&as among two profitable firms are more likely to focus on the market value of the firm, market power, financial synergies or asset divesture (capron et al., ; fluck and lynch, ; mueller and sirower, ; rubinowitz, ; walter and barney, ) , rather than efficiency. fluck and lynch ( ) argued that if firms harvest valuable intangible assets, they may avoid coordination costs by divesting much of the tangible assets of one firm. building on this aforementioned work we contend that divestment is more likely to occur in m&as if both firms are profitable and intangible assets valuable. if the target controls valuable resources sought by the bidder and the acquirer is bidding for a firm in a strong bargaining position it is likely to j o u r n a l p r e -p r o o f pay a premium (capron and pistre, ) . in such mergers, absorption of valuable complementary resources takes place followed by the divestiture of redundant assets (capron and pistre, ) or even part of the workforce if the value is vested in tangible rather than intangible assets (krishnan et al., ) . hence, we predict that m&as involving profitable firms are different from m&as involving unprofitable firms and are likely to involve both reduction in fc (asset divestiture) and an increase in vc as the incentive to improve efficiency is lacking at the same time that merger related costs and inefficiencies are sustained. our predictions are tested on a sample of firms engaged in horizontal m&as in the airline industry (a highly competitive industry). a merger between firms that belong to the same industry and sell the same products is considered horizontal. since significant horizontal mergers can reduce competition and are scrutinized by the competition authorities, the cost structure effects of such mergers pose a particular interest to m&a scholars. we conduct our modeling in two stages. first, we decompose costs into variable and fc. then we test our predictions about the two merger types, potent efficiency improvement incentive versus weak efficiency improvement incentive. this second stage helps identify if prior profitability typologies are associated with the distinctive ex-post incentives to realize m&as efficiencies affecting the cost structure. our sample of international airlines has inherently similar strategic fit characteristics, a sampling approach suggested by lubatkin ( : ) . the airline industry benefits from a relatively uniform international statistics program standardized by the international civil aviation organization (icao). what is more, icao sets worldwide standards regarding operations and safety making airline operations within and between countries highly homogeneous, process wise. however, airlines in most countries around the world are considered a national strategic resource, similar to the defense industries, precluding or limiting foreign direct investment and cross-border mergers. thus, all but one of the m&as in our dataset are considered single country or single market (eu) m&as. in addition, icao issues international regulation and standards to countries around the world that assures high degree of relatedness in airline operations from one country to another and one airline to another. what is more, anti-trust authorities scrutinize m&as in the airlines mostly on network overlaps, implying that airlines are concerned with complementarity for both strategic (shaw and ivy, ) and policy reasons (morrison, ) . thus, we deem the combination potential among firms in our dataset inherently high because of joint standards (relatedness/fit) and network separation (complementarity) (wang and zajac, ) . what is more, m&as with high degree of network overlap are rare and routinely challenged by anti-trust authorities. gayle and le ( ) point out the difficulties finding decomposed cost data for firms explaining the scarcity of studies focusing on the effects of m&as on the cost structure. in fact, previous research decomposing vc and fc effects used a cost estimation method not requiring actual cost data (i.e. gayle and le, ). what makes our paper different and original is testing our predictions with actual cost data from the international airline industry. the data was collected from annual reports and government statistical sources spanning the period to . table illustrates the m&as considered in our analysis, including information on the year of the transaction, the geographical regions involved, and the number of years under observation. to estimate the necessary sample size for our study we used an approach suggested by overall and doyle ( ) taking into account its longitudinal design. in our sample size analysis we assumed a range of statistical power levels (cohen, ) , two groups, time-points, and a % probability level, resulting in a minimum required sample sizes of firms for power level . , for level . , and for level . . since our sample was composed of ( firms) firm-year observations our study was in the proximity of power level . . as the firms do not exist throughout the data collection period, we worked with an unbalanced sample. however, given the longitudinal characteristics of the sample we capture both the timing of m&as and external industry conditions. although the study can be considered a small sample study the effect sizes (cohen, ; lakens, ) are large. the sample firms were first analyzed as a single group and then divided based on ex-ante profitability. as mentioned before the profitability segregation was based on the assumed strength of incentives to improve efficiency. a further justification in the context of m&as is that the airline industry is characterized by intense rivalry depicted by a steady decline in revenues per output unit over many decades implying that average prices decline along with average costs (maillebiau and hansen, ) . in other words, cost savings are reflected in prices. what is more, the industry shows constant returns to scale (gillen et al., ), a relationship that holds following an instant enlargement through m&as (gudmundsson et al., ) . these industry characteristics, high rivalry and lack of economies to scale, cause high fluctuations in industry profitability (and variance in the number of unprofitable firms) over time, induced by industry shocks like the aftermath of deregulation, changes in input costs, and innovations (mitchell and mulherin, ) . the dependent variable we use in our models is total annual operating costs of the firm, composed of total fc and total vc. the arguments so far are that different types of m&as will cause different cost structure effects ex-post, in other words, a different impact on fc and vc. however, the line between fixed and vc is often blurred and information from the profit and loss accounts not always specific enough to permit this separation. as a result, we employ a regression approach to separate fc and vc (hansen et al., ) . in this way, our approach estimates fc independent of the firm's annual unit output, typically including overhead and administrative costs. meanwhile, vcs are dependent on output, typically associated with serving one more customer (including a fuel cost component in our industry setting). the model classifies other operating costs depending on the presence of excess operating capacity, in which case a higher portion of costs would be identified as fixed, and uncorrelated with output. aviation fuel is one of the most important input costs affecting vc in airlines (e.g. merkert and swidan, ) , whereas inflation affects fc associates with items such as property. in fact, changes in fuel prices affect arguably most of the vc of an airline and usually immediately (with either short-term negative or positive comparable cost effect if the airline hedges its fuel costs), whilst catering, ramp, and air traffic charges will lag behind. although fixed costs remain constant with airline output level, it is affected by inflation. airlines incur fixed administration, reservations, marketing, group services, route costs, aircraft costs (financing and depreciation), maintenance, station costs, and offices and ramp handling j o u r n a l p r e -p r o o f costs. inflation affects all of these costs, including aircraft financing and leasing costs. for example, private equity firms financing aircraft leases usually require inflation protection by resetting regularly the financial terms. for these reasons, the value of aircraft whether bought or leased is influenced by the rate of inflation. in our models we control for cost increases associated with changes in input prices and inflation versus the merger effect on costs. changes in fuel costs and inflation as external shocks can influence the number of m&a deals (haleblian et al., ; gort, ) . thus, to control for external shocks when decomposing costs (over the full sample period) we used an industry index of the main input cost (∆ jet fuel prices -in $ per btu) obtained from the energy information agency (eia). to control for changes in the external economic environment on a national basis, we measured inflation through average consumer prices (∆ consumer price index -cpi) coming from the international monetary fund (imf). the full sample was divided into two sub-samples, one for unprofitable firms and another for profitable firms. in the former sub-sample, firms were included if at least one firm incurred an operating loss one year before the m&a (t- ), and in the latter firms were included if both firms made operating profits one year before the merger (t- ). m&as in the unprofitable sub-sample involved mostly unprofitable targets, although in very few cases there were unprofitable acquirers, and also both an unprofitable acquirer and a target (see table ). in merger models it is common to test various control variables, such as level of growth, annual average change in consumer prices, and inflation. in our case, these control variables became integral part of the model rather than endogenous controls. this approach is appropriate when modeling cost effects of multiple m&as measured over longer period of time with various event dates. external conditions can influence the post-merger cost structure and must therefore remain an integral part of the cost model to separate between cost changes associated with the merger itself and other cost changes. we employed a dyad-level design approach (kenny et al., ) to evaluate whether the individual m&as brought significant cost variations to the merged firms. using the dyad-level approach the observations of each firm pair were grouped together by summing their respective output units (available seat miles) and total operating costs. within a relevant range in terms of a firm's output (anthony et al., ) , we assume that the cost function has a linear form (we also tested size effects but those models did not yield any robust results). the linear form allows separation between fixed and vc and thus enables us to test the cost-structure effects resulting from m&as. this approach computes total costs as follows: where tc i,t denotes the total annual operating cost of airline i at time t, fc i,t represents annual fc, vc i,t the variable unit cost, and ou i,t stands for annual output in available seat miles, a typical measure of airline output. to allow for heterogeneity in both the intercept and the slope coefficients, we extend the fixed effect methodology of wooldridge ( ) and o'connell ( ) . we account for industry-specific effects in accordance with the hausman's ( ) specification test. the traditional fixed effect panel model is not appropriate because not only are the intercepts (fc) heterogeneous, but the slopes (vc) can vary among individual firms. we account for longitudinal dynamics (annual variations) by assuming that fc change with inflation (∆ consumer prices) and unit vc change with key input costs (∆ fuel costs). to describe these relationships we employ the following model: , = + * , + * , * , + + * , * , + * , * , * , + , where tc i,t is the total operating cost for firm i in year t; in i,t and ic i,t are indices related to inflation (∆ consumer prices) and input costs (∆ fuel costs) for firm i at time t standardized to for every firm in the year of the m&a; ou i,t denotes output units measured as asms for firm i at time t; and sd i,t symbolizes a step dummy taking a value of for firm i, in the years before the event, and starting from the year of the event; and finally , is the estimation error for firm i at time t. we converted all operating costs into u.s. dollars by employing the average exchange rates of the related fiscal year. model ( ) has three functions: ) one to separate total costs tc i,t into fixed and variable cost using the output unit ou i,t as independent variable (see equation ( )); ) another to introduce a year by year dynamic for the fixed and variable cost components, where the former changes with the inflation index in i,t and the latter changes with the fuel cost index ic i,t ; and ) a function to estimate whether the fixed and variable cost components change after the year in which the m&a takes place through a step dummy sd i,t that interacts with the estimated fixed and variable cost components. to test the models we use least square dummy variable regression (lsdvr). see the appendix for a discussion about the methodology aspects related to endogeneity. the estimated coefficient denoted as α is the constant of the model and represents the average fixed cost common to all firms; β i are the fc that are specific to each firm i. to account for variations in the firm specific fc over time (i.e., inflation), β i is multiplied by average annual inflation (in) measured as change in the consumer price index. to account for the variations of vc over time, the coefficient γ i for firm i is multiplied by output units (ou), and by the input costs (fuel costs (ic)) since airline vc are more closely correlated with fuel costs than consumer prices (in). we could have employed real (inflation-adjusted) costs, by using the country-specific consumer prices (in) to adjust the annual figures. however, as explained above, our model has two different cost dynamics: ) inflation for fc; and ) fuel for vc. if we employ inflation-adjusted costs, we need to adjust input costs accordingly, leading to a complication in the model. furthermore, since our panel includes m&as spanning three decades, using real costs by adjusting all costs to a reference year would introduce a bias into the model. we therefore employ nominal (unadjusted) costs and consider inflation as an independent variable. finally, in order to test if cost variations depend on the m&as included in the sample rather than general industry cost trends, we consider an alternative model, in which the dependent variable is the total costs in excess of the world airlines average costs, i.e.: where wac i,t is the world average cost index per asm for the world airlines based on icao's world financials. means, standard deviations, and correlations can be found in table . all correlations are as expected. table shows the results of the models evaluating the effects of m&as on the cost structure, with total costs as the dependent variable. in the model results, negative coefficients signify reduction in costs, and positive coefficients depict increase in costs. the effect sizes (f ) (cohen, (cohen, , for all three models are very large, ranging from . (profit model) to . (loss model). referring to table , we see in the base model a negative variable cost variation, signifying that this type of costs decreased on average expost (δ = - . , p < . ), with a confidence interval for the coefficient entirely below zero (- . to - . ). conversely a positive coefficient for fc means that this type of costs increased ex-post (δ = . , p < . ), with a confidence interval for the coefficient entirely above zero ( . to , . ). the r of the regression is . , and the effect size f is very large, . . the results demonstrate that horizontal airline m&as, on average, are associated with a decrease in vc and an increase in fc. however, we maintain that this global relationship can be misleading as the effect of m&as on the cost structure may be influenced by the ex-post incentive to realize cost efficiencies depending on the ex-ante profitability of the firms, necessitating a finer grained picture, we depict in two additional models: a loss and a profit model. for the loss model we argued that horizontal m&as involving at least one unprofitable firm are associated with a decrease in vc and an increase in fc. in the loss model, a reduction in vc occurred ex-post (δ = - . , p < . ) and an increase in fc (δ = . , p < . ). a confidence interval for the vc coefficient was entirely below zero (- . to - . ) and the fc coefficient was entirely above zero ( . to , . ). the r of the regression is about . , and the effect size f is very large about . . hence, horizontal m&as involving at least one unprofitable firm show reduced vc and increased fc. for the profit model we proposed that horizontal m&as involving only profitable firms are associated with a decrease in fc and an increase in vc. in the profit model the coefficient for fixed cost variation is negative (δ = - . , n.s.) and the coefficient for variable cost variation is positive (δ = . , n.s.). however, the confidence intervals for vc (- . to . ) and fc (- , . to . ) include zero, suggesting that m&as among profitable firms do not affect the cost structure of merged firms in one particular direction. the r of the regression is about . , and the effect size f is very large about . . thus, although the signs of the coefficients were as expected, the coefficients are not significant and the results do therefore not support our prediction. horizontal m&as involving profitable firms do not show statistically significant effects on variable-or fc. to test if the cost effects are different between the two types of m&as we carried out an ad-hoc test of the significance of the difference between the slopes of the regression lines for the two groups (cohen et al., ) . the results confirm a statistically significant difference for both fc (loss-δ : profit-δ , t= . , p < . ) and vc (loss-δ : profit-δ t= . , p < . ). based on these results we can state with confidence that the two groups of horizontal m&as are not the same. in order to account for the high volatility in the airlines' yearly operating results, we also considered a second, much more demanding criterion to identify whether an airline belongs to the loss group or not. as such, the m&as are classified as belonging to the loss group if at least one of the merging airlines incurred operating losses in all the three years before the transaction occurred. the results shown in table confirm our hypothesis suggesting a significant reduction of variable costs for m&as of the loss group, whereas m&as among two profitable firms do not impact on the cost structures of merged firms. we also show that the increase in fixed costs for the profit group is not statistically significant. in order to separate out m&as ex-post cost changes from the industry average cost changes we ran separate models using as a dependent variable the total costs in excess of the world airlines average costs (see table ). the resulting models supported all our predictions except m&as involving profitable firms that do show a significant positive increase in vc unlike our previous findings. table shows the results of the alternate models evaluating the effects of m&as on the cost structure using total costs in excess of the world airline average costs as the dependent variable. in particular, we confirmed our prediction that horizontal m&as involving at least one unprofitable firm (loss model) bring a decrease in vc (δ = - . , p < . ) and an increase in fixed cost (δ = . , p < . ). in this case, all main results of the basic model of table are validated. regarding our prediction that horizontal m&as involving only profitable firms (profit model) are associated with a decrease in fc and an increase in vc taking into account average cost trends of the worlds' airlines, we find the coefficient for fixed cost variation to be negative and not significant (δ = - . , n.s.). however, when considering the coefficient for the variable cost variation, we find it positive and statistically significant (δ = . , p < . ). the r of the regression is about . , higher than in the basic model ( . ). based on the analysis in table describing three alternative models that take into account the cost developments for the world's airlines, our results appear for the most part robust. controlling for the industry cost trends all our models hold except the profit model that shows significant increase in vc in line with our original predictions. the second part of table shows the results of applying the alternative criterion for classifying the m&as based on ex-ante profitability. the significant reduction of variable costs for m&as of the loss group is confirmed, whereas m&as among profitable firms have no significant impact on the cost structure of merged firms. our results appear robust even when choosing an alternative ex-ante profitability condition. academics and practitioners have long sought explanations based on robust empirical evidence on the determinants of ex-post m&as performance. one pungent question in past research has been why so many firms cite efficiency motives when m&as are initiated, even if these gains are rarely realized in practice according to the lion's share of the literature that has examined total cost effects. setting legal reasons aside, our results indicate that m&as do influence the cost structure under certain conditions but not total costs. this finding supports williamson ( ) suggesting that firms, for competitive reasons, may seek reduction in vc to strengthen their competitive advantage regardless of an increase in fc, so far as lower vc are translated into lower prices. our research lends strong support to this analogy offering an explanation why some mergers under certain conditions can have a beneficial cost structure impact strengthening the competitive position of the merged firm in an industry. our findings contribute to the m&as literature in an important way. first, we demonstrate that potential impact on the cost structure is influenced by the ex-ante profitability of firms, an important finding for antitrust and competitive policy in view of airliner mergers often getting approved on the hope that a financial healthy airline merging with a financially failing airline will save jobs. we explain this improvement of cost competitiveness of the merged entity based on the ex-post incentive to act on efficiency improvements. second, our methodological contribution is that we decompose the cost structure by using a novel approach and thereby circumvent the data availability difficulties common in this research domain. as such, our research fills a critical gap in the literature dealing with the cost structure effects of m&as using larger samples. previous research on the cost structure effects of m&as was limited due to lack of detailed cost data but also the challenge of linking efficiency incentives and outcomes under competing perspectives. our findings suggest that horizontal m&as have different efficiency incentives based on their performance/profitability ex-ante the transaction suggesting diverse strategic objectives across two exante profitability groups. our findings further contribute to the growing transport and competition policy literature around airlines. in contrast to the extant literature which argues that airline market structure should have an impact on how airlines (oliveira and oliveira, ) are governed through policy measures (e.g., choo, et al., ) taking cost structures into account. while there may be a reduction in capacity and diseconomies of scale (merkert and morrell, ) , it may still be reasonable for antitrust authorities and regulators to approve the m&a transaction particularly when one of the airlines involved in the transaction is ex-ante in a weak financial position as the resulting ex post merged entity may enjoy significant improvements in cost competitiveness. this will not only to save jobs but may actually foster competition (assuming that there will be other players in the market ex post the transaction). what is more, airline m&a transactions that will suffer from negative effects in the cost structure of the merged firm will be at the other end of the spectrum more likely to be incentivized to exercise their created market power. as such we argue that m&a transactions impact not only market structures but also cost structures and thereby market behavior and ultimately performance. all of that could potentially require transport policy responses such as foreign ownership restrictions that exists in many markets today (walulik, ) . the findings of this study should be considered in light of its limitations, which also provide directions for future studies. first of all, we draw data from a single industry that has inherently high fitness, complementarity and competitiveness (cost savings run in tandem with reduced average prices) among firms. although this sampling approach provides a well-defined context for our predictions, it limits the generalizability of the findings. even though our study uses horizontal m&as among service firms in a network industry, we argue that our approach can be applied to any competitive industry where reduction in variable costs strategically matters. while this research provides important evidence on how horizontal m&as impact on the cost structure of firms, it is limited to merger insiders (merging entities). research could explore potential impacts on merger outsiders and test the preemption hypothesis (fridolfsson and stennek, ) as we do understand that firms do not embrace being merger outsiders (not to engage in mergers while key competitors merge) as it is synonym to potential competitive disadvantage in the minds of managers. thus, the preemption hypothesis assumes that firms will race between themselves to acquire other firms even if it means giving up surplus to the targets (laamanen and keil, ) . future research may extend the framework developed in this paper to include serial acquisitions (laamanen and keil, ) . serial acquisitions may help develop competence in managing acquisitions that in combination with acquisition timing may help the acquirer derive superior cost structure benefits along the lines demonstrated in this paper. our study focused on the cost structure without investigating scale efficiencies directly. future studies may examine unit costs in the context of scale and cost structure using a similar sample setting and contingencies. although, we have demonstrated in this research that pre-merger profitability plays a role in the incentives of merging firms to improve efficiency, further exploration is needed on m&as incentives among profitable firms. finally, this study could be extended by examining merger conditions, incentives, and timing more directly with larger samples of m&as to help pre-evaluate efficiency gains from potential mergers and acquisitions (bai et al., ) in the airlines. in conclusion, this present paper shows that a finer grained cost structure analysis offers added insights into m&as cost performance which would otherwise remain underexplored. the conditions surrounding cost structure effects established in this paper help guide future studies and antitrust/ competition authorities/ regulators on m&a cost structure effects, taking into account ex-ante performance improvement incentives based on merger types and external conditions surrounding the deals. increase prices to cover all their costs, and are eventually forced out of the market or to restructure. another possible source of endogeneity is between the dependent variable tc i,t , operating costs, and quantity ou i,t measures in terms in available seat miles asms. in this case, quantity could be (circularly) determined by operating costs, in the sense that airlines with higher operating costs may decide to reduce capacity to become more efficient. we tested endogeneity in the explanatory variable ou i,t by deploying the hausman specification test (hausman, ) and used the asms industry index as the instrumental variable, necessary to perform the test. a good instrumental variable needs to be correlated with the instrumented variable (asms of single airlines), but unrelated with the dependent one (operating costs of single airlines). in our case, the asms industry index is correlated with the asm of single airlines, as both would depend on the growth of the industry, or to industry shocks (september th, sars, etc). however, the operating costs of specific airlines could not significantly affect the overall asm industry index. by performing the hausman specification test we verified the validity of the exogeneity assumption. please find enclosed the nd revision of the paper "cost structure effects of horizontal airline mergers and acquisitions". pricing mergers with differential synergies. strategic change delta's decision the effects of megamergers on efficiency and prices: evidence from a bank profit function new evidence and perspectives on mergers accounting: texts and cases does it pay-off to capture intangible assets through mergers and acquisitions antecedents of m&as success: the role of strategic complementarity, cultural fit, and degree and speed of integration bank scale economies, mergers, concentration, and efficiency: the us experience performance of acquisitions of unprofitable firms the long-term performance of horizontal acquisitions asset divestiture following horizontal acquisitions: a dynamic view when do acquirers earn abnormal returns? mergers, takeovers, and economic efficiency: foresight vs. hindsight scale and scope: the dynamics of industrial capitalism types of synergy and economic value: the impact of acquisitions on merging and rival firms joint impact of airline market structure and airport ownership on airport market power and profit margin statistical power analysis for the behavioral sciences statistical power analysis. current directions in control specification tests in econometrics the efficiency of firms: what difference does competition make? the market for corporate control and firm innovation acquisitions as a means of restructuring firms in chapter mergers without cost advantages mergers and profitability: a managerial success story the modern industrial revolution, exit, and the failure of internal control systems estimation of cost efficiencies from mergers: application to us radio the job satisfaction-job performance relationship: a qualitative and quantitative review acquisitions, acquisition programs and acquisition capabilities dyadic data analysis meta-analyses of postacquisition performance indications of unidentified moderators performance of serial acquirers: toward an acquisition program perspective calculating and reporting effect sizes to facilitate cumulative science: a practical primer for t-tests and anovas boarding the aircraft: trust development amongst negotiators of a complex merger allocative efficiency vs. x-efficiency mergers and the performance of the acquiring firm merger strategies and stockholder value demand and consumer welfare impacts of international airline liberalisation: the case of the north atlantic market power at the seattle-tacoma international airport: the case of alaska airlines disappointing marriage: a study of the gains from merger mergers and acquisitions in aviation: management and economic perspectives on the size of airlines flying with(out) a safety net: financial hedging in the airline industry the impact of industry shocks on takeover and restructuring activity disentangling the performance effects of efficiency and bargaining power in horizontal growth strategies: an empirical investigation in the global retail industry preemptive horizontal mergers: theory and airline mergers: a longer view creating value in the face of declining performance: firm strategies and organizational culture the causes of mergers: tests based on the gains to acquiring firms shareholders and the size of premia antitrust analysis of airlines mergers and acquisitions, human resource issues and outcomes: a review and suggested typology dealing with panel data in accounting and managerial finance research what drives effective competition in the airline industry? an empirical model of city-pair market concentration estimating sample sizes for repeated measurement designs incentive asymmetries in the mergers and acquisitions process competitive advantage competitive advantage: logical and philosophical considerations debt, mergers and acquisitions, institutional arrangements and change to the multilayered subsidiary form european merger control: do we need an efficiency defence? the role of fixed cost savings in merger analysis the impact of airline mergers and hub reorganization on aviation kilometer consumption industrial market structure and economic performance perception of airline industry structure by regulators and managers value creation in acquisitions: a re examination of performance issues airline mergers and their effect on network structure bankruptcy avoidance as a motive for merger the neglected importance of leadership in mergers and acquisitions synergy, agency, and the determinants of premia paid in mergers value of merger and acquisition deals worldwide from the relative effect of operational hedging on airline operating costs the competitive effects of mergers between asymmetric firms what do we know about the success and failure of mergers? merger motives and merger prescriptions attributional tendencies in cultural explanations of m&as performance at the core of airline foreign investment restrictions: a study of countries acquisition or alliance? a dyadic perspective on interfirm resource combinations decline in organisations: a literature integration and extension the entry inducing effects of horizontal mergers: an exploratory analysis generic business strategies, organizational context and performance: an empirical investigation economies as an antitrust defense: the welfare tradeoffs fixed effects and related estimators for correlated random coefficient and treatment effect data models. working paper pre-evaluating efficiency gains from potential mergers and acquisitions based on the resampling dea approach: evidence from chinas railway sector airline horizontal mergers and productivity: empirical evidence from a quasi-natural experiment in china choosing the right target: relative preferences for resource similarity and complementarity in acquisition choice * p < . , **p < . , ***p < . .we thank the reviewers for excellent comments that have greatly improved the paper.we sincerely hope to have the paper published in tp.on behalf of all the authors, prof. sveinn vidar gudmundsson reykjavik university one of the most common sources of endogeneity for this kind of study, is between price and quantity. yet, in our model, specified in equation ( ), the dependent variable is operating costs and not prices or revenues. in a highly competitive industry such as the airline industry, prices are not strictly related to costs. due to competition, highly inefficient airlines, to offer a given product/quality mix, cannot key: cord- - glw pir authors: lloyd, helen m.; ekman, inger; rogers, heather l.; raposo, vítor; melo, paulo; marinkovic, valentina d.; buttigieg, sandra c.; srulovici, einav; lewandowski, roman andrzej; britten, nicky title: supporting innovative person-centred care in financially constrained environments: the we care exploratory health laboratory evaluation strategy date: - - journal: int j environ res public health doi: . /ijerph sha: doc_id: cord_uid: glw pir the cost cares project aims to support healthcare cost containment and improve healthcare quality across europe by developing the research and development necessary for person-centred care (pcc) and health promotion. this paper presents an overview evaluation strategy for testing ‘exploratory health laboratories’ to deliver these aims. our strategy is theory driven and evidence based, and developed through a multi-disciplinary and european-wide team. specifically, we define the key approach and essential criteria necessary to evaluate initial testing, and on-going large-scale implementation with a core set of accompanying methods (metrics, models, and measurements). this paper also outlines the enabling mechanisms that support the development of the “health labs” towards innovative models of ethically grounded and evidenced-based pcc. the world health organisation defines universal health care as that 'which all citizens can access without incurring financial hardship' [ ] . many nations fail to provide this as a basic human right, health promotion as an approach aims to inform, influence, and support people, communities, and organisations to improve health. supporting people to increase control over their health is in essence health promoting, both for the individual and society [ , ] . hp activities can work in synchronicity with pcc if developed in partnership with the person, taking into consideration their life context and socioeconomic conditions [ , ] . cost action 'cost cares' was funded by the eu commission to create the impetus in both the research and development required to design and test innovative exploratory health laboratories (ehls) to implement pcc and hp across the eu. this paper sets out a strategy for evaluating them. to understand how the ehls might work to deliver pcc, hp, and cost outcomes, it was first necessary to develop programme theories (pts). pts describe how interventions (service, treatment, policy) are thought to work by specifying the ways in which they produce outcomes. they are a set of causal relationships often referred to as "if-then" statements. they can also be written or represented graphically to show the relationships between cause and effect. pts are also useful for understanding both the positive and negative impacts that can occur when interventions are implemented. they are often accompanied by logic models, which help plan and evaluate interventions based on their internal logic, and the role of context in supporting successful delivery and evidence acquisition. we created evidenced-based pts to specify how ehls would deliver pcc, hp, and cost outcomes through the critical enablers detailed in the we-care roadmap (see figure ). repeated here for clarity the critical enablers are ( ) information technology (it), which describes the use of computers or other computerized devises to store, transmit, and receive data to support pcc planning and care coordination, for handling and communicating health and evaluation data, and for delivering pcc and hp interventions. ( ) quality measures, such as organizational processes, that ensure health services increase the likelihood of the desired health outcomes consistent with current scientific knowledge, which take into consideration an individual's preferences, and ensure that health services are effective, affordable and accessible to all citizens. ( ) infrastructure to create the necessary resources and structures that support the shift from health systems that are excessively hospital-centric and biomedically-oriented, to those which value continuity, responsiveness, and multidimensionality in community care, e.g., shifts in staffing, training, and delivery of care. ( ) incentive systems that reward pcc processes and outcomes, such as personal health goals, pcc plans, improvements in patient self-efficacy and experiences of care, and hp activities. this will require an expansion and critical revision of existing system-based biomedically driven performance indicators. ( ) contracting strategies that define and endorse pcc incentive systems and infrastructural support and efficiencies for ehls, purchasing strategies and contracts between payers and providers of healthcare that promote the alignment in organisational goals based on pcc, hp, and cost containment. ( ) cultural change that represents shared assumptions, values, and beliefs that govern how people behave in an organisation. receptiveness or readiness to change is considered a prerequisite for ehls. as other critical enablers are modified within a given ehl, cultural change towards pcc, hp, and cost containment may present as either a pre-requisite and/or a natural consequence of development. the addition of this sixth critical enabler represents the importance of organizational culture in achieving pcc and cost stability. upon establishing agreed definitions of the above enablers, the next step was to hypothesize how these might work to support the aims of an ehl. following this step, the literature was searched to detect evidence for the hypothesized statements, referred to as 'if-then' statements. to expedite this process, tables of 'if-then' statements were compiled, which, in keeping with the evaluation methods of critical realism [ , ] , permitted the compilation of patterns of causal chains within the ehl. for example, if condition x is in place (e.g., practitioners are incentivized to engage in shared decision making with patients), then outcome y might follow (e.g., patients will feel like they are taking an active role in rehabilitation planning), thus improving service user experiences of pcc [ ] . this task facilitated exploration of how the critical enablers interacted with pcc and hp to improve quality pcc and cost containment (see figure ). the points at which pcc and hp intersect with each of the critical enablers in figure are referred to as intersection points (e.g., pcc and information technology (it)). this section describes the considerations and necessary steps for evaluating and implementing ehls to improve quality pcc and cost containment. first, the practice of pcc is explored, and then the role of critical enablers is illustrated. a number of controlled studies have been performed comparing pcc to usual care [ , [ ] [ ] [ ] . the core components in the interventions have been to listen carefully to the patient's illness narrative and to mutually agree on a health plan. the true case story (see figure ) previously published in a position paper demonstrates how the patient narrative can open up and reveal information needed for the patient and the professionals to be able to agree on a relevant health plan [ ] . this is concordant with the theory and philosophy that pcc is based on starting with each person's capability and wish to take responsibility for their own health. the true case below is a vignette based on a real person to illustrate how pcc can be applied in practice through a worked example. upon establishing agreed definitions of the above enablers, the next step was to hypothesize how these might work to support the aims of an ehl. following this step, the literature was searched to detect evidence for the hypothesized statements, referred to as 'if-then' statements. to expedite this process, tables of 'if-then' statements were compiled, which, in keeping with the evaluation methods of critical realism [ , ] , permitted the compilation of patterns of causal chains within the ehl. for example, if condition x is in place (e.g., practitioners are incentivized to engage in shared decision making with patients), then outcome y might follow (e.g., patients will feel like they are taking an active role in rehabilitation planning), thus improving service user experiences of pcc [ ] . this task facilitated exploration of how the critical enablers interacted with pcc and hp to improve quality pcc and cost containment (see figure ). the points at which pcc and hp intersect with each of the critical enablers in figure are referred to as intersection points (e.g., pcc and information technology (it)). this section describes the considerations and necessary steps for evaluating and implementing ehls to improve quality pcc and cost containment. first, the practice of pcc is explored, and then the role of critical enablers is illustrated. a number of controlled studies have been performed comparing pcc to usual care [ , [ ] [ ] [ ] . the core components in the interventions have been to listen carefully to the patient's illness narrative and to mutually agree on a health plan. the true case story (see figure ) previously published in a position paper demonstrates how the patient narrative can open up and reveal information needed for the patient and the professionals to be able to agree on a relevant health plan [ ] . this is concordant with the theory and philosophy that pcc is based on starting with each person's capability and wish to take responsibility for their own health. the true case below is a vignette based on a real person to illustrate how pcc can be applied in practice through a worked example. pcc for mr. g was facilitated via various critical enablers, detailed in the following: information technology: the medical documentation and information (in the patient records) as well as the commonly formulated treatment and health plan are digitalized and accessible to mr. g and his providers in a way that he comprehends and can agree or ask questions about. in formulating the health plan, mr. g was supported by a digital patient decision aid [ ] . health information technology (it) systems support the smooth flow of information between services, and to and from citizens and their families. artificial intelligence might facilitate this and help improve interactions with patients [ ] . quality measures: mr. g was invited to download an app after his first myocardial infarction where he can follow the development of symptoms and well-being and contact health care services for help and support with formulating his personal health plan. contracting strategies, incentives, and infrastructure: the infrastructure supported cumulative documentation according to the criteria for pcc. this was linked to incentive payments for the whole team. this type of incentive payment includes quality measures (care plans) that are sanctioned and contracted between the provider and commissioner organisation. program theories (pts) are useful ways in which to facilitate an understanding of how complex interventions work; in this case, how the critical enablers could work with pcc and hp interventions to generate cost containment and quality pcc outcomes. table provides worked examples of pts referred to as 'if-then' statements with explanatory 'because' statements and associated suggestions for assessment or measurement. instruments and methods to assess pts should be carefully selected and the use of mixed methods is advised. knowledge base and practical constraints will add to the existing complexity of measurement and evaluation. the type of design employed can help remedy some of these issues. for example, beginning with small-scale and qualitative assessment will help determine what to measure and how to measure it, and what improvements to expect. ensuring measures or assessments capture professional and patient partnership work in care planning is key for emphasizing the importance of this for pcc. the following pts are presented here as examples of a larger body of work (available from the first author) conducted to inform the design and evaluation of ehls. table presents seven different types (a-g) of evidence-based pts that could shape the design of an ehl. type a (contracting strategies for quality and cost outcomes) pts represent how contracting strategies could operate at macro and meso levels to support quality pcc and contain cost. in the two examples provided, 'alliance' or 'partnership' models contract to deliver an ehl based on shared or co-designed pcc and hp objectives to improve quality pcc and costs. this fosters trust and productivity based on collective ownership and the sharing of risk and reward within ehl. a mixture of quantitative and qualitative measures of delivery and management team dynamics, and progress towards aligned goals (e.g., pcc health plans), and costs over time could be used to ascertain the success of the contracting strategy. these enablers provide causal mechanisms for cost and quality outcomes at macro and meso levels within the ehl. type b (incentives and contracting strategies for quality pcc resulting in cultural change) pts represent the potential for contracting strategies combined with incentives to improve cost and quality outcomes by providing incentives at multiple levels across the ehl. for example, if cost effectiveness is measured across the whole care chain with the savings provided to all participants, this creates the potential to act as an incentive towards aligned pcc and cost goals. to combat perceptions of unfairness in the equal distribution of savings across the system, objective measures of effort will need to be employed. these measures, however, should to be balanced against the knowledge of the operational context. for example, settings low on staff resources may seem to have contributed less towards the achievement of savings across the chain. ensuring that contextual knowledge supports objective measurement will help communicate the conditions of contributors towards the savings gained and shared. long-term planning and monitoring, active communication, and shared goals will help mitigate against perceptions of unfairness. redistributing resources based on savings can help achieve the stated organization goals and thus improve the sector's efforts where these are perceived to be lacking. these seemingly radical shifts align to the principles of fair division and social choice [ ] . over time, resultant cultural change across the system could be operationalized as permanent transformation of routines/habits. measures of pcc and hp routines/habits, savings distribution, and measures of patient experience of care could help establish if this strategy is beneficial. type c (contracting strategies, incentives, and quality measures for cost and quality) pts combine contracting strategies with incentives and quality measures to effect change in quality and costs. these build on type a and b pts by, for example, suggesting that if contract payments are made at the same time to all providers and tied to measures of pcc and hp, this fosters trust and productivity by reducing the misalignment and unproductive competition between partners and reduces transaction hazards operating at macro and meso levels within the system. type d (incentives for quality pcc) pts work at the micro level with incentives applied equally to all delivery staff irrespective of hierarchy or professional grouping [ , ] (e.g., patient feedback forms at clinic and ward levels). for quality pcc outcomes to be achievable, incentives must ensure that the reward system motivates individuals to align their own goals with those of the organization (ehl) [ , ] . as the pcc approach is based on qualitative changes, financial incentives may not be the best type of incentive to test. it has been long recognized that financial incentives are positively related to quantitative performance (e.g., number of tasks completed) but not necessarily with performance quality [ ] [ ] [ ] [ ] . thus, ideally, particularly since pcc is based in an aristotelian ethics of virtues, the incentive systems should be a combination of financial and non-financial rewards (e.g., recognition, positive feedback from leaders, promotions, money, as well as target setting and performance evaluation itself) [ , ] . these rewards would be directed to all ehl members, since in "a complex network of interdependent relationships" [ , ] necessary for pcc implementation, it is difficult to identify an individual contribution. the success of micro-level incentives can be measured by carefully selected patient experience measures and focus groups. type e (incentives, quality measures for cost, and quality pcc) pts work by combining incentives with quality measures at macro and meso levels. for example, if a pcc quality measure is linked to an ehl accounting system and able to deliver cost containment information resulting from pcc processes, then the measure itself becomes the incentive. quality measures therefore act as both an aligned incentive and measurement of implementation. a pre-and post-comparison of costs associated with pcc quality processes analyzed against quality measure scores would provide an assessment of effectiveness. a benchmarking strategy against non-ehl settings may be an example of a measurement process being itself an incentive. it is important to note that the cost containment may not be immediate, as some costs may be incurred upfront and/or it may take time for outcomes to stabilize or become apparent. ehls employing longitudinal designs can help to account for these potential delays. type f (information technology for quality) pts provide examples of how it has the potential to improve quality. these pts work to support patient self-management through mobile technology, for example, through symptom monitoring or appointment reminders, to help people manage their own health [ ] . they may also operate to support the adoption of pcc electronic health records and care plans, which provide teams with the tools to maintain and share pcc information. measurement and evaluation of these mechanisms would be tailored to detect changes in patient self-management activities, team effectiveness, and resultant health system impact (e.g., reviews, appointments attended, etc.). in the current covid- context, remote monitoring of patients, video-linked consultations, and e-health interventions could provide an exciting opportunity to test the delivery of person-centred care remotely, with the potential to calculate costs compared to previous standard practice [ ] . type g (infrastructure for quality pcc) pts provide examples of how components of an organization's infrastructure could help result in quality care at meso and micro levels. at a meso level, if staff training is provided to enhance professional skills to support patient empowerment and enhance professional communication skills, this then has the potential to improve pcc delivery and experience of care. furthermore, using patient-reported measures to shape care planning and use of the feedback from these measures to improve staff training has the potential to embed the patient voice in quality improvement practices and shape equitable person-centred relationships between professionals and patients [ ] . a multitude of measures are available to measure these outcomes [ ] and for use in care planning in this way. however, sampling care plans with patient-reported outcome measures (prom) and interviews with professionals and patients would be insightful. these examples of pts are not comprehensive, but they illustrate how those developing ehls can use these and other mechanisms to design their interventions and corresponding evaluation strategies. for further guidance on the use of evaluation metrics and measures, see p c.org.uk. if "quality measures" are linked to pcc ideas and information systems (e.g., accounting system) and able to deliver information about cost containment or other quantitative indicators improvement against non-ehl settings (benchmarking), then the measurement process itself will be an incentive [ ] the measurement process has also the function of ex-ante control applied "quality measures" enabler the evaluation of ehls should address questions that will enable commissioners of health services and delivery organizations to implement, sustain, and scale up the innovations. key evaluation questions for the ehls will include those that probe pcc processes, practices, and patient experiences of pcc care as markers of quality pcc. the health outcomes measured should be relevant to the patient and their family, health care provider, and other decision-makers. key areas of interest in the implementation of pcc are changes in functional ability, experiences of care, self-efficacy, and cost. ehls will also be informed by wilson and cleary's [ ] model for integrating concepts of biomedical outcomes and measures of health-related quality of life: (i) biological and physiological factors, (ii) symptoms, (iii) functional status, (iv) general health perceptions, and (v) overall quality of life). specific questions (see figure ) will also probe the mechanistic relationship between the critical enablers and pcc and hp. these are referred to as intersection points. irrespective of the type of intervention, commissioners and policy makers require proof that the additional health care resources needed to make the procedure, service, or program available to those who could benefit from it are justified [ ] . the purpose of economic evaluation is to inform such funding decisions. an economic evaluation deals with both inputs and outputs (costs and consequences) of alternative courses of action, and is concerned with choices and consideration of the costs and benefits at multiple levels. ehls will therefore have to evaluate the main costs involved in the change of a healthcare system towards pcc and hp. weinstein [ ] identifies costs related to changes in the use of healthcare resources, changes in the use of non-healthcare resources, changes in the use of informal caregiver time, and changes in the use of patient time (for treatment). in a similar way, drummond et al. [ ] identifies health sector costs, other sector costs, patient/family costs, and productivity losses. measurement within economic evaluation expands beyond the healthcare system under study. according to weinstein [ ] , direct health care costs include all types of resource use, including professional, family, volunteer, or patient time, as well as the costs of tests, drugs, supplies, healthcare personnel, and medical facilities. non-direct health care costs include the additional costs related with the intervention, such as those for childcare (for a parent attending a treatment), the increase of costs required by a dietary prescription, and the costs of transportation to and from the clinic; they also include the time family or volunteers spend providing home care. citizen time costs include the time a person spends seeking care or participating in or undergoing an intervention or treatment. time costs also include travel and waiting times as well as the time receiving treatment. productivity costs include ( ) the costs associated with a lost or impaired ability to work or to engage in leisure activities due to morbidity and ( ) lost economic productivity due to death. the world health organization (who) recognizes quality health care in those organizations that have a high degree of professional excellence, with minimum risks, good health outcomes for patients, and efficient use of resources [ , ] . to promote the health of the population, the who recommends key objectives for continuous quality improvement in health care. these include the structuring of health services, the rational and efficient use of both human and financial resources, and the guarantee of professional competence to citizens in order to meet their needs. measures or questions relating to quality are likely to overlap and complement those relevant for cost containment (see figure ). is the ehl coordinating its activities around the person and their carers/family? are carers supported? are community assets are being deployed, including peers, social networks, and the voluntary sector? the evaluation of the ehls must contain the most suitable measures and approaches to answer the questions. quantifiable measures or questions can either be aggregated (single criterion analysis) or handled separately (multi-criteria analysis). careful consideration of the combination of qualitative and quantitative approaches is advised, particularly since different health systems display different capabilities in this regard. in terms of minimum design standards, at least two data collection points-pre-and post-intervention/implementation-are recommended. this is the minimum standard advised. should the availability of knowledge, skills, and resources be forthcoming, more complex experimental and implementation-focused designs could be undertaken upon careful consideration of the amount of preexisting evidence for pcc in that particular context or condition [ ] . ideally, monitoring and data collection will be continuous and with feedback to practice, with long follow-up periods to capture lasting changes in care delivery and outcomes. to account for the variance in ehls, a core minimum data set from each site with three categories of data is recommended: routinely collected audit data or similar (e.g., collected at country or hospital level); questionnaire data specifically collected for the ehl; and qualitative data to support implementation development. examples of suitable measures, depending on the focus of the changes in the healthcare system that are implemented, are given in table . as an ehl is scaled out in practice, it may be necessary to add new measurements to capture unanticipated and/or unintended changes. a metrics framework provides the structure for planning the sampling and timing of data collection during the evaluation of an ehl. it is likely that data could flow from different sources, e.g., routinely collected data and quantifiable data, surveys, and qualitative data. the pt will guide the sampling strategies for data collection, the timing of data collection, and the various units of analysis. qualitative approaches will always necessitate careful sampling because they are resource and time intensive. in contrast, an ehl may decide on a questionnaire to measure the experiences of all those using a service to canvas a broad view. the trade-off between qualitative approaches and more structured approaches involves considerations of depth versus breadth; different sampling strategies are required for different forms of data. as qualitative approaches are effective for determining "how and why" the ehl is working, it will be important to consider a range of perspectives. sampling should therefore aim for diversity in terms of ethnicity, social and economic status, age, disability, and health conditions. services may also decide to film or record care interactions for ongoing implementation and quality improvement activities, using purposeful samples or random selection. convenience and pragmatism will also play a role in any sampling procedure, which is common in applied health care research and evaluation, where time and resources are limited. the phasing of data collection will likely include baseline data and follow-up data to mirror the timeframes of the intervention. it might also be necessary (providing sufficient justification and acceptability from practitioners and patients) that focused data capture on a specific element of the delivery is added into the core set of measures at particular times. for example, if communication or shared decision-making was an improvement target, implementing a tool that specifically addresses this issue of relational care could be used as both the intervention and data collection [ ] . the potential to link health and social care data to understand an individual's pathway following exposure to an ehl will be determined by local ethical restrictions, data flow, and governance guidelines. linked data sets (or even unified data sets) allow for a longitudinal exploration of the impact of the intervention on service utilization (costs) and health using time series analysis or similar [ ] . analysis will be more powerful if compared to a control cohort (tracked by a unique identifier following explicit consent) of people who are part of a health lab. the use of techniques, such as propensity scoring, to identify and match control groups of service users are particularly helpful for this type of evaluation and service development [ ] . the analysis plan should be informed by the pts and shaped by the evaluation framework. in principle, three main stages of analysis are envisaged. the first stage will commence with univariate analysis to examine each variable or source of data (for example, acceptability of services as a measure of quality or use of care plans as a measure of it) independently. this could explore the time trends in say routinely collected data and the statistical properties of the data, e.g., the distribution of the data. parallel qualitative analysis could seek to surface emerging themes. in the second stage, for each ehl, the pt will be tested to check if it is functioning as expected. in the third stage, findings both within and across the ehls will be compared to answer the higher-order questions about the relationships between the quality of care and cost containment. working to understand trends in the data and other potential factors influencing outcomes (i.e., closure of a community hospital, or lack of out-of-hours primary care) will be a necessary effort. collaboration between academic and health science partners will facilitate a robust evaluation, linking efforts to capture patient experiences and outcomes with cost indicators. the ultimate result will be a more nuanced story of how the intervention is delivered, experienced, and the extent to which it is achieving change. in this regard, it is important to note that change may not be immediate. even if change is achieved quickly, impact on outcomes may require longer-term follow-up, especially, for instance, to demonstrate the cost-benefit ratio. to convince european societies and key decision-makers at a national and an eu level that the we-care roadmap is viable, reliable evidence from the ehls based on robust evaluation and implementation is required. many barriers and uncertainties may threaten the implementation of pcc. the first is the quality and accuracy of the pt that underpins the ehl model; whether it includes all key aspects needed to provide pcc, if it examines quality care and/or cost, and the extent to which it includes the enablers within the ehl. the model should also be appealing and promise significant benefits, in order to convince key stakeholders of the potential ehl. however, not only is the quality of the theoretical model important, the legitimacy and reliability of the person or organization presenting the model to its future users is also crucial [ ] . the engagement of authoritative local leaders who endorse the model to a range of stakeholders will be important to achieve early on in the process. this is likely to affect stakeholders' perception of its quality and validity [ ] , as well as its advantage over alternative solutions [ , ] . the ehls will affect people, their families, health professionals, and employees throughout the organization, including managers. thus, a bundle of incentives for different groups will probably be required. varied incentives, not only financial, as pay-for-performance, but also prospects of increased external recognition or legitimacy for participant organizations should be considered. the title of "the best provider", achieved by public benchmarking, could be an example. this requires accurate outcome measurement. incentive bundles can apply to three enablers of we-care roadmap: incentive systems, quality measures, and contracting strategies. the case-mix systems that are used in many european countries to finance hospital care are motivating providers to admit more patients, because the more patients they serve, the higher their income. if a hospital or a hospital ward agrees to become an ehl, the issue of contradictory incentives is likely to arise and must be overcome. for example, if, by implementation of an innovative community care ehl, more patients are cared for in the community, then the hospital will not receive money from the payer for those patients. the fixed costs of the hospital will remain, creating a deficit in the hospital system. a risk-reward sharing framework between the hospital and community provider could agree to cover hospital losses over the course of the project, but provisions for who will pay the fixed costs afterwards would need to be considered. involving key stakeholders from across the system will be important to provide strategies to overcome these conflicting issues. there should also be a distinction made between the average and the marginal cost of in-patient care. for example, the costs of a hospital ward (e.g., general medicine) are unlikely to differ significantly between a -or -patient occupancy. this means, that even if a treatment of a group of patients was organized outside of a hospital and the hospital infrastructure remained unchanged, the cost savings would be meagre or illusory. if, after introduction of the innovative care system, the medical infrastructure seemed unnecessary, then ehl employees would need to be motivated to support the ehl to ensure sustainability. the extent to which the organizational climate is favorable for ehl implementation must also be considered [ ] . the implementation climate is more evident and less stable than the organizational culture and is thus more susceptible to amendments. policies, procedures, and reward systems are those incentives that may effectively affect the implementation climate [ ] . the other fundamental ingredient of a positive implementation climate is the extent to which important actors perceive the current healthcare delivery model as intolerable or unsustainable and are motivated for change, defined as cultural readiness [ ] . the proposed model of the ehl should be compatible with stakeholders' own norms and values (culture), as well as with their priorities [ , ] . to maintain a positive climate for the implementation of ehls, important indicators related to citizen health, well-being, quality, costs, and other important factors should be presented to stakeholders. thus, both the climate for change and incentives and reward systems call for accurate, objective, and verifiable measures viable to reflect the real performance in pivotal areas. if measures do not meet these requirements, this could undermine implementation [ ] . measures must clearly communicate pcc goals and feedback to participants indicating the degree of goal achievement. to support pcc implementation and address potential barriers, each pt should be linked to a strategy with its own resources. resources include knowledge, time, money, training, and in some cases physical space. especially important is the access to widely understandable and convincing information and knowledge about pcc implementation, specifically about new work processes for the staff and the nature of care provided to patients and their social environment. if resources are not available, this creates a further barrier to implementation that must be effectively managed. organizational change begins with changes in individual behavior, although as numerous studies have shown, this is complex and challenging [ ] . ensuring the main actors do not perceive implementation of ehls as threats to their own interests is a critical issue to address. subjective interests are, however, not often easy to identify. powerful actors, in particular leaders who at multiple levels across the system represent the core activity of the pcc implementation, must include physicians, nurses, allied health and social care professionals, people, and their communities. in ehls, leadership should be transformational and innovative to create teams working to develop a workplace that is person centred. this is a key factor in the delivery and sustainability of pcc [ ] . if this is achieved, it will promote cultural change and the upskilling of existing employees. having several key people within the organization take on this role will ensure leadership sustainability. although these groups should support every change to augmenting healthcare quality, such as pcc, in reality, however, explicit or latent resistance can be a common problem [ ] . the medical and health professions are built on an ideology to protect and care for humanity over economic profitability and self-reward [ ] , but contradictions between altruism and professional self-interest have been established. the excessive self-interest of individual doctors or groups of physicians should be mitigated by professional self-regulation and self-control [ ] . since large-scale testing is the ultimate aim, it is assumed that a significant number of enabler elements will be in place when an ehl begins. as suggested earlier, the ehl will be underpinned by the pt that describes how the central work processes and independent actions of actors should be coordinated to deliver high-quality pcc. to be effective, the model, once elaborated, will require continuous adjustment not only to local environment factors but also to external and internal uncertainties emerging over time in each setting. thus, some feedback and regulatory mechanisms should be an integral part of the model. the development and improvement of ehls will be facilitated by a commitment to formative learning in response to the feedback from the evaluation data (data-driven improvement). there is a long tradition of using these methods to improve practice, and good evidence to suggest benefit [ ] . learning will vary by organization and setting. however, it will usually require a "plan-do-study-act" (pdsa) cycle or a similar process [ ] . this will typically involve action learning sets [ ] using quality improvement methodology [ ] . action learning sets are particularly suited to iterative complex intervention development as they focus on learning from interactions, thus providing a mechanism to reflect and problem solve. these skills are particularly important for health and social care professionals who are being asked to work in a different way, where this is likely to be challenging. . . co-design and participatory action for pcc emancipatory research designs have been a core feature of community development and strengths-based approaches in social care. such approaches value the lived experience and partnership with patients and the public in developing and evaluating services [ ] [ ] [ ] [ ] . research approaches based on these principles have in the past been subject to much derision but are now becoming recognized as critical to citizen-relevant and humanistic healthcare planning and evaluation, and align well with the philosophy of pcc. the uk standards for patient and public involvement in the planning and evaluation of health and social care are supported by academic, research, and government policy. involving patients and the public in the consultation and shaping of ehls is a core and fundamental standard we advocate. this paper laid out a comprehensive plan for the evaluation of exploratory health laboratories that aims to improve the quality of health care in the eu whilst also containing costs. the plan was developed by members of the we-care fp -funded project and cost cares cost action from a range of academic and professional backgrounds and different countries. this process identified pcc and hp as the solution, along with critical enablers to facilitate implementation. examination of the intersections among and between these enablers, as well as the impact on quality of care and cost of care, via evidence-based pts provides the justification for the design and incorporation of particular components into an ehl. furthermore, the paper also described how these components and ehls might be evaluated as complex interventions at micro, meso, and macro levels. this work and the resources it produced (www.costcares.eu) are intended to serve as a reference material for those considering setting up ehls or similar initiatives beyond the scope of this cost action. author contributions: h.m.l. was instrumental in leading the work of working group . this included development and supervision of the plan of work, data curation, preparation of the original manuscript and subsequent revisions. i.e. was core to working group and supported the writing, reviewing and editing of the manuscript. i.e. led the funding acquisition and contributed to the work of working group and critical development of the manuscript. h.l.r. is a member of working group and was responsible for data curation, manuscript revision with re-conceptualization of some aspects presented, and critical review of the final manuscript. v.r., p.m. and v.d.m. are members of working group and were responsible for data curation, manuscript review and critical revision. s.c.b. read versions of the manuscript and was responsible for reviewing the final manuscript. e.s. contributed to data curation, critical review of the manuscript and was responsible for editing and adding citations. r.a.l. helped to write the manuscript, supported the literature review and subsequent critical revisions of the final version of the manuscript. n.b. supported the work of working group through conceptualization and methodological input and was instrumental to the development and writing the original manuscript and subsequent reviews. all authors have read and agreed to the published version of the manuscript. the authors declare no conflict of interest. world health universal health coverage (uhc) world bank and who: half the world lacks access to essential health services, million still pushed into extreme poverty because of health expenses goranitis, i. health care financing and the sustainability of health systems a european late starter: lessons from the history of reform in irish health care. health econ the impact of the financial crisis on human resources for health policies in three southern-europe countries integrating and rationalizing public healthcare services as a source of cost containment in times of economic crises pharmaceutical policies under economic crisis: the greek case. j. health policy outcomes res the fiscal crisis in the health sector: patterns of cutback management across europe health policy in times of austerity-a conceptual framework for evaluating effects of policy on efficiency and equity illustrated with examples from europe since presenting cost and efficiency measures that support consumers to make high-value health care choices pvs group feasibility of an implementation strategy for the integration of health promotion in routine primary care: a quantitative process evaluation is integration of healthy lifestyle promotion into primary care feasible? discussion and consensus sessions between clinicians and researchers prioritizing lean management practices in public and private hospitals steccolini, i. performance management in the public sector: the ultimate challenge health care privatization processes in europe: theoretical justifications and empirical classification the health gap: the challenge of an unequal world social justice, epidemiology and health inequalities national health expenditure projections, - : despite uncertainty, fundamentals primarily drive spending growth health at a glance: europe : state of health in the eu cycle commentary: swedish initiative on person centred care collaborative action for person-centred coordinated care (p c): an approach to support the development of a comprehensive system-wide solution to fragmented care person-centred care for patients with chronic heart failure-a cost-utility analysis ekman, i. effects of person-centred care after an event of acute coronary syndrome: two-year follow-up of a randomised controlled trial person-centred care after acute coronary syndrome, from hospital to primary care-a randomised controlled trial creating and facilitating change for person-centred coordinated care (p c): the development of the organisational change tool (p c-oct). health expect capability and well-being world health organization. ottawa charter for health promotion: health promotion; world health organization: ottawa new perspective on the health of canadians a working document; minister of supply and services: ottowa the global obesity pandemic: shaped by global drivers and local environments the need for a new medical model: a challenge for biomedicine critical realism and causal research theory and methods in sport policy and politics research process evaluation of complex interventions: medical research council guidance effects of person-centred care in patients with chronic heart failure: the pcc-hf study effects of person-centred and integrated chronic heart failure and palliative home care. prefer: a randomized controlled study effectiveness of person-centred care after acute coronary syndrome in relation to educational level: subgroup analysis of a two-armed randomised controlled trial person-centred care-ready for prime time the impact and utility of encounter patient decision aids: systematic review, meta-analysis and narrative synthesis using artificial intelligence to assess clinicians' communication skills the dark side of alliances managers at work: are you satisfied with your technology alliances? res. manag unpacking strategic alliances: the structure and purpose of alliance versus supplier relationships make it count management control systems: performance measurement, evaluation and incentives getting the most from strategic partnering optimal incentives for teams the sociology of groups and the economics of incentives: theory and evidence on compensation systems social norms and economic incentives in firms effective organizational control: a framework, applications, and implications the role of health it and delivery system reform in facilitating advanced care delivery electronic health records and support for primary care teamwork guidance to good practice: aria digitally-enabled, integrated, person-centred care for rhinitis and asthma what are the core elements of patient-centred care? a narrative review and synthesis of the literature from health policy, medicine and nursing patient-centred approaches to health care what is person-centred care and why is it important? health innovation network functionality and feedback: a realist synthesis of the collation, interpretation and utilisation of patient-reported outcome measures data to improve patient care pay for performance and contractual choice: the case of general practitioners in england an empirical study of the role of accounting data in performance evaluation target payments in primary care: effects on professional practice and health care outcomes the effect of financial incentives on the quality of health care provided by primary care physicians paying for performance to improve the delivery of health interventions in low-and middle-income countries are financial incentives related to performance? a meta-analytic review of empirical research selecting effective incentive structures in health care: a decision framework to support health care purchasers in finding the right incentives to drive performance the effects of monetary incentives on effort and task performance: theories, evidence, and a framework for research. account world health guidelines: incentives for health professionals new ways of setting rewards: the beyond budgeting model effects of a person-centred telephone-support in patients with chronic obstructive pulmonary disease and/or chronic heart failure-a randomized controlled trial methods for the economic evaluation of health care programmes world health organization. handbook for national quality policy and strategy: a practical approach for developing policy and strategy to improve quality of care scoping implementation science for the beginner: locating yourself on the "subway line" of translational research can practitioners use patient reported measures to enhance person centred coordinated care in practice? a qualitative study statistical process control and interrupted time series: a golden opportunity for impact evaluation in quality improvement challenges, solutions and future directions in the evaluation of service innovations in health care and public health diffusion of innovations in service organizations: systematic review and recommendations. milbank q role of the organization in translating research into evidence-based practice fostering implementation of health services research findings into practice: a consolidated framework for advancing implementation science developing and testing a model to predict outcomes of organizational change measurement of organizational culture and climate in healthcare a practical, robust implementation and sustainability model (prism) for integrating research findings into practice communicating and controlling strategy: an empirical study of the effectiveness of the balanced scorecard person-centred practice in nursing and health care: theory and practice why some employees adopt or resist reorganization of work practices in health care: associations between perceived loss of resources, burnout, and attitudes to change professionalism, the third logic: on the practice of knowledge physician professionalism and accountability: the role of collaborative improvement networks key driver : implement a data-driven quality improvement process to integrate evidence into practice procedures systematic review of the application of the plan-do-study-act method to improve quality in healthcare the kings fund using action learning to develop a more strategic approach to quality improvement at oxleas what is "quality improvement" and how can it transform healthcare? qual. saf. health care bringing user experience to healthcare improvement: the concepts, methods and practices of experience-based design experience-based co-design and healthcare improvement: realizing participatory design in the public sector patients and staff as codesigners of healthcare services point of care foundation ebcd: experience-based co-design toolkit this article is an open access article distributed under the terms and conditions of the creative commons attribution (cc by) license key: cord- -oy hsrpt authors: beutels, philippe p.a. title: economic aspects of vaccines and vaccination: a global perspective date: journal: the grand challenge for the future doi: . / - - - _ sha: doc_id: cord_uid: oy hsrpt nan a basic principle of economic decision-making is that in an environment of scarce resources choices have to be made in the allocation of these resources. this principle also applies to the provision of health care. the share of health-care expenditures in the gross domestic product (gdp) of most industrialised countries has increased from %- % in the early sixties to %- % in (from % to % in the usa) [ ] this rise has been attributed to medical advances (increasing the number and technological complexity of medical interventions), population aging, sociological changes (more, but smaller families and less familial support for the elderly) and insufficient productivity increases in the services sector. in less wealthy economies, medical decision-makers are faced with a smaller margin, and such a rise in health-care spending has not been observed yet. basically, the richer a country, the more it can afford (in nominal and in relative terms) to spend on health care. the two-way interaction between health and economic development is generally explained as follows. the healthier the population, the more adults can contribute to society by productive activity (i.e., work creating a surplus value in terms of capital gains and human resources), as well as by raising children in a stable environment, thus ensuring continued economic development. the process of economic development itself creates conditions (education, employment, infrastructure (including safe water and sanitation)), which provide a basis for continued improvement in longevity and health-related quality of life [ ] . individual good health can be seen as the product of some unknown complex function to which health care is only one of the inputs. other important inputs are: life-style variables (eating habits, smoking, etc.), environmental factors (urbanisation, climatic conditions, etc.), income, education and genetic predestination. furthermore, expenditures on health are not necessarily put to use in the most efficient economic aspects of vaccines and vaccination: a global perspective way. in this respect a distinction can be made between technical efficiency (providing maximal health care for a given cost, or delivering a certain service at minimal cost) and allocative efficiency (the distribution of resources across alternative services so as to maximise health gains, in accordance with preferences). finally, though much may be spent on health care, not all people may have equal access to health care of the same quality. indeed inequities in the consumption of health care may also interfere with the overall allocative efficiency of the system, and create inequities in health per se. therefore greater expenditures on health care are no guarantee for more global health. it should be noted that these observations do not plead for a reduction or containment of health-care budgets, but rather for a way of spending that ensures that societal goals are met. in order to achieve this, welfare economists focus research on two broad topics: efficiency and equity. efficiency relates to choosing options that maximise utility from marginal expenditures (i.e., by optimising the production process of health, for which health care is one of the inputs). equity relates to the fair distribution of all aspects related to health across members in society (e.g., equal access to care). clearly, there may exist a trade-off between efficiency and equity and giving priority to either of them is a normative issue that should be decided by social and political debate. vaccination is undoubtedly one of the major contributors to health improvements in the last three centuries. during this period, the impact of vaccination on longevity is undeniable, despite the fact that its partial contribution is difficult to distinguish from that of improved hygienic conditions and nutrition, and the discovery of penicillin [ , ] . all of these combined provided the basis for the so-called "epidemiological transition" in industrialised countries. at the same time, infectious diseases remain the main cause of death in many developing countries. despite the continuing expansion of the vaccine portfolio, implementing financially sustainable basic vaccination programs in poor countries remains problematic. though this is not so much an economic as a financial aspect, we will return to this issue in the section "financing vaccines". a number of peculiar characteristics set vaccination apart from other interventions in health care [ ] : ( ) since vaccination is (usually) a form of primary prevention, it intervenes in people (often children) who are generally in good health. but unlike most other prevention programs, the interven-tion itself can cause harm to the vaccine recipient, because in rare cases vaccine-associated adverse events (vaae) occur. this means that people make trade-offs between risks of vaccine-preventable disease and risks of vaae. the perception of these risks is quintessential to the individual demand for vaccines (if left to free-market mechanisms), and dominates the influence of other factors such as price [ ] . ( ) vaccination not only protects vaccine recipients, but it also reduces exposure of unvaccinated people, due to the reduced circulation of the infectious agent (if the transmission of infection occurs from human to human). this is not always beneficial for public health as the reduced risk of transmission leads to an increased average age at infection (with many "childhood" infections being more severe if contracted in adulthood) [ , ] . together with the first characteristic, this means that people generally have an interest in having everyone else vaccinated, but not themselves (or their children). ( ) a number of infections can be eradicated in the long run if vaccination efforts are sustained at sufficiently high coverage levels around the world. in other words, sometimes vaccination has the potential of making itself redundant. choices about eradication are closely linked to the welfare of future generations and societal time preference (see also below) [ , ] . since the perceptions outlined above are usually distorted by insufficient or biased information, government intervention (in the form of subsidies, or coercion) is desirable to ensure that vaccine uptake remains optimal. indeed, as uptake increases the risk of vaae remains constant, but individuals may perceive it to increase. at the same time the absence of vaccine-preventable disease may create a false sense of security, and lure people into believing that their risk of disease has reduced to zero as well, while this is highly dependent on historical and future rates of exposure and vaccination in the rest of the population. the vaccine market represents only . % of the global pharmaceutical market, but has high growth potential (estimated at - % per year by various sources, mainly due to new combination, new prophylactic and new therapeutic vaccines) [ ] . for a manufacturer, the contribution margins of vaccines are low compared to those of other products in both the developing and industrialised world (due to price and licensing regulations). the few suppliers of vaccines now aim to limit production to the projected global needs in any given year (unicef bought about % of "traditional" vaccine supply in , compared to about % in ). thus the market is very vulnerable to capacity problems: a problem with a single batch of vaccines by a single producer can have severe knock-on effects across the globe. this may also explain why, for some of the old vaccines, the price fluctuates, and has had the tendency to rise over the last years. close co-operation between demanders (governments or agencies) and suppliers is essential to ensure continued availability at the right time. vaccines are supplied under a tiered price system, with % of sales volume in developing countries and countries in transition, but % of sales revenue in industrialised countries [ ] . it is therefore not surprising that global vaccine manufacturers (with three big producers (glaxosmithkline, aventis and merck) occupying % of the global market) tend to focus on products for the industrialised world. it is to be expected that more combination vaccines will become available and existing combinations extended. examples of this include the hexavalent diphteria-tetanus-pertussis-inactivated polio virus -haemophilus influenza type b -hepatitis b (dtp-ipv-hib-hepb) vaccine, and the quadrivalent measles-mumps-rubella-varicella-zoster (mmrvz) vaccine. the research and development costs for these vaccines are high due to technical and regulatory complexity. the technicality, the multiple patents and requirements in terms of clinical trials (all demanding great time and money investments) increase barriers to enter the vaccine market. this may lead to more monopolistic behaviour, with risks to supply, choice and price. clearly, the benefits of combination vaccines are many. for instance, reductions in the number of injections and associated administration costs (including reduced money, time and pain costs for children and their parents), and reduced transmission by contaminated needles benefit recipients and the public health bodies. free-rider effects (important and not-soimportant vaccines can hook up with established vaccines, irrespective of how recipients perceive their importance) and economies of scope benefit manufacturers and perhaps public health bodies. these benefits will have to be traded off versus the higher price of combination vaccines. because governments, health insurers or agencies (unicef, paho) typically buy vaccines directly from producers, there is also little diversity on the demand side of the market. all of this implies that there is little competition on both sides of the market and that global societal goals (development and supply of affordable vaccines for poor countries as well as rich countries) are unlikely to be met by relying entirely on free-market mechanisms (particularly since these are hampered by (necessary) regulation with regards to quality control and licensing). by using economic evaluation we are essentially trying to answer the following questions [ ] : ) is the vaccination program under study worth doing compared to alternative ways of using the same resources? in other words: should the (health care) resources be spent on such a vaccination program, and not on something else? ) more specifically, if we are deciding to vaccinate against a particular disease, whom should we vaccinate, at which age, with which vaccine and how should the vaccine be delivered and administered in order to deploy our scarce resources in the most efficient way? most economic evaluations of vaccination are model-based, because the alternative, empirical analysis, is usually impractical, very time-consuming (for most vaccines it takes decades for the full effects to unfold), very expensive and potentially unethical. a complete economic evaluation should compare different options for an intervention, in terms of economic costs as well as health consequences. there may be several options to prevent an infectious disease, some of which are mutually exclusive, while others are complementary. the relevant costs and benefits need to be collected for each option, and calculated relative (incremental) to another option. the choice of the reference strategy against which the other options are evaluated can be highly influential for the results of the evaluation. unless it is a cost-ineffective strategy, current practice is the preferred strategy of reference. when a new vaccine is introduced, the reference strategy is often referred to as "doing nothing" (no vaccination), although in this case "doing nothing" usually means the treatment of cases as they arise, with the corresponding public health measures. a generalised distinction between the costs and benefits of vaccination is presented in table . the intervention costs dominate the cost side. these are the costs necessary to implement the vaccination program. additionally there are costs incurred to receive the vaccine. the benefits of vaccination are the gains in health and the avoided costs. direct costs can be avoided because less treatment is needed for curing or nursing the disease against which the vaccination program is aimed. additionally indirect costs can be avoided because vaccination may partly prevent people having to interrupt their normal activities in society because of their illness or the illness of their relatives. from the health-care payer's point of view, only direct medical costs need to be taken into account. however, from society's viewpoint, indirect non-medical costs are also relevant. other viewpoints can be those of patients, hospitals, travel clinics, insurance companies, employers, etc. (see fig. ). for each of these perspectives different costs and benefits may be relevant. this implies that it is possible for an intervention to be relatively cost-effective for one party involved, while it is not for another. different cost categories are listed in table . the listings in italics are often not taken into account, because they can be relatively small in comparison to the other costs and/or because they are difficult to estimate. sometimes their inclusion is not relevant to express the viewpoint of the analysis. however, if they are relevant for the viewpoint of the analysis, their impact on the results could be tested in a sensitivity analysis and their existence should be mentioned when the results are presented. some diseases affect expectations and behaviour beyond one degree of separation from the pathogen. for instance the global impact of the sars outbreak in was modest in disease burden ( probable cases, deaths) and associated health-care costs, but it had an impressive impact on the global economy (us$ - bn, or $ - m per case) in macro-economic terms (when the impact on consumption and investments are considered) [ , ] . a similar situation could arise for pandemic influenza, or any other disease that affects risk perceptions of consumers and investors (e.g., variant creutzfeld jacobs disease). however, for most currently vaccine-preventable diseases, micro-economic evaluation would provide an appropriate analytical framework, preferably adopting a societal perspective. in reality, decisions about universal vaccination are often taken from the perspective of the national health service (nhs) or the ministry of health and at best from the health-care payer's perspective (which in addition to the nhs costs also includes direct co-insurance and co-payments by the patient). indeed, decision-makers in health care tend to focus primarily on direct costs since these are most indicative of their immediate budgets, even if their decision has bearings on society at large. when it comes to estimating unit costs or prices, it should be noted that costs in an economic sense are opportunity costs: they represent a sacrifice of the next best alternative application [ ] . this entails that costs in an eco- philippe p.a. beutels health gains (physical and psychological) source: [ ] nomic sense are not necessarily the same as financial expenditures, and that they can also represent goods and services that are not expressed in monetary terms. however, market prices are often used as a proxy. if particular goods and services are not traded on a market, ("shadow -") prices of a similar activity can be used instead. for example, work of volunteers can be approximated by wages of unskilled labour. similarly, patients' leisure time could be based on average earnings or average overtime earnings. average costs per unit of output are the total costs of producing a quantity divided by that quantity. marginal costs constitute the additional costs of producing one additional unit of output. since decisions are made at the margin, marginal costs should be used where they are substantially different from average costs [ ] . for vaccination, this distinction is most relevant for estimating the costs of the program [ ] . the costs of adding a particular vaccine to the existing program depends on how well the schedule of the new vaccine fits in with the other schedules, whether specific precautions need to be taken, whether potential vaccinees need to be screened prior to vaccination or whether a specific target group is envisaged. the costs that are most heavily affected by adding a new vaccine to the existing program are the variable costs of the program (time spent per vaccinee, number of vaccines bought, etc.), whereas the influence on the fixed intervention costs (buildings, general equipment, etc.) is usually small (unless a new vaccine requires a substantially different infrastructure in terms of storage and transport). a good example of this is provided by hall et al. who examined the immunisation program in the gambia (more recently these results were confirmed by a similar analysis in addis ababa, ethiopia) [ , ] . they found that the additional costs of adding hepatitis b vaccine to the existing expanded program on immunisation (epi) vaccines (measles, polio, dtp and bacille calmette-guérin (bcg)), would be for % recurrent costs (of which % for purchasing hepatitis b vaccine (hepb)). still, the introduction of a new expensive vaccine could more than double the costs of the program in some countries because of its sheer price compared to other vaccines in the program. the main objective of vaccination is to prevent disease. the most important benefits from a public health point of view are therefore the health gains (see tab. ). these are both physical (avoiding illness, suffering, mortality, etc.), and psychological (avoiding distress, anxiety, etc.). specific vaccinerelated psychological health gains include the general feeling of well-being and security of vaccine recipients from knowing that they are protected against disease. this could evidently lead to behavioural changes (e.g., a vaccine against hiv/aids could have a large influence upon the sexual behaviour of vaccine recipients). the valuation of health outcomes has far-reaching consequences for the methodology and study design of applied analyses. generally, a distinction is made between four different methods, depending on the way in which health gains are measured [ ] . a cost-minimisation analysis compares the costs of equally effective alternatives, without quantifying the health gains. it differs from a pure cost philippe p.a. beutels source: [ ] analysis in that there is always more than one option analysed and that the effectiveness of the different alternatives is known to be equal. in a cost-effectiveness analysis, health gains are measured in one-dimensional natural units (e.g., infections prevented, hospitalisations prevented, deaths averted, life-years gained…), implying that only one aspect of effectiveness is considered (e.g., postponing the time of death) and other related aspects are not (e.g., the quality of life). the results of cost-effectiveness analyses (cea) are usually presented as a ratio. a cost-effectiveness ratio (cer) is a measure of the incremental costs, which are necessary to obtain one unit of a health gain by implementing a strategy j instead of a strategy i (expressed in incremental costs per life-year gained, incremental costs per infection prevented, etc.). the lower the ratio, the more efficiently strategy j gains health compared to strategy i. the units in which health gains are expressed should represent the final results or clinical endpoints of an intervention as adequately as possible, in order to enable comparison between different interventions [ ] . if, hypothetically, the cost-effectiveness of hepatitis b vaccination were $ per infection prevented, whereas hib vaccination is estimated at $ , per infection prevented, it is wrong to conclude that hepatitis b vaccination is more cost-effective (because it is less costly to prevent one infection). to make that judgement, the avoided effects would need to be expressed in a more comparable endpoint, like life-years saved. to make the comparison even more relevant, different health states should be weighed by their quality (scaled from (meaning death) to (meaning perfect health)). this approach is used in cost-utility analysis (cua), where health gains are measured in quality-adjusted life-years (qalys) saved or another combined measure of morbidity and mortality (e.g., disability-adjusted life-years (daly)). a cost-utility ratio (cur) is similar to a cer, except that the denominator contains the difference in qalys (or dalys), instead of the difference in natural units, such as cases avoided or life-years gained. the main advantage of cua over cea is that it allows comparison of very different health-care interventions, for instance, those that predominantly extend lives (e.g., flu vaccination of the elderly) with those that improve the quality of life (e.g., drugs against erectile dysfunction). in cost-benefit analysis (cba), the health gains are converted into monetary units, which, in theory, allows the many dimensions that are associated with an improvement in health status (over and above the length and health-related quality of life) to be included. there are benefits beyond the health outcomes such as information, caring, regret, anxiety reduction, communication and process utility (benefits from health-care use). further-more, option value (i.e., benefits derived from needing care in the future) and non-use value (i.e., externalities related to caring for the health of others) can also be (potentially) elicited [ ] . the results of a cba can be presented as the difference between costs and benefits (the net costs (or net savings)) or as a ratio. the benefit-cost ratio (bcr) expresses to which extent an investment in an intervention can be recovered by the consequences of that intervention (expressed as a unitless number or a %). cost-benefit analysis allows for comparisons between totally different projects in society (e.g., comparing a vaccination campaign with the construction of a new bridge). when budgets are very limited and many urgent interventions compete, as in developing countries, such cross-sector comparisons may actually be used in practice. clearly, the potential of cba to make such comparisons possible is a major advantage to aid decision-making. the strength of cba in theory, i.e., the explicit monetary valuation of health gains, has up till now been also its weakness in practical decisionmaking. in theory it seems preferable that the valuation of health gains (and of life) is done in an explicit, transparent and representative way as in cba, instead of the implicit, inconsistent and arbitrary way it is often done in today's decision-making. however, in a health-care environment the monetary valuation of health (and particularly of life) is often rejected on an emotional basis [ ] . additionally, economists have few credible arguments to counter these objections, as the current methods which place a monetary value on health (human-capital and willingness-to-pay methods) can hardly be called consistent and reliable in practice [ ] [ ] [ ] . in view of this, most economic evaluations in health care are based on cea or cua. the literature on these has increased exponentially since the s, for vaccines at least as much as for other interventions in health care, underlining the importance of a sound theoretical framework for these analyses [ ] . individuals (and societies), in general, prefer to receive benefits as early as possible and incur costs as late as possible. this so-called time preference means that the same amount of wealth or health would have a different value to a decision-maker in the present, if this amount is gained at different points in time. note that time preference has nothing to do with inflation. a vaccination program is an investment made in the present (i.e., the costs of buying and administering vaccines) to gain benefits spread out over the future (i.e., avoided costs of treatment, avoided morbidity and mortali-ty). discounting is a technique by which future events (e.g., costs and health outcomes) are valued less the further in the future they arise. the degree to which they are valued less is determined by the discount rate (frequently assumed to be constant through time): the higher the rate the less future benefits and costs are valued. although there is general agreement on the discounting of costs, the arguments for discounting non-monetised health outcomes are contradictory [ ] . discounting costs without discounting benefits leads, amongst others, to the paradoxical situation that any eradication program will yield infinite benefits [ ] . this would imply that all current resources should be spent on research of eradicable diseases, and the implementation of eradication programs, and not a single penny on cure. such paradoxes, and the observation that individuals generally have a positive discount rate for health, clearly indicates that health too should be discounted at a positive rate. but there is no general agreement on how the discount rate for health should compare to that of wealth. there are arguments to apply an equal discount rate to both costs and health effects [ , ] . the underpinnings and relevance of these are questionable, so that a lower discount rate for health effects than for costs has also been proposed [ , ] . because of the very long time spans over which benefits accrue, the analysis of most vaccination programs is very sensitive to discounting (of costs as well as health effects). nonetheless, this is no cause for a different approach to discounting for vaccination. still, further empirical research is needed to strengthen or to change the basis for conventional discount rates (mostly %, or %) and discount models (mostly stationary) [ ] . a slight decrease in discount rate (from, say, % to %) could change the cost-effectiveness of some vaccination programs from unattractive to attractive. also, it is likely that time preference in developing countries is substantially different (i.e., higher) from that in industrialised countries, particularly for those countries that have decreasing health (e.g., life-expectancy due to hiv/aids) or wealth (e.g., real gdp) expectations [ ] . in theory, decisions are made by interpreting the results of economic evaluation as follows. in figure a new program is plotted in terms of costs and effectiveness versus the reference strategy in the origin. if the new program is less costly and more effective than the reference, then the new program (a "dominant" strategy) should be implemented. likewise, if the new program is more costly and less effective than the reference, it should be rejected. in the other quadrants the decision is more complex, because it depends on a value judgement. if the incremental cer (or cur) is smaller than a given willingness to pay (or threshold cost-effectiveness criterion), "k", it would be acceptable. the question then is, how to determine k? this could be determined by social debate or by comparing it to what is widely accept-ed in practice. the most widely cited k in industrialised countries is $ , per qaly gained. there may also be a grey zone for k in which some interventions are implemented and others are not (e.g., between $ , per qaly gained and $ , per qaly gained), whereas under and above that grey zone all and none of the interventions are implemented, respectively. however, the greater the analytical uncertainty and the burden of disease, the more decisions are likely to deviate from such clear cut-off practices [ ] . different societies should have a different willingness to pay, though there are few instances in which societies (or their decision-makers) have tried to determine what the appropriate value of k is. the world bank has suggested using gnp per capita as a benchmark for k. note that in cba, k has already been given an explicit value. in league tables, many vaccination programs rank with the most costeffective interventions in health care in industrialised countries [ , [ ] [ ] [ ] . it is tempting to try and estimate the global historical value of vaccination. however, due to scarcity of data in most parts of the world such an exercise philippe p.a. beutels figure . the cost-effectiveness plane. cer: cost-effectiveness ratio, i.e., incremental costs divided by incremental effects; k = willingness to pay, or a cost-effectiveness ratio of acceptable magnitude. all points on a line in this plane have identical cost-effectiveness ratios. would be, by necessity, extremely crude. it seems clear, though, that the smallpox eradication program and the establishment of the epi have generated enormous benefits, not only by directly protecting against important vaccine-preventable diseases, but also by providing opportunities for health education and infrastructure in developing countries [ ] . yet the associated disease reduction in smallpox, measles and tetanus alone is bound to have been a cost-saving enterprise around the world (i.e., in the lower right quadrant of fig. ) , currently averting over million deaths per annum, compared to a "never having vaccinations" situation. however, when we are making choices today, we have to consider what additional benefits we will achieve by making additional investments, and this is bound to vary between countries at different stages of economic development, different epidemiologies of disease, and different historical vaccine-uptake levels. hence data from one country cannot always be simply extrapolated to another. in practice, there are many factors that come into play when decisions are made about new health-care interventions (see fig. ). in a democracy, a decision-maker receives a temporary, renewable mandate from the public to allocate a given budget. that person is well aware of the public perceptions of public health problems, and the impact of decisions thereon. at the same time, pressure groups may try to influence decision-makers or the public's perception. these pressure groups have vested interests in the decisions (be it as sellers of vaccines, or sellers of services for the cure of vaccine-preventable diseases). societal goals with regards to the decision can only be met by considering its medical, social, ethical and cost implications. the theoretical foundation of economic evaluation (so-called "pareto opti- figure . factors influencing decision-making in practice mality") addresses efficiency, without concern for distributional aspects (equity). therefore, economic evaluation combines the medical/epidemiological and cost implications, but does not consider the social and ethical implications depicted in figure (though in cba these aspects could theoretically be included, if a willingness-to-pay approach is used, and it is possible to weight quality-of-life gains to help achieve equity goals, as is commonly performed in dalys). therefore economic evaluation should be seen as an additional type of analysis that cannot stand on its own in its current form (it is an aid to decision-making, not a decision-maker in itself). at the same time, ideally, the influence of pressure groups, and the public's perceptions (rather than the public's true preferences) should be minimised in this process. it is noteworthy that most vaccination programs are likely to be equitable according to prevailing theories of justice [ ] . indeed, an analysis for bangladesh indicated that socio-economic inequalities in mortality of under- -year-olds were eliminated by measles vaccination [ ] . in the past, vaccination interests of poor and wealthy nations seemed more in tune than today. moreover, the research and development costs of the new generation of vaccines, based on biotechnology, are greater and the regulatory hurdles higher, meaning that new vaccines are much more expensive than the basic package of "traditional" vaccines. the first new expensive vaccine for global use was the hepatitis b vaccine, which became available in . the main reason why it was not immediately included in universal vaccination programs was its price, because initially the hepatitis b vaccine cost more than the other six epi vaccines put together. with the advent of more expensive vaccines, the introduction of a new vaccine is not as straightforward as it used to be in the industrialised world. in contrast to some of the "older" vaccines (e.g., measles, pertussis), newer vaccines may not result in net savings to the health-care system. nonetheless, if considered desirable, industrialised countries have no difficulty in financing the introduction of new vaccines, and ensuring the continuing uptake of old ones. for developing countries, the main difficulty is not so much to determine whether it would be cost-effective to introduce a vaccine, but to ensure that the introduction is financially sustainable. when external donors sponsor vaccination programs the sustainability takes the form of a partnership with shared responsibility and the promise by the receiver of the financing to create the conditions to become self-sufficient in the long run, either alone or by attracting further external funding. global immunisation efforts came under pressure as the epi, which was launched in the s as a way of building on the success of the smallpox eradication program, lost its momentum in the s, and failed to attain the year goal of % global vaccination coverage. indeed, global child-hood immunisation coverage against the six main target diseases (polio, dtp, measles and tuberculosis), which was less than % in , decreased from about % in to % in [ ] . coverage for the complete schedule of dtp remains well below % in tens of developing countries, mostly in sub-saharan africa. these countries are traditionally bottlenecks in the epi because of great financial constraints, the evolution of the hiv epidemic, logistical difficulties, poor governance and general socio-economic conditions (sometimes aggravated by war). as these factors evolved unfavourably in the s, international alliances shifted their efforts from reducing general global inequalities in health ("health for all") to more selective strategies, like the polio eradication program and the introduction of new and improved vaccines. the discrepancy between the developing and the industrialised world is likely to become greater, as private vaccine development focuses primarily on diseases that affect the wealthy. indeed, only about % of world drug sales is for african countries. it has been estimated that of all expenditures on health research (over $ billion per year), % is for diseases that affect % of the world's population [ ] . using recent examples, kaddar et al. assert that financing vaccination should be affordable by all countries, at least for the basic vaccines [ ] . the cost of fully immunizing a child with the basic vaccines is $ to $ , which typically represents % to % of public health expenditures, <$ . per capita or about . % of gdp. most vaccination costs are fixed costs of personnel and infrastructure, and the marginal costs of an additional vaccine may often be bearable for the domestic budget (though still highly dependent on vaccine price). as the th century drew to a close, the landscape of external vaccine financing underwent dramatic changes with the inception of the global alliance for vaccines and immunization (gavi) and the vaccine fund, with the aims to stimulate research and development for developing world problems, strengthen immunization systems, and promote and support the introduction of new and underused vaccines. gavi is an alliance of financiers (development banks, aid agencies, foundations), agencies (unicef, who), vaccine developers and manufacturers, as well as developing country governments, whereas the vaccine fund manages private financial resources, such as those from the bill &melinda gates foundation, and public contributions from a small number of wealthy countries. the first generation of vaccines, such as measles and oral polio vaccines, was used against common and serious childhood diseases afflicting all countries in the world. few of these vaccination programs were subject to economic analysis before introduction, and for good reason: the benefits were obvious and the costs low. indeed, they were probably amongst the most effective and cost-effective public health programs of the th century. this is no longer necessarily the case with new vaccine introductions. new vaccines are generally higher priced and unlikely to fall in price to the level of the first generation of vaccines. furthermore, they are often aimed at less common or less serious diseases (particularly in the industrialised world). thus, whether these vaccines are worth introducing is less clear. vaccine financing has recently changed with important initiatives stimulating development and use of vaccines for the developing world. these are to be welcomed as they may further alleviate the disease burden in developing countries at affordable cost, correct market imperfections with regard to research and development, and reduce inequalities in health. nonetheless, the introduction of new vaccines demands cautious planning. if it comes at the expense of the uptake of the first generation of vaccines, it may have a detrimental influence on the effectiveness and cost-effectiveness of the whole program. in view of all these developments, the role of economic evaluation in vaccine program design is only likely to increase in the future. world development indicators, online publication available on cd rom commission on macroeconomics and health: investing in health for economic development the conquest of smallpox an interpretation of the modern rise of population in europe economic evaluation of vaccination programmes in humans: a methodological exploration with applications to hepatitis b, varicella-zoster, measles, pertussis, hepatitis a and pneumococcal vaccination economic epidemiology and infectious disease infectious diseases of humans -dynamics and control increase in congenital rubella occurrence after immunisation in greece: retrospective survey and systematic review economics of eradication vs control of infectious diseases lecture at the advanced course in vaccinology, international vaccine institute assessing the economic impact of communicable disease outbreaks: the case of sars globalization and disease: the case of sars methods for the economic evaluation of health care programmes the cost of integrating hepatitis b virus vaccine into national immunization programmes: a case study from addis ababa cost-effectiveness of hepatitis b vaccine in the gambia providing health care. the economics of alternative systems of finance and delivery methodological issues and new developments in the economic evaluation of vaccines the economics of health and medicine estimating costs in costeffectiveness analysis evaluation of life and limb: a theoretical approach theory versus practice: a review of "willingness to pay" in health and health care discounting of life saving and other non-monetary effects foundations of cost-effectiveness analysis for health and medical practices discounting costs and effects: a reconsideration discounting for health effects in cost-benefit and cost-effectiveness analysis does nice have a cost-effectiveness threshold and what other factors influence its decisions? a binary choice analysis cost-effectiveness analysis and the consistency of decision making: evidence from pharmaceutical reimbursement in australia a comprehensive league table of cost-utility ratios and a sub-table of "panel-worthy" studies fivehundred life-saving interventions and their cost-effectiveness the societal value of vaccination in developing countries measles vaccination improves the equity of health outcomes: evidence from bangladesh the state of the world's children. early childhood the / gap report financial challenges of immunization: a look at gavi the author is grateful to dr john edmunds (health protection agency, uk) for constructive comments on an earlier version, and to the flemish fund for scientific research (fwo g. . ) and the european union funded project polymod (eu fp ) for financial support. key: cord- -gilnlwms authors: nahar, nazmun; asaduzzaman, mohammad; sultana, rebeca; garcia, fernando; paul, repon c.; abedin, jaynal; sazzad, hossain m. s.; rahman, mahmudur; gurley, emily s.; luby, stephen p. title: a large-scale behavior change intervention to prevent nipah transmission in bangladesh: components and costs date: - - journal: bmc res notes doi: . /s - - - sha: doc_id: cord_uid: gilnlwms background: nipah virus infection (niv) is a bat-borne zoonosis transmitted to humans through consumption of niv-contaminated raw date palm sap in bangladesh. the objective of this analysis was to measure the cost of an niv prevention intervention and estimate the cost of scaling it up to districts where spillover had been identified. methods: we implemented a behavior change communication intervention in two districts, testing different approaches to reduce the risk of niv transmission using community mobilization, interpersonal communication, posters and tv public service announcements on local television during the – sap harvesting seasons. in one district, we implemented a “no raw sap” approach recommending to stop drinking raw date palm sap. in another district, we implemented an “only safe sap” approach, recommending to stop drinking raw date palm sap but offering the option of drinking safe sap. this is sap covered with a barrier, locally called bana, to interrupt bats’ access during collection. we conducted surveys among randomly selected respondents two months after the intervention to measure the proportion of people reached. we used an activity-based costing method to calculate the cost of the intervention. results: the implementation cost of the “no raw sap” intervention was $ , and the “only safe sap” intervention was $ , . the highest cost was conducting meetings and interpersonal communication efforts. the lowest cost was broadcasting the public service announcements on local tv channels. to scale up a similar intervention in districts where niv spillover has occurred, would cost between $ . and $ . million for one season. placing the posters would cost $ , and only broadcasting the public service announcement through local channels in districts would cost $ , . conclusions: broadcasting a tv public service announcement is a potential low cost option to advance niv prevention. it could be supplemented with posters and targeted interpersonal communication, in districts with a high risk of niv spillover. electronic supplementary material: the online version of this article (doi: . /s - - - ) contains supplementary material, which is available to authorized users. transmission may reduce the risk of a potentially large outbreak. based on previous pilot studies on interrupting bats access to sap [ ] [ ] [ ] , and on the government of bangladesh's recommendation to abstain from drinking raw sap, we developed and implemented a behavior change communication intervention using two different approaches to reduce the risk of niv transmission. after the intervention, local residents' knowledge of niv increased, and people reported changing their behavior to reduce the risk of niv transmission through date palm sap [ ] . thus, understanding the intervention development, process and logistics will help plan scaling it up. calculating the approximate cost of the intervention, and the proportion of people to be reached, is useful to make investment decisions [ ] [ ] [ ] between potential interventions to prevent not just niv, but other emerging zoonoses. the objective of our paper is to describe and calculate the cost of an already implemented behavior change communication intervention, and estimate the cost of scaling it up to districts where niv spillover was identified in bangladesh, using risk-based scenarios. we developed a behavior change communication intervention using two separate approaches, targeting rural areas from two niv endemic districts: rajbari and faridpur, where date palm trees are harvested and residents drink raw date palm sap (fig. ) . we selected these districts because both have been repeatedly affected by niv outbreaks, both are from the same geographical region, neighboring each other, and have similar raw sap collection and consumption practices. within those districts, we selected two sub-districts that do not border each other to avoid interference between the interventions. the population of rajbari and faridpur study sites was approximately , and , respectively. following the government of bangladesh's recommendation of abstaining from drinking raw sap, we developed an intervention discouraging people from drinking raw date palm sap in rajbari district, herein referred to as the "no raw sap" intervention. some people continued to drink raw sap though they were aware of the risk [ ] , thus we developed an "only safe sap" intervention in faridpur district, discouraging drinking raw sap but offering the option of drinking sap protected by a skirtlike barrier locally called bana (fig. ) . during collection, banas can stop bats from accessing and contaminating the sap with niv [ ] . we worked with a bangladeshi communication organization to develop posters, calendars, yearly planners, stickers, sweatshirts and tv public service announcements. our qualitative research data collection team pretested the materials conducting focus group discussions with audiences similar to our target audience. based on these results, we revised and fine-tuned the messages and illustrations. we also developed training guides for the staff implementing the intervention. the communication organization designed and printed the final training guides. we developed the "no raw sap" intervention, including production of the communication materials, from june to october, and the "only safe sap" intervention from august to september, (fig. ) . we visited local ngos from both districts to assess their experience and capability to implement the interventions in the selected sub-districts. using a competitive bidding process, we selected one local ngo from each district. we assessed their experience with similar interventions, knowledge of the areas to be covered and qualifications of their key personnel. we also compared the size of the organizations, as an indicator of their capacity to implement the intervention, and the budget required to carry it out. the selected ngos visited villages and talked to villagers to get an estimate on the number of households, and identify opinion leaders and local sap harvesters (gachhis). we provided training to the ngos' staff on interpersonal communication, on organizing and conducting meetings with opinion leaders and community residents, and on key intervention messages. in both intervention areas, the ngos conducted one opinion leaders and one community meeting per households approximately. prior to conducting the meetings, the ngos affixed niv prevention posters in public places such as health centers, bazaars, and areas with heavy traffic of people. we provided calendars or yearly planners, with niv prevention messages, to the opinion leaders, and broadcast-quality public service announcements, in the form of dvds, to the local tv channels. in the "only safe sap" area, the ngo trained gachhis on making banas, and encouraged using them on trees used for raw sap consumption. we also provided sweatshirts as an incentive to those gachhis who made and used banas. date palm sap is harvested during cold months from november to march [ ] . we implemented a full "no raw sap" intervention from december , to march , in villages in rajbari district (fig. ) . during the next sap harvesting season, from november , to january , , we only broadcast the tv public service announcement. we implemented a full "only safe sap" intervention from october , to january , in villages in faridpur district, including a gachhi training component. we started the "only safe sap" intervention slightly before the sap season because we needed to train gachhis on making and using banas before they started collecting sap. during the intervention implementation period, we received ngo weekly reports with photographs of the meetings. our monitoring team visited randomly selected villages to confirm placement of at least one poster, watched the tv public service announcements at least in one tea stall, and observed one meeting per village incognito. tea stalls with a television set exist in almost every village, and serve as gathering places where men drink tea, watch television and chat with others. since most of the villagers do not have television at home, this communication channel was used to target men. we also recruited tea stalls with television access in each study area to monitor the number of times the tv public service announcement was broadcast daily. we collected written weekly reports from those tea stalls, indicating dates and times when the announcements were broadcast. after the intervention, during april-may , our quantitative data collection team interviewed adult male and female respondents from randomly selected villages from each "no raw sap" and "only safe sap" district. we described the sampling procedure for this study elsewhere [ ] . our data collection team asked about niv knowledge, sap consumption behavior, use of banas and exposure to the interventions. in this manuscript, we only present data about the respondents' direct exposure to the intervention. we used an activity-based costing approach to compare health interventions [ ] [ ] [ ] [ ] [ ] . we identified, costed out, and quantified all development and implementation activities. we reviewed timelines and deliverables to confirm activities performed, transport requisition emails, and budgets submitted to the donor. we calculated the cost per activity performed using person time, with the exception of ngo activities that were calculated using per activity cost instead of person time cost. we separated the start-up cost from the intervention implementation cost ( table ). the start-up cost covered the development of materials before the implementation, from the period of time between the decision to implement, to the start of its delivery to the beneficiaries [ ] . because we developed some of the materials for both interventions, we were not able to completely separate the cost of developing all the materials for each intervention. thus, we could not add the start-up cost to the implementation cost to determine the total cost per intervention. the implementation cost included ngo cost, mass media dissemination expenditures (local tv channel, dvds copies and printing posters) and intervention monitoring cost. the cost of training ngo staff included training manuals, personnel, snack allowance, venue, electricity, photocopies, and transportation. in the "only safe sap" area, we also included the cost of bana-making materials and the allowance and transportation cost of a banamaking expert as part of the ngo staff training cost. the cost of training the ngo staff and printing the materials would be incurred before any future implementation, thus we included them in the implementation cost. we calculated the amount of money the ngos spent as cost of the meetings and gachhi training. since ngo staff affixed posters while visiting the villages for meeting purposes, the ngos did not include the cost for placing posters separately in their reporting. to estimate this cost, we assumed that one person could visit four villages per day, to affix posters per village, and estimated the cost of affixing one poster based on the daily salary, meal allowance and transportation costs. we deducted these costs from the meetings cost to calculate the cost per meeting. we calculated costs in us dollars, using a rate of . bangladeshi takas per us$ , the conversion rate used on the original budget. we did not include the cost of the research study in this analysis. we calculated the start-up cost first, followed by the implementation cost of the interventions. we calculated cost per meeting by dividing the total cost to conduct all meetings, provided by the ngos, by the total number of meetings conducted; and the cost per gachhi training by dividing the total training cost provided by the ngos by total number of gachhis trained. from our survey data, we calculated the percentage of people directly reached or exposed to each communication channel used during the intervention [ ] . we found that a lower percentage of respondents from the "no raw sap" area reported that they were directly exposed to the intervention than the respondents from the "only safe sap" area ( % vs. %). also a lower percentage of respondents reported exposure to each intervention component: tv public service announcement ( % vs. %), saw a poster ( % vs. %) and attend a meeting ( % vs. %) in the "no raw sap" area than the "only safe sap" area [ ] . we calculated the cost per person reached per channel by dividing the implementation cost by the total population ( , in the "no raw sap" area and " , in the "only safe sap" area) times the percentage of people reached per channel. we estimated the future start-up cost and intervention implementation cost in all districts where at least one niv spillover has been identified in the past. we added person-day cost for activities, including the cost of revising the intervention and materials, identifying cable operators, cost for transportation and phone communication. using different risk-based scenarios in all affected districts where niv spillovers were identified from • six districts with six or more spillovers ( % of all spillovers) • thirteen districts with two to five spillovers ( % of all spillovers) • eleven district with one spillover, ( % of all spillovers). we estimated the implementation cost at the district level, based on implementation expenditures during the - interventions. to estimate the cost of the meetings for a future intervention, we estimated the number of rural households in all sub-districts using census data [ ] . we projected conducting one opinion leaders meeting and one community meeting per every rural households, using the cost-per-meeting from the "only safe sap" area. we projected the approximate number of gachhis using ngo data from the "only safe sap" area ( gachhis per village or within households). to estimate the cost of training the gachhis we used the per-gachhi training cost from the "only safe sap" intervention. we assumed two cable operators per sub district to estimate the cost of broadcasting the tv public service announcement. we incurred most of the start-up costs developing the intervention, including expenditures on national and international experts and local staff, materials' pretesting, revisions and production, districts and ngo selection, and training of trainers ( table ). the second highest cost was the production of the tv public service announcements, followed by the cost of creating and producing the other communication materials. the ngos conducted opinion leaders and community meetings in the "no raw sap" area, and opinion leaders and community meetings in the "only safe sap" area. they affixed posters in the "no raw sap" area and posters in the "only safe sap" area. local channels broadcast the tv public service announcements times daily. in addition, in the "only safe sap" area, the local ngo conducted gachhi training sessions on how to make and use banas. our implementation cost was lower in the "no raw sap" intervention than in the "only safe sap" intervention ($ , vs. $ , ) ( table ). the cost of the intervention components, broadcasting the tv public service announcement ($ vs. $ ), promoting posters ($ vs. $ ) and conducting community meeting costs ($ , vs. $ , ) was lower in the "no raw sap" intervention than in the "only safe sap" intervention ( table ) . the cost per person directly reached by at least one intervention component was also lower in the "no raw sap" area than in the "only safe sap" area ( cents vs. cents). the cost to reach one person per communication channel was lower in the "no raw sap" area than in the "only safe sap" area: tv public service announcement was . cents versus . cents, poster was . cents versus . cents, and community meetings was cents versus cents. the cost of the gachhi training program in the "only safe sap" area, including the incentive of providing a sweatshirt to those observed using banas during follow up visits, was $ , . the per gachhi cost with incentive was $ . with no incentive was $ . ( table ) . to scale up the intervention, we estimated the start-up cost at $ , (table ; additional file ). our future estimated implementation cost of meetings, posters and the public service announcement was the same for both the "no raw sap" and the "only safe sap" intervention (table ). however, the gachhi training component increased the cost of the "only safe sap" intervention. thus, the implementation cost of a future intervention covering districts would be $ . million using an "only safe sap" approach, and $ . million using a "no raw sap" approach ( table ). the cost of printing and affixing the posters in districts would be $ , . broadcasting the tv public service announcement in districts would cost $ , . to implement an "only safe sap" intervention with community meetings, gachhi training, poster and the tv public service announcement in the six districts with % of all spillover would cost $ , . to implement it in the second most affected area, thirteen districts with % of all spillover, would cost $ . million and in eleven districts with % of all spillover, would cost $ . million. to implement a full "no raw sap" intervention with community meetings, posters and the tv public service announcement in the six most affected districts would cost $ , . in the second most affected thirteen districts it would cost $ million and another $ , to implement it in the other districts. we spent $ , implementing the "no raw sap" intervention and $ , on the "only safe sap" intervention. to scale these interventions up to districts in bangladesh where human infections with niv have been identified, we estimated a cost of $ . million us$ for the "no raw sap" and $ . million us$ for the "only safe sap" intervention. niv usually affects impoverished rural communities in bangladesh, thus, affected families often experience a severe social and financial crisis [ , ] . niv kills people and leaves survivors with permanent neurological sequelae, similar to those experienced by some survivors of japanese encephalitis [ , ] . sixty-one percent of niv cases affected males with a mean age of [ ] who could be the main wage earners of the family. most died [ ] , and those that survived could not continue to work due to the neurological effects of niv. in addition, niv is a disease that requires special care. hospitalization and illness episodes can last a week [ ] . the financial burden associated with hospitalization translates into reduced monthly food and children education expenditures, having to borrow money, taking loans with high interests, and selling assets [ ] [ ] [ ] . prevention could reduce the risk of disease transmission as well as save poor families from social degradation. despite the severity of nipah illness, since an average of fewer than niv cases are identified annually in bangladesh [ ] , the cost of niv prevention is unlikely to meet the traditional criteria for cost-effective interventions to prevent cases [ ] . however, in addition to causing sickness and death, outbreaks have social consequences including fear, social unrest, violence and economic loss [ ] [ ] [ ] [ ] . for diseases with moderate to high perceived severity, such as pandemic influenza, sars or ebola, investing and intervening earlier in the outbreak can be cost effective [ ] . niv is a deadly disease that can transmit from person to person and represents a global pandemic threat [ , ] . estimating niv prevention costs is of interest to local and global health communities, helping to make informed decisions on funding interventions to prevent this disease. if we prevent a large high-mortality niv pandemic, an effective intervention would be remarkably cost-efficient. disaster preparedness reduces the impact of disasters and associated costs, compared to a scenario without preparedness [ ] . initiatives to mitigate low probability, high catastrophic risks are not uncommon. nasa spends millions of dollars each year to track asteroids, though chances of dying from an asteroid impact are very low for the average person in the united states [ ] . investing in active surveillance activities for zoonotic infections, implementing effective ecological health interventions, improving modeling capabilities, increasing evaluations of health systems and public health needs and policies, and implementing better risk communication can improve the preparedness to respond to emerging infectious diseases [ ] . for example, taiwan established a nationwide emergency department, based on a syndromic surveillance system, that collaborated with hospitals for better public heath response to improve their pandemic flu preparedness and disease control capabilities [ ] . similarly, investing in preventing niv could provide an important benefit. health intervention studies from bangladesh, focusing on cost, find some similarities with our study [ ] [ ] [ ] . a study on neonatal and child health reported a lower cost per person reached through local tv channels than other intervention components [ ] . in our intervention, the cost of interpersonal communication was around times higher than broadcasting the televised public service announcement in the "only safe sap" area. the estimated cost of posters was also low and could be integrated in future interventions. findings from our trial suggested more behavior change resulted from a one season "only safe sap" intervention than a two-season "no raw sap" intervention [ , ] . this could be because the "only safe sap" intervention offered the option of drinking safe sap by promoting the use of banas among gachhis, an already existing behavior [ ] that still allowed people to enjoy drinking sap. the gachhi training component might also have contributed to increased exposure to the intervention. although its estimated scale up cost was higher than the "no raw sap" intervention, for upcoming seasons, the "only safe sap" intervention should be considered. spending us$ . million annually on an "only safe sap" intervention would be prohibitively costly for a low-middle income country like bangladesh that currently spends only $ . per capita per year for healthcare [ ] and . % of gross domestic product in total health expenditures [ ] . the high cost of the meetings used in this intervention makes it impossible to scale up and sustain this intervention without external funding. reducing meetings and interpersonal communication would reduce costs and so increase the feasibility of scaling it up. we could achieve a lower cost intervention by including community health workers [ ] and health workers from the expanded program of immunization (epi), as well as health workers from ngos such as brac [ , ] . they could conduct meetings in the areas immediately surrounding their offices, affix posters, provide leaflets, and disseminate messages to people receiving their services during the sap harvesting season, adding a minimal cost. in addition, eliminating the gachhi incentive for using banas would reduce the cost of the gachhi intervention by more than one-third. our intervention findings provide a framework to calculate costs of a future intervention to prevent niv. however, the following limitations of our findings require consideration. we did not include the intervention impact data in the results of this cost manuscript, therefore, we cannot calculate cost-effectiveness. the complexity of the impact data required a separate manuscript to be properly presented. nevertheless, this cost analysis, conducted from a provider's perspective, enables future providers to weight the costs of taking on this intervention against those of other interventions [ ] . better understanding of the cost, from intervention providers and recipients, would provide an understanding of costrelated potential barriers and obstacles to implementing the intervention. although we calculated the separate cost of each intervention component, we cannot interpret the separate impact of each component. since communication campaigns often rely on a synergistic effect, all of its components may need to run in parallel for maximum impact [ ] [ ] [ ] . therefore, although deploying only a single component markedly reduces cost, this body of work does not provide direct evidence that the standalone components will alter behavior. to reduce costs, we proposed engaging government and other health workers to conduct meetings within their locality. since, they already have other tasks to accomplish, small-scale pilot efforts could help identify practical strategies to integrate niv prevention messages into health worker activities. the government already broadcast the "no raw sap" public service announcement during the - season. continuing to measure the prevalence of raw sap consumption as these messages are disseminated more widely can provide useful guidance on adjusting interventions and messages going forward. exploring low cost strategies to communicate prevention messages in frequently affected districts, such as broadcasting the public service announcement on local channels, combined with health workers visiting communities to spread messages and affix posters in districts with high risk of niv spillover, may be an effective way to reduce the risk of niv. continuous monitoring efforts may help to further develop and refine the intervention components for more effective communication. person-to-person transmission of nipah virus in a bangladeshi community nipah virus outbreak with person-to-person transmission in a district of bangladesh recurrent zoonotic transmission of nipah virus into humanbangladesh nipah virus infection outbreak with nosocomial and corpse-tohuman transmission date palm sap linked to nipah virus outbreak in bangladesh foodborne transmission of nipah virus date palm sap collection: exploring opportunities to prevent nipah transmission a randomized controlled trial of interventions to impede date palm sap contamination by bats to prevent nipah virus transmission in bangladesh piloting the promotion of bamboo skirt barriers to prevent nipah virus transmission through date palm sap in bangladesh piloting the use of indigenous methods to prevent nipah virus infection by interrupting bats' access to date palm sap in bangladesh the impact of a behavior change communication intervention to prevent nipah virus transmission in bangladesh. iceid: atlanta achieving universal coverage with health interventions economic analysis for evidence-based policy-making on a national immunization program: a case of rotavirus vaccine in thailand prevention of cardiovascular disease in high-risk individuals in lowincome and middle-income countries: health effects and costs raw sap consumption habits and its association with knowledge of nipah virus in two endemic districts in bangladesh an activity-based cost analysis of the honduras community-based, integrated child care (ain-c) programme. health policy plan filmless versus film-based systems in radiographic examination costs: an activity-based costing method evaluation of economic efficiencies in clinical retina practice: activity-based cost analysis and modeling to determine impacts of changes in patient management applying activity based costing (abc) method to calculate cost price in hospital and remedy services activity-based costing in services: literature bibliometric review programme costs in the economic evaluation of health interventions the effect of two different intervention approaches on a nipah virus prevention intervention in bangladesh. in: th european congress on tropical medicine and international health population and housing census statistics and informatics division, ministry of planning effect of parent's death on child survival in rural bangladesh: a cohort study socioeconomic and health implications of adult deaths in families of rural bangladesh longterm neurological and functional outcome in nipah virus infection japanese encephalitis clinical presentation of nipah virus infection in bangladesh economic burden of influenza-associated hospitalizations and outpatient visits in bangladesh during . influenza other respir viruses sauerborn r. health coping strategies of the people vulnerable to climate change in a resource-poor rural setting in bangladesh coping strategies for financial burdens in families with childhood pneumonia in bangladesh guide to cost-effective analysis. world health organization, geneva contract no biological impact of social disruption resulting from epidemic disease a time of fear": local, national, and international responses to a large ebola outbreak in uganda epidemic psychology: a model. sociol health illn economic impact of ebola virus disease in west african countries: a call for national and regional containment, recovery and prevention united: united nations development group-western and central africa cost-effective control of infectious disease outbreaks accounting for societal reaction the pandemic potential of nipah virus infection and disease in reservoir and spillover hosts: determinants of pathogen emergence investing in disaster management capabilities versus pre-positioning inventory: a new approach to disaster preparedness no reduction in risk of a massive asteroid impact lessons from the ebola outbreak: action items for emerging infectious disease preparedness and response establishing a nationwide emergency department-based syndromic surveillance system for better public health responses in taiwan costs of integrating demandbased reproductive health commodity model into the government and ngo service delivery systems in bangladesh: a supply side perspective economic evaluation of neonatal care packages in a cluster-randomized controlled trial in sylhet cost of behavior change communication channels of manoshi -a maternal, neonatal and child health (mnch) program in urban slums of dhaka a controlled trial to reduce the risk of human nipah virus exposure in bangladesh the world bank data on health expenditure per capita (current us$) bangladesh health system review avenue appia assessing community based improved maternal neonatal child survival (imncs) program in rural bangladesh the effect of payment and incentives on motivation and focus of community health workers: five case studies from low-and middle-income countries collecting and analysing cost data for complex public health trials: reflections on practice increases in self-reported consistent condom use among male clients of female sex workers following exposure to an integrated behaviour change programme in four states in southern india impact of a behaviour change intervention on long-lasting insecticidal net care and repair behaviour and net condition in nasarawa state using behavior change communication to lead a comprehensive family planning program: the nigerian urban reproductive health initiative additional file . start-up cost to prepare the intervention covering nipah-affected districts with at least one nipah spillover, bangladesh, supporting document for table . the authors declare that they have no competing interests. all relevant data have been presented in the main paper and an additional supporting file has been uploaded. the ethical review committee of international centre for diarrhoeal disease research, bangladesh and family health international ′s institutional review board reviewed and approved the study protocol. the data collection team obtained written informed consent from the respondents before conducting interviews. support for this study was provided by fhi with funds from usaid cooperative agreement ghn-a- - - - . springer nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. key: cord- -b s es authors: kelso, joel k; halder, nilimesh; postma, maarten j; milne, george j title: economic analysis of pandemic influenza mitigation strategies for five pandemic severity categories date: - - journal: bmc public health doi: . / - - - sha: doc_id: cord_uid: b s es background: the threat of emergence of a human-to-human transmissible strain of highly pathogenic influenza a(h n ) is very real, and is reinforced by recent results showing that genetically modified a(h n ) may be readily transmitted between ferrets. public health authorities are hesitant in introducing social distancing interventions due to societal disruption and productivity losses. this study estimates the effectiveness and total cost (from a societal perspective, with a lifespan time horizon) of a comprehensive range of social distancing and antiviral drug strategies, under a range of pandemic severity categories. methods: an economic analysis was conducted using a simulation model of a community of ~ , in australia. data from the pandemic was used to derive relationships between the case fatality rate (cfr) and hospitalization rates for each of five pandemic severity categories, with cfr ranging from . % to . %. results: for a pandemic with basic reproduction number r( ) = . , adopting no interventions resulted in total costs ranging from $ per person for a pandemic at category (cfr . %) to $ , per person at category (cfr . %). for severe pandemics of category (cfr . %) and greater, a strategy combining antiviral treatment and prophylaxis, extended school closure and community contact reduction resulted in the lowest total cost of any strategy, costing $ , per person at category . this strategy was highly effective, reducing the attack rate to %. with low severity pandemics costs are dominated by productivity losses due to illness and social distancing interventions, whereas higher severity pandemic costs are dominated by healthcare costs and costs arising from productivity losses due to death. conclusions: for pandemics in high severity categories the strategies with the lowest total cost to society involve rigorous, sustained social distancing, which are considered unacceptable for low severity pandemics due to societal disruption and cost. results: for a pandemic with basic reproduction number r = . , adopting no interventions resulted in total costs ranging from $ per person for a pandemic at category (cfr . %) to $ , per person at category (cfr . %). for severe pandemics of category (cfr . %) and greater, a strategy combining antiviral treatment and prophylaxis, extended school closure and community contact reduction resulted in the lowest total cost of any strategy, costing $ , per person at category . this strategy was highly effective, reducing the attack rate to %. with low severity pandemics costs are dominated by productivity losses due to illness and social distancing interventions, whereas higher severity pandemic costs are dominated by healthcare costs and costs arising from productivity losses due to death. conclusions: for pandemics in high severity categories the strategies with the lowest total cost to society involve rigorous, sustained social distancing, which are considered unacceptable for low severity pandemics due to societal disruption and cost. keywords: pandemic influenza, economic analysis, antiviral medication, social distancing, pandemic severity, case fatality ratio background while the h n virus spread world-wide and was classed as a pandemic, the severity of resulting symptoms, as quantified by morbidity and mortality rates, was lower than that which had previously occurred in many seasonal epidemics [ ] [ ] [ ] . the pandemic thus highlighted a further factor which must be considered when determining which public health intervention strategies to recommend, namely the severity of symptoms arising from a given emergent influenza strain. the mild symptoms of h n resulted in a reluctance of public health authorities to use rigorous social distancing interventions due to their disruptive effects, even though modelling has previously suggested that they could be highly effective in reducing the illness attack rate [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] . had the h n influenza strain been highly pathogenic, more timely and rigorous responses would have been necessary to mitigate the resultant adverse health outcomes. furthermore, there is continuing concern that a highly pathogenic avian influenza a(h n ) strain may become transmissible between humans. this scenario is highlighted by the large reservoir of influenza a(h n ) in poultry in south-east asia [ ] , and recent experimental results which have shown that the a (h n ) virus may be genetically modified to become readily transmissible between ferrets, a commonly used animal model for human influenza transmission studies [ ] [ ] [ ] . the severity of a particular influenza strain directly impacts on the cost of any pandemic; increased severity increases health care costs and escalates productivity losses due to a) absenteeism arising from increased illness and b) increased mortality rates. in this study, the role which pandemic severity has on the total cost of a pandemic for a range of potential intervention strategies is analysed, and for highly pathogenic influenza strains inducing significant morbidity and mortality, as occurred during the pandemic [ , ] , the results suggest which intervention strategies are warranted in terms of reduction of illness and total pandemic cost. this study adopts a societal perspective on the cost of a pandemic, with the time horizon being the lifetime of individuals experiencing the pandemic. we used a detailed, individual-based simulation model of a real community in the south-west of western australia, the town of albany with a population of approximately , , to simulate the dynamics of an influenza pandemic. comparing simulations with and without interventions in place allowed us to analyse the effect which a range of interventions have on reducing the attack rate and on the health of each individual in the modelled community. epidemic outcome data produced by the simulation model were used to determine health outcomes involving hospitalisation, icu treatment, and death. in turn, these healthcare outcomes, together with productivity losses due to removal from the workforce, were used to estimate the overall cost of interventions. figure provides an overview of this analysis methodology, showing each of the processes that make up the methodology, their input parameters and the resulting data generated by the process. the simulation model captures the contact dynamics of the population of albany, western australia using census and state and local government data [ ] . these data allowed us to replicate the individual age and household structure of all households in this town of approximately , individuals, and also allowed for the construction of an explicit contact network linking households, schools, workplaces and other meeting places by allocating individuals to workplaces and schools. the modelled community was chosen so as to be representative of a developed world population, and selfcontained in the sense that all major locales for interpersonal mixing were represented within the community. the model includes both urban and rural components, a central commercial core, a complete set of schools (covering all age groups), and a mix of large and small employers. the community is also of a size where public health interventions could be uniformly implemented based on local information. the model captures explicit person-to-person contact with the contact network describing population mobility occurring between households, schools, workplaces and the wider community as shown in figure . the virus spreads through the community due to this mobility, as transmission occurs between individuals when they are co-located, possibly following a move from one location to another. for figure overview of pandemic cost analysis methodology. input parameters are shown on the left in boxes with blue text, with arrows indicating to which part of the cost analysis methodology they apply. boxes with white text represent different processes of the methodologyeach process is described in the methods section under a subsection of the same name. boxes with green text appearing at the bottom and on the right represent results generated by the analysis. example, an infectious child moves from household to school on a given day, and infects two further children; they return to households and and, following virus incubation, become infectious and may infect other household members in their households. note that these households may be geographically separate, but are connected via contact of children at school. each household contains uniquely identified individuals. children and adults were assigned by an allocation algorithm to school classes and workplaces, respectively. the assignment of children to classes was based on age, school class size data, and proximity between schools and households; the assignment of adults to workplaces was based on workplace size and commuter survey data. in addition to contact occurring in households and mixing hubs, community contact was introduced to capture mixing which occurs outwith these locales and in the wider community. the number of contacts made by each individual each day in school, work and community settings were adjusted to reproduce the proportion of cases occurring in different settings as reported by empirical studies, specifically % of infections occurred in households, % in schools and workplaces, and % in the wider community [ ] [ ] [ ] . contacts within schools and workplaces occurred in fixed-size mixing groups of maximum size . within mixing groups contact was assumed to be homogeneous. community contacts occurred between randomly selected individuals, weighted toward pairs of individuals located in neighbouring households. a simulation algorithm, realised in the c++ programming language, manipulates the underlying demographic model and captures both population mobility and the time-changing infectivity profile of each individual. each individual has their infectivity status denoted by one of the four (susceptible, exposed, infectious, recovered) states at any time point during the duration of the simulated period. the simulation algorithm captures the state of the whole population twice per day, a daytime pointin-time snapshot and an evening snapshot, with individuals (possibly) moving locations between successive day or night periods, such as household to school or workplace for the day phase, returning to home for the night period. individuals come into contact with other individuals on a one-to-one basis in each location, with possible influenza transmission then occurring. individuals in each household and contact hub make contacts within a close-contact mixing group, taken to be the entire household or a subset of larger hubs, and also make additional non hub-based random community contacts. the attributes of the various locations in which individuals come into potentially infectious contact are summarized in table . using the contact, mobility and transmission features described above, stochastic simulations of influenza spread were conducted. all simulations were repeated times with random numbers controlling the outcome of stochastic events (the locality of seeded infected individuals and the probability of transmission) and the results were averaged. analysis of this simulation model has shown that the -run mean attack rate is highly unlikely ( % confidence) to differ by more than . % from the mean attack rate of a much larger set of experiment repeats. one new infection per day was introduced into the population during the whole period of the simulations, and randomly allocated to a household. this seeding assumption of case per day was chosen to reliably begin a local epidemic in every stochastic simulation. for the transmission characteristics described above, analysis shows that seeding at this rate for days results in a sustained epidemic in > % of the simulation runs and % with two weeks of seeding, with higher percentages for the higher transmissibility scenarios. seeding at this rate is continued throughout the simulation in order to capture the case where an epidemic may be initially suppressed by a rigorous intervention strategy, but may subsequently break out if intervention measures are relaxed. after the beginning of a sustained local epidemic, any subsequent variation in the amount of seeding has very little effect on the progress of the local epidemic, as the number of imported cases is much smaller than those generated by the local epidemic. preliminary analyses using the present model have shown that even if the seeding rate is increased to infections per day, after days the number of infections generated from the selfsustained local epidemic is twice the number of imported infections, and by days local infections outnumber imported infections by a factor of . the simulation period was divided into hour day/ night periods and during each period a nominal location for each individual was determined. this took into consideration the cycle type (day/night, weekday/weekend), infection state of each individual and whether child supervision was needed to look after a child at home. individuals occupying the same location during the same time period were assumed to come into potential infective contact. details of the simulation procedure are presented in [ ] . in the simulation model, we assumed that infectious transmission could occur when an infectious and susceptible individual came into contact during a simulation cycle. following each contact a new infection state for the susceptible individual (either to remain susceptible or to become infected) was randomly chosen via a bernoulli trail [ ] . once infected, an individual progressed through a series of infection states according to a fixed timeline. the probability that a susceptible individual would be infected by an infectious individual was calculated according to the following transmission function, which takes into account the disease infectivity of the infectious individual i i and the susceptibility of susceptible individual i s at the time of contact. maximum group size is . tertiary and vocational education institutions, number and size determined from state education department data. weekdays during day cycle. young adult and adult individuals who are allocated into the hub if they are active*. maximum group size is . workplace number and size of determined for local government business survey data. weekdays during day cycle. adult individuals who are allocated into the hub if they are active * . maximum group size is . community represents all contact between individuals in the community that is not repeated on a daily basis. everyday during day cycle. all individuals make contacts if they are active*, contact is random but weighted towards pairs with nearby household locations. * all individuals are active during day cycles unless: he/she is symptomatically infected and chooses to withdraw to household ( % chance for adults, % for children); or if his/her school or workplace is affected by social distancing interventions; or if he/she is a parent of a child who is inactive (only one parent per family is affected this way). the baseline transmission coefficient β was initially chosen to give an epidemic with a final attack rate of . %, which is consistent with seasonal influenza as estimated in [ ] (in table three of that paper). to achieve simulations under a range of basic reproduction numbers (r ), β was increased from this baseline value to achieve epidemics of various r magnitudes; details of the procedure for estimating β and r are given in [ ] . a reproduction number of . was used as a baseline assumption, and the sensitivity of results to this assumption was gauged by repeating all simulations and analyses for alternative reproduction numbers of . and . . a pandemic with a reproduction number of . corresponds to some estimations of the basic reproduction number of the pandemic [ ] [ ] [ ] [ ] , while a reproduction number of . corresponds to an upper bound on estimates of what may have occurred in the pandemic, with most estimates being in the range . - . [ , ] . the disease infectivity parameter inf(i i ) was set to for symptomatic individuals at the peak period of infection and then to . for the rest of the infectivity period. the infectiousness of asymptomatic individuals was also assumed to be . and this applies to all infected individuals after the latent period but before onset of symptoms. the infection profile of a symptomatic individual was assumed to last for days as follows: a . day latent period (with inf(i i ) set to ) was followed by day asymptomatic and infectious, where inf(i i ) is set to . ; then days at peak infectiousness (with inf(i i ) set to . ); followed by . days reduced infectiousness (with inf(i i )set to . ). for an infected but asymptomatic individual the whole infectious period (of . days) was at the reduced level of infectiousness with inf(i i ) set to . . this infectivity profile is a simplification of the infectivity distribution found in a study of viral shedding [ ] . as reported below in the results section for the unmitigated no intervention scenario, these assumptions regarding the duration of latent and infectious periods lead to a mean generation time (serial interval) of . days which is consistent with that estimated for h n influenza [ , , ] . following infection an individual was assumed to be immune to re-infection for the duration of the simulation. we further assume that influenza symptoms developed one day into the infectious period [ ] , with % of infections being asymptomatic among children and % being asymptomatic among adults. these percentages were derived by summing the age-specific antibody titres determined in [ ] . symptomatic individuals withdrew into the home with the following probabilities; adults % and children %, which is in keeping with the work of [ , ] . the susceptibility parameter susc(i s ) is a function directly dependent on the age of the susceptible individual. it captures age-varying susceptibility to transmission due to either partial prior immunity or age-related differences in contact behaviour. to achieve a realistic age specific infection rate, the age-specific susceptibility parameters were calibrated against the serologic infection rates for seasonal h n in - in tecumseh, michigan [ ] . the resulting age-specific attack rates are consistent with typical seasonal influenz, with a higher attack rate in children and young adults (details of the calibration procedure may be found in [ ] ). the antiviral efficacy factor avf(i i ,i s ) = ( -ave i )*( -ave s ) represents the potential reduction in infectiousness of an infected individual (denoted by ave i ) induced by antiviral treatment, and the reduction in susceptibility of a susceptible individual (denoted by ave s ) induced by antiviral prophylaxis. when no antiviral intervention was administrated the values of both ave i and ave s were assumed to be , indicating no reduction in infectiousness or susceptibility. however, when antiviral treatment was being applied to the infectious individual the value of ave i was set at . , capturing a reduction in infectiousness by a factor of % [ ] . similarly, when the susceptible individual was undergoing antiviral prophylaxis the value of ave s was set to . indicating a reduction in susceptibility by a factor of % [ ] . this estimate is higher than most previous modelling studies [ , , ] , which assume an ave s of %. this common assumption appears to stem from an estimate made in [ ] based on - trial data. our higher value is based on a more comprehensive estimation process reported in [ ] , which also incorporated data from an additional study performed in - [ ] . it is also in line with estimates of %- % reported in [ ] . we examined a comprehensive range of intervention strategies including school closure, antiviral drugs for treatment and prophylaxis, workplace non-attendance (workforce reduction) and community contact reduction. these interventions were considered individually and in combination and social distancing interventions were considered for either continuous periods (that is, until the local epidemic effectively ceased) or periods of fixed duration ( weeks or weeks). antiviral drug interventions and social distancing interventions were initiated when specific threshold numbers of symptomatic individuals were diagnosed in the community, and this triggered health authorities to mandate the intervention response. this threshold was taken to be . % of the population. this threshold was chosen based on a previous study with this simulation model, which found that it represents a robust compromise between early, effective intervention and "premature" intervention, which can result in sub-optimal outcomes when limited duration interventions are used [ ] . it was assumed that % of all symptomatic individuals were diagnosed, and that this diagnosis occurred at the time symptoms appeared. for continuous school closure, all schools were closed simultaneously once the intervention trigger threshold was reached. for fixed duration (e.g. weeks or weeks) school closure, schools were closed individually as follows: for a primary school the whole school was closed if or more cases were detected in the school; in a high school only the class members of the affected class were isolated (sent home and isolated at home) if no more than cases were diagnosed in a single class; however if there were more than cases diagnosed in the entire high school the school was closed. note that these school closure policies were only activated after the community-wide diagnosed case threshold was reached; cases occurring in schools before this time did not result in school closure. this policy of triggering school closure based on epidemic progression avoids premature school closure which can reduce the effectiveness of limited duration school closure [ , , ] ; see [ ] for a detailed description of school closure initiation triggering strategies. two primary antiviral drug strategies have been examined; antiviral drugs used solely for treatment of symptomatic cases (strategy t), and treatment plus prophylaxis of all household members of a symptomatic case (strategy av). a further strategy was also examined, in which prophylaxis was also extended to the contact group (school or workplace contacts) of a symptomatic case (strategy t + h + e). due to the logistical resources required, it is unlikely that this extended strategy could be implemented throughout a pandemic, and we do not report the results of this strategy in the main paper; full results are however given in (additional file ). antiviral treatment (and prophylaxis for household or work / school group contacts) was assumed to begin hours after the individual became symptomatic. it was assumed that an individual would receive at most one prophylactic course of antiviral drugs. further details of antiviral interventions are given in [ , ] . workforce reduction (wr) was modelled by assuming that for each day the intervention was in effect, each worker had a % probability of staying at home and thus did not make contact with co-workers. community contact reduction (ccr) was modelled by assuming that on days when the intervention was in effect, all individuals made % fewer random community contacts. the most rigorous social distancing interventions considered in this study, which we denote as strict social distancing, involve the combined activation of school closure with workforce reduction and/or community contact reduction, and for this to occur for significant time periods; continuous and weeks duration were considered. in the present study we simulated a total of intervention scenarios (for each of three reproduction numbers . , . and . ). to simplify the results, we only present those interventions that reduce the unmitigated illness attack rate by at least %. we defined five severity categories based on those proposed by the cdc [ ] . the cdc pandemic index was designed to better forecast the health impact of a pandemic, based on categories having cfrs ranging from < . % to > = . %, and allow intervention recommendations to match pandemic severity. the discrete cfrs used are listed in table . we extend the cdc categories to further include rates of hospitalisation and icu treatment, as described below using data collected during the pandemic in western australia, by the state department of health. these data permit case hospitalisation (icu and non-icu) and case fatality ratios (cfr) to be related, as described below. the least severe pandemic considered (category ) has cfr of . % which is at the upper end of estimates for the pandemic. initially, the pandemic cfr was estimated to be in the range . % - . % [ ] ; however recent reanalysis of global data from suggest a cfr (for the - age group) in the range . % - . % [ ] . cost analysis results for a pandemic with h n characteristics using a similar simulation model to the one described here can be found in [ ] . calculation of costs arising from lost productivity due to death and from hospitalisation of ill individuals requires that individual health outcomes (symptomatic illness, hospitalisation, icu admission, and death) be estimated for each severity level. the pandemic data from western australia was used to provide this relationship between the mortality rate and numbers requiring hospitalisation and icu care. these data indicated a non-icu hospitalisation to fatality ratio of : and an icu admission to fatality ratio of : . these values align with those in a previous study by presanis et al. in [ ] , which estimated the ratios in the ranges - to and . - . to , respectively. the economic analysis model translates the age-specific infection profile of each individual in the modelled symptomatic infectiousness timeline . day latent (non infectious), day asymptomatic; days peak symptomatic; . days post-peak symptomatic [ ] asymptomatic infectiousness timeline . day latent; . days asymptomatic [ ] asymptomatic infectiousness . [ ] peak symptomatic infectiousness . post-peak symptomatic infectiousness . [ ] probability of asymptomatic infection . [ ] probability average school closure cost (per student per day) $ . [ ] average gp visit cost $ . [ ] average hospitalization cost (per day) $ [ ] average icu cost (per day) $ [ , ] population, as derived by the albany simulation model, into the overall pandemic cost burden. total costs involve both direct healthcare costs (e.g. the cost of medical attention due to a gp visit, or for hospitalisation) and costs due to productivity loss [ , ] . pharmaceutical costs (i.e. costs related to antiviral drugs) are also estimated. all costs are reported in us dollars using consumer price index adjustments [ ] . us dollar values are used to make the results readily convertible to a wide range of countries. age-specific hospitalisation costs are achieved by multiplying the average cost per day by average length of stay for each age group [ , ] . hospitalisation costs, including icu costs, those involving medical practitioner visits, and antiviral drug (and their administration) costs are taken from the literature and are presented in table [ , , ] . the antiviral costs include the costs of maintaining an antiviral stockpile. this was calculated by multiplying the antiviral cost per course (but not the dispensing cost per course, which was included separately) by the expected number of times each antiviral course would expire and be replaced between pandemics, assuming a mean inter-pandemic period of . years (based on the occurrence of pandemics in , , and ) and an antiviral shelf life of years [ ] . treatment costs, lengths of stay in hospital (both icu and non-icu), and other cost data used in establishing the overall cost of mitigated and unmitigated epidemics in the modelled community are given in table . productivity losses due to illness and interventions (e.g. necessary child-care due to school closure and workforce reduction) were calculated according to the human capital approach, using average wages and average work-days lost; the latter being determined from day-to-day outbreak data generated by the simulation model. assumed average wages are given in table . school closure is assumed to give rise to two costs. the first, following the work of perlroth et al. [ ] , is a $ per student school day lost. this is intended to approximate the cost of additional education expense incurred in the futurewhich might occur for example in the form of additional holiday classes. the second component is lost productivity of parents staying at home to supervise children. the simulation model calculates whether this occurs for every day for every household, based on what interventions are in force (school closure and/or workforce reductions), whether children or adults are ill, the number of adults in the household, whether it is a school day, etc., and accumulates the cost accordingly. indirect production losses due to death were also derived using a human capital approach, based on the net present value of future earnings for an average age person in each age group. this was calculated by multiplying the age-specific number of deaths due to illness by the average expectancy in years of future earnings of an individual by an average annual income [ ] . we assumed a maximum earning period up to age . productivity losses due to death were discounted at % annually, which is a standard discounting rate used to express future income in present value [ ] . to provide an alternative analysis, total costs were also calculated without this long-term productivity loss due to death component. overview figure presents the final attack rate (ar) and the total cost of the epidemic for each intervention strategy applied, for a pandemic with a basic reproduction number of r = . . although costs are calculated from the whole-of-society perspective, total costs are presented as a cost per person in the community, calculated by dividing the simulated cost of the pandemic by the population of~ , , in order to make the results more easily transferable to communities of various sizes. strategies are ordered from left to right by increasing effectiveness (i.e. their ability to decrease the attack rate), and only intervention strategies that reduce the attack rate by at least % are included. figure shows three distinctive features. firstly, for an epidemic with basic reproduction number r = . , no single intervention is effective in reducing the attack rate by more than %, and thus do not appear in figure . this finding is consistent with previous modelling studies which found that layering of multiple interventions is necessary to achieve substantial attack rate reductions [ ] [ ] [ ] [ ] [ ] [ ] , , ] . secondly, higher severity pandemics have higher total costs. total costs of unmitigated pandemics range from $ to $ per person for pandemics from category to category (see table ). thirdly, for high severity pandemics total costs are lower for the more effective intervention strategies. figure presents the constituent components that make up the total cost of each intervention and severity category, measured in terms of cost per person in the modelled community. three distinctive features can be seen in figure . firstly, for high severity pandemics costs are dominated by productivity losses due to death and health care costs. secondly, for low severity pandemics costs are dominated by social distancing and illness costs. thirdly, for all severity categories antiviral costs are comparatively low when compared with all other cost components of antiviral based intervention strategies. antiviral costs never constitute more than % of the total cost, and for all severity categories greater than (cfr > . %) antiviral costs are always the smallest cost component. below we report on effectiveness, total costs and cost components of interventions for pandemics with high and low severity. these cost data are presented in table . figure summarises the characteristics of key intervention strategies. for high severity pandemics (categories and , with case fatality rates above . %) the least costly strategy combines continuous school closure, community contact reduction, antiviral treatment and antiviral prophylaxis. at category this strategy has a total cost of $ , per person, a net benefit of $ per person compared to no intervention. this strategy is also the most effective intervention strategy, reducing the attack rate from % to . %. the results indicate that strategies with the lowest total costs are also the most effective. for a category pandemic the most effective strategies, all of which reduce the attack rate to less than %, have total costs ranging from $ , to $ , per person, which is less than one-third the cost of the unmitigated pandemic ($ , ), showing the substantial net benefit of effective interventions for high severity pandemics. these strategies all feature continuous school closure, with either continuous community contact reduction or antiviral treatment and prophylaxis. the ability of highly effective interventions to reduce the total cost of a high severity pandemic is due to the largest component of the overall cost being productivity losses arising from deaths. this is illustrated in figure which shows the cost components for each intervention. it can be seen that the majority of the cost for an unmitigated pandemic of severity category and is due to death-related productivity losses (shown in purple). although highly effective interventions incur large intervention-related productivity losses (shown in green), for high severity pandemics these intervention costs are more than outweighed by the reduction in medical costs and death-related productivity losses. the most costly intervention considered (i.e. which still reduced the attack rate by at least %) is continuous school closure combined with continuous workforce reduction, which costs $ , per person. for low severity pandemics (in category , having cfr < = . %) the intervention strategy with the lowest total cost considered is weeks school closure combined with antiviral treatment and prophylaxis, costing $ per person which represents a net saving of $ per person compared to no intervention. however, this strategy is not as effective as other intervention strategies, reducing the attack rate to only %. the most effective intervention (combined continuous school closure, community contact reduction, and antiviral treatment and household prophylaxis), which reduces the attack rate to . %, costs $ per person, a net benefit of $ per person compared to no intervention. figure shows that for category and pandemics, although highly effective intervention measures reduce medical costs and death-related productivity losses, they incur larger costs due to intervention-related lost productivity. the most costly intervention considered is continuous school closure combined with continuous workforce reduction, which costs $ , per person, a net cost of $ per person compared to no intervention. this is due to the large cost associated with % workforce absenteeism. an important subset of intervention strategies are those consisting of purely social distancing interventions. in the case that antiviral drugs are unavailable or ineffective, only these non-pharmaceutical interventions strategies will be available. the most effective non-pharmaceutical strategy is the continuous application of the three social distancing interventions, school closure, workforce reductions, and community contact reduction, which reduces the attack rate to %. this intervention has a total cost ranging from $ , to $ , per person for severity categories ranging from to respectively. the least costly non-pharmaceutical strategy omits workforce reduction, resulting in a slightly higher attack rate of %. this intervention has a total cost ranging from $ to $ , per person for severity categories ranging from to respectively. the costing model used for this analysis includes future productivity losses from deaths caused by the pandemic. this long-term cost is often not included in cost-utility analyses. the inclusion of death-related productivity losses greatly increases the total costs of severe pandemics. however, even if these costs are not included, medical costs (due to hospitalisation and icu usage) play a similar, although less extreme, role. if long-term productivity losses due to death are not included in the costing model, the total cost of the pandemic is not surprisingly lower. however the effectiveness and relative total costs of intervention strategiesthat is, the ranking of intervention strategies by total cost -remains the same whether or not death-related productivity losses are included (spearman's rank correlation coefficient r = . , p = . for a null hypothesis that rankings are uncorrelated). full cost results of an alternate analysis that omits death-related productivity losses is contained in an additional file accompanying this paper (additional file ), and is summarised below. for category , when death-related productivity losses are not included the total cost of intervention strategies ranges from $ to $ , . this range is much smaller than if death-related productivity losses are included, in which case total cost ranges from $ , to $ , . for lower severity pandemics with lower case fatality ratios, the contribution of death-related productivity losses is naturally smaller. for category , when death-related productivity losses are not included total cost ranges from $ to $ , ; with death-related productivity losses the range is $ to $ , . if death-related productivity losses are not included, social distancing and illness costs dominate the total cost of each intervention strategy for low severity pandemics, while health care costs dominate the cost profile for high severity pandemics. sensitivity analyses were conducted to examine the extent to which these results depend upon uncertain model parameters that may impact on the cost or effectiveness of interventions. the methodology adopted was to identify assumptions and model parameters known to have an effect on intervention outcomes, taken from previous studies with this simulation model [ , , , , , ] , and to perform univariate analyses on each, examining parameter values both significantly higher and lower than figure breakdown of pandemic cost components. breakdown of pandemic costs shown as horizontal bar, for each intervention strategy and each severity category. coloured segments of each bar represent cost components as follows: (blue) health care; (red) antiviral drugs, including dispensing costs; (green) productivity losses due to illness and social distancing interventions; (purple) productivity losses due to deaths. note that horizontal scale is different for each severity category. values are for a pandemic with unmitigated transmissibility of r = . . interventions abbreviated as: scschool closure; ccr - % community contact reduction; wr - % workforce reduction; , intervention duration in weeks; contcontinuous duration; avantiviral treatment of diagnosed symptomatic cases and antiviral prophylaxis of household members of diagnosed symptomatic cases. the baseline values. alternative parameter settings were analysed for transmissibility (as characterised by the basic reproduction number r ), voluntary household isolation of symptomatic individuals, antiviral efficacy, compliance to home isolation during school closure, degree of workforce reduction, and degree of community contact reduction. a common finding across all sensitivity analyses was that alternative parameter settings that rendered interventions less effective resulted in strategies that not only had higher attack rates, but also had higher total pandemic costs, with this effect being most pronounced for pandemics of high severity. further details and results of the sensitivity analysis can be found in an additional file accompanying this paper (additional file ). the need for an unambiguous, extended definition of severity has been noted in the world health organization report on the handling of the pandemic [ ] , which highlights the impact pandemic severity has on health care provision and associated costs. in the absence of such definitions, an extended severity metric is presented. this extends the case fatality ratio (cfr) severity scale devised by the cdc [ ] , with hospitalisation and intensive care unit (icu) data collected in australia during the pandemic. these data have been used to generate a more extensive notion of pandemic severity, relating actual age-specific attack rates with agespecific hospitalisation and mortality rates, thereby contributing to the realism of both the simulation model and the economic analysis. this pandemic severity scale together with a pandemic spread simulation model allows the calculation of the total cost of a pandemic, and to estimate the relative magnitude of all the factors that contribute to the pandemic cost, including not only pharmaceutical and medical costs, but also productivity losses due to absenteeism and death. the severity of a future pandemic is shown to have a major impact on the overall cost to a nation. unsurprisingly, high severity pandemics are shown to be significantly more costly than those of low severity, using a costing methodology which includes costs arising from losses to the economy due to death, in addition to intervention and healthcare costs. a key finding of this study is that at high severity categories, total pandemic costs are dominated by hospitalization costs and productivity losses due to death, while at low severities costs are dominated by productivity losses due to social distancing interventions resulting from closed schools and workplaces. consequently, findings indicate that at high severity, the interventions that are the most effective also have the lowest total cost. highly effective interventions greatly reduce the attack rate and consequently the number of deaths, which in turn reduces productivity losses due to death. although highly effective interventions incur significant intervention-related productivity losses, for severe pandemics having high cfr, these intervention costs are more than compensated for by the reduction in death-related productivity losses, resulting in lower overall costs. conversely, for low severity pandemics, although highly effective intervention measures do reduce medical costs and death-related productivity losses, these savings can be smaller than costs incurred due to intervention-related lost productivity, resulting in total costs that are higher than the unmitigated baseline. antiviral strategies alone are shown to be ineffective in reducing the attack rate by at least %. however, the addition of antiviral case treatment and household prophylaxis to any social distancing strategy always resulted in lower attack rates and lower total costs when compared to purely social distancing interventions. the cost of all antiviral interventions constitutes a small fraction of total pandemic costs, and these costs are outweighed by both the healthcare costs prevented, and productivity gained, by their use in preventing illness and death. it should be noted that the lowest severity category considered, pandemic category , has a cfr of . % which is at the upper end of cfr estimates for the pandemic, which has been estimated to have a cfr of between . % and . % [ ] . thus, the cost results are not directly applicable to the pandemic. vaccination has been deliberately omitted from this study. the effectiveness and cost effectiveness of vaccination will depend crucially on the timing of the availability of the vaccine relative to the arrival of the pandemic in the communityvaccination cannot be plausibly modelled without considering this delay, and how it interacts with the timing of introduction and relaxation of other, rapidly activated interventions. the examination these timing issues for realistic pandemic scenarios that include both vaccination and social distancing / antiviral interventions is an important avenue for future work. as they stand, the results of this study, specifically the "continuous" duration social distancing strategies, can be considered to be models of interim interventions to be used prior to a vaccination campaign. the results are based on the community structure, demographics and healthcare system of a combined rural and urban australian community, and as such may not be applicable to developing world communities with different population or healthcare characteristics. although the cost and effectiveness results are directly applicable to pandemic interventions in a small community of , individuals, we expect that the per-capita costs and final attack rate percentages derived in this study can be extended to larger populations with similar demographics, provided a number of conditions are met. for the results to be generalisable, it needs to be assumed that communities making up the larger population implement the same intervention strategies, and instigate interventions upon the arrival of the pandemic in the local community (according to the criteria described in the methods section). the assumption is also made that there are no travel restrictions between communities. it should be noted that the single-community epidemic results do not predict the overall timing of the pandemic in the larger population. the simulation model used in this study has been used in previous studies to examine various aspects of social distancing and pharmaceutical (antiviral and vaccine) pandemic influenza interventions [ , , , , , ]. this simulation model shares characteristics with other individual-based pandemic influenza simulation models that have been employed at a variety of scales, including small communities [ , , , , , ] , cities [ , ] , countries [ , , , ] and whole continents [ ] . several related studies which also used individualbased simulation models of influenza spread coupled with costing models are those of those of sander et al., perlroth et al., brown et al., and andradottir et al. [ , , , ] . the current study extends upon the scope of these studies in several ways: five gradations of pandemic severity are considered, more combinations of interventions are considered, social distancing interventions of varying durations are considered, and probabilities of severe health outcomes for each severity category are based on fatality, hospitalization and icu usage data as observed from the pandemic. also in contrast with those models, we have chosen to include a cost component arising from productivity loss due to death, though a similar costing without death-related productivity losses has been included in (additional file ). for a pandemic with very low severity, with a cfr consistent with mild seasonal influenza, and that of the pandemic, previous results with the simulation and costing model used for this paper coincide with the studies mentioned above [ ] . specifically, they showed that antiviral treatment and prophylaxis were effective in reducing the attack rate and had a low or negative incremental cost, and that adding continual school closure further decreased attack rates, but significantly increased total cost. for high severity pandemics the inclusion of productivity loss following death, as presented in this study, leads to a markedly different assessment of total costs when compared to the two studies quoted above that considered severe pandemics [ , ] . for example, perlroth et al. found that the incremental cost of adding continuous school closure to an antiviral strategy was always positive, even for pandemics with high transmissibility (r = . ) and a cfr of up to %, meaning that adding school closure always increased total costs. similarly sander et al. found that the addition of continuous school closure to an extended antiviral strategy also increased total costs, including pandemics with a % cfr. in contrast, we found that adding continuous school closure to an extended prophylaxis strategy reduced total costs where the cfr was . % or greater (i.e. category and above), for a pandemic with r = . . the study of smith et al. estimated the economic impact of pandemic influenza on gross domestic product for a range of transmissibility and severity values [ ] . consistent with our study was the finding that at low severity the largest economic impacts of a pandemic would be due to school closure (effective but costly) and workplace absenteeism (largely ineffective and costly). like the other two studies mentioned above, the study of smith et al. did not include future productivity losses due to death. as a result, in contrast to our findings, they did not find that, for severe pandemics, the high short-term costs of rigorous social distancing interventions were outweighed by future productivity of people whose lives were saved by the intervention. in this study we considered the case of a pandemic that infects a significant proportion of the population, and thus incurs significant direct costs stemming from medical costs and productivity losses. however, in the case of a pandemic perceived by the public to be severe, there are likely to be additional indirect macroeconomic impacts caused by disruption of trade and tourism, consumer demand and supply, and investor confidence [ , ] . in the case of a pandemic of high severity (i.e. high case fatality ratio) but low transmissibility, these indirect effects and their resulting societal costs may constitute the main economic impact of the pandemic, an effect seen with the sars outbreak in [ ] . the results of this study are relevant to public health authorities, both in the revision of pandemic preparedness plans, and for decision-making during an emerging influenza pandemic. recent modelling research has shown that combinations of social distancing and pharmaceutical interventions may be highly effective in reducing the attack rate of a future pandemic [ , , , , , , , , ] . public health authorities are aware that rigorous social distancing measures, which were used successfully in some cities during the pandemic [ , ] , when pharmaceutical measures were unavailable, would be highly unpopular due to resulting societal disruption, and costly due to associated productivity losses [ ] . the results of this study give guidance as to the pandemic characteristics which warrant the use of such interventions. the results highlight the importance of understanding the severity of an emergent pandemic as soon as possible, as this gives guidance as to which intervention strategy to adopt. in the likely situation where the severity of an emerging pandemic is initially unknown (but is suspected to be greater than that of seasonal influenza), the results indicate that the most appropriate intervention strategy is to instigate school closure and community contact reduction, combined with antiviral drug treatment and household prophylaxis, as soon as transmission has been confirmed in the community. if severity is determined to be low, public health authorities may consider relaxing social distancing measures. in the case of a category pandemic (cfr approximately . %), little is lost by the early imposition and subsequent relaxation of social distancing interventions: results indicate that even if schools are closed for weeks while severity is being determined, the total cost of the pandemic is lower than if no interventions had been enacted. if severity is determined to be high, extending the duration of social distancing interventions results in both net savings to society and reduction in mortality. anzic influenza investigators: critical care services and h n influenza in australia and new zealand europe's initial experience with pandemic (h n ) -mitigation and delaying policies and practices mortality from pandemic a/h n influenza in england: public health surveillance study analysis of the effectiveness of interventions used during the h n influenza pandemic developing guidelines for school closure interventions to be used during a future influenza pandemic mitigation strategies for pandemic influenza in the united states targeted social distancing design for pandemic influenza reducing the impact of the next influenza pandemic using household-based public health interventions effective, robust design of community mitigation for pandemic influenza: a systematic examination of proposed us guidance a small community model for the transmission of infectious diseases: comparison of school closure as an intervention in individual-based models of an influenza pandemic simulation suggests that rapid activation of social distancing can arrest epidemic development due to a novel strain of influenza school closure is currently the main strategy to mitigate influenza a(h n )v: a modelling study modelling mitigation strategies for pandemic (h n ) nature outlook: influenza airborne transmission of influenza a/h n virus between ferrets experimental adaptation of an influenza h ha confers respiratory droplet transmission to a reassortant h ha/h n virus in ferrets the potential for respiratory droplet-transmissible a/h n influenza virus to evolve in a mammalian host statistics of influenza morbidity with special reference to certain factors in case incidence and case fatality emerging infections: pandemic influenza australian census a bayesian mcmc approach to study transmission of influenza: application to household longitudinal data estimating the impact of school closure on influenza transmission from sentinel data strategies for mitigating an influenza pandemic papoulis a: probability, random variables and stochastic processes tecumseh study of illness xiii. influenza infection and disease, - pandemic potential of a strain of influenza a (h n ): early findings pandemic (h n ) influenza community transmission was established in one australian state when the virus was first identified in north america estimating the reproduction number of the novel influenza a virus (h n ) in a southern hemisphere setting: preliminary estimate in new zealand epidemiological and transmissibility analysis of influenza a (h n ) v in a southern hemisphere setting: peru time lines of infection and disease in human influenza: a review of volunteer challenge studies early transmission characteristics of influenza a (h n ) v in australia: victorian state serial intervals and the temporal distribution of secondary infections within households of pandemic influenza a (h n ): implications for influenza control recommendations influenzavirus infections in seattle families, - containing pandemic influenza at the source design and evaluation of prophylactic interventions using infectious disease incidence data from close contact groups temporal factors in school closure policy for mitigating the spread of influenza the impact of case diagnosis coverage and diagnosis delays on the effectiveness of antiviral strategies in mitigating pandemic influenza a/h n containing pandemic influenza with antiviral agents management of influenza in households: a prospective, randomized comparison of oseltamivir treatment with or without postexposure prophylaxis neuraminidase inhibitors for influenza simulating school closure strategies to mitigate an influenza epidemic interim pre-pandemic planning guidance: community strategy for pandemic influenza mitigation in the united states cost-effectiveness of pharmaceutical-based pandemic influenza mitigation strategies health outcomes and costs of community mitigation strategies for an influenza pandemic in the united states economic evaluation of influenza pandemic mitigation strategies in the united states using a stochastic microsimulation transmission model direct medical cost of influenza-related hospitalizations in children responding to pandemic (h n ) influenza: the role of oseltamivir world health organization: making choices in health: who guide to cost-effectiveness analysis. geneva: world health organization the severity of pandemic h n influenza in the united states estimated global mortality associated with the first months of pandemic influenza a h n virus circulation: a modelling study cost-effective strategies for mitigating a future influenza pandemic with h n characteristics vaccination against pandemic influenza a/h n v in england: a real-time economic evaluation usa historical consumer price index economics of neuraminidase inhibitor stockpiling for pandemic influenza economic analysis of pandemic influenza vaccination strategies in singapore reactive strategies for containing developing outbreaks of pandemic influenza measures against transmission of pandemic h n influenza in japan in : simulation model who: report of the review committee on the functioning of the international health regulations ( ) in relation to pancemic (h n ) . geneva: world health organization an influenza simulation model for immunization studies individual--based computational modeling of smallpox epidemic control strategies modeling targeted layered containment of an influenza pandemic in the united states mitigation measures for pandemic influenza in italy: an individual based model considering different scenarios the role of population heterogeneity and human mobility in the spread of pandemic influenza would school closure for the h n influenza epidemic have been worth the cost?: a computational simulation of pennsylvania the economy-wide impact of pandemic influenza on the uk: a computable general equilibrium modelling experiment globalization and disease: the case of sars. asian economic papers global macroeconomic consequences of pandemic influenza. sydney australia: lowy institute for international policy the effect of public health measures on the influenza pandemic in u.s. cities quantifying social distancing arising from pandemic influenza economic analysis of pandemic influenza mitigation strategies for five pandemic severity categories additional file : additional results and sensitivity analyses. "milne pandemiccostadditionalfile .doc". competing interests gjm has received a travel grant from glaxosmithkline to attend an expert meeting in boston, usa; mjp has received travel grants from glaxosmithkline and wyeth to attend expert meetings in reykjavik, iceland, boston, usa and istanbul, turkey. jkk and nh have no potential competing interests. key: cord- -j navhku authors: zhang, sicui; genga, laura; dekker, lukas; nie, hongchao; lu, xudong; duan, huilong; kaymak, uzay title: towards multi-perspective conformance checking with aggregation operations date: - - journal: information processing and management of uncertainty in knowledge-based systems doi: . / - - - - _ sha: doc_id: cord_uid: j navhku conformance checking techniques are widely adopted to validate process executions against a set of constraints describing the expected behavior. however, most approaches adopt a crisp evaluation of deviations, with the result that small violations are considered at the same level of significant ones. furthermore, in the presence of multiple data constraints the overall deviation severity is assessed by summing up each single deviation. this approach easily leads to misleading diagnostics; furthermore, it does not take into account user’s needs, that are likely to differ depending on the context of the analysis. we propose a novel methodology based on the use of aggregation functions, to assess the level of deviation severity for a set of constraints, and to customize the tolerance to deviations of multiple constraints. nowadays organizations often define procedures describing how their processes should be performed to satisfy a set of constraints, e.g., to minimize the throughput time or to comply with rules and regulations. a widely used formalism to represent these procedures consists in so-called process models, that are graphic or logic formalism representing constraints defined on organization processes, e.g., by the order of execution of the activities. however, it is well documented in literature that real process behavior often deviates from the expected process, which often leads to performance issues or opens the way to costly frauds [ ] . in recent years, the increasing use by organizations of information systems (e.g., erp, sap, mrp and so on) to support and track the execution of their processes enabled the development of automatic, data-driven techniques to assess the compliance level of the real process behavior. among them, conformance checking techniques have been gaining increasing attention both from practitioners and academics [ , , , , ] . given an event log, i.e., a log file tracking data related to activities performed during process executions, conformance checking techniques are able to pinpoint discrepancies (aka, deviations) between the log and the corresponding model. while classic conformance checking techniques only deal with the control-flow of the process, i.e., the activities execution order, in recent years also some multi-perspective conformance checking, aimed to deal also with data constraints, have become more and more relevant [ , ] . nevertheless, there are still several open challenges to implement multiperspective conformance checking. among them, here we focus on the lack of appropriate modeling mechanisms for dealing with the uncertainty and graduality often characterizing human-decisions in real-world processes. state of the art techniques implement a crisp approach: every execution of an activity is considered as either completely wrong or completely correct. [ , , ] . while this assumption is well grounded to deal with the control-flow (indeed, each activity is either executed at the right moment, or it is not), when addressing data constraints it can easily lead to misleading results. a well-known example of this issue can be found in the healthcare domain. let us assume that a surgery department implements a guideline stating that the systolic blood pressure (sbp) of a patient has to be lower than to proceed with a surgery. it is reasonable to expect that sometimes clinicians will not refuse to operate patients whose sbp is , since this is quite a small deviation and delaying the surgery could be more dangerous for the patient. clearly, surgeries performed with this value of sbp are likely to be much less problematic than surgeries performed with a sbp equal to, e.g., . however, conformance checking techniques would simply mark both these cases as 'not compliant ', without allowing for any distinction. this behavior is undesirable, since it is likely to return in output a plethora of not-interesting deviations, at the same time hiding those which could deserve further investigation. we investigated this issue in our previous work [ ] , where we proposed to use fuzzy sets, which are used to present the flexibility in the constraints and the goals in fuzzy optimization [ ] , to determine the severity of violations of a single soft constraint per activity. however, the previous work used basic strategy of standard conformance checking techniques for dealing with multiple constraints deviations; namely, the total degree of data deviations of that activity is computed by summing up the costs for all the violated constraints. this strategy poses some important limitations when investigating the data compliance. first, it introduces an asymmetry in the assessment of control-flow and data deviations. while controlflow deviations for each activity express the level of compliance of the activity to control-flow constraints (either fully compliant or wrong), in the presence of multiple data constraints the obtained value does not give an indication of the overall level of compliance to the constraints set. furthermore, no customization to the user's needs is provided. first, in this setting data violations tend to be considered more severe than control-flow ones, even if this might not fit with user's intention. furthermore, different contexts might require tailored functions to assess multiple data deviations severity. in this paper, we address this issue by proposing a novel fuzzy conformance checking methodology based on the use of aggregation functions, which have been proved feasible for modeling simultaneous satisfaction of aggregated criteria [ ] . with respect to previous work, the approach brings two main contributions: a) it applies fuzzy aggregation operators to assess the level of deviation severity for a set of constraints, and b) it allows to customize the tolerance to deviations of multiple constraints. as a proof-of-concept, we tested the approach over synthetic data. the remainder of this paper is organized as follows. section introduces a running example to discuss the motivation of this work. section introduces basic formal notions. section illustrates our approach, and sect. presents results obtained by a set of synthetic experiments. section discusses related work. finally, sect. draws some conclusions and presents future work. consider, as a running example, a loan management process derived from previous work on the event log of a financial institute made available for the bpi challenge [ , ] . figure shows the process in bpmn notation. the process starts with the submission of a loan application. then, the application passes through a first assessment of the applicant's requirements and, if the requested amount is greater than euros, also through a more thorough fraud detection analysis. if the application is not eligible, the process ends. otherwise, the application is accepted, an offer to be sent to the customer is selected and the details of the application are finalized. after the offer has been created and sent to the customer, the latter is contacted to discuss the offer with her. at the end of the negotiation, the agreed application is registered on the system. at this point, further checks can be performed on the application, if the overall duration is still below days and the amount is larger than , before approving it. let us assume that this process is supported by some system able to track the execution of its activities in a so-called event log. in practice, this is a collection of traces, i.e., sequences of activities performed within the same process execution, each storing information like the execution timestamp of the execution, or other data element [ ] . as an example, let us consider the following traces showing two executions of the process in fig. (note that we use acronyms rather than complete activity names) : σ = (a s, {amount = , . both executions violate the constraints defined on the duration and the amount of the loan, according to which the activity w f a should have been anyway skipped. conformance checking techniques also attempt to support the user in investigating the interpretations of a deviation. in our case, the occurrence of the activity w f a could be considered either as a ) control-flow deviation (i.e., data are corrected but the activity should not have been executed) or as a ) data-flow deviation (i.e., the execution of the activity is correct but data have not been properly recorded on the system). in absence of domain knowledge in determining what is the real explanation, conformance checking techniques assess the severity (aka, cost) of the possible interpretations and select the least severe one, assuming that this is the one closest to the reality. in our example, conformance checking would consider σ as a control-flow deviation, since the cost would be equal to , while data-flow deviation would correspond to , having two violated constraints; for σ , instead, the two interpretations would be equivalent, since only one data constraint is violated. in previous work [ ] we investigated how to use fuzzy membership function to assess severity of data deviations taking into account the magnitude of the deviations. however, the approach still comes with some limitations when considering multiple constraints. indeed, with this approach the overall severity of the data deviation for an activity is assessed by a simple sum operation. for example, let us suppose that with the method in [ ] we obtained a cost of . , . for the violations of amount and duration in w f a in σ , thus obtaining a total cost of . , and . and in σ , thus obtaining, a total cost of . . in this setting, activities involving multiple constraints will tend to have an interpretation biased towards control-flow deviations, since the higher the number of constraints, the higher the the data-deviation cost. furthermore, it is worth noting that the comparison between the two traces can be misleading; in one case, constraints are violated, even if one only slightly deviated; while in the second case only one constraint is violated, even if with quite a strong deviation. however, the final numerical results are quite similar, thus hiding the differences. this example shows how the use of the simple sum function can impact the results significantly, without the user realizing it and, above all, without providing the user with any customization mechanism. for example, the user might want to assess the data-compliance level in terms of the percentage of satisfied constraints, or by considering only the maximum cost, and so on. however, current techniques do not allow for this kind of customization. this section introduces a set of concepts that will be used through the paper. conformance checking techniques detect discrepancies between a process model and the real process execution. here we define the notion of process model using the notation from [ ] , enriched with data-related notions explained in [ ] . m = (p, p i , p f , a m , v, w, u, is a guard function, i.e., a boolean formula expressing a condition on the values of the data variables. w : a m → v is a write function, that associates an activity with the set of variables which are written by the activity. finally, v alues : is a function that associates each state with the corresponding pairs variable=value. the firing of an activity s = (a, w) ∈ a m × (v → u ) in a state p is valid if: ) a is enabled in p ; ) a writes all and only the variables in w (a); ) g(a) is true when evaluate over v alues(p ). to access the components of s we introduce the following notation: vars(s) = w, act(s) = a. function vars is also overloaded such that vars the set of valid process traces of a model m is denoted with ρ(m ) and consists of all the valid firing sequences σ ∈ (a m × (v → u )) * that, from an initial state p i lead to a final state p f . figure provides an example of a process model in bpmn notation. process executions are often recorded by means of an information system in event logs. formally, let s n be the set of (valid and invalid) firing of activities of a process model m ; an event log is a multiset of traces l ∈ b(s * n ). given an event log l, conformance checking builds an alignment between l and m , mapping "moves" occurring in the event log to possible "moves" in the model. a "no move" symbol " " is used to represent moves which cannot be mimicked. for convenience, we introduce the set s n = s n ∪ { }. formally, we set s l to be a transition of the events in the log, s m to be a transition of the activities in the model. a move is represented by a pair (s l , s m ) ∈ s n × s n such that: lm such that the projection of the first element (ignoring ) yields σ l , and the projection on the second element (ignoring ) yields σ m . let us consider the model in fig. and the trace σ in sect. . table shows two possible alignments γ and γ for activity w f a. for alignment γ , the pair (w f a, w f a) is a move in both with incorrect data, while in γ the move (w f a, ⊥) is matched with a , i.e., it is a move on log. (in remaining part, amount and duration are abbreviated to a and d). as shown in example , there can be multiple possible alignments for a given log trace and process model. our goal is to find the optimal alignment, i.e., the alignment with minimum cost. to this end, the severity of deviations is assessed by means of a cost function. aggregation operations (aos) are mathematical functions that satisfy minimal boundary and monotonicity conditions, and are often used for modeling decision making processes, since they allow to specify how to combine the different criteria that are relevant when making a decision [ , ] . in literature, many aos have been defined (see [ , , ] for an overview), with different level of complexity and different interpretations. a commonly used class of aggregation operators are the t-norms, which are used to model conjunction of fuzzy sets. in compliance analysis, one often tries to satisfy all constraints on the data, and so t-norms are suitable operators for modeling soft constraints in compliance analysis. widely used t-norms are the minimum, product and the yager operators [ ] . in addition to the t-norms, other aggregation operators could also be used, depending on the goals of the compliance analysis. we do not consider other types of aggregation operators in this paper, but, in general, one could use the full flexibility of different classes of fuzzy set aggregation operators that have been used in decision making (see, e.g. [ ] ). we introduce a compliance checking approach tailored to dealing with decision tasks under multiple guards, to enhance the flexibility of the compliance assessing procedure. to this end, we investigate the use of aos. compliance checking in process analysis is based on the concept of alignment between a process model and a process trace that minimizes a cost of misalignment. the computation of an optimal alignment relies on the definition of a proper cost function for the possible kind of moves (see sect. ). most of stateof-the art approaches adopt (variants of) the standard distance function defined in [ ] , which sets a cost of for every move on log/model (excluding invisible transitions), and a cost of for synchronous moves. multi-perspective approaches extend the standard cost function to include data costs. elaborating upon these approaches, in previous work [ ] we defined our fuzzy cost function as follows. this cost function assigns a cost equal to for a move in log; plus the number of variables that should have been written by the activity for a move in model; finally, the sum of the cost of the deviations ( -μ i ) for the data variables if it's a move in both. note that the latter consider both the case of move with incorrect and incorrect data. as discussed in sect. , summing up all the data cost presents important limitations to assess the conformance of multiple constraints. therefore, in the present work, we propose a new version of our fuzzy cost function with the goal of standardize every move within the range ( , ) and allow the user to customize the cost function to her needs. let π(μ , μ , ..., μ n ) be an userdefined aggregated membership function of multiple variables. then ( − π) is the overall deviation cost of a set of variables. the cost k(s l , s m ) is defined as: ( ) the problem of finding an optimal alignment is usually formulated as a search problem in a directed graph [ ] . let z = (z v , z e ) be a directed graph with edges weighted according to some cost structure. the a* algorithm finds the path with the lowest cost from a given source node v ∈ z v to a node of a given goals set z g ⊆ z v . the cost from each node is determined by an evaluation function f (v) = g(v) + h(v), where: -g : z v → r + gives the smallest path cost from v to v; -h : z v → r + gives an estimate of the smallest path cost from v to any of the target nodes. if h is admissible,i.e. it underestimates the real distance of a path to any target node v g , then a* finds a path that is guaranteed to have the overall lowest cost. the algorithm works iteratively: at each step, the node v with lowest cost is taken from a priority queue. if v belongs to the target set, the algorithm ends returning node v. otherwise, v is expanded: every successor v is added to the priority queue with a cost f (v ). given a log trace and a process model, to employ a* to determine an optimal alignment we associate every node of the search space with a prefix of some complete alignments. the source node is an empty alignment γ = , while the set of target nodes includes every complete alignment of σ l and m . for every pair of nodes (γ , γ ), γ is obtained by adding one move to γ . the cost associated with a path leading to a graph node γ is then defined as g(γ) = k(γ) + |γ|, where k(γ) = sl,sm ∈γ k(s l , s m ), with k(s l , s m ) defined as in ( ), |γ| is the number of moves in the alignment, and is a negligible cost, added to guarantee termination. note that the cost g has to be strictly increasing. while we do not give a formal proof for the sake of space, it is straight to see that g is obtained in our approach by the sum of all non negative elements. therefore, while moving from an alignment prefix to a longer one, the cost can never decrease. for the definition of the heuristic cost function h(v) different strategies can be adopted. informally, the idea is computing, from a given alignment, the minimum number of moves (i.e., the minimum cost) that would lead to a complete alignment. different strategies have been defined in literature, e.g., the one in [ ] , which exploits petri-net marking equations, or the one in [ ] , which generates possible states space of a bpmn model. example . let us analyze possible moves to assign to the activity w f a in σ . let us assume that the memberships of the variables are μ a = . and μ d = . . according to ( ) and p roduct t-norm we get the fuzzy cost function k(s l , s m ). ( ) figure shows the portion of the space states for the alignment building of σ . at node # , f = , since no deviations occurred so far. from here, there are two possible moves that could be selected, one representing a move on log (on the left), one a move on model (on the right) and finally a move in both (in the middle). since using the p roduct aggregation the data cost is equal to . , the algorithm selects the move in both, being the one with the lowest cost. this section describes a set of experiments we performed to obtain a proof-ofconcept of the approach. we compared the diagnostics returned by an existing approach [ ] and our new cost functions with three t − norm aggregations. more precisely, we aimed to get the answer to the question: what is the impact of different aggregation operations on the obtained alignments? in particular, we assess the impact of the aggregation function in terms of a) differences in the overall deviation cost, and b) difference in terms of the interpretation, i.e., the moves selected by the alignment algorithm as the best explanation for the deviation. in order to get meaningful insights on the behavior we can reasonably expect by applying the approach in the real world, we employ a realistic synthetic event log, consisting of , introduced in a former paper [ ] , obtained starting from one real-life logs, i.e., the event log of the bpi challenge . we evaluated the compliance of this log against a simplified version of the process model in [ ] , to which we added few data constraints (see fig. ). the approach has been implemented as an extension to the tool developed by [ ] , designed to deal with bpmn models. our process model involves two constraints for the activity w f a, i.e., amount ≥ and duration ≤ . here we assume that amount ∈ ( , ) and duration ∈ ( , ) represent a tolerable violation range for the variables. since we cannot refer to experts' knowledge, we derived these values from simple descriptive statistics. in particular, we considered values falling within the third quartile as acceptable. the underlying logic is that values which tend to occur repeatedly are likely to indicate acceptable situations. regarding the shape of the membership functions for the variables, here we apply the linear function μ, as reported below. for the classic sum function, we use the cost function provided by ( ); while for the new approach with aos, we apply the cost function in ( ) . we tested the t − norms: minimum, p roduct, and y ager. when data deviations and control-flow deviations show the same cost, we picked the control-flow move. this assumption simulates what we would do in a real-world context. indeed, without a-priori knowledge on the right explanation, it is reasonable to assume that it is more likely that the error was executing the activity, rather than accepting out-of-range data deviations. note that here we focus on the activity w f a, since, in our log, is the only one involving multiple data constraints. table shows differences in terms of number and type of moves, as well as in terms of costs. the columns #move in log, #move in data show the number of traces in which the alignment has selected for the activity w f a a move in log or a move in data, respectively. the column "average costs" shows the average alignment cost. the conformance checking algorithms selects for each activity the move corresponding to the minimum cost. therefore, the differences among the chosen move depend on the different costs obtained on w f a when applying different operators. to provide a practical example of the impact of the aggregated cost on the obtained diagnostics, below we discuss the results obtained for one trace. table shows the cost of possible moves for w f a according to the aggregation functions. table shows the move picked by each function to build the alignment. using the sum function, the data cost is . , so that a move-in-log is chosen as an optimal alignment. in the other cases, instead, the move in data is the one with the lowest cost. since both the deviations fall in the acceptable range, this interpretation is likely to be more in line with the user's expectations. the observations made for the example can be generalized to the overall results of table , which shows a set of traces whose interpretation is heavily affected by the chosen cost function. as expected, the sum function is the most biased towards the choice of move in log interpretation. it selects moves in log more that product and min, and more than yager. one can argue that this choice is likely not one the human analyst would have expected. indeed, we are using yager with ω = [ ] , that means that when both the variables show severe deviations, we expect the data cost to be and move-in-log to be picked. this means that at least of the aligned traces were marked as move-in-log also if both the constraints did not show severe deviations. we argue that this behavior can be misleading for the analyst or, anyway, not being in line with her needs. the product function marks other traces as move-in-data, in addition to the ones marked by the yager. this was expected, since the product function relaxes the requirements on the full satisfaction of the set of constraints. nevertheless, this implies that in all these traces the deviations always fell in the tolerance range. therefore, also these situations might have been better represented as data deviations, depending on the analysts' needs. as regards the min function, it returns a full data deviation in the presence of at least one deviation outside the deviation range, which explains why it returned the same alignments of the product function. the overall alignments costs are in line with the expectations. the sum function returns the highest average cost, as expected, the min the lowest, while the yager and the product behave similarly, and the difference can likely be explained with the traces of difference discussed above. while the absolute difference among the costs is not very relevant, these results show that both the alignments and the assessment of the deviations are impacted by the choice of the cost function, thus highlighting once again the need for a more flexible approach to compliance assessment allowing the user to tailor the cost function to her context. during the last decades, several conformance checking techniques have been proposed. some approaches [ , , ] propose to check whether event traces satisfy a set of compliance rules, typically represented using declarative modeling. rozinat and van der aalst [ ] propose a token-based technique to replay event traces over a process model to detect deviations, which, however, has been shown to provide misleading diagnostics in some contexts [ ] . recently, alignments have been proposed as a robust approach to conformance checking based on the use of a cost function [ ] . while most of alignment-based approaches use the standard distance cost function as defined by [ ] , some variants have been proposed to enhance the provided diagnostics, e.g., the work of alizadeh et al. [ ] , which computes the cost function by analyzing historical logging data. besides the control flow, there are also other perspectives like data, or resources, that are often crucial for compliance checking analysis. few approaches have investigated how to include these perspectives in the analysis: [ ] extends the approach in [ ] by taking into account data describing the contexts in which the activities occurred. some approaches proposed to compute the control-flow first then assessing the compliance with respect to the data perspective, e.g. [ ] . these methods gives priority to check the control flow, with the result that some important deviations can be missed. [ ] introduces a cost function balancing different perspectives, thus obtaining more precise diagnostics. the approaches mentioned so far assume a crisp evaluation of deviations. to the best of our knowledge, the only work which explored the use of a fuzzy cost function is our previous work [ ] which, however, did not consider multiple constraints violation. in this work, we investigated the use of fuzzy aggregation operations in conformance checking of process executions to deal with multiple data constraints for an activity. the proposed approach enhances significantly the flexibility of compliance checking, allowing the human analyst to customize the compliance diagnostic according to her needs. we elaborated upon the relevance of this aspect both theoretically and with some examples. as a proof of concept, we implemented the approach and tested it over a synthetic dataset, comparing results obtained by cost functions with classic sum function and three different aggregations. the experiments confirmed that the approach generates more "balanced" diagnostics, and introduces the capability of personalizing the acceptance of deviations for multiple guards. nevertheless, there are several research directions still to be explored. in future work, first we plan to test our approach with real-world data. furthermore, we intend to investigate the usage of different aggregation functions, as well as the possibility of extending the notion of aggregation to take into account also other kinds of deviations. finally, we intend to investigate potential applications, for example in terms of on-line process monitoring and support, with the aim of enhancing the system resilience to exceptions and unforeseen events. process mining manifesto replaying history on process models for conformance checking and performance analysis mining process performance from event logs towards robust conformance checking memory-efficient alignment of observed and modeled behavior alignment based precision checking constructing probable explanations of nonconformity: a data-aware and history-based approach history-based construction of alignments for conformance checking: formalization and implementation conformance checking and diagnosis for declarative business process models in data-aware scenarios comprehensive rule-based compliance checking and risk management with process mining model predictive control using fuzzy decision functions data-and resource-aware conformance checking of business processes aligning event logs and process models for multi-perspective conformance checking: an approach based on integer linear programming generalized best-first search strategies and the optimality of a discovering anomalous frequent patterns from partially ordered event logs predicting critical behaviors in business process executions: when evidence counts fuzzy measures and integrals in mcda aggregation functions: construction methods, conjunctive, disjunctive and mixed classes aggregation functions: means weighted constraints in fuzzy optimization operations on fuzzy numbers extended by yager's family of tnorms fuzzy sets and fuzzy logic: theory and applications balanced multiperspective checking of process conformance conformance checking of processes based on monitoring real behavior dependence-based data-aware process conformance checking compliance checking of data-aware and resource-aware compliance requirements modeling decisions: information fusion and aggregation operators aligning event logs to task-time matrix clinical pathways in bpmn for variance analysis towards multiperspective conformance checking with fuzzy sets the research leading to these results has received funding from the brain bridge project sponsored by philips research. key: cord- -utpg p u authors: erdmann, anett; ponzoa, josé m. title: digital inbound marketing: measuring the economic performance of grocery e-commerce in europe and the usa date: - - journal: technol forecast soc change doi: . /j.techfore. . sha: doc_id: cord_uid: utpg p u this research investigates the cost-result relationship of the inbound marketing actions used by grocery e-commerce. the analysis is based on the application of the dorfman and steiner ( ) model for optimal advertising budget, which is adapted by the authors to digital marketing and verified with empirical statistical analysis. considering leading companies in six countries over a time horizon of six years, an analysis of the mix of seo and sem techniques aimed at the attraction and conversion of internet users to the web pages of their companies is carried out. the results confirm that e-commerce is optimizing digital inbound marketing in line with the established model. differences are identified depending on the type of format (pure player versus brick and mortar) and at the country level (uk and usa versus others). the market entry of pure players (pp), with its innovative management and marketing techniques (philipp, ; verhoef et al., ) , has seen a revival in retailing. click and mortar (cm) have adapted (or are adapting) processes associated with different areas of digital commerce: tangible and intangible services and resources (cronin, ; beitelspacher et al., ) , logistics and product delivery (hänninen et al., ) , brand influence (zarantonello and pauwels-delassus, ) , product selection, presentation format and demand forecast (boyd and bahn, ; cenamor et al., ; chong et al., ) , dynamic price setting (petrescu, ; cebollada et al., ) , and management of offers, promotions and recommendations by other users (chong et al., ; breugelmans and campo, ) . digital inbound marketing (dim) has also been studied in the marketing literature from a perspective that contemplates its conceptualization, its different techniques, its function within the marketing system in general, and its own application or management (bleoju et al., ; halligan and shah, ; hernández et al., ; opreana and vinerean, ; vieira et al., ; patrutiu-baltes, ) . hence, digitalization of marketing fuels the buying process in all its stages (dahiya, ) , from attraction to loyalty (baye et al., ; seitz et al., ; jun et al., ; melis et al., ) , but concrete analytical guidance is sparse. in particular, regarding the optimization of investment in advertising and marketing, which has been addressed by several authors (bagwell, ; corfman and lehmann, ; cooper and nakanishi, ; eryigit, ; kienzler and lischka, ; wierenga, ) , the nature of dim changes the cost structure of advertising and marketing (frohmann, ) . the techniques of sem, backlinks (in the form of external links), and display require an explicit marketing expense (goldfarb, ) . the amount of the budget to be invested by the marketing department depends mainly on the selected internet media type in which it is advertised, type of content (video, image, text) , characteristics (size, position on the website), selected keywords, and the obtained website visits ("performance-based") for the e-commerce (melis et al., ; halbheer et al., ) . seo and backlinks (in the form of internal links) involve internal costs generated from the company's own structure (ziakis et al., ) . this triggers our research question on how the standard theory of optimal advertising can be used in the digital environment understanding its performance and optimal composition and whether we see this reflected in firm behavior. that is, our goal consists of analytically exploring the relationship between dim and performance in terms of economic efficiency analysis (green, ) . concretely, in the environment of grocery e-commerce (ge) and from a classic economic perspective of evaluation of the optimal marketing budget of dorfman and steiner ( ) , which has been adapted by the authors of this research to the nature of digital marketing, an analytical model is proposed which allows to optimize the investment in marketing based on a marginal analysis and respond to a fundamental question in dim: the analysis of its economic performance. the model accounts for conversion as well as the cost structure of dim techniques and therefore goes beyond just focusing on the sales outcome and allows to address efficiency (desired result at minimum costs). in the context of ge, using search traffic data, we conduct an econometric analysis to test the hypothesis that the firm behavior of the leading grocery e-commerce meets the established economic optimality of dim. additionally, with the purpose of revealing possible differences across firms and the state of the question on dim performance in europe and the usa, the following three objectives are established: first, measurement of the efficacy (desired result) of the dim in terms of visits (capture) of users to e-commerce and in terms of sales volume (conversion) achieved by e-commerce and representation of the conversion technology of a firm. second, measurement and analysis of the variable cost structure of dim and possible cost advantages. third, analysis of firms, in terms of standard performance criteria (like conversion advantage and cost advantage) and the introduced dim-efficiency measure, as cost -marketing efforts optimization relationship, which is presented using position mapping. this comparative analysis allows us to describe the observed differences regarding the way in which pp and cm optimize their investments in dim. given the percentage of the marketing budget invested in dim by digital business managers, which, according to statista ( ) , is more than percent, implications for management are considered equally relevant. we provide operational conclusions for data-based marketing managers in terms of mixing digital advertising and positioning through search engine optimization (seo) and search engine marketing (sem) to optimize digital marketing costs. the analysis of the optimal dim mix of seo and sem oriented towards the generation of website traffic is conducted for a total of leading grocery e-commerce firms in europe (uk, france, germany, netherlands, and norway) and the usa over a time horizon of six years ( - ) at a monthly level to evaluate the relationship between dim and economic performance. data on organic positioning and paid positioning are extracted from the web analytics tool semrush ( ), to construct the main dataset for the study. originally used by companies and digital media planning agencies, the tool has recently found application by academic researchers in the field of digital marketing huang and shih, ) . complementary data on online sales generated by each of the e-commerce companies is provided by the lzretailytics ( ) database and ecommercedb ( ). considering grocery e-commerce in particular, "food retail business or food retailing is a collective term for retailers, which primarily carry food products in their assortment" (seitz et al., (seitz et al., , p. , especially fast-moving consumer goods (fmcg), which have been studied in this sense by several authors (e.g. kureshi and thomas, ; barile et al., ; elms et al., ; wilson-jeanselme and reynolds, ) and is considered a strategic sector of the retail assortment due to the loyalty and the reiteration of purchase it generates among customers (sieira and ponzoa, ) . grocery e-commerce is a subset of this retail segment with the integration of internet technology reflected in different online business formats, which has become an integral part of the grocery industry in some countries (e.g., kureshi and thomas, ) . the most common formats observed are websites of retailers born on the web (pure players) and those of a physical nature that have incorporated e-commerce as a business unit (brick and mortar initially, click and mortar today) or offer a click and collect format (bleoju et al., ; davies et al., ) . the transformative process in which the retail sector is immersed goes through digitalization, a megatrend that affects both digital and physical commerce (bleoju et al., ) . in this sense, today, an establishment that only has one type of format or sales channel may not be competitive, since multichannel is becoming one of the strategic factors from which business is derived (breugelmans and campo, ) . a clear tendency exists to mix the physical and virtual store, thus making digitalization and virtualization intermingle (hänninen et al., ) . this phenomenon is especially significant in grocery, with marketing managers facing different challenges, including: digitalization without denaturing the product or removing its healthy or fresh features (communication challenge), delivering the product at the consumer's house in optimal conditions (logistical challenge), and continuously adapting the sale of food and drink to demand (customer behavior), especially in formats, product development or the combination of items directed at different consumer segments. the evolutionary process of e-commerce and marketplace in relation to the inclusion of new product categories (boyd and bahn, ) is assumed to incorporate grocery into its assortment. it is worth mentioning that the literature regarding the evolution of online grocery sales is heterogeneous. while some sources report a slow increase for general e-commerce (between % and %), others provide larger magnitudes up to % (emarketer, ; statista, ; eurostat, ; ecommercedb, ) . for the respective online sales of food and drinks, the analysis shows a flat evolution over time that contrasts with the rest of the categories. in terms of sales participation, depending on the source and period analyzed, the results vary between % (in the case of general e-commerce) and % (in the case of e-commerce with a clear bias in the sale of grocery). the process of maturing electronic commerce, its penetration as a purchase option for younger audiences (kureshi and thomas, ) , and the search for differential value in the assortment make it one of the key categories. moreover, this evolution has been fast-forwarded through the recent coronavirus disease (covid- ) pandemic when many consumers were triggered the first time to buy groceries online (statista, ; coresight research, ) . the strategic nature, its transformative process, and the challenges the firms face make e-commerce grocery an area of particular interest for academics and marketing professionals. the results confirm that e-commerce is optimizing digital inbound marketing in line with the established model and identifies differences across countries and by the type of firms. emphasis is placed on the marginal analysis for the use and readjustment of the set of dim techniques, which in general depends on the starting point of the company. as the analysis is not constant, it has to adjust to the individual situation of each company to identify the investments in dim with the highest return (individually and as a mixture of techniques). in this sense, we are not looking for a single governing rule based on competition, but rather individual adjustment guidelines based on an optimization condition. in this process, three main components are discussed: the management of the technology on which seo and sem techniques are based, the opportunity cost derived from the market entry, and the different customer management strategies applied between the marketing managers of the retail companies of the pure player and click and mortar retailers. within the context of data-driven retail management of e-commerce websites, we systematically survey the literature on optimal digital marketing actions within a conceptual framework of the stages of the online buying process-from attraction to loyalty-under the roof of the optimization of the marketing budget, which is illustrated in fig. (this framework lays out the basis for the methodological design of the analysis, explained in detail in section ). that is, the research relates to both marketing literature on the buying process and the economic literature on return and analytical models of optimization within the digital environment. table provides an overview of the most relevant preceding studies and the a. erdmann and j.m. ponzoa technological forecasting & social change ( ) positioning of our research, which are explained in detail in the following within the conceptual framework. considering the current level of development of the information system of firms, an important part of the communicative interactions of users with firms' websites can be measured (molodchik et al., ; opreana and vinerean, ; sandvig, ; seitz et al., ) . the concept of service-dominant logic (sdl) is restructuring the vision of research and business practice. transactional interactions are likely to be used for service optimization, with customer orientation being a critical operant resource that can lead to superior market performance of retailers, especially when leveraging different operant resources in the supply chain (beitelspacher et al., ) . a concrete example of a customer-oriented service-based resource is providing a direct link from the corporate website homepage to the e-commerce site of the firm, which is still observed with a certain delay in some concrete markets for food and beverage products (festa et al., ) . e-commerce retailers should take advantage of the amount of data available to optimize their web activity (barile et al., ) and consider that their competitors will also do so (croll and yoskovitz, ) . there is a need for a systematic process to define and readjust the use of new digital marketing techniques (goldfarb, ; clarke and jansen, ) . retail is one of the sectors in which digitalization has had the greatest impact. for their strategic decision-making, the marketing managers of the sector require precise studies of consumer behavior that consider cross-platform access to their sales channels (hänninen et al., ) . it is possible to identify and control kpis and business indicators through new digital tools and measurement techniques for audiences on the website, both as quantitative (saura et al., ) and qualitative analysis (aulkemeier et al., ) . this implies a new way of management, in which fast and precise access to information plays a fundamental role (breugelmans and campo, ) . for the analysis of traditional marketing, there are several analytical models of decision-making regarding marketing expenses. one of the best known is the dorfman and steiner ( ) model, based on a marginalist analysis with a microeconomic approach. concretely, their analytical model shows that if the demand is sensitive to advertising, the optimality condition for maximizing profits is based on the marginal return, and the marginal cost of an additional dollar spent on ads, which has established as a fundamental theory in management and industrial organization (wierenga, ; froeb et al., ; bellflamme and peitz, ; bagwell, ) . considering the marketing literature, after the publication of the fundamental theorem of kotler ( ) , which proposed a proportional relationship between the marketing effort and the sales or market share, a wide variety of methods -theoretical, econometric or rules of thumb -emerged with the objective to improve the efficacy of marketing actions and the allocation of the marketing budget (jones, ; corfman and lehmann ) . with the digitalization of marketing, the relation between effort and sales is used to analyze the conversion rate (moe and fader, ) and the allocation and interaction of resources used in the offline and online channels (wiesel et al., ; banerjee and bhardwaj, ) . the study by wiesel et al. ( ) offers an empirical example of how different online and offline marketing activities (flyer, adwords, discount, etc.) affect purchase funnel metrics. their findings help firms in the decision of allocating resources from a sales perspective, considering the response of one activity on the other, and interestingly find that online funnel metrics have a unidirectional effect on offline funnel metrics. similarly, breugelmans and campo ( ) empirically identify for the largest online grocery retailer in the uk, an asymmetric cannibalization effect of promotions in the online channel on the offline channel. these studies emphasize the need to optimize the resource allocation in online marketing, controlling continuously for an optimal response in terms of investment adjustment. regarding dim, and specifically in the stages of attraction and conversion, there have been multiple approaches with different focuses erdmann and j.m. ponzoa technological forecasting & social change ( ) of analysis. considering the literature on attracting search traffic (organic or paid), many studies have focused on paid traffic, which motivated baye et al. ( ) to explicitly analyze organic traffic through seo for online retailers, confirming the importance of seo strategies in attracting consumers to retailers' e-commerce websites, in particular, the benefits of rank improvement and brand awareness. saura et al. ( ) conduct a study to understand digital marketing based on the identification of the ratios and metrics used in the professional world. the authors highlight the benefits of web analytics for digital marketing depending on the context, instead of a general rule. concretely, in a systematic review focused on seo and sem techniques, they identify the most relevant kpis to control efficacy of dim: conversion rate, user differentiation between new and returning visitors, type of traffic source, and keyword type and ranking. since a successful dim strategy needs to be both effective and efficient, in the present paper, we focus on efficiency, in the context of a contraction model where advertisers pay when the ad is clicked (pay-per-click (ppc)), and consider both paid and organic search traffic. seitz et al. ( ) center their study on the consumer interest in the german grocery market and empirically identify, from the perspective of practical application, the consumer type and incentives of attraction to the website, which allows to differentiate digital marketing content and actions by consumer groups. concretely, they find that working mothers and young professionals show a significantly higher interest in online grocery shopping than other groups and the most important reason is convenience (independence from opening hours, easy ordering, no queueing, time saving) while the main obstacle was the lack of trust in grocery e-commerce and digital marketplaces. such preceding studies show that the observed search traffic stems from a variety of heterogeneous consumer interests that can be explicitly targeted. if it comes to the specific content of interest, halbheer et al. ( ) , examine the optimality ratio of offering free samples to disclose quality and offering paid content only, within the dorfman and steiner ( ) framework. the authors conclude that the optimal decision depends on the sensitivity of consumers' quality expectations with respect to free samples, reflecting the trade-off between market expansion through learning and cannibalization of their own sales. with the aim to make data on search traffic useful to forecast technological adaptation in terms of sales volume, jun et al. ( ) use a time series analysis based on keywords used. the authors were able to identify that branded keywords can be used for predicting the purchase behavior of website visitors. another study that models conversion in detail, rather than providing only aggregated measures, is moe and fader ( ) . they propose and estimate a structural model on the purchase probability based on clickstream data of amazon, differentiating by shopper motivation (directed buyers, search visitors, hedonic browsers, and knowledge-building visitors). the authors highlight the dynamics of the individual purchase-threshold of consumers, which for returning consumers may diminish over time due to a higher frequency of visits before making the purchase decision but at the same time may increase due to prior positive purchase experience. these studies show once more that search traffic can be exploited to measure performance and increase understanding of the purchase process. however, since these papers focus on the demand side in terms of consumer attraction, their objective differs from our paper as we focus on economic optimality in terms of the minimization of digital marketing costs for a given sales objective. the process of customer engagement is reconsidered by bowden ( ) and naumann and bowden ( ) , extending the understanding of customer engagement to a variety of brand-focused activities and suggesting new measures of loyalty (as a crucial outcome of engagement) in terms of satisfaction, affective commitment or rapport. melis et al. ( ) empirically analyze the optimal store choice with a focus on loyalty. differentiating between variable and fixed shopping utility, they find that online shopping choice is initially determined by table research contribution embedded in the literature on dim. a. erdmann and j.m. ponzoa technological forecasting & social change ( ) the preferred offline retail brand, but with the online experience, the online store loyalty dominates the consumer choice online. on the other hand, as mentioned earlier, the work by moe and fader ( ) , suggests a dynamic and ambiguous result of loyalty on the conversion rate. these dynamics of decision-drivers in the online grocery choice of consumers can be regarded as an indication that optimal digital marketing choice is also relevant at the loyalty stage and suggests a continuous readjustment of marketing techniques. from a cost perspective, the main challenge of the home delivery business model are the high delivery costs in the "last mile" (shipping costs from the local platform to the consumer's home or work) and the high expectations of consumers on the internet in terms of fast and correct delivery on time. zissis et al. ( ) study possible collaboration between companies in urban areas to manage the distribution challenges of last-mile delivery. all this reduces the profit margin in the online segment (suel and polak, ) . considering the costs associated with marketing techniques or attraction of potential consumers, reinares and ponzoa ( ) analyze the optimization of the marketing budget based on the cost of contact through different direct marketing channels (mailing, email, telemarketing, and sms). however, in the literature on digital marketing techniques, the cost as profit driver is in general neglected or considered in isolation in the context of a contraction model where advertisers pay when the ad is clicked (saura et al., ) . finally, considering the evolution of the marketing mix, jackson and ahuja ( ) provide an overview of the increased relevance of customer-centric marketing in a changing technological environment, that provides new analytical tools, automatization of the sales force, and data mining. in this context, the authors suggest to move away from a marketing-mix understood as "demand-impinging instruments", but instead redefine it towards a set of adjustable tools to gain competitive advantage and maximize profits in the long run. from the beginnings of the application of the dim techniques, new marketing metrics and data access the marketing landscape has changed and is still changing considerably, with the new resources being beneficial to firms as well as to customers. digital marketing research related to electronic commerce identifies digital platforms as the most outstanding digital growth strategy (verhoef et al., ) . in this line, cenamor et al. ( ) analyze the effect of selling through digital platforms, for example, in the marketplace alibaba (www.alibaba. com), identifying the performance implications for firms and the creation of competitive advantages. these business-to-consumer platforms imply changes in the cost structure of digital marketing (although different from pure e-commerce), shifting advertising costs from the marketing or advertising budget to transaction fees or provision (frohmann, ) . the firm's beliefs about entry and positioning in the online market for grocery have been investigated by kureshi and thomas ( ) . the authors identify, based on firm interviews, positive outcome beliefs in terms of business expansion and increased visibility and reputation, but also concerns about increasing and restructuring inventory management or increased costs for store helpers or computer assistants. additionally, social and peer pressure with respect to customer expectations, suppliers, and rivals, drives the firm's belief in gaining first-mover advantage entering the online business, with no entry or exit cost. given these positive and negative beliefs about the outcome of online market participation, the performance of market leaders in the sector may provide some guidance. on the other hand, from a consumer perspective, aponte ( ) identifies perceived security, privacy, risk, and website quality as determinants of consumer confidence towards e-commerce. all these studies reveal the existing uncertainty and importance in regards to the information structure and data-based decision in the new online business models. the literature on strategic interactions between companies in digital transformation processes is sparse. the first research in this line has been conducted by zutshi et al. ( ) , estimating a management model based on game theory to identify the number of potential customers registered on the web ("online leads") that a company must achieve in order to compete efficiently in a market. in this context, the operational optimization of companies in terms of dim techniques is considered fundamental for further steps of competition based strategies. the literature that uses or verifies economic theories within the framework of digital transformation is still scarce. fedoseeva et al. ( ) verify the economic theory that suggests that a market with better-informed consumers (able to compare prices online without incurring substantial opportunity costs) reduces price dispersion. considering the retail sale of grocery on the internet, these authors identify that there is no price convergence; that is, they conclude that there are still significant differences in the prices of online commerce. similarly, the digitalization of the whole buying process triggers the question of whether firms meet the economic theory of optimal marketing spendings within a digital environment. "the need for actionable knowledge" in the adaptation of digital marketing in this new market is emphasized by bleoju et al. ( ) , the closest paper to the present work, which provides practical insights on how to switch between a focus on inbound marketing and outbound marketing. concretely, the authors identify based on a firm-level survey on the use of digital inbound and outbound activities, including firms with different degree of integration of internet technologies, that the combination of creating content and interaction commitment explains the propensity toward inbound marketing (while the combination of loyalty profiling and client interests are identified as causal for outbound marketing.) we complete this path for actionable knowledge focusing on dim and when to switch between seo and sem techniques. finally, a multidisciplinary literature review by verhoef et al. ( ) reveals that the digital transformation of incumbents (in general traditional brick and mortar firms) is especially relevant in terms of redefining value-creation, investment in new resources and analytical capabilities. in this context, they state a variety of kpis and intermediate results, which are important for fine-tuning the new business and observe that while traditional firms stick to financial profitability, pure digital firms focus on the growth of users, customers or sales. likewise, baye et al. ( ) find that pure online retailers receive, on average, % more organic traffic than their click and mortar competitors. our paper investigates this difference between click and mortar and pure online firms further in terms of the economic efficiency of digital marketing as essential intermediate performance metrics. thus, based on the underlying mechanisms of the purchasing process online and the digitalization of marketing discussed in previous studies, we extend the literature on data-based marketing management, theoretically and evidence-based. our adaptation of a classical analytical framework for optimal advertising explicitly considers the structure of marketing costs online as profit drivers, neglected in the literature, which allows performance measure in terms of efficiency. based on the model, the usability is verified based on search traffic data for leading firms in the sector. additionally, we study the existence of differences in the use of dim techniques between companies and, in particular, the difference between pure players and the new click and mortar players, in terms of performance measures and conversion technology and cost structure. the results provide implications for data-based management decisions on the adjustment of dim techniques optimizing the marketing budget, in particular, the readjustment of seo-and sem-generated visits. this article studies the development on the internet of leading ecommerce organizations in six different countries (usa, uk, france, germany, netherlands, and norway) during a time horizon of six years a. erdmann and j.m. ponzoa technological forecasting & social change ( ) ( - ), comparing the use of dim techniques of firms operating physical as well as virtual stores with retail companies born in the digital environment. four different paths or access routes to the website can be differentiated: (i) visits derived from search engines through paid positioning or sem; (ii) access to the web of e-commerce from search engines through organic positioning or seo; (iii) visits through media links, social networks, web pages or publications that include references and links that direct the user to the respective e-commerce websites (backlinks) and (iv) the visits derived from banners, interstitials, megabanners, billboards, skyscrapers, pop-ups or other graphic formats used to support internet advertising (known by the generic term display ads). from this starting point, we investigate dim techniques used by ecommerce in the process of capturing internet users or addressing them from free web browsing to their websites (see fig. ). in particular, this article focuses its research on seo and sem marketing techniques that make it possible to position e-commerce in web search engines. in order to analyze the economic performance of these dim techniques, the classic economic model of evaluation of the optimal marketing budget of dorfman and steiner ( ) is adapted to the nature of the dim (fig. ) . based on this analytical framework and in particular, the optimization condition to achieve allocative efficiency of the dim effort, the behavior observed by the online grocery retailers is used to verify, using time series analysis, whether grocery e-commerce firms indeed optimize marketing effort in line with the model. in this context, we also analyze potential differences between click and mortar grocery ecommerce compared to pure players in terms of performance and positioning, in particular, the internet access to different e-commerce and the corresponding marketing costs with the objective to achieve a certain level of consumer attraction at minimum costs. using the attraction, interest, desire and action (aida) model fig. proposed by american publicist elias st. elmo lewis (cited by barry, ) , adapted to the digital environment by rowley ( ) we propose a design of research based on two of the fundamental stages of the digital marketing funnel: the interest stage, which involves attracting the customer to the web (measured in number of visits) and the action stage, which involves converting the visit into purchases (measured in sales). the model has been updated by the authors of this article after including the commercial objectives, the marketing results, the inbound marketing techniques used and a new stage of loyalty (motivated by several studies on the impact of dim on engagement and loyalty, discussed in section ) within the process. the stage of attracting the consumer to the web through any of the marketing techniques is associated with explicit or implicit costs, which are analyzed together with the visits and sales generated in the proposed analytical model. as shown in fig. , to obtain the information necessary to create the database of the study, sources referred to dim (provided by semrush), and online sales (provided by lzretailytics) have been used, both being analytic tools of wide acceptance and use in the academic as well as professional world. the database lzretailytics (https://www.retailytics.com/) provides the sales data for the analysis. this data source, provided directly by the german market research firm of the same name that is specialized in the grocery retail sector, has been used to identify for five european countries (uk, france, germany, netherlands, and norway) a total of leading grocery e-commerce firms (tesco, asda, e. leclerc, rewe, ahold, among others) and amazon (the world's leading e-commerce firm with a turnover close to , million us dollars, according to data of the company itself) and its corresponding grocery sales online. firms have been selected based on sales in a decreasing order such that the joint market share covers at least % of the corresponding national grocery e-commerce market and provides a broad assortment that allows the purchase of a standard shopping basket (excludes pure frozen distributors, pet food stores, drugstores). for the us market, the list of companies and sales data comes from ecommercedb (https://ecommercedb.com/). the web analytics tool semrush ( ), which has been used to extract the main data for this study, considers both sem costs (positioning paid at the closing price provided by google in a keyword auction) and seo costs (an estimate based on an extrapolation based on the sem cost of each keyword indexed by e-commerce on google). in this way, and using data very close to the real investment made by companies, this study proposes a marginal analysis (dorfman and steiner, ) of the investment in dim. this web analysis tool that has been used in several academic articles in recent years (molodchik et al., ; sandvig ; huang and shih, ) . semrush allows, through its own algorithm, access to a large number of dim indicators to be visualized through a dashboard. based on a license provided by semrush for research purposes, we created a dataset for this study extracting the necessary variables for the considered ecommerce. by focusing the study on the economic efficiency of dim with a focus on seo and sem, special attention was given to the web traffic generated and the corresponding costs of each of these techniques. the constructed panel dataset includes more than , observations per variable and about , records. for the purpose of replicability, it is considered necessary to state that semrush distinguishes between domain analytics (seo, sem, backlinks, and display data, based on keyword positions) and traffic analytics (clickstream data, by traffic source: direct, reference, unpaid search, paid search, and access from social networks). in this work, we use the data from domain analytics, which provides direct cost estimates for sem and estimated opportunity costs for seo. the historical traffic data and monthly costs are available from january . however, limited by the availability data horizon of economic data from the complementary database lzretailytics, the horizon from january to december is considered. it is important to specify that from january , semrush began to track additional mobile data; however, until april , mobile traffic sources were merged and undefined. to extract data from a consistent time series on the website, we focus on monthly traffic from the computer and within the respective country. for the analysis, we use the merged dataset at the annual level, and the created dim-panel at the monthly level. table summarizes the sample, with the countries ordered from highest to lowest by total sales in food & beverage e-commerce (according to statista, ), indicating online sales and acquisition costs through seo and sem for . in general, a set of metrics (dashboard) are used to evaluate a particular marketing activity. choosing the most relevant metric for decision making and control of objectives depends on the purpose of the analysis. for example, from a financial perspective, the return on investment (roi) is usually used as a standard measure of profitability. alternatively, from a commercial perspective, we could consider the advertising elasticity of demand (aed) as a measure of the efficacy of a given campaign. in this study, both approaches are considered using an economical approach with the objective of minimizing costs and/or maximizing profit. based on classical microeconomics, economic efficiency is defined as follows: "producers are characterized as efficient if they have produced as much as possible with the inputs they have actually employed or if they have produced that output at minimum cost" (green, , p. ). in the present study, we use the dorfman and steiner model ( ) the workhorse of optimization of marketing spending, which is adapted to dim techniques. a recent example of an application of the model is halbheer et al. ( ) , which analyzes the optimal strategy and advertising revenues for a given digital content strategy. in the dim context, with a focus on seo and sem, we define the profit of the company as follows: where demand, and therefore sales, depends on the number of website some sales and cost data were not available in eur and have been extracted in usd applying the average annual exchange rate reported by statista ( ) to convert into eur. a. erdmann and j.m. ponzoa technological forecasting & social change ( ) visits generated and digital marketing costs depend directly and indirectly on the number of website visits. thus, an important part of the cost structure of dim is variable costs, instead of a fixed amount of advertising investment. we differentiate in visits generated by seo (v seo ) and visits attributed to campaigns through sem (v sem ). this process of transformation of visits into sales is illustrated conceptually in fig. (a) . in this sense, a superior "conversion technology" is understood as converting m additional visits in more additional sales (marginal revenues, mr) than an inferior technology (with marginal revenue, mr′). the costs are the sum of the production costs (depending on the quantity sold online) and the dim costs, which are mostly based on performance and, therefore, are directly associated with the visits generated to the website. we could additionally include a fixed amount of investment in advertising (a) as in the classic model, which can be omitted for simplicity. since the visits generated by different dim techniques have different costs, we differentiate between the explicit costs associated with sem traffic (c sem ) and the implicit costs associated with seo (c seo ). the relevant costs for the company are the variable costs (explicit or implicit). here, in order to readjust the marketing-mix of grocery ecommerce, in particular, the relevant costs are the variable economic costs of the dim: note that a low average cost of either technique implies a competitive advantage for the firm. fig. (b) illustrates the relationship between visits and the corresponding marketing costs of website traffic attraction. note that the cost of generating additional traffic of m visits (marginal cost, mc) may depend on the level of the website traffic of the firm as well as on the cost structure relative to competitors. given the individual conversion and cost structure of a firm, there is no general rule but a firm specific optimality condition, determined by the marginal cost and marginal return of additional website traffic: the empirical literature on evaluating the operation of the firm above minimum costs in a debreu-farrell style (green, ) differentiates in technical efficiency (here: conversion of visits into sales) and allocative efficiency (here: misallocation of advertising budget; i.e., using seo and sem in the wrong proportion). this approach requires the estimation of the process of generating sales from different marketing techniques and the estimation of the cost function. at this point, we acknowledge this approach of empirically estimating efficiency in a structural way, which inspired our analytical approach but is not required here. in the interest of brevity, we abstain from a more extensive review of this type of analysis. instead, given the availability of cost data at the disaggregated level (costs for seo, costs for sem), we can directly test the optimality condition. here we use the identifying assumption that marginal revenues per visit are independent of the traffic source (imposed by the availability of sales data at the annual level only). the marginal cost can be calculated directly from the data as the ratio of discrete changes in costs and visits between the two following months. other studies that use directly available cost information (which, for most industries, is from their own information) for statistical testing or empirical analysis come from the electricity industry, and consider the estimation of markup (wolfram, ) or the poor performance of neio estimates versus the use of actual marginal costs data (kim and knittel, ) . the variable that measures the discrepancy from optimality is calculated for each firm at a monthly level for periods. we investigate this hypothesis graphically and econometrically. time series analysis is used to check the hypothesis of optimization based on the discrepancy from optimization as a random variable that should follow a white noise process (e.g., zhang, ; zhang et al., note that higher values imply an inefficient dim-mix while low values suggest the firm operates close to the minimum possible dim spending for a given objective of website visits or sales. however, the absolute discrepancy from optimality doesn't allow comparison across countries. hence, in order to make the measure comparable across firms and markets, we set up a dim-efficiency measure, reformulating the discrepancy from optimality in relative terms and normalized as follows: a. erdmann and j.m. ponzoa technological forecasting & social change ( ) with = c uk france germany netherlands norway usa { , , , , , }and c c. note that this transformation of the discrepancy from optimality is analogous the error formulation in the ordinary least square approach and has been chosen for the following properties: i. account for the total deviation in a market, without positive and negative deviations compensating each other. ii. relative measure, which allows us to rank firms within a country according to their dim-efficiency. iii. normalized metric within the interval [ , ], with rd= implying that the firm operates completely efficient while rd→ implies a misallocation of dim effort, which is a waste of the advertising budget. apart from the presented measures of dim-efficiency, we consider average kpis in terms of average costs per visit (acpv) and average sales per visit (aspv), measuring potential cost advantage and advantage in conversion technology, respectively. finally, we analyze the positioning of the firms given these metrics in national and global contexts. first, we consider the efficacy in terms of the technical ability of firms to generate sales and the associated costs on a visit-by-visit basis to the website. appendix a provides the graphical representation of an estimated visits-sales relationship that captures the company's conversion technology and the cost structure in terms of the relationship visits -dim costs at the country and company levels. note that the positive relationship between online sales and website visits, as well as variable dim costs and website visits, is in line with the model assumptions (fig. ) . moreover, many of the firms show a concave sales function, suggesting a diminishing conversion rate as website visits increase. table presents the dim-efficiency measures based on the adapted dorfman-steiner approach, as well as average revenues per visit (arpv) and average costs per visit (acpv seo and acpv sem ). the data are reported for , as the latest available complete data for sales and search traffic with the associated costs, with the purpose of a descriptive comparison of the state of the market leaders (the relation holds in a similar way for previous years). three interesting observations are apparent: i. the pure player shows the highest dim-efficiency in almost all considered countries. ii. the existence of differences across countries in the level of dimefficiency. uk grocery e-commerce retailers optimize the dim-mix, minimizing the advertising budget, which is in line with the model. likewise, us firms show a high level of optimal resource allocation. iii. the cost advantage in seo in terms of average dim-costs per visit is, in general, held by the national market leader in grocery e-commerce. there is no clear ranking detected for sem costs. in the econometric analysis, we consider the time series of a firm's discrepancy from the optimal marketing budget for seo and sem activities and test whether the observed firm behavior is in line with the proposed model. appendix b (left column) plots the evolution of the absolute measure of dim-efficiency by country and firm. the adf test suggests the discrepancy from the optimality condition is a stationary, random variable with an expected value close to optimality. further, the correlogram and ljung box test do not allow us to reject the hypothesis of independence (no autocorrelation) of meeting the condition each period. hence, our hypothesis has been confirmed, that is, the observed firm behavior is in line with the adapted dorfman-steiner model, with fluctuations around the optimal value in dim adjustment. the descriptive analysis of the time series reveals that some firms show much more volatility in meeting the efficiency criteria than others (heteroskedasticity). to be precise, we find that: i. pure players show less volatile behavior compared to click and mortar firms in the adjustment of the dim-mix. appendix b (right column) expresses the positioning of the firms in terms of cost advantage, conversion advantage, and dim-efficiency. (for all european countries and firms, the revenues and cost measures are in euros, while results for the us are reported in dollars). while the average measures are assigned to the axis, the marginal approach in terms of dim-efficiency is captured by the size of the corresponding circle, with a larger circle implying a more efficient dim-mix. here we pay particular interest to the comparison with the pure player, which is represented with dashed lines. complementarily, a global positioning of all considered firms is provided. ii. in general, the market leader positions in the upper left corner and the pure player in the lower right corner. while the market leaders are highly positioned in terms of revenues per visit and show a cost advantage, the retailers born in the digital environment are most efficient in terms of dim-mix. iii. in norway, where there is no pure player with direct market presence, click and mortar firms show a relatively high cost per visit compared to other markets where physical retailers face direct competition from pure players. we observe in the data that some grocery retailers are gaining market share through their online sales, and in some cases, even position themselves in the online sales ranking in front of traditional industry leaders. in order to understand which metrics have been the drivers of this evolution, we present the dynamic evolution of the positioning strategy of the new grocery retailers and the pure players with arrows for different years (appendix d), showing in different shades of gray the countries included in the study and the retailers analyzed in each one of them. the relationship between customer attraction and sales achieved in the different retailers, and the relationship between pp and c&m (represented by amazon) can be observed. the repositioning of the firms suggests that: iv. in europe, traditional, established retailers are generating fewer sales from the received website visits than earlier, and at the same time, show an increase in dim costs per website visit. v. the pure players, in turn, start operating with high dim costs per visit and gradually increase the average sales per visit and decrease cost per visit. at any time, they operate very close the optimal marketing-mix from an economic perspective, minimizing advertising costs for a given sales objective. the conversion technology and cost structure of the firms suggest the following: vi. for the considered european grocery retailers: a. the technology and cost structure for digital marketing is firmspecific. the marginal revenues and marginal costs associated with k or m more visits differ across firms and depend on the level of visits. a. erdmann and j.m. ponzoa technological forecasting & social change ( ) b. the data suggest that market leader(s) use a superior conversion technology. c. in general, market leaders(s) work with a dim cost structure that is lower than competitors. vii. for the considered us grocery retailers: a. the dim cost structure seem to be similar across firms, such that the difference in marginal costs depends primarily on the level of generated visits where the firm operates. dim efficiency allows to quantify the trade-off of the allocation of the marketing budget to seo and sem exploiting search traffic data, which complements the existing dashboard of firm specific kpis or rules online (saura et al., ) , the isolated optimization of seo (baye et al., ) or sem (balseiro and gur, ) and interrelations between online and offline promotions (breugelmans and campo, ) . the difference between pure players and click and mortar players in managing dim is supposed to be driven by the origin of the firms. entering the online market, brick and mortar firms behave as though in the offline world. note that this observation is analogue the first experience of consumers buying online and taking as reference their behaviour offline with respect to their favourite brick and mortar brand (melis et al., ) and the successive adaptation of behaviour online (moe and fader, ) . hence, as cm firms become more experienced in dim the differences are expected to vanish. this evolution seems to be speed up by the market presence of pp in the country, which may be considered a potential threat by incumbent brick and mortar firms. accomodating the new rival in the country implies a potential loss of sales, but profit effects may be mitigated (or even avoided) by decreasing the necessary advertising budget and hence contribute to stay competitive through dim optimization. a further accelerator of the evolution towards dim implementation and optimization, combining the two previous arguments, is supposed to be the current covid- pandemic. as outlined in the introduction, especially in the grocery industry online sales have started to skyrocket. this provides the firm with plenty of experience, and at the same time all retail firms have speed up the digital transformation implying "new players" in the online market. with experience on the firm-and consumer-side, that is, a consolidated e-commerce in the grocey industry, also differences in the dim cost structure across firms are expected to vanish, as we see it already in the us market. likewise, in terms of the conversion at a visit by visit basis, the considered pure players, born in the usa, show a lower rate, which is notation: the marginal cost cannot be calculated when the firm does not generate any traffic through sem activities during the two following years. ⁎ amazon reported data are for the retailer in total, not the grocery segment. • amazon has no own website. shipping of some products from amazon.com. + outlier in . in general, tesco shows a high dim-efficiency ( : , ). a. erdmann and j.m. ponzoa technological forecasting & social change ( ) interpreted as a result of diminishing returns in the sense of the established model as website traffic increases. this is coherent with the findings in the literature that pure players show a higher organic traffic share than brick and mortar firms (baye et al., ) . currently we are also observing pure players becoming click-andmortar players (example amazon), such that it would be interesting in the future to follow up whether the pp keep on advertising at a stable dim efficiency level or whether the brick and mortar business implies more idiosynchratic shocks. the application of marginal analysis to the optimization of the dim budget depends on the identification of marginal costs, where three main challenges arise: i. the requirement to identify, in addition to explicit costs, the opportunity costs. this is especially relevant in the case of seo, for which an explicit variable cost (or direct cost) cannot be attributedor simply through indirect labor and structural costs. however, from an economic perspective, we argue that website traffic generated from seo could be alternatively generated through a paid search and, therefore, implies an implicit cost (opportunity cost), which is estimated through software solutions such as semrush and should be considered in the optimization of the use of dim techniques. ii. consider that dim activities may have a fixed component and a variable component (cost per click or access to a link) that should be separated from the total advertising budget (hu et al., ; phippen et al., ) . the use of an optimization model for the dim cost is influenced by three main components, in addition to the choice and application of the economic control and measurement model. these components can be deduced from the analysis of the results: iii. the knowledge and management of the technology on which seo and sem techniques are based, influenced by the application of keyword search algorithms in search engines, and, in general, by the internet ecosystem. the fact that pps obtain a competitive advantage in the efficient application of dim techniques with respect to cm indicates a better management knowledge of dim techniques and, therefore, their better application. it is known that innovation and technology adoption, unlike other sectors, has never been one of the main assets of traditional retailing (now cm), greatly influenced by its strong investments in infrastructure and direct customer service through people. this attitude of departure has influenced a lower initial predisposition to investment in technology applicable to e-commerce in professional profiles capable of applying it and in digital marketing budgets assigned for this purpose. vi. the opportunity cost derived from market entry. the pp boosted ecommerce, to which the cms joined some time later. although the study presents a time horizon of six years of e-commerce activity, the commitment of companies in terms of investment in seo and sem positioning has been uneven. our conjecture is that this is due to the need to apply applicable resources and the slowness with which seo investments begin to deliver results. while pps have been very clear from the beginning that positioning keywords through content is crucial for the future development of e-commerce in terms of opportunity cost, the bms have been slow to understand the process of the accumulating value of marketing (in number, type, and position of keywords in search engines). the fact that in the us and the uk (markets with a higher e-commerce penetration rate), this fact is more present confirms this opportunity cost. v. the strategy of customer management by marketing managers in the application of dim techniques has been unequal between cms and pps. in the case of pp, and in a very special way in the case of amazon, its tendency has been to push the client towards more advanced stages (which is reflected in a high but efficient cost per visit): action and loyalty influencing the generation of customer databases, repetition of purchase, increase in average value per purchase, and recommendation to third parties. these are strategies and processes that are very present in direct sales. in contrast, physical retailers (present in the study through the cm) seem to have chosen to replicate themselves on the internet through ecommerce using the strategy followed in physical commerce in order to influence the attraction of clients and to influence sales. both behaviors (pp versus cm) have a direct impact on the way in which the dim is set up. appendix c presents the global perspective of economic performance of dim implementation. the optimization of investment (in terms of marketing) or cost (in economic terms) of marketing actions based on the established objectives of the firm is one of the main concerns of marketing managers. in fact, good professional marketing practice is often associated with its efficacy (achievement of objectives, mainly sales, regardless of cost) and efficiency (in terms of the relationship between costs and results). this study provides new information and raises questions for reflection for marketing professionals regarding the following issues: i. dim is explored and focuses on two of its key techniques: seo and sem. this analysis allows a comparative description of the market situation of the countries included in the study. it offers, in this sense, an intentional outlook based on data on the situation of grocery e-marketing in general and on the situation of the dim in particular. ii. a reflection on decision making in marketing in dim and its future impact can serve as a reference to markets with lower levels of development. iii. the need to apply analytical models of econometric control of investment in dim, based on the large amount of data available, and proposes a specific model of marginalist analysis, which is replicable by the company. iv. the study is based on professional tools and databases (semrush, lz retailytics, and ecommercedb); in this same sense, it highlights the functionality of these solutions and proposes their use through this research design. v. from the theoretical questions outlined in the previous point, conclusions of application in business praxis can be derived, due to the novelty of the study and its own foundation in business reality. deductible management implications of this research are issues such as: a. the company's disposition of the appropriate planning and control tools of dim. b. the recruiting of professional people (internal or external) capable of optimizing investments in seo and sem. c. the importance of considering the opportunity cost that can be derived from highlighted positions in web search engines and their subsequent generated access to e-commerce. d. the need for managers of the e-commerce portals to have a clear, strategic focus on the part of the conversion of generated website visits into sales. e. the interest in generating customer databases from which to establish direct relationships with customers. vi. the model can serve as a starting point for software development a. erdmann and j.m. ponzoa technological forecasting & social change ( ) companies, in a complementary way to the metrics that are already provided, allowing to facilitate cost-dynamic dim readjustment solutions. one of the main limitations of the study is of technical origin. the use of external databases to the data sources of the companies included in the study may introduce error margins between the actual data and those recorded by the tracking tools (semrush in this study). according to information provided by semrush ( ), there are differences between information registered by the tool when processing millions of network interactions and the interactions registered on the server itself, which hosts the web (the computer code) that supports the e-commerce. however, since it is such a high volume of information, and without being able to verify the difference between one data source and another in this research, a very high level of statistical validity is assumed without indecency, or with minimal impact on the results. the data provided and used to identify e-commerce sales volume and the market share of grocery e-commerce come from different sources for the european (lzretailytics, ) and american (ecommercedb, ) market. both databases are based on obtaining information through annual reports from retailers and e-commerce and may apply a broader or narrower market definition. while this may have implications for the identification of the respective market leaders covering the established cumulative market share to ensure representativeness, the advantage of using data from local market experts is an expected higher level of precision in the data and complementary information. moreover, being able to access online sales data on a monthly basis and differentiating by the traffic source would allow us to accurately estimate the marginal effect of additional visits on revenues and disregard the identification assumption that the conversion is independent of the source of web traffic. with regard to future research lines focused on the economic performance of digital marketing, the possibility of applying similar studies to the use of other dim techniques, such as display (or advertising in the networks) or backlink (or generation of links from social networks, blogs and other content support on the internet), in which case it would imply accounting for incomplete information on private costs of companies regarding these backlink and display techniques. finally, it would be desirable to contemplate the interaction between companies in terms of best response based on game theory, within the framework of the analysis of economic efficiency of the dimmix and the richness of the available data with respect to the measurement of dim. the motivation, results, and contribution of the study converge on the idea of providing a new step in the improvement of marketing decisions made in the digital environment. determinantes de la confianza del consumidor hacia el comercio electrónico: una aplicación al caso de puerto rico a service-oriented e-commerce reference architecture the economic analysis of advertising. handbook of industrial organization learning in repeated auctions with budgets: regret minimization and equilibrium aligning marketing and sales in multichannel marketing: compensation design for online lead generation and offline sales conversion grocery retailing in the i . era *. symphonya emerg the development of the hierarchy of effects: an historical perspective search engine optimization: what drives organic traffic to retail sites? retail service-based operant resources and market performance industrial organization: markets and strategies exploring organizational propensity toward inbound-outbound marketing techniques adoption: the case of pure players and click and mortar companies the process of customer engagement: a conceptual framework when do large product assortments benefit consumers? an information-processing perspective cross-channel effects of price promotions: an empirical analysis of the multi-channel grocery retail sector online category pricing at a multichannel grocery retailer how entrepreneurial smes compete through digital platforms: the roles of digital platform capability, network capability and ambidexterity predicting online product sales via online reviews, sentiments, and promotion strategies: a big data architecture and neural network approach predicting consumer product demands via big data: the roles of online promotional marketing and online reviews conversion potential: a metric for evaluating search engine advertising performance market share analysis: evaluating competitive marketing effectiveness us online grocery survey . coresight research. https:// coresight.com/research/us-online-grocery-survey- -many-more-shoppersbuying-more-categories-from-more-retailers/, accessed date the prisoner's dilemma and the role of information in setting advertising budgets lean analytics: using data to build a better startup faster retrospective: a cross-sectional test of the effect and conceptualization of service value revisited a research paper on digital marketing communication and consumer buying decision process: an empirical study in the indian passenger car market buy online collect in-store: exploring grocery click&collect using a national case study optimal advertising and optimal quality internet or store? an ethnographic study of consumers' internet and store-based grocery shopping practices accessed date marketing models: a review of the literature was the economics of information approach wrong all the way? evidence from german grocery r(e)tailing the service-dominant logic perspective for enhancing the e-commerce of wine -a test/application on the italian wine sector digitales pricing: strategische preisbildung in der digitalen wirtschaft mit dem -level-modell analyzing website choice using clickstream data the econometric approach to efficiency analysis. the measurement of productive efficiency and productivity change choosing a digital content strategy: how much should be free? inbound marketing: get found using google, social media, and blogs digitalization in retailing: multi-sided platforms as drivers of industry transformation customer behavior in electronic commerce: the moderating effect of e-purchasing experience incentive problems in performance-based online advertising pricing: cost per click vs. cost per action remaining on current social network sites: an unconscious and conscious perspective do i book at exactly the right time? airfare forecast accuracy across three price-prediction platforms dawn of the digital age and the evolution of the marketing mix ad spending: maintaining market share a study of the method using search traffic to analyze new technology adoption planung von werbeausgaben: Überblick zu den methoden der werbebudgetierung und ihren einflussfaktoren. handbuch werbeforschung biases in static oligopoly models? evidence from the california electricity market online grocery retailing -exploring local grocers beliefs marketing management: analysis, planning, and control the impact of the multi-channel retail mix on online store choice: does online experience matter? dynamic conversion behaviour at e-commerce sites digital relational capital of a company exploring the process of customer engagement, selfbrand connections and loyalty a new development in online marketing: introducing digital inbound marketing inbound marketing -the most important digital marketing strategy online price dispersion-more than imperfect information the case of amazon.com: towards a conceptual framework of online customer service experience (ocse) using the emerging consensus technique (ect) a practical evaluation of web analytics evaluation of cost per contact and cost per response in interactive and direct media planning: a spanish case study. direct mark. an information marketing in a digital world mobilization techniques utilized by leading global e-commerce sites understanding the digital marketing environment with kpis and web analytics online grocery retailing in germany: an explorative analysis & personal care e-commerce report coronavirus: impact on the fmcg market worldwide development of joint models for channel, store, and travel mode choice: grocery shopping in london in pursuit of an effective b b digital marketing strategy in an emerging market digital transformation: a multidisciplinary reflection and research agenda the past, the present and the future of marketing decision models. handbook of marketing decision models marketing's profit impact: quantifying online and off-line funnel progression understanding shoppers' expectations of online grocery retailing measuring duopoly power in the british electricity spot market handbook brand management scales fulfilment time performance of online retailers -an empirical analysis white noise testing and model diagnostic checking for functional time series important factors for improving google search rank collaboration in urban distribution of online grocery orders a game theory approach to online lead generation for oligopoly markets the authors thank lzretailytics®, semrush®, and statista® for providing access to their databases and complementary information to carry out the study. special thanks to josé luis hervás oliver (universitat politècnica de valència) and abel monfort de bedoya (esic business & marketing school) for their careful reading and suggestions. supplementary material associated with this article can be found, in the online version, at doi: . /j.techfore. . . key: cord- -fn zlutj authors: nan title: abstracts of the th annual meeting of the german society of clinical pharmacology and therapy: hannover, – september date: journal: eur j clin pharmacol doi: . /bf sha: doc_id: cord_uid: fn zlutj nan grapefruit juice may considerably increase the systemic bioavailability of drugs as felodipine and nifedipine. this food-drug interaction has potential practical importance because citrus juices are often consumed at breakfasttime when drugs are often taken. it is likely that a plant flavonoid in grapefruit juice, naringenin, is responsible for this effect (inhibition of cytochrome p- enzymes in the liver or in the small intestinal wall). ethinylestradiol (ee ), the estrogen of oral contraceptive steroides, shows a high first-pass-metabolism in vivo. therefore, the purpose of this study is to test the interaction between grapefi-uite juice and ee , the area under the serum concentration-time curve (auc _ h) ofee was determined in a group of young healthy women (n = ) on day + ofmenstruale cycle. to compare intraindividually, the volunteers were randomly allocated to two test days. the female volunteers took lag ee together with either ml of herb tea or with the same amount of grapefruit juice (content of naringenin mg/ ). furthermore, on the day of testing the women drank times ml of the corresponding fluid every three hours up to four times. the auc . h of ee amounts to . + . pg x mi- x h after the administration of the drug with grapefruit juice; that means % higher in comparison to + . pg x m - x h after concomitant intake of tea. also, the mean cmax-value increases to %, p _< . ( . + . pg x m - and . + . pg x m - , respectively). this result shows that the systemic bioavailability ofee increases after intake of the drug with grapefruit juice. the extent of this effect is lower than the extent of known interindividual variability. procarbazine is a tumourstafic agent widely used m hodgin's disease, non-hodgldn's lymphomas and mmours of brain and lung. procarbazine is an inactive prodrug which is converted by a cytochrome p mediated reaction to its active metabolites, in the first step to azoprocarbazine. the kinetics of both procarbazine and azoprocarbazine is not described in humans up to now. on turnout patients we have investigated the plasma kinetics of both procarbazine and azoprocarbazine after oral adminislxation of mg procarbazine in form of capsules and drink solution, respectively. a hplc method with uv-detection ( nrn) and detection limits of and ng/ml was developed for procarbazine and azoprocarbazine respectively. after both the capsules and drink solution the parent drug could be detected in plasma only for h. in contrast the tl/ of terminal elimination of azoprocarbazine was estimated in the range of , to , h with a mean of , h - + , h. the auc of procarbazine was less than % of that of azoprocarbazine. cma x values of azoprocarbazine were determined in the range of , to ,l gg/ml. in comparison to the drink solution we determined on the basis of the plasma levels of azoprocarbazine a bioavailability of the therapeutic used procarbazine capsules of , + , %. prostaglandin e (pge ) is used for the treatment of patients with peripheral arterial disease, and probably effective due to its vasodilator and antiplatelet effects. l-arginine is the precursor of endogenously synthesized nitric oxide (no). in healthy human subjects, larginine also induces peripheral vasodi]ation and inhibits platelet aggregation due to an increased no production. in the present study the influence of a single intravenous dose of l-arginine ( g, min) or pge ( p.g, min) on blood pressure, peripheral hemodynamics (femoral artery duplex sonography), and urinary no -and cgmp excretion rates was assessed in ten patients with peripheral arterial disease (fontaine iii -iv). blood flow in the femoral artery was significantly increased by l-arginine by % (p < . ), and by pge by % (p < . ). l-arginine more strongly decreased systolic and diastolic blood pressure than pge . plasma arginine concentration was increased -fold by l-arginine, but unaffected by pge . urinary excretion of no -increased by % after l-arginine (p < . ), and by % after pge (p = n.s.). urinary cgmp excretion increased by % after l-arginine and by % after pgei (each p = n.s.). we conclude that intravenous l-arginine decreases peripheral arterial resistance, resulting in enhanced blood flow and decreased blood pressure in patients with peripheral arterial disease. these effects were paralleled by increased urinary no -excretion, indicating that systemic no production was enhanced by the infusion. increased no -excretion may be a sum effect of no synthase substrate provision (l-arginine) and increased shear stress (pge and l-arginine). it is weli established that the endothelial edrf/no-mediated relaxing mechanism is impaired in atherosclerotic and in hypertensive arteries. recently it was suggested that primary pulmonary hypertension might be another disease in which the endothelial edrf/no pathway is disturbed. we tested the hypothesis that intravenous administration of l-arginine (l-arg), the physiological precursor of edrf/no, stimulates the production of no, subsequently increasing plasma cgmp levels and reducing systemic and / or pulmonary vasular resistance, in patients with coronary heart disease (chd; n = ) and with primary pulmonary hypertension (pph; n = ). l-arg ( g, min) or placebo (nac ) was infused in chd patients, and l-arg was infused in pph patients undergoing cardiac catheterization. mean aortic (pao) and pulmonary (ppul) arterial pressures were continuously monitored. cardiac output (co; by thermodilution), and total peripheral resistance (tpr) were measured before and during the infusions. plasma cgmp was determined by ria. in chd patients, pao decreased from . + . to . + . mm hg during l-arg (p< . ), whereas ppul was unchanged. tpr decreased from . -+ . to . + . dyne sec cm - during l-arg administration (p< . ). co significantly increased during l-arg (from . + . to . + . /min, p< . ). placebo did not significantiy influence any of the haemodynamic parameters, cgmp slightly increased by . + . % during l-arg, but slightly decreased during placebo (- . + . %)(p < . for l-arg vs. placebo). in pph patients, l-arg induced no significant change in pao, tpr, and co. mean ppul was . + . mm hg at the beginning of the study, but was only slightly reduced by l-arg to . + , mm hg (p = n.s.). plasma cgmp was not affected by l-arg in these patients. we conclude that l-arg stimulates no production and induces vasorelaxation in chd patients, but not in patients with primary pulmonary hypertension. thus, the molecular defects underlying the impaired no foimation may be different m both diseases. institutes of clinical pharmacology, *cardiology, and **pneumology, medical school, hannover, germany. the influence of submaximal exercise on the urinary excretion of , -dinor-pgflc, (the major urinary prostacyclin metabolite), , dinor-txb (the major urinary thromboxane a metabolite), and pge (originating from the kidney), and on platelet aggregation was assessed in untrained and endurance-trained male subjects before and after days of rag/day of aspirin. urinary , -dinor-txb excretion was significantly higher in the athletes at rest (p < . ). submaximal exercise increased urinary , -dinor- -keto-pgfl~ excretion without affecting , -dinor-txb or pge excretion or platelet aggregation. aspirin treatment induced an - % inhibition of platelet aggregation and , -dinor-txb excretion in both groups. however, urinary , -dinor- -keto-pgfl~ was inhibited by only % in the untrained, but by % in the trained group (p < . ). urinary pge was unaffected by aspirin in both groups, indicating that cyclooxygenase activity was not impaired by a systemic aspirin effect. after low dose aspirin administration, the same selective stimulatory effect of submaximal exercise on urinary , -dinor- -keto-pgfl~ excretion was noted in both groups as before. the ratio of , -dinor- -keto-pgfld , -dinor-txb was increased by exercise; this effect was potentiated by aspirin (p < . ). our results suggest that the stimulatory effect of submaximal exercise on prostacyclin production is not due to an enhanced prostacyclin endoperoxide shift from activated platelets to the endothelium, but rather the result of endothelial prostacyclin synthesis activation from endogenous precursors. mg/day of aspirin potentiates the favorable effect of submaximal exercise on endothelial prostacyclin production by selectively blocking platelet cyclooxygenase activity. institute of clinical pharmacology, medical school, hannover, germany. soluble guanylyl cyclases (gc-s) are heterodimeric hemeproteins consisting of two protein subunits ( kda, kda). the enzyme is activated by nitric oxide (no) and catalyzes the formation of the signal molecule "cgmp" (cyclic guanosine- 's'-monophosphate) from gtp. numerous physiological effects of cgmp are already very well characterized. however, detailed insights in the no-activation mechanism of this enzyme have been described to date only in a hypothetical model ( ). recently, this concept was supported by experimental data using sitedirected mutagenesis to create a no-insensitive soluble guanylyl cyclase mutant ( ). it is generally accepted that the prostethic heine-group plays a crucial role in the activation mechanism of this protein. nonetheless, some interesting questions with regard to structure and regulation of soluble guanylyl cyclases still need to be uncovered (e.g. activation with other free radicals, such as carbon monoxide). since this kind of studies is limited so far by isolating large quantities of a biologically active enzyme with conventional purification techniques, the recombinant protein was expressed in the baculovirus / insect cell system. we describe here the construction and characterization of recombinant baculoviruses, harboring the genes that encode both protein subunits of the soluble guanylyl cyclase. insect cells infected with these recombinant baculoviruses produce between - % (as related to total cell protein) of functional soluble guanylyl cyclase. positive infection was monitored as a change in morphology of the cells and by production of the respective recombinant viruses detected by polymerase-chain-reaction (pcr). so far examined, the recombinant enzyme exhibits similar physicochemical characteristics as the "natural" protein. exogenous addition of several heme analogues to the infected cells is able to either stimulate or inhibit the enzymatic activity of gc-s. we are confident to purify milligram amounts of the recombinant protein in the near future. pet studies of myocardial pharmacology have principally concerned the sympathetic nervous system and u'acers have been developed to probe the integrity of both pre-and post-synaptic sites. the sympathetic nervous system plays a crucial role in the control of heart rate and myocardial contractility as well in the conlrol of the coronary circulation. alterations of this system have been implicated in the pathophysiology of a number of .cardiac disorders, in particular, heart failure, ventricular arrhythmogenesis, coronary artery disease, idiopathic dilated and hypertrophic cardiomyopathy. several beta blockers have been labelled with carbon-ll for imaging by pet. the most promising of these is cgp which is a non-selective beta adrenoceptor anatagonist particularly suited for pet studies due to its high affinity and low lipophilicity, thus enabling the functional receptor pool on the cell surface to be studied. studies in our institution in a group of young healthy subjects have yielded bmax values of . _+ . pmol/g myocardium. these data are consistent with literature values of bmax for beta adrenoceptors in human ventricular myocardium determined by a variety of in vitro assays. a recent study in patients with hypertrophic cardiomyopathy has shown that myocardial beta adrenoceptor density is decreased by approximately - % relative to values in normal subjects. the decrease in receptor density occurs in both hypertrophied and nonhypertrophied portions of the left ventricle. these data are consistent with the hypothesis that sympathetic overdrive might be involved in the phenotypic expression of hypertrophic cardiomyopathy. a further decrease of myocardial beta adrenoceptor density (to levels well below _ - . pmol/g) has been observed in those patients with hypertrophic cardiomyopathy who procede to ventricular dilatation and heart failure. cyp a hydroxylates polycyclic aromatic hydrocarbons such as benzo(a)pyrene occurring e.g. in cigarette smoke. two hereditary mutations are discovered: ml, a t to c transition , bp downstream of exon ; m , located at position , in exon representing an a to g transition resulting an isoleucine to valine substitution in the heme-binding region. recently we could demonstrate in caucasians that carriers of the m -mutation possess an increased risk towards lung cancer (drakoulis et al clin.lnvestig. : , ) , whereas the ml-mutation shows no such association. the phasg-ii enzyme gstm catalyses the conjugation of glutathione to electrophilic compounds such as products of cyp ai. gstm is absent in . % of the caucasian population due to base deletions in exon and of the gene. we found no contrariety in the gstm distribution, including frequencies of type a (p.) and type b (v) among lung cancer patients (odds ratio = . , n = ; cancer res. : res. : , . lung cancer patients and reference patients were investigated for mutations of cypia and gstm by allele-specific pcr and rflp. a statistically significant higher risk for lung cancer among carriers of the m trait was found (odds ratio = . , p = . ). interestingly, amid lung cancer, m -alleles were less often linked to ml than in controls (odds ratio = . , %-confidence limits = . - . , p = . ). however, the frequency of cypia mutations did not differ among active and defective gstm types. consequently, we could not confirm in the caucasian population the synergistic effects of cypia mutations (especially m ) and deficient gstm as combined susceptibility factors for lung cancer as described among the japanese (cancer res. : , in healthy subjects the effect of gastrointestinal hormones like somatostatin and glucagon on splanchnic hemodynamics is not well defined due to the invasiveness of the direct measurement of e.g. portal vein (pv) wedged pressure. methods : now, we applied duplex sonography ( . m~z) and color coded flow mapping to compare the effects of ocreotide (i ~g sc), a long acting somatostatin agonist, and glucagon (i mg iv) on the hemodynamics of the pv, superior mesenteric artery (sma) and common hepatic artery (ha) in healthy volunteers ( g,i q; ± y; x ± sem). basal values of pv flow ( . ± . cm/s), pv flow volume ( ± ml/min), sma systolic (sf: ± cm/s) and diastolic flow (df: ± cm/s), sma pourcelot index (pi) ( . ± . ), ha sf ( ± cm/s) and df ( ± cm/s) and ha pi ( . ± . ) well agreed with previously reported results. within min ocreotide resulted in a decrease of sma sf (- ± %) sma df (- ± %), ha sf (- ± %) and ha df (- ± %). maximum drop of pv flow (- ± %) and flow volume (- ± %) occurred at min. all effects diminished at min. no significant change of vessel diameter and pi was seen. min following its application glucagon caused a highly variable, only short lasting increase of pv flow volume (+ ± %) and sma df (+ ± %). ha fd (+ ± %) showed a tendency to rise (ns). we conclude that in clinical pharmacology duplex sonography is a valuable aid for measuring effects of hormones and drugs on splanchnic hemodynamics. pectanginal pain and signs of silent myocardial ischemia frequently occur in hypertensives, even in the absence of coronary artery disease (cad) and/or left ventricular hypertrophy, probably due to a reduced coronary flow reserve. since the oxygen extraction of the heart is nearly maximal during rest, increases of oxygen demand cannot be balanced by increases of myocardial perfusion: to assess the frequency of ischemic type st-segment depressions in this patients and to determine the influence of heart rate (hr) and blood pressure (bp), simultaneous h hoher-and h ambulatory bp monitoring were performed in hypertensives (age - years, f, m) without cad before and after four weeks on therapy with the -blocker betaxolol. episodes of significant st-segment depressions (> . mv,> min) of a total length of min could be demonsu'ated in / patients ( %) without antihypertensive therapy_ systolic bp significantly increased from + . mmhg (mean + sd, p < . ) min before to a maximum of + . mmhg during the ischemic episodes, hr and rate-pressure product (rpp) increased from + . min -t and . + . mmhg x rain -t x to _+ . min-: and . + . mmhg x min - x (p < . ). the extent of st-segment depressions significantly correlated with hr and rpp (p < . ). drug therapy with - mg/d betaxolol for weeks significantly decreased mean hr, systolic' and diastolic bp (p < . ). ischemic episodes of a total length of min were recorded only in of hypertensives ( . %; p < . ; x -test). in conclusion, increases of hr and systolic bp seem to be the most important factors which induce myocardial ischemia in hypertensives without cad. as silent ischemia is a independent risk factor for sudden cardiac death and other cardiac events, specific antihypertensive therapy should not only be aimed to normalize blood pressure, but should also address reduction of ischemic episodes as demonstrated here. phosphodiesterase inhibitors exert their positive inotropic effects by inhibiting camp degradation and increasing the intracellular calcium concentration in cardiomyocytes. an identical phosphodiesterase type i[ has been demonstrated in platelets and vascular smooth muscle cells. we studied the influence ofpiroximone on platelet function in vitro and ex vivo and the hemodynaimc effects of a bolus application of piroximone in patients with severe heart failure (nyha iii-iv) using a swan -ganz-catheter. in order to study the influence ofpiroximone on platelet function in vitro, platelet rich plasma from healthy volunteers was incubated with piroximone ( - ~tmol/l) from minute to hottrs and aggregation was induced by addition of adp. for the ex vivo experiments platelet rich plasma was obtained from patients, who received piroximone in doses of . , . , . or . mg/kg bw. blood samples were drawn immediately before and , , , and minutes after bolus application. the adp-induced platelet aggregation was inhibited time-and dosedependently. the ic value for piroximoue in vitro amounted to + omol/ . in the ex vivo experiments the maximal inhibition of adp-induced aggregation was obtained in prp from patients who had received mg/kg bw piroximune minutes before. the admitdstration ofpiroximone resulted in a marked hemodynamic improvement with a dose-dependent increase in cardiac index and decreases in pulmonary artery pressure and resistance. to treat conditions associated with acute and chronic multiorgan dysfunction. studies indicate patients receive approximately ten drugs, on average during their icu stay, from several drug classes. commonly prescribed drugs include narcotics, sedatives, antibiotics, antiarrhythmics, antihypertensives, drugs for stress ulcer prophylaxis, diuretics, vasopressors, and inotropes. reports suggest surgical icu patients cost the hospital an average of $ , /patient in un-reimbursed costs under fixed-price reimbursement. furthermore, patients with the greatest drain in revenue received catecholamines, triple antibiotics, or antifungal agents. thrombolytics, antibiotics, plasma expanders, and benzodiazepines account for nearly twothirds of the cost of drugs prescribed in medical and surgical icus. agents with considerable economic impact include biotechnology drugs for sepsis. pharmacoeconomic data in icu patients suggest increased attention should be directed towards several areas, including patients with pneumonia, intraabdominal sepsis, nosocomial bloodstream infections, optimizing sedation and analgesic therapy, preventing persistent paralysis from neuromuscular blockers, preventing stress ulcers, treating hypotension, and providing optimal nutritional support. studies are needed to assess the impact of strategies to improve icu drug prescribing on length of stay and quality of life. if expensive drugs are shown to decrease the length of icu stay, then their added costs can have positive economic benefits to the health care system. the responses to min iv. infusions of the -and -adrenoceptor agonist isoprenalin (iso) and the -(and c~-) adrenoceptor agonist adrenalin (adr) at constant rates of ijg/min were evaluated noninvasively after pretreatment (pre-tr) with placebo (pl), mg of the -selective adrenoceptor antagonist talinolol (tal) and mg of the non-selective antagonist propranolol (pro) in healthy subjects. the following were analysed: heart rate (hr, bpm), pre-ejection time (pep, ms), ejection time (vet, ms), hr-corrected electromechanical systole (qs c, ms), impedance-cardiographic estimates of stroke volume (sv, ml), cardiac output (co, i/min) and peripheral resistance (tpr, dyn.s.cm - ) calculated from co and mean blood pressure (sbp and dbp according to auscultatory korotkoff-i and -iv sounds this indicates that ) about half the rise of hr and co and half the shortening of pep is -respectively ~ -determined, ) that predominant -adrenergic responses, whilst not affecting vet, take optimal benefit from the inodilatory enhancement of pump performance, ) that an additional -adrenergic stimulation is proportionally less efficient, as vet is dramatically shortened, thus blunting the gain in sv so that the rise in co relies substantially on the amplified increase of hr and ), vet is more sensitive than qs c in expressing additional -adrenoceptor agonism and ) prime systolic time intervals provide a less speculative and physiologically more meaningful represenation of cardiac pump dynamics than hr-corrected ones. zentrum flit kardiovaskul~re pharmakologie, mathildenstral e , mainz, brd a regression between blunting of ergometric rise of heart rate and l~ladrenoceptor occupancies in healthy man c. de mey, d. palm, k. breithaupt-grsgler, g.g. belz the hr-responses to supine bicycle ergometry ( min at appr. watt) were investigated at several time points after the administration of propranolol (pro: , , mg), carvedilol (car: . , , , mg), talinolol (tal: , , , mg), metoprolol (met: mg) and celipro-iol (cel: mg) to healthy man. the effects of the agents (= difference of the ergometric response for active drug and placebo) were analysed for both the end values (end) and the increments (inc) from resting values immediately before ergometry up to end. the effects were correlated with the %-~l-adrenoceptor occupancies estimated using a standard emax-model (sigmoidicity=l) from the concentrations of active substrate in plasma determined by i~l-adrenoceptor specific radioreceptor assay. the respective intercepts (i), slopes (s) and correlation coefficients (r) are detailed here below : inhibition of leukotrienes is a promising approach to the treatmer~t of several diseases because excess formation of these lipid mediators has been shown to play an important role in a wide range of pathophysiological conditions. since until recently we were not able to obtain specific drugs suppressing leukotriene biosynthesis or action for clinical practice, we started investigating the effects of putative natural modulators of leukotriene biosynthesis such as fish oil. healthy male volunteers were supplemented for days with fish oil providing mg eicosapentaenoic and docosahexaenoic acid per kg body weight and day. the urinary concentration of leukotriene e plus n-acetyl leukotriene e served as a measure for the endogenous leukotriene production, treatment resulted in a significant increase in the eicosapentaenoate concentration in red blood cell membranes. fish oil reduced the endogenous leukotriene generation in of the volunteers. the effect was associated with a decrease in urinary prostaglandin metabolites, determined as tetranorprostanedioic acid. in contrast to what was expected from published in vitro and ex vivo experiments, no endogenously generated cysteinyl leukotrienes of the series could be identified. the inhibitory effect of fish oil on the endogenous leukotriene generation was not synergistic to the effect of vitamin e, which also exhibited some suppressive activity. early clinical data on the effects of fish oil on teukotriene production in patients with allergy or rheumatoid arthritis are not yet conclusive. we conclude that fish oil exhibits some inhibitory activity on leukotriene production in vivo. the effectivity of fish oil may be attenuated by concomitant modulation of other mediator systems e.g. up-regulation of tumor necrosis factor production. • the number and affinity of platelet thromboxane (txa ) and prostacyclin (pgi )-receptors are regulated by several factors. we studied the influence of oral intake of acetylsalieylic acid (asa) on ex-vivo binding studies with human platelet membranes on the binding of the specific thromboxane a antagonist h-sq- and the pgi agunist h-l]oprost. the number of receptors (bmm) and the binding affinity (kd) were calculated using scatchard's plot analysis. in healthy male volunteers o significant difference was seen following intake of mg/d of asa for days (mean -+ sem): the potency of meloxicam (mel), a new anti-inflammatory drug (nsaid), in the rat is higher than that of well-known nsaids, in adjuvant arthrtitis rats, mel is a potent inhibitor of the local and the systemic signs of the disease. mel is also a potent inhibitor of pg-biosynthesis by leukocytes found in pleuritic exudate in rats. conversely, the effect of mel on pg-biosynthesis in isotated enzyme preparations from bull seminal vesicle in vitro, the effect on intragastric and intrarenal pg-biosynthesis and the influence on the txb - evel in rat serum is weak. in spite of the high antiinflammatory potency in the rat, mel shows a low gastrointestinal toxicity and nephrotoxicity in rats. -cyclooxygenase- (cox- ) has been recently identified as a isoenzyme of cyclooxygenase. nsaids are anti-inflammatory through inhibition of pg-biosynthesis by inducible cox- and are ulcerogenic and nephrotoxic through inhibition of the constitutive cox- . we have investigated the effects of mel and other nsaids on cox- of non stimulated and on cox- of lps-stimulated guine pig peritoneal macrophages. cells were cultured with and without lps for hrs together with the nsaid. arachidonic acid was then added for further mins, the medium removed and pge measured by ria. bimakalim, emd , is a new investigational k+-channel activator with vasod[lating properties. single pereral doses of . mg bimakalim, mg diltlazem, either alone or in combination, were investigated in healthy male supine volunteers ( to years of age) [n a placebo-controlled, periodbalanced, randemised, double-blind, way cross-over design. point estimates of the global effects of bimakalim [k] , di]tiazem [d] and their interaction [kxd, = in case of mere additivity] incl. % confidence intervals (ci) were analysed for systolic and diastolic blood pressure (sbp, dbp; mmhg), heart rate (hr; bpm), pq (ms), systolic time intervals (pep, qs c, lvetc; ms), cardiac output (co; i.min- ), total peripheral resistance (tpr; dyn.s.cm- ), heather index (hi; q.s- ); , h after dosing, *statistically significant at a= . : - to - - to - to - to . to . - . to , . to . * - . to , - . to - . * - . tol, - . to , - . to . - . to . - . to . - . to . - . to . - to - to -& to . - . to . afterload reduction and a drop in dbp occurred with bimakalim associated with a rise in hr and mild increase in cardiac performance, diltiazem (slightly) decreased afterload and bp with little (reflectory) accompanying changes and had a negative dromotropic effect. the combination caused additive effects. center for cardiovascular pharmacology, zekapha gmbh, mathildenstr. , mainz, germany. rheumatoid arthritis (ra) is characterized by an immunological mediated inflammatory reaction in affected joints. infiltration of granulocytes and monocytes is the pathophysiological hallmark within the initial phase of inflammation. these cells are able to synthesize leukotrienes. ltb is a potent chemotactic factor and therefore could be responsible for the influx of granulocytes from the circulation. cysteinyl leukotrienes ltc , d and e augment vascular permeability and are potent vasoconstrictors. ltb and cysteinyl leukotrienes have been detected in synovial fluid of patients with ra. however, these results are difficult to interprete, because the procedure is invasive and artificial synthesis cannot be excluded. we used a different, noninvasive approach by assessing the excretion of lte into urine. studies with hltc have demonstrated that lte is unchanged excreted into urine and is the major udnary metabolite of cysteinyl leukotrienes in man. udnary lte was isolated from an aliquot of a hour urine collection by solid phase extraction followed by hplc and quantitated by ria. nine patients were enrolled in the present study. all met the american college of rheumatology criteria for ra. patients were treated with nonsteroidal inflammatory drugs and disease modifying drugs. therapy with prednisolon was started after collection of the initial hour urine sample. disease activity was assessed by crp (mean + mg/l) and esr (mean _+ mm/hour platelet aggregation is mediated by the binding of an adhesive protein, fibrinogen, to a surface receptor, the platelet glycoprotein lib/ilia. gpiib/llla is one of a family of adhesion receptors, integrins, which consist of a ca++-dependent complex of two distinct protein subunits. under resting conditions, gpiib/llla has a low affinity for fibrinogen in solution. however, activation of platelets by most agonists, including thrombin, adp and thromboxane results in a conformational change in the receptor and the expression of a high affinity site for fibrinogen. binding of fibrinogen to platelets is a common end-point for all agonists and therefore is a potential target for the development of antiplatelet drugs. these have included chimeric, partially humanised antibodies ( e ), peptides and peptidomimetics that bind to the receptor and prevent fibrinogen binding. the peptides often include the sequence rgd, a sequence that is present in fibrinogen and is one of the ligand's binding sites. when administered in vivo, antagonists of gpiib/llla markedly suppress platelet aggregation in response to all known agonists, without altering platelet shape change, a marker of platelet activation. they also prolong the bleeding time in a dose and perhaps drug dependent manner, often to more than rain. in experimental models of arterial thrombosis, gpllb/llla antagonists have proved highly effective and are more potent than aspirin. studies in man have focused on coronary angioplasty, unstable angina and coronary thrombolysis and have given promising results. e given as a bolus and infusion combined with aspirin and heparin reduced the need for urgent revascularisation in patients undergoing high-risk angioplasty, although bleeding was more common. some compounds have shown oral bioavailability raising the possibility that these agents could be administered chronically. antagonists of the platelet gpiib/llla provide a novel and potent approach to antithrombotic therapy. drug databases on computers are commonly textfiles or consist of tables of generic-names or prices for example. until now pharmacokinetic data are not easily available for regular use, because searching parameters in a textfile is time consuming and personal intensive. on the other hand these pharmacokinetic data are the fundamental background of every dosage regimen and individual dosage adjustment. for many drugs elimination is dependent on the patients renal function. renal failure leads to accumulation, possibly up to toxic plasma concentrations. therefore, the decision was to build up a pharmacokinetic database. the aim is to achieve simplicity and effectiveness by using the basic rules. only three parameters are needed to describe the pharmacokinetics: clearance (ci), volume of distribution (vd) and half-life (t~). moreover, with two parameters the third can be calculated ancl'controlled by the equation: cl = vd * , / t½ according to the dettli-equation and the bayes' theorem estimation of individual pharmacokinetic parameters will be done by a computer program. the advantage is that the impact of therapeutic drug monitoring can be increased. using the population data and the bayesian approach, only one measurement of serum drug concentrations might be enough to achieve an individual dosage regimens (el desoky et al., ther drug monitor , : ) higher therapeutic security for the patient can be achieved. there is also a major pharmacoeconomic aspect: adapting drug dosage reduces costs (susanka et al., am j hosp pharm , : ) the basic database for future pharmacokinetic clinical desicions is going to be built up. the pharmacokinetic interactions with grape#uit juice reported for many drugs are attributed to the inhibition of cytochrome p enzymes by nanngenin, which is the aglycene of the bitter juice component nadngin. however, only circumstantial evidence exist that naringenin is indeed formed when grapefruit juice is ingested, and the lack of drug interaction when naringin solution is given instead of the juice is still unexplained. we investigated the pharmacokinetics of naringin, naringenin and its conjugated metabolites following ingestion of ml grapefruit juice per kg body weight, containing ijm naringin, in male and female healthy adults. urine was collected - , - , - , - , - , - , - and - hours alter juice intake. naringin and naringenin concentrations were measured by reversed phase hplc following extraction using ethyl acetate, with a limit of quantitation of nm. conjugated metabolites in urine were transformed by incubation with glucuronidase ( u/ml) / sulfatase ( u/ml) from abalone entrails for h at ph . and determined as parent compounds. additionally, naringin and naringenin concentrations were measured in plasma samples from grapefl'uit juice interaction studies conducted previously. neither naringin nor its conjugated products were detected in any of the samples. naringenin was not found in plasma. small amounts of nanngenin appeared in urine alter a median lag time of hours and reached up to . % of the dose (measured as nanngin). after treatment with glucuronidase / sulfatase, up to % of the dose was recovered in urine: the absence of naringin and its conjugates and the lag time observed for naringenin to appear in urine suggests that cleavage of the sugar moeity may be required before the flavonoid can be absorbed as the aglycone. naringenin itself undergoes rapid phase ii metabolism. whether the conjugated metabolite is a potent cytochrome p inhibitor is unknown but not probable. the pronounced variability of naringenin excretion provides a possible explanation for apparently contradictory results in grapefruit and/or naringin interaction studies. grapefruit juice increases the oral bioavailablity of almost any dihydropyridine tested, presumably due to inhibition of first-pass metabolism mediated by the cytochrome p isoform cyp a / . the mean extent of increase was up to threefold, observed for felodipine, and more pronounced drug effects were also reported. thus, a such interaction may be of considerable clinical relevance. no data are yet available for nimodipine. we conducted a randomized cross-over interaction study on the effects of concomitant intake of grapefruit juice on the pharmacokinetics of nimodipine and its metabolites m (pyridine analogue), m (demethylated) and m (pyridine analogue, demethylated). healthy young men ( smokers / nonsmokers) were included into the investigation. nimodipine was given as a single mg tablet (nimotop e) with either ml of water or ml of grapefruit juice (d~hler gmbh, darmstadt, mg/i naringin). concentrations ef nimodipine and its metabolites in plasma withdrawn up to hours p.ostdose were measured by gc-ecd, and model independent pharmacokinetic parameters were estimated. the study was handled as an equivalence problem, and anova based % confidence intervals were calculated for the test (=grapefruit period) to reference (= water period) ratios. the absence of a relevant interaction was assumed if the ci were within the . to . range: grapefruit juice was reported to inhibit the metabolism of a variety of drugs, including dihydropyridines, verapamil, terfenadine, cyclosporine, and caffeine. these drugs are metabolized mainly by the cytochrome p isoforms cyp a (caffeine and, in part, verapamil) and cyp a (others). theophylline has a therapeutic range of - mg/i and is also in part metabolized by cyp a . therefore, we conducted a randomized changeover interaction study on the effects of concomitant intake of grapefruit juice on the pharmacokinetics of theophylline. healthy young male nonsmokers were included. theophylline was given as a single dose of mg in solution (euphyllin e ), diluted by either ml of water or ml of grapefruit juice (d hler gmbhi darmstadt, mg/i nadngin). subsequently, additional fractionated . i of either juice or water were administered until hours postdose. theophylline concentrations in plasma withdrawn up to hours postdose were measured by hplc, and pharmacokinetics were estimated using compartment model independent methods. the study was handeled as an equivalence problem, and anova based % confidence intervals were calculated for the test (=grapefruit period) to reference (= water period) ratios (trnax: differences thus, no inhibitory effect of grapefruit juice on theophylline pharmacokinetics was observed. the lower contribution of cyp a to primary theophylline metabolism or differences in naringin and/or naringenin kinetics are possible explanations for the apparent contradiction between the effects of grapefruit juice on caffeine and on theophylline metabolism. the physical stability of erythromycin stearate film tablets was studied according to a factorial design with experimental variables temperature, relative humidity, and storage time. after one half year of storage at oc and % relative humidity, the fraction of dose released within min in a usp xxl paddle apparatus under standard conditions decreased from % for the reference stored at ambient temperature in intact blister packages to % for the stress-tested specimens. chemical degradation of the active ingredient did not become apparent before months of storage. under all other storage conditions, no effects of physical aging upon drug release were found. the bioequivalence of reference and stress-tested samples was studied in six healthy volunteers. the extent of relative bioavailability of the test product was markedly reduced (mean: . %, range: - %), mean absorption times of the test product were significantly prolonged. the results indicate that the product tested can undergo physical alterations upon storage under unfavourable conditions, and lose its therapeutic efficacy. it can be expected that this phenomenon is reduced by suitable packaging, but the magnitude of deterioration may cause concern. on the other hand, incomplete drug release is in this case easily detected by dissolution testing. whether similar correlations exist for other erythromycin formulations remains to be demonstrated. the efficacy of a drug therapy is considerably influenced by patient compliance. within clinical trials the effects of poor compliance on the interpretation of study results frequently leads to underestimating the efficacy of the treatment. in the evaluation of the "lipid research clinics primary coronary prevention trial" and the "helsinki heart study" special attention was focused on compliance with medication. the strong influence of compliance on clinical outcome and the dilutional effect of poor compliance on the efficacy of the respective drugs occured in both these trials. there are indirect (e.g. pill-count, patient interview) and direct methods (e.g. measurement of drugs, metabolites or chemical markers in body fluids) used to assess compliance with drug therapy. the indirect methods mentioned are commonly considered as unreliable. the direct methods can prove dose ingestion a short time before the sample is taken, however, they cannot show the time history of the drug use. an advanced method of measuring compliance is to use electronic devices. the integration of time/date-recording microcirculty into pharmaceutical packaging, so as to compile a time history of package use, provides real-time data as indicative of the time when dosing occurred. this method supports a precise, quantitative definition of "patient compliance" as: the extent to which the actual time history of dosing corresponds to the prescribed drug regimen. by taking real-time compliance data into account the results from clinical trials show not only clearer evaluations of drug efficacy and dese-reponse-relationship but also a better understanding of dose dependant adverse drug reactions. in the present study, we examined the usefulness of eroderm- and eroderm- . seventy five impotent men, to years old, participated in the present trial. the patients were classified into groups, patients each. the first group was treated by cream containing only co-dergocrine mesilate (eroderm- ), the second received a cream containing isosorbide dinitrate, isoxsuprine hcl and co-dergocrine mesilate (eroderm- ), while the third used a cream containing placebo. the cream was applied to penile shaft and gland / - hr before sexual stimulation and intercourse. the patients were asked to report their experience via questi'onnaire after one week. the results of treatment are as follows: seven patients ( %) who applied eroderm- indicated a full erection and successful intercourse. the use of eroderm- restored potency in patients ( %) of the second group. three men ( %) of psychogenic type reported overall satisfaction with placebo cream. treatment of impotence with eroderm cream was most successful in patients with psychogenic disorders which are often coincident with minor vascular or neurological disorders. fair results were reported by patients afflicted by moderate neurological disorders. except for one case of drug allergy following the use of eroderm- , no side effects were reported. we believe that eroderm cream has obvious advantages and may be a suitable treatment before the use of non-safe method as intracavernous medication. a new type of topically applied drugs (eroderm creams) for impotence is presented. eroderm creams contain vasoactive drugs. these drugs have ability to penetrate the penile cutaneous tissue and facilitate erection. in the present study, we examine the usefulness of eroderm- in the treatment of erectile dysfunction. eroderm- contains tiemonium methylsulfate, a.f. piperazine and jsosorbide dinitrate. a randomized, double blinded control trial on patients was performed. the etiology of impotence was investigated. all patients received eroderm- and placebo cream. the patients randomized into groups of . the first group received eroderm- on day and placebo cream on day , however, group two received placebo on day . the patients were advised to apply the cream on the penile shaft / - hr, before sexual stimulation and intercourse. the patients reported their experience via questionnaire. overall percent of patients demonstrated a response with eroderm- . the other responders reported a partial erection and tumescenous. three men ( %) reported a full crection and satisfied intercourse with either cream. these patients were psychogenic impotence. neither eroderm- nor placebo cream produced marked response in patients. four patients were venous leakage which were advised to use tourniquet at the base of penis after / hr. of cream application. only one of them indicated a good response. the highest activity proved to occur in psychogenic impotence. less rate of success was observed in patients with minor to moderate neurological and/or arterial disorders. no marked side effects were recorded. for these reasons eroderm- may be proposed as first line therapy of erectile dysfunction. control of cell proliferation is a basic homeostatic function in multicellular organisms. we studied the effects of some prostaglandins and leukotrienes and of their pharmacological inhibitors on cell proliferation in murine mast cells and mast cell lines, in a human promyelocytic cell line (hl- cells) and in burkitt's lymphoma cell lines. in addition, prostaglandin and leukotriene production was investigated in mast cells, representing putative endogenous sources of these lipid mediators. murine mast cells were derived from bone marrow of balb/c mice. proliferation of cells was estimated using a colorimetric assay (mtt-test). production of prostaglandin d (pgd ), pgj , delta- -pgj , leukotriene c (ltc ) and ltb by mast cells was determined by combined use of high performanceliquid chromatography and radioimmunoassay. pgd and its metabolites pgj and delta- -pgj exhibited significant antiproliferative effects in the micromolar range in mast cells, mast cell lines, hl- and burkitt's lymphoma cell lines whereas inhibition of cyclooxygenase by indomethacin was without major effects. ltc and ltb had a small stimulatory effect on cell proliferation in hl- cells. degradation and possibly induction of cell differentiation may have attenuated the actions of leukotrienes. the leukotriene biosynthesis inhibitors aa- and mk- reduced proliferation of hl- and lymphoma cells significantly but had no major effects on mast cell growth. on the other hand, mast cells stimulated with calcium ionophore produced pgd and its metabolites, as well as ltb and ltc in significant amounts. from our data we conclude that prostaglandins and leukotrienes may play an important role in the control of cell proliferation. we compared the pattern of drug expenditures of several hospitals in (size: to beds). a, b are university hospitals in the ,,old"and c,d,e are university hospitals in the ,,new" german countries, f is a community based institution in an ,,old" german country. main data source were lists comprising all drags according to their expenditures in a rank order up to %. items were classified into i) pharmaceutical products including immunoglobulines, ii) blood and -derived products (cell concentrates, human albumin, clotting factors) and iii) contrast media (x-ray). with regard to group i) the highest expenditures nccured in hospitals a and b whereas drug costs in c -e were / less and came to only % in hospital f. the main groups of drugs which together account for > % of these expenditures are shown in the table. ) products were about % up to % of group i and highest in hospitals a, b and e, but about / lower in hospitals c and d. these results suggest meaningful differences in the drug utilization between the old and new countries as well as betv,,een university institutions and community based hospitals. however, although all hospitals provide oncology and traumatology services and all university hospitals offer ntx, differences in other subspecialities e.g bone marrow and liver transplantation and treatment of patients with haemophilia must be considered, too. dr.medsebastian harder, dept c]inicai pharmacology, university hospital frankfurt, theodor stern kai , frankfurt/main frg m. hgnicka, r. spahr, m. feelisch, and r. gerzer organic nitrates like glyceryl trinitrate (gtn) act as prodrugs and release nitric oxide (no), which corresponds to the endogenously produced endothelium-derived relaxing factor. in the vascular tissue, no induces relaxation of smooth muscle cells, whereas in platelets it shows an antiaggregatory effect. both activities are mainly mediated via stimulation of soluble guanylyl cyclase (sgc) by no. in contrast to compounds which release no spontaneously, a membrane-associated biotransformation step is thought to be required for no release from organic nitrates. glutathione-s-transferasea and cytochrome p- enzymes have been shown to metabolize organic nitrates in the liver, but little is known as to whether these enzymes are involved in the metabolic conversion of organic nitrates in the vasculature. furthermore, it is still unclear whether or not platelets are capable of metabolizing organic nitrates to no. we isolated the microsomal fraction of bovine aorta in order to characterize the activities towards organic nitrates using the guanylyl cyclase reaction as an indirect and the oxyhemoglobin-technique as a direct measure for no liberation. gtn was metabolized to no by the microsomal fraction under aerobic conditions already in the absence of added cofactors. this activity was not influenced by the cytochrome p- inhibitors cimetidine and metyrapone. in contrast, the glutathione s-transferase substrate -chloro- , -dinitrobenzene and the glutathione s-transferase inhibitors sulfobromophthalein and ethacrynic acid did not affect no release, but potently inhibited sgc activity. blocking of microsomal thiol-groups resulted in a decreased no release from gtn. homogenates of human plateles isolated by thrombapheresis and stabilized by addition of mm n-acetylcysteine did not show no-release from gtn as determined by the stimulation of the platelet sgc even after addition of the possible cosubstrates glutathione and nadph. these data demonstrate ( ) that bovine aortic microsomes exhibit an organic nitrate metabolizing and no-releasing activity whose properties are clearly different from classical cytochrome p- enzymes and from glutathione s-transferases, and ( ) that human platelets itself are not capable of bioactivating organic nitrates and therefore require organic nitrate metabolism in the vessel wall for antiaggregation to occur. bioavailability of acesal ®, acesal ® extra, micristin ® (all mg acetylsalicylic acid -asa), and miniasal ® ( mg asa), opw oranienbufg, relative to respective listed references was studied in female and male healthy volunteers (age - y, weight - kg, height - cm). asa and salicylic acid (sa) were measured using an hplc method validated from ng/ml to pg/ml. extent of absorption was assessed by auc (bioequivalence range . - . ), rate by cr~/auc (bioequivalence range . - . ). geometric means and %-confidence limits of the ratios test/reference (multiplicative model) are shown in the acesal ® and micdstin ® were bioequivalent in rate and extent of absorption with the reference formulations. the fast liberating acesal ® extra was bioequivalent with respect to extent only. asa from miniasal ® was absorbed more slowly than from an asa solution (cm= ( %-range): - ng/ml and - ng/ml; t~ (min-max): . - . h and . - . h). asa from micdstin ® and the corresponding reference was absorbed more slowly than from acesal ® and acesal ® extra. this was accompanied by decreased aucasa (increase of first pass metabolism) and increased apparent trrz (absorption being rate limiting). all ratios of aucsa/aucasa after administration of mg asa were markedly higher than after mg asa. thus, the formation of salicyludc acid from sa might be capacity limited at doses of mg asa. in the study >>physicians' assessment of internal practice-conditions and regional health-services-conditions in accordance with ambulatory patient-management<< a sampie of primary care physicians -comprising gps and internists -provide data for continuons analyses of arnbulatory health care quality and structure. focussing on the physicians' drug prescription, the impacts of reform law (gesundheitsstralcturgesctz, gsg) upon primary care providers and their therapeutic decisions were examined in . four different surveys were carded out during the year, dealing with frequent patients' reasons for encounter in gps' offices. after a pretest was carried out, physicians reported on patient-physician-encounters, basing on mailed questionnaires. for every therapeutic change patients received, the reasons for the change were recorded (e.g, reform law, medical indication) and above the physicians' expectations towards three criteria to measure the quality: ) physicians' assessment of the patients' satisfaction, ) adverse drug effects, ) therapeutic benefit. according to therapeutic changes due to reform law (drag budgets, blacklist) it can be stated: ) therapeutic changes due to reform law were carried out with relevant frequency. ) the reform law was of different concern regarding the different reasons for encounter we investigate& ) the impacts' strangth of the legal control mechanisms differed among several groups of physicians: those who already have been liable to recourse before more often carried out therapeutic changes according to fixed drug budget. different multivariate logistic regression-models yield an estimation of the odds-ratio of about . ) therapeutic changes in accordance with the reform law having been carried out at the beginning of the year more often suffered from negative expectations towards the therapeutic quality then changes during the actual encounter, e.g. >>joint pains<c % decrease of the therapeutic benefit in contrast to %. during the year a decrease of negative expectations is noticable. ) patients overestimating the severity of their own health problem in comparison to the physician more often suffered from negative expectations towards the therapeutic change than other patients, e.g. concerning the therapeutic benefit ( % in contrast to %). generally speaking, firstly the reform law led to insecurities -particularly in the introduction phase -among susceptible subgroups of physicians and patients, regarding the four investigated reasons for encounter. secondly, there is evidence of a loss of quality in ambulatory health care -predominandy in terms of subjective criteria. these impacts considerably subsided within the field phase duration from may to november, . the presented analyses give evidence of the practicability of studies assessing effects of legally-initiated mechanisms to control health costs upon primary health care quality. postansehrifi: iseg, bissendorfer sm , d- hannover the cellular and biochemical mechanisms that lead to the longerm complicatiorts of diabetes mellitus are poorly understood. glycation is a multireaetion beginning with a condensation reaction between reducing sugars and reactive amino groups of proteins. the accumulation of the glycated products of various proteins may contribute to the development of the complications of diabetes and aging. in order to find suitable non-toxic glycation inhibitors we used sulfur deriva-tives (penicillamine, eaptopril and alpha-lipoic acid) and amino derivatives (aminoguani-dine and aminoguanidine propionic acid) and postulated the mechanisms involved.the inhibitou effect of the drags on the early (amadori product) and advanced glycation products (age). determined by affinity chromatography, deglycation method. i-iplc and fluorescence spectroscopy, differs markedly depending on the drag used and is greatest (up to %) with aminoguauidine and penicillamine ( mm) and smallest with alpha-lipoic acid. aminoguanidine propionic acid had no inhibitory effect on the glycation process. the ability of penicillamine to inhibit both the formation of amadori products and age was demonstrated by isolation of the adducts of penicillamine with ribose, glucose, glucosa-mine and p-tolylglucosamine, a standard for glycated protein. their structures were characterized using ir-and h-nmr and c-nmr spectroscopy. the oxidative deg-radation of glycated human serum albun{in in the presence of copper (ii), forming the carbox jmethyllysine, was inltibited by the sulfur derivatives. the in vitro formation of pentosidine, a crossliuking product was influenced by all drugs used however captopril, alpha-lipoic acid and aminoguanidine propionic acid do not condense with sugar derivatives.the mechanism by which captopril and alpha-lipoic acid inhibit the formation of age may be involve radical formation, where a spontaneus thiol/disul-fide interchange between drug and protein takes place. complications can appear at any time in diabetic patients.the cause of these complica-lions in some cases may be due glycation and/or oxidation. the clinical applications of substances which are able to prevent these events may have advantages. these results and the mechanism proposed show that the amino and sulfl~ydryl groups of drags reduce glycation and postglycation of proteins and thus have potential value in this regard. non-enzymatic glycation starts by covalent binding of single sugars to o~ or e amino groups of proteins. in subsequent reactions a number of early glycation products like schiff bases or amadori products, and of advanced glycation endproducts (ages) is formed. glycation is increased in a variety of proteins in diabetic patients [schleicher e ( ) diabetes und stoffwechsel : ] including human serum albumin (hsa), the red cell membrane, hemoglobin, collagen, laminin, immunoglobulin g, low-and high-density lipoproteins. the process is supposed to be a key mechanism in the pathogenesis of diabetic complications and in tisstie ageing [vlassara het al. ( ) lab invest : ] . inhibition of glycation may therefore be important for the prevention of late diabetic complications. we investigated the in vitro glycation of human serum albumin (hsa) and its inhibition by aminoguanidine, penicillamine, captopril and c~-lipoic acid ( ram). hsa ( g/l) was glycated by incubation with mm glucose in . m phosphate buffered saline, ph . , at °c for days under sterile conditions in the presence of . % natrium azide. the glycation rate was determined by two different methods: deglyeation -measures amadori products based on the colorimetry of -ketoglucose which is released from the glycated protein (ketoamine) the effect of aniseylated #asminogen streptokinase activator complex (apsac) on left ventricuiar (ll r) function ;,.,as investigated in ! patien~ with acute myocardial infarction (am!). a dose of q units apsac was given intraveneusiy to patien~ with ami within hours after onset of pain, patients underwent cardiac angiegraphy at minutes and at day . wail motion ,.~s measured by the global ejection fraction and by the centefline method in :'=he infarct and neninfar~ regions and expressed in units of sf~ndard deviations beiew normai, as defined in normai patient. group a with successfui ',th, rombo!ysis (timi and ) was compared with group b without successfui thrombolysis (timi and !). at minutes there vras no significant difference in giebal ejection fraction between both groups, by using the regiona! wa[i motion by the centerline method group a showed a significant improvement of lv function (p ~ . ). the benefit of successeli thrombeiysis was more marked in the peripheral infarct region than in the centrai infarct region. thus early repeffusion is an important factor in maintaining good lv function after ami. the re=gionai wai! motion in the infarct region seems to be more sensmve man me g!oba! eiection fraction to show ear!y differences in lv function after ami. the quinoline derivative bay x ((r)- -[ -(quinoliu- -yl-metho.,qc)phenyl]- cyclopentyl acetic acid) is a selective inhibitor of ltb -biosynthesis in various in vitro and in vivo models. it acts via binding to the integral nuclear membrane protein five-lipoxygenase activating protein (flap), which has been shown to be an essential component of the cellular leukotriene synthesis machinary. the function of flap is only poorly understood, but recently it has been shown that a radiolabeled photoaffinity-analog of arachidonic acid binds specifically to flap. consequently, a role of flap in the substrate transfer to -lipoxygenase ( -lox) has been suggested. using intact and fractionated human polymorphonuclear leukocytes (pmnl), we present data showin~ that exogenously added arachidonie acid diminishes dose-dependently binding of [ c]-bay x to its target. this effect correlates with the reduction of bay x mediated inhibition of ltb synthesis after stimulation of intact pmnl with tile calcium ionophore a in the presence of increasing extracellular arachidonic acid concentrations in addition, enhancing the release of intracellular arachidonic acid from phospholipids after activation of phospholipase a (pla ) with increasing a concentrations or after prolonged incubation of the cells with the ionophore results in a loss of bay x potency. to further study" the effect of arachidonic acid on the bay x mediated ltb synthesis inhibition, human pmnl were stimulated with paf, c a or fmlp in the presence of exogenous arachidouic acid. under such conditions, ltb synthesis is significantly enhanced in response to paf, c a or fmlp. which are only weak activators of cell-associated pea the effect of extracellular arachidonic acid is dosedependently inhibited by bay x . coincubation of the cells with arachidonic acid and paf, c a or fmlp, respectively, enhances synergistically the translocation of -lox from the cnosol to the membrane fraction. again, this effect is inhibited by bay x . in summary, using bay x , we provide evidence, that flap is involved in the utilization of the -lox substrate araehidouic acid from exogenous or endogenous sources for the agonist-induced leukotrieue biosynthesis. switzerland, but their effect on the expression and function of the multiple drug resistance gene product has not been studied in detail. we have therefore examined the effect of gm-csf and g-csf on the uptake and toxicty of vincristine with and without the chemosensitizers r-verapamil and norverapamil in the pglycoprotein-containing vincristine-resistant cell-line, hl- /vi,o. following a h pre-incubation with ng/ml g-csf or gm-csf, cells were washed and reincubated in rpmi medium with hvincristine and the modulators for hours. the uptake of hvincristine was increased up to -fold and this effect was equally marked when the vincristine-uptake was stimulated with the chemosensitizers in the range of . pg/ml to pg/ml. although cytokine-pretreatment increased the accumulation of vincristine, we could not show an increase in vincristine toxicity over h with the same concentrations of vincristine and chemosensitizers. it is concluded that the effects of cytokines on vincristine accumulation are compensated by a positive effect on cell-survival or cell-growth. department of clinical pharmacology, university clinic, frankfurt, theodor stern kai , frankfurt, germany. from the economic point of view people often tend only to regard the price of a product. savings for the health care system as a whole which could be gained by use of these drugs and which could more than compensate the high costs of medication are not considered. the cost-benefit-analysis enables a comprehensive analysis of the economic effectiveness of a pharmaceutical. the relevant cost factors such as the costs for in-and out-patient treatment depending on the therapeutic success and the inability to work are compared to the costs of medications. using the example of stroke prevention, it was possible to show that prophylaxis with an expensive but effective pharmaceutical is economically better than standard medication. according to the tass study ticlopidine can prevent at least more strokes per i treated patients than ass. on the basis of this study cost savings of up to dm , per i treated patients can be reached including higher costs of ticlopidine. projected onto i , patients up to dm , per year can be saved for the german health care system by the use of ticlopidine in secondary prophylaxis after tia. the naphthodianthrones hypericin (h) and pseudohypericin (ph) were proved to be elfectire in mild depressive disorders. recently they were discovered to exhibit high in vitro activity against enveloped viruses fostering clinical trials in infections with hiv or herpes viruses, for studying the pharmacokinetics of h and ph a solidified extract from st. john's wort (hypericum pefforatum l) was orally administered to healthy male volunteers in three single doses of , , and ~g of h and , , and p.g of ph, followed by a -days course of multiple dosing of p.g h and pg of ph tds. plasma concentrations were determined by hplc, the quantification limit was . #g/i (staffeldl et al., nervenheilkunde ( ) : ). single,dose kinetics, h showed a median lagtime of . (range: , - . ) h, which was significantly longer compared to , (range: . - , ) h of ph. h was eliminated with a median half-life of (range: - ) h, and in case of ph the hail-life ranged between . and (median: , ) h. median crux levels were . , . , . ~g/i for h and . , . , . pg/i for ph for the three single-dose levels given above, respectively. lowest dose crre, x and auc of h was lower then expected from medium and highesl dose, which was statistically significantly for crnax (p= . , anova; p= . , onferroni/dunn), indicating nonproportionality of pharmacokinetic response may exist in the lower dosing range. mul!iple-dose kinetics. during long-term dosing steady-state trough levels (ctreugh) of . (range: . - . ) ~g/i for h and . (range: -to. ) ~g/i for ph were reached after days for both. cmax after days of treatment were . (range: . c. kori-lindner, r. eberhardt the healthcare restructure act ( ) changed the evaluation of drugs from efficacy to cost effectiveness: therapeutic necessity, utility, purpose; medical progress and economic utility have to be considered. criterias and suroundings of pharmacooeconomic evaluations and studies according to the bealthcare restructure act are demonstrated by an example: "costs of nsaid-induced gastro-duodenal ulcers and prevention with miso~rostol". methods: cost-utility, cost-benefit, legal requirements, decision-model and decision -tree, basic calculation and the transfer to the socialsystem of the frg. premissions: rate of ulcers without prevention , %, with prevention , %; rate of compliance %, rate of hospitalisation % and duration of hospitalisation , days. rheumatic disorders have become a frequent reason for services m ambulatory care. therefore the use of drugs which are administered in rheumatic complaints is widespread. data of the sientific institute of statutory health insurances (arzneimittelindex ) demonstrate that the group of antirheumatics/analgetics held the first rank due to the number of prescriptions. with regard to the costs thts drug class were amongst the top three. apart from anurheumatics and analgetics different other drug classes are used for the treatment of rheumatic disorders for example corticoids. to estimate the total expenditure for drug utilization in out-patients suffer from musculoskeletal disorders data of the statutory health insurance dortmund (aok dortmund) were investigated. a % random sample of insurees (n= ) was analysed for diagnoses and drug therapy. the data were collected in a individual-related way for an observation period of one year ( ). the results demonstrate the high frequency of rheumatic disorders among out-patients: the -month prevalence of musculoskeletal disorders was about % in the studied population of whom one half suffered from back pain. approximately % of the patients were prescribed any drugs for treatment of rheumatic disorders. as might be expected nonsteroidal antiinflammatory drugs were used most frequently. the extensive use of topical products was a considerable factor of costs. more than one half of the patients received drugs for topical application. further details will be presented. diachronous data of statutory health insurences that have been gathered in a person-related way permit the estimation of drug expenditure in outpatients with regard to the several musculoskeletal disorders. misleading results of endocrine tests in patients suffering from liver diseases promted us to investigate the influence of liver function on the pharmacokinetics of dexamethasone (dex). it was the aim of this study to find out whether routine parameters of liver function may be usefull in roughly predicting the clearance of dexamethasone. methods : in patients liver function was rated on the basis of their plasma cholinesterase activity (che) and prothrombin time (pt). blood samples were withdrawn , , , , , , , and min following an iv bolus injection of mg of dexamethasone- -dihydrogenphosphate corresponding to , mg of dex for the determination of dex pharmacokinetics by ria. results : computer assisted best fit analysis ( compartment model) of the plasma dex decay did not show any influence of liver function on the distribution half-life (t½-a) and on the initial volume of distribution of dex, but a % increase in the terminal phase, "metabolic" half-life (t½-b) with decreasing liver function. while the t½-b increased from ± min in patients with a che activity > . ku/l to ± min in those with a che s . ku/l, the metabolic clearance rate (mcr) decreased from ± ml/min to iii ± ml/min. in patients on phenytoin the t½-b was reduced to <i min. che activity (r=- . ) and pt (r=- . ) were significantly (p< . ) correlated with the t½-b of dex. we conclude that the dosage of dexamethasone given for diagnostic or therapeutic purpose needs to be adjusted in patients on phenytoin therapy and in patients with liver disease causing a che s ku/l. etofenamate is one of the most frequently applied topical nsaids in germany. it can be regarded as a lipophilic prodrug of flufenamic acid, which is liberated after transdermal absorption and cleavage of the ester. we compared several topical nsaids in patients suffering from osteoarthrosis or rheumatoid arthritis with concomitant knee joint effusions. the products contained between and % of etofenamate and were administered according to the recommended dosage schedule. following multiple applications of defined doses we obtained synovial fluid and plasma specimen in order to measure the concentrations and to compare the bioavailabilily of several products. the analysis of synovial fluid and plasma samples was achieved by hplc. following the addition of other n-phenylanthranilic acid derivatives as internal standards the samples were extracted twice with a -fold volume of acetonitrile/methanol ( : ). the diluted extracts were concentrated with the use of preconditioned ods-columns and subsequently eluted with methanol. the separation of the products was perlormed on a reversed phase column with an elution medium of methanol/water/acetic acid ( : . : . ). the highest synovial fluid and plasma concentrations were measured after hours following the final administration of etofenamate. the concentrations reached average maximum values of . ng/ml of flufenanamic acid in plasma and . ng/ml in synovial fluid. the bioavailability could be estimated from the area under the concentration time interval between . and hours, during which samples were obtained. they served for the comparison of the topical products. from in vitro data of cyclooxygenase or lipoxygenase inhibition with n-phenylanthranilic acid derivatives it can be concluded, that the synovial fluid levels of etofenamate and its metabolites are unsufficient to produce effects relying on this mechanism. it can not be excluded, however, that other effects contribute to the clinical improvements observed by other investigators. in animal studies it has been shown to have excellent anti-arthritic efficacy and an improved tolerability profile over other nsaids. early preclinical studies have shown that mel preferentially inhibits cyclooxygenase- (cox- ). the aims of this three-week, parallel, multicentre trial were to assess the efficacy anct safety of two different doses of mel in comparison to placebo (pla) in patients with active ra. the specific question was, whether cox- inhibition relates to therapeutic efficacy and safety in the clinical'setting. methods: patients ( f, m, mean age years) were randomised (mel . mg= , mel mg = , pla= ) and analysed by an intention-to-treat analysis. all trial drugs were given orally once daily. results: patients ( %) in the pla group, patients ( %) in the . mg mel group and patients ( %) in the mg mel group discontinued the trial prematurely due to lack of efficacy or adverse events. mg mel was significantly superior (p< . ) to pla in of the primary endpoints: disease activity assessed by the investigator and by the patient and reduction of the tender/painful joint count. the difference between . mg mel and pla reached statistical significance in of the primary endpoints: disease activity assessed by the patient and reduction of the tender/painful joint count. in none of the primary endpoints a statistically significant difference between mg and . mg mel occured. global tolerance assessed by the patient and investigator at the last trial visit was similar in all three treatment groups. adverse events were slightly more frequent with meloxicam than with placebo. in none of the patients gastric or duodenal ulcers, perforations or bleeds were observed. conclusion: the efficacy of meloxicam . mg and mg per day was significantly superior to placebo in the treatment of patients with ra. there were no relevant differences regarding safety parameters between the meloxicam doses and placebo. the preferential cox- inhibition may be the reason for the favorable efficacy and safety profile of meloxicam. the major metabolites of morphine are morphine - -b-d-glucuronide (m g) and morphine - - -d-glucuronide (m g). in humans plasma aucs of m g and m g exceed those of morphine depending on form and route of application. m g and m g penetrate the blood-brain-barrier due to an unexpectedly high lipophilicity, m g is an antagonist able to provoke with&awl symptoms even in opioid naive animals. m g is an agonist markedly more potent than morphine. methods: pig brains were homogenized and the homogenate centrifuged to give a microsomal preparation free of nuclei and coarse debris. aliquots were incubated at °c in buffer (ph . ) containing . nm [ h]damgo. results: the ic for morphine, m g and m g were . nm, nm and . pm, respectively. m g and m g, respectively, shifted the ic for morphine markedly to the right. discussion: our results on the single substances confirm data from the literature indicating that m g and m g inhibit binding to the ~t-or less potent than morphine. from the fact that m g and m g shift the ic for morphine to the right we infer binding of m g and m g to the morphine binding site at the ~t-or. bypassing activity and producing cost of dm . ,-was not included in the above shown results. • blood products, coagulation factors and anttmierobial agents are the most expensive substances, however their use could be more restricted. • chemotherapy represents also a high-cost factor, but alternatives do not exist. [ntensiv care units are the most expensive wards consuming - % of the hospital budget, drug use reviews on sicu's of the university hospitals frankfurt ( beds) and jena ( beds) were compared after recording the indication related administration of pharmaceuticals and blood products over a months investigation period using notebook-pc's. all substances were included, wich produced % and % of total pharmacentical cost in the year before, respectively. in frankfurt and jena and patients were registered, causing total cost of dm . ,-and dm . . ,-and causing cost per patient and day of average dm ,and dm ,-, respectively. in jena most expensive diagnoses were severe trauma ( % of total cost; n= ), gastre-intestinal bleeding ( . %; n= ) and malignant growth ( . %; n= ). in frankfiart nigh cost diagnoses were liver transplantation ( % of total cost; n= l), severe trauma ( , %; n= ) and coronary bypass craring ( . %" n= ). although the patient populations were different, the percentage cost of the substance groups on total drug cost were the same in both sicu's: blced-products ± %, antibiotic and chemotherapeutics io, %, immuna-ghibulines , :t , % anyway, inside these groups drug usage patterns differed distinctly. for example, in jena the nigh-cost antibiotics were tobramycin, cefuroxim, amoxicillin and ambisome ® causing . %, . %, . % and % of total antibiotic cost, respectively. corresponding data for franlffurt are ambisome* . %, imipedem . %, cefotaxim , % and cefamandol . %. severe sepsis was the most abundant complication occuring in beth sicu's in patients and producing cost of dm . ,-in frankfurt and dm . ,-in jena. in sepsis blood products were the most expensive substances with . % (frankfurt) and . % (jena) followed by antibiotics with . % and . % and immuneghibulines with . % and . %, respectively conclusion: * recording the indication related drug administration drug usage patterns may be identified and targets for economical intervention may be found. • blood and blood-products must be included because they produce about % of total cost. • complications, especially sepsis, produce nigh cost, which should be included in the discussion with the health assurances concerning the payment system using diagnosis related groups. in a drug-utilisafion-review-group (durg) was founded by clinical pharmacologists. the task of this group is to identify drug usage patterns, to record indication related drug administration including cost and to analyse these data. as a result recommendations concerning rational and economical use of pharmaceuticals and blood-products were given (e.g. ht -receptor-antagonists, human albumin, g-csf the most expensive substances are blood-products ( %), antimicrobial agents ( %) and immunoglobulines ( %). in conclusion, at the universitiy hospital frankfurvmain the activities of the drug-utilisation-review-group seem to be successful and are at least inpart responsible for the reduction of drug-cost of more than dm mio last year. the aim of this study was to determine the extent and the time course of fenoterol induced tachycardia and hypokalemia during and after a continuous intravenous infusion with three different rates within the therapeutic range for tocolysis. furthermore we compared the point estimators and precision of parameters obtained with a standard two step approach and a population model. according to a randomized double blind multiple cross over design, healthy females received an intravenous infusion of either placebo or fenoterol ( . ijg/min, ijg/min, ljg/min) over hours. the plasma concentration of fenoterol, heart rate, blood pressure, plasma potassium and plasma glucose were monitored during and up to hours after the infusion. the plasma concentrations of fenoterol and potassium and the heart rate were fitted to an integrated pkpd model. both individual parameter estimates and population estimates were obtained with nonmem. both effects correlated best with the concentration in the central compartment. the relationship between heart rate and concentration was adequately described by an emax model. modelling of the dose dependent hypokalemic effect of fenoterol required an negative sigmoid emax model with an additional term allowing for passive diffusion of potassium from the intra-to the extracellular compartment. during the observation period, the concentration response relationship remained stable for both effects. as to be expected, the estimates of the population parameters were much more stable than those of the individual patients. it is concluded that the population method might be superior to the sts approach even in a data rich situation. volunteers received different controlled release formulations (a, b, c, containing mg is- -m ", ) in a randomised cross-over design. plasma samples were taken from to hours after application. one week wash-out phase was kept between investigations. is- -mn were determined in plasma using a gc-method. in vitro dissolution of a, ] , c was investigated by using the rotating bottle method according nf xiv. mean absorption time (mat) was calculated as the difference of mrt and ke - , the mean dissolution time (mtdiss) was calculated as proposed by brockmeier ( ). additionally half value duration (hvd) and plateau time opt) are reported. the difference of the dissolution profiles between the product is not reflected in the parameters mat, hvd and pt. although preparation c has the longest mtdiss it has the largest cmax value and shortest hvd. the origin of this apparent discrepancy between in vivo/in vitro profiles deserves further elucidation ( ) acta pharm. technol. ( ) calciumantagonists (ca) are lipophilic molecules which, however, have no common structural similarities. nevertheless, ca are characterized by the ability to attach to various receptors and binding places in cardiac and smooth muscle preventing calcium overload, necrosis and cellular death. in order to study the physiological effects of ca on cerebral cortex of normal and vit d calcified rats, i.v. infusion of ca have been applied. cerebral blood flow (via hydrogen clearance) and oxygen supply (via surface oxygen tension) have been measured in normal and sclerotic albino rats in order to differentiate cerebral and systemic effects. miniaturized surface oxygen and hydrogen electrodes were placed on top of the exposed cerebral cortex of ketamine~xylazine anesthetized rats. nifedipen, vcrapamil, diltiazem and flunarizine induced hypotension in normal and monckebcrg type sclerotic rats. nifedipen and verapamil showed comparable improvements in cbf and surface po , the initial hypotension recovered completely under nifedipen but not under verapamil. compared with vcrapamil, the hypotension was even more pronounced under flunarizine ( : %). nevertheless, in response to both ca cbf was improved to about %. however, diltiazem did not increase cbf due to hypotension. sclerotic rats showed a map-decrease from to mm hg, the po histogram was left-shifted. diltiazem did not improve the cortical supply of sclerotic animals. the influence of intracoronary enoximone at a dose of . mg/kg/lo min on global and regional wall motion, myocardial perfusion (contrast echo) and diastolic lv function during pacing-induced ischemia was investigated in pts. with significant lad stenoses. results: enoximone did not change either heart rate at "rest ( ± vs ± min -i) or blood pressure (lvsp: ± vs ± mmhg) . in the postpacing ischemic period, lvedp fell from a mean of . to . mmhg after enoximone (p< . ), dp/dtmax increased from to mmhg/s (p< . ) and regional ef from % to % (p< . ) with global ef remaining unchanged ( % vs %). st segment depression was reduced significantly from . to . mm (p< . ). enoximone induced an increase in myocardial perfusion by % (p< . ) in the stenosis-depending myocardial areas with shortening of the wash-out half-time of the echo contrast medium from a mean of to s (p< . ) . the isovolumetric relaxation (echo) was shortened by % (p< . ), the e wave by %, and dp/dtmin was increased by % (p< . treated with pivaloyl-pivaloyl-substituted - actamantibiotics belong to a new group of prodrugs with an improved bioavailability following oral administration. after metabolic cleavage the pivalic acid and the free - actam are eliminated separately. pivalic acid is subjected to enzymatic coupling with endogenous camitine leading to carnitinepivaloylesler, which can be found in large amounts in the urine. free camitine is reduced under the treatment which may lead to neurological side effects in patients with metabolic predisposition. we showed, that the pattern of other acylcarnitineesters excreted with the urine was not changed significantly after administration of a pivaloyl.substituted cephalosporine in healthy subjects. in order to compare the pattern of excreted carnitineesters we analyzed urine samples of pediatric patients treated with cefetamet-pivoxil. the analysis of carnitineesters in the urine was achieved by hplc following precolumn derivatization. as an internal standrad we used undecanoyl-carnitine aster, which does not occur in human plasma or urine. after the extraction using ion exchange columns the samples were incubated with p-bromo-phenacyl reagent. the mixture of derivatives was separated on a reversed phase column applying a gradient of different buffers. with this method the whole spectrum of carnitineesters can be quantitatively analyzed. the percentage of the pivalic acid ester eliminated with the urine in pediatric patients did not differ from that observed in healthy subjects, therefore an increased risk in this patient group cannot be inferred. institut for klinische pharmakologie, universit~.tsklinikum charit& humboldt-universit&t zu berlin, schumannstral~e , d- berlin; pharmacokinetics and bioavailability of bupranolol-l"rs and oral bupranolol in patients with impaired liver function compared to healthy volunteers k.-h. molz*, w. cawello**, g. haug-pihale*", r. bonn** w. hammes**, a. weil*** and g. popescu*** bupranolol is nearly completely absorbed after oral administration, but more than % of the parent compound are metabolized by first pass through the liver. it can be expected that transdermal application of the substance will bypass the high first pass metabolism resulting in higher plasma concentrations or lower therapeutic doses, respectively. a reduced capacity of metabolizing liver enzymes might mimic this effect. the study was performed to investigate if metabolisation by hepatic enzyme systems results in liver function dependent pharmacokinetics. special interest was to be given to possible differences between oral and transdermal application. in an open sequential single dose-study including patients with impaired liver function and age/sex/weight-matched healthy volunteers a bupranolol transdermal therapeutic system (containing mg/ h) and bupranolol tablets ( mg) were administered to each participant separated by a washout period of a week. phenotyping for mephenytoin and sparteine was done to exclude poor metabolizers. blood samples for the quantification of bupranolol plasma concentrations were drawn before and , , , , , , , , (before patch removal), , , , , , and hours after patch application as well as before tablet administration and . , , . , , , , , , , , , , , , and hours post dose. urine samples were collected in ('its) and (tablets) intervals after dosing. blood pressure and heart rate were measured before and , , , and hours after dosing, ecg's were recorded before and , and hours after dosing. the pharmacokinetic results indicate only minor differences between patients and healthy controls after patch applications. for example the main parameter auc( -~) were . ± . ng/ml.h (mean±sd) in healthy controls and . ± . ng/ml.h in patients. completely other situation was given in comparison of data from oral administration: auc( -=)decreased dramatically to . ± . ng/ml-h in healthy controls and increased to . ± . ng/ml.h in patients. vascular complications are a common fmding in patients with diabetes mellitus. it has been reported that elevated glucose impairs acetylcholine-but not nitruprusside-mediated relaxation of rabbit aorta in vitro suggesting an abnormal endothelium-mediated vasodilator mechanism (cohen, circulation ; (suppl. v): ). the purpose of the study was to investigate whether this phenomenon . implies general impairment of the nitric oxidemediated relaxation (or only specific interference with muscarinic receptors), and . can be also demonstrated for coronary arteries. porcine coronary arteries were cut in rings and were set up in organ baths for isometric tension recordings. rings were randomized and incubated with different glucose concentrations ( . , , , or . mm) for hours according to cohen and coworkers. rings were preconstricted with prostaglandin f ~t ( . - ~m); the nitric oxidemediated relaxation was induced by either substance p ( . - nm), bradykinin ( . - nm), or calciumionophor a ( - nm), the endothelium-independent relaxation by nitroglycerin ( - nm). in addition, the effects of glucose were studied on acetylcholine-( . - ~tm) and serotonin-( . - ~m) induced vascular responses. results: surprisingly, elevated glucose concentrations ( and mm) had neither an effect on the potency nor the efficacy of any of the agonists tested. it is concluded that this previously reported impairment of endothelium-mediated vasodilation of rabbit aorta by elevated glucose concentrations may not be important in other species or types of arteries. nitric oxide production may be induced by cytokine-mediated immune responses. we tested the hypothesis that graft rejection in cardiac transplanted patients is associated with an increased production of nitric oxide. the study group consisted of patients in whom one or several myocardial biopsies (total biopsies) were obtained to assess for graft rejection. at the day of biopsy, urine was collected for a time period of at least hours ( % isopropanol was added for antimicrobial activity). nitrate was measuerd in urine samples by gaschromatography. nitrate excretion was expressed as the quotient of nitrate and creatinine content (gmol nitrate/mmol creatinine). since nitric oxide is oxidized in biologic systems to nitrite and finally nitrate, it was assumed that nitrate excretion reflects total nitric oxide release. graft rejection was classified according to the hannover classification system (a /i= no/minor rejection, a -a =moderate/severe rejection). results: nitrate excretion showed a large interindividual variation ( to i.tmol nitrate/mmol creatinine). /£ tough not significant, the nitrate excretion tended to increase with the degree of rejection (a : + ; a : + ; a : + ; a / : + ; mean + se). in a subgroup of patients (n= ) who underwent several biopsies with and without rejection, moderate/severe rejection was significantly associated with increased nitrate excretion (a /ai: + vs. a -a : + , p< . ). conclusion: the urinary nitrate excretion is increased in patients with moderate/severe graft rejection. because of large variation in excretion, this marker may be not suitable for rejection diagnosis. the combination of thremboxane a~/prostaglandin endoperoxide (epo) receptor antagonism with thromboxane synthetase inhibition (trasi) is designed to both abolish the epos' prothrombotic role and direct epos from activated platelets to adjacent cells for the generation of antithrombotic prostaglandins. thus, we have investigated the effect of the potent trasi dt-tx on thrombus formation in models of vascular injury (of the media versus selective deendothelialization) in man ex vivo. healthy volunteers were treated with an oral dose of vehicle (placebo; n= ) or , , , mg; n= per dose). blood was taken before and , , and h after the ingestion, anticoagulated and allowed to flow over a thrombogenic cell-free subendothelial matrix (cfm) or a layer of human venous smooth muscle cells (smc) for min. the "en face" area of platelets/thrombi deposited was determined by microscopic morphometry. in the absence of cells of the vessel wall (cfm) the thrombus formation was inhibited by + % (+sd) after mg, +_ % after mg, + % after mg and +p % after mg dt-tx versus + % after placebo. the subclass of the large thrombi (= % of all platelets/thrombi but incorporating > % of the platelet mass) was much stronger affected by the dt-tx treatment: the mean area was reduced by +p % after rag, + % after mg, _+ % after mg, + % after mg and _+ % after mg dt-tx versus - + % after placebo. in the presence of cells of the vessel wall (smc) the overall thrombus formation was reduced by up to + % after only mg, + % after mg, +_ % after mg, _+ % after rag and -+ % after mg dt-tx versus +_ % after placebo. dt-tx , a molecule combining potent and specific th romboxane synthetase inhibition with prostaglandin endoperoxide/thromboxane a receptor antagonism, has been examined in healthy male subjects. collagen-induced platelet aggregation in platelet rich plasma prepared from venous blood was measu red photometrically before and up to hours after a single oral dose of , , , or mg dt-tx in a placebo-controlled, double-blind study. platelet aggregation was induced in the ex vivo samples by collagen in concentrations between . and p.g/ml to evaluate platelet aggregation in relation to the strength of the proaggregatory stimulus. the ecs , i.e. the concentration of collagen required for a half-maximal aggregatory response (defined as the maximal change of the optical density), was determined. in the placebo-treated control group, the mean ecso was + ng/ml collagen (+ se; n= ) before treatment. it then varied between + and +_ ng/ml collagen after treatment. the'ratio of the post-to the individual pre-treatment ecso values was . _+ . (n= ) at . h, . _+ . at lh, . _+ . at h, . - . at h, . + . at h and . + . at h. this indicates that the sensitivity of the platelets to collagen was not affected by the placebo treatment. oral treatment with dt-tx , however, strongly inhibited the aggregatory response of the platelets to collagen stimulation. the ecs -ratio was increased to a maximum of . the detection of endogenous opioids suggested the opinion that in case of the presence in the organism of a receptor for an exogenous substance there is probably a similar endogenous substance.the occurrence in the blood of persons, who were not treated with cardiac glycosides, of endogenous digoxin-like or ouabain-like [actors confirms that opinion. in our study we took up the research of other drug-like [actors in the blood serum of healthy people. in two hundered and twenty-five healthy volunteers (llom,ll f) non-smokers not receiving any treatment before or during the test and aged between ib and y(mean age y) the occurrence of drug-like [actors in blood serum was studied.the examinations were carried out with the use of the fluorescence-polarization-immunoassay (fpia)-tdabbott. th e presence of the following endogenous drug-like foctors in the blood serum was evaluated: quinidine,phenytoin, earbamazepine,theophylline, cyclosporineand gentamicin. the presence of endogenous phenytoin-like, theophyllinelike and cyclosporine-like [actors has been demonstrated. the drug-like [actors were not found in the case of quinidine ,carbamazepine and gentamicin. the phenytoin-like factor was found in , ~, theophylline-like [actor , ~ and cyclosporine-like [actor in , ~ of examined volunteers.the mean value of the drug-like [actors were as follow : phenytoin , ~ , pg/ml,theophylline , ~ o,ll pg/ml and cyclosporine , z , ng/ml. the supposition may be proponued that organism produces drug-like substances according to its needs. the acetylation and oxidation phenotypes were studied in healthy volunteers ( m, [) aged between ib and years (mean y) in the wielkopolska region in poland. the acetylation phenotype was studied with the use of sulphadimidine which was given in a dose of mg/kg b.w. per os.sulphadimidine was determined by a spectrophotometric method.the border value of m.r. was ~ in urine. the oxidation phenotype was studied with the use of sparteine which was given in a dose of , mg/kg b.w.per de. sparteine was determined by the gas chromatographic method in urine. if m~ was <hr these are values for intensive metabolism, if mr was , <mr( these are values for poor metabolism. the determinated phenotypes were bimodally distributed. our studies have short that healthy volunteers ( . ~) were slow and ( , ~) were fast metabolizers. intensive oxidation phenotype could be ascertained in volunteers ( , ~) and poor oxidation phenotype in volunteers ( , ~). these findings are comparable with the values typical for the caucasian population. the genes coding for rapid acety!ation were present in . % of the cases, genes coding for slow acetylation were detected in . %. the frequency of specific nat alleles were: r , . %; r , . %; sla, . %; slb, . %; slc, . %; s , . %; s , . %; and s , . %. there was no mutation found at nt. in cases ( . %) we have stated nat genoand phenotype concordance, but in cases ( . %), genotype and phenotype deviated from each other. seven individuals were genotyped as slas , one as slasle, and one as slas but were phenotypically rapid acetylators. one child was genotyped as r s a but proved a slow acetylator. we can confirm that in about % nat genotyping allows to properly predict nat phenotype. discrepancies may derive, i. a., from yet unknown mutations which we try to detect by dna-sequencing. several studies demonstrated that cardiac glycosides reduce diameter of epicardial coronary arteries. in patients (ffs) with coronary artery disease (cad) this vasoconstrictive side effect of glycosides may cause ischemic complications in the presence ot high-grade stenoses. this study evaluated quantitatively the effect of an i.v. injection of digoxin on diameter of epicardial coronary arteries in pts with cad. eleven grs (group ) were treated with aspirin mg/d, nine frs (group ) were treated with aspirin and rag nisoldipin hours prior to angiography. coronary angiograms were taken in identical projections before ( ) and , , and minutes after i.v. injection of . mg digoxin, at rain .ling of nitroglycerin (ntg) was given. the mean diameters of normal segments and the minimum diameters of stenoses were analysed with cms (medis, leiden). a total of normal and stenotic segments were analysed. figure shows the change of coronary diameter (% predrug-vahie; mean + se) for group i (left) and ii (right): ( -( -cyclohexylureido-phenoxy)- -hydroxy- -tert.-butylaminopropan; awd, dresden) is frequently used as a cardioselective g-adrenoceptor antagonist. talinolol metabolites in urine were detected by hplc and gc/ms with previously characterized reference compounds and were quantified by hplc with a normal phase silica column (oertel et al., br. j. clin. pharmacol. ( ) ) . less than one per cent of administered dose was recovered in urine as bydroxylated talinolol. other metabolites could be excluded. in serum after therapeutic doses talinolol metabolites were not traceable. however, in a suicidal patient with a parent compound serum concentration of ng/ml (manifold higher as after usual therapeutic dose) two hydroxy metabolites at concentrations near the detection limit ( ng/ml) were found. obviously only a small amount of the talinolol dose is metabolized in human. in this study the relation between structure, polarity and urinary recovery of the metabolites was examined. four different isomers of mono-hydroxylated talinolol were identified in urine of volunteers after a single dose of talinolol. eight isomers are possible: six hydroxy-cyclohexyl-compounds and two hydroxy-phenoxy-compounds. there was a low degree of specificity of biotransformation seen with respect to the point of attack on the cyelohexyl ring. hydroxylation of the phenyl ring could be excluded in human. using hplc methods very small quantities of reference compounds are sufficient to estimate and compare the polarity and hydrophily of similar compounds. long retention times in a normal phase system and short retention times in a reversed phase (rp) system are found for polar and hydropbilic compounds. the shortest retention times of all hydroxylated talinolol compounds in the normal phase system as well as the longest retention times in the rp system were found for the main metabolites -trans-and -cis-hydroxycyclohexyl-talinolol. accordingly, the main metabolites are the most polar and hydrophilic mono-hydroxylated talinolol isomers. lysine clonixinate is a non-steroidal antiinflarnmatory drug with marked analgesic effects. it is used for treatment of acute pain such as postoperative, dental or menstrual pain. we have determined the biosynthesis of thromboxane (tx)b z and prostaglandin (pg)f ~ in clotting whole blood ex vivo as well as collagen-induced platelet aggregation before and . , . (only txb and pgf ,~ biosynthesis), , . , , and hours after oral intake of mg lysine clonixinate. the results were compared with data obtained after oral intake of mg acetylsalicylic acid (asa). while both txb and pgf ,~ biosynthesis measured radioimmunologically were inhibited significantly (p< , and p< . , respectively) . hours, but not hours after lysine clonixinate, inhibition by asa was much greater and still highly significant (p< , ) after hours. the mean concentrations of lysine clonixinate and of asa inhibiting txb biosynthesis in vitro by % (ecs ) were . + . ~tm and . _+ . gm, respectively. aggregation of . ml aliquots of platelet-rich plasma induced by gg/ml collagen was inhibited by lysine clonixinate significantly (p< . ), but not completely for up to hours, while aggregation induced by gg/ml collagen was inhibited for up to hours only (p< . ). platelet aggregation induced by both collagen concentrations was inhibited by asa almost completely (p< . ) for at least hours. the ecs for in vitro inhibition of submaximal collageff-induced platelet aggregation was + btm for lysine clonixinate and + lam for asa. it remains to be investigated, whether other mechanisms than inhibition of cyclooxygenase contribute to the analgesic activity of lysine clonixinate. the profile of side effects including only moderate inhibition of platelet function by lysine clonixinate as compared to asa might be an advantage with regard to its use as an analgesic. increased cell proliferation induced by ras-oncogenes is mediated by growth factors. growth factors bind to cell membrane surface receptors and activate tyrosine kinases leading to protein phosphorylation. this triggers posttranslational modification of the function of low molecular mass gtp binding ras proteins, e.g. p ras, including heterotrimetric g proteins. th~se proteins carry a c-terminal amino acid sequence, caax where c is cysteine, which is the target of isoprenylation by famesyl protein transferases (fpptases) and geranyl-geranyl protein transferases (ggptases). the amino acids of this sequence determine whether the ras protein is substrate for fptases or ggptases. this biochemical pathway is related to mevalonate metabolism, therefore, cholesterol biosynthesis shares several steps with isoprenylation. the active compounds include a series of isoprenoid inter-mediates, such as geranyl pyrophosphate (c- ), farnesyl pyrophosphate (fpp, c- ) and geranyl-geranyl pyrophosphate (ggpp, which are critical for cell proliferation. inhibitors of isoprenylation may (i) compete with fpp to its binding site on the alpha subunit of the enzyme, (ii) inhibit the zn protease of fptase beta subunit, and (iii) interfere with the binding of the caax peptide to its binding site on the beta subunit. the inhibitors include microbial products like manumycin, gliotoxin, acetylgliotoxin and pepticin-namins, fpp analogues, tetrapeptides and related peptidomimetics that may compete for the caax sequence, products of plant origin like limonene as well as synthetic compounds. all these compounds have been known to retard cell proliferation, and their effect can be abolished by mevalonic acid. inhibitors of hmg-coa reductase, such as lovastatin, which have been developed as antihypeflipidemic drugs, may also effective as anticancer agents. the recognition of the importance of mevalonate metabolism in cancerogenesis provides possibility to test simple chemical comp unds against tumor growth, and gives promise for pharmacological exploitation of inhibitors of isoprenylation and hmg-coa reductase inhibitors for cancer chemotherapy. dopamine (da) and epinine (epi, the active metabolite of ibopamine) are da-and ~-and ~-adrenergic receptor (r) agonists. to study whether da-r mediated effects can be separated from ~-and ~-ar effects we assessed in male volunteers effects of i.v. da and epi ( . ; . ; . and . ~g/kg/min for min each) on changes in serum prolactin (prl, da-r mediated), blood pressure (p) and heart rate (hr, ~-and ~-ar mediated), natriuresis (urinary excretion of sodium), diuresis (urine flow rate) and plasma noradrenaline (na) levels. in the . and . ~g doses da and epi did not affect psyst, pdiast and hr, but significantly decreased prl (da: from . ± . to . ± . ng/ml; epi: from . ±i.i to . ± . ng/ml, p<o. ). in the ~g dose da and epi increased psyst by mmhg, pdiast by and mmhg, and hr by and beats/min but further decreased prl. in addition, both da and epi significantiy increased diuresis and natriuresis; in contrast, only da, but not epi, dose-dependently increased plasma na ievels, an effect that was not inhibited by ~-ar antagonists. we conclude that in . and . ~g doses (plasma levels of - nmol/ ) oa and epl act only at da-r. in higher doses, however, da and epi can activate ~-and ~-ar whereby epi has a stronger ~-ar component than da. moreover, part of da-effects are indirect via release of endogenous na whereas epl-effects do obviously not include an indirect component. j. raderrnacher, r.mliller-bartels, f. barrels, h. baurngarten, k. thon, d. tsikas and j.c. fr ich, patients with essential hypertension were ireated with cicletanin ( rag/d) for weeks. compliance was verified by measuring urinary cicletanin excretion. before entering the study anti_hypertensive medication had been withdrawn for at least weeks (wash out phase). the study was preceeded by a placabo phase of week duration. blood pressure measurements and urine collections for measurements of , dinor- -keto-prostaglandin f a (a stable prostacyclin metabolite) were performed at the start and end of placebo phase, atter week of treatment with cicletanin and in monthly intervalls thereafter. cicletanin decreased systolic and diastolic blood pressure, the effect becoming significant at weeks. urinary excretion of the prostacyclin metabolite , dinor- keto pgfla.increased by %. hypertensive patients and control subjects (n= ; , dinor- -keto-pgfla: + pg/rnl] did not differ significantly in urinary excretion of , dinor- keto pgf ct. neither systolic nor diastolic blood pressure showed a significant correlation with , dinor- keto-pgf alpha excretion (systolic bp: r-- , , p-- , l; diastolic bp: r= , , p=l, ). these data show that cicletanin is effective in reducing blood pressure. increased prostacyclin production is probably not responsible for the observed reduction of blood pressure. in the past the realization of trials on the pharmacokinetics of nicergoline in human beings break down according to analytical difficulties. from animal studies with labelled compounds it is known, that nicergoline is rapidly absorbed from the gastrointestinal tract and two main metabolites (mdl, i-mmdl) can be detected.the mean elimination half-life for nicergoline is given with , h, for mdl - h and for i-mmdl with - h. the compound is excreted mainly by the kidney. due to known problems of the application of labelled drugs in human beings there are two approaches for the determination of these compounds: first the detection by monoclonal antibodies by using a ria-method, secondly a hplc-analysis. although there are promising in-vitro results of the immunological method, quantitative analyses of serum samples show deceptive results. the hplcassay reveales that in human serum samples after oral administration of nicergoline only the metabolites mdl and i-mmdl can be detected. further analytical development lower the limit of detection which should be i/i of c~x. with an improved hplc-method it will be demonstrated that only the measurement of the metabolite mdl is suitable to give sufficient results in pharmacokinetic trials and in studies on bioavailability with formulations of nicergoline. antiplatelet treatment with aspirin is well established as secondary prophylaxis after a transient ischemic attack or minor ischemic stroke, but the effect of aspirin treatment on the course of carotid atheroselerosis is unknown. we tried to investigate the effect of aspirin on the initial stages of carotid atherosc}erosis. patients were recruited from a prospective, randomized, double blind clinical trial to compare two doses of aspirin ( mg vs mg claity) with respect to restenoses after lower limb angjoplasty. of the patients admitted to the angioplasty trial, patients showing small carotid atheroma (< % lumen narrowing) were examined at entry and after year of aspirin treatment using a high-resolution ultrasound duplex system. disease progression and regression were defined by a change of maximal plaque area (as measured by longitudinal ultrasound sections) of more than standard deviations of the method. the change in plaque area was significantly different for the treatment groups: average plaque size remained'unchanged after treatment with mg aspirin daily but increased markedly after treatment with mg aspirin daily (p = . ). there were significantly more lesions in the -mg group showing progression than in the -mg group ( plaques [ %] vs plaques [ %], p = . ). ultrasonic disappearance of a lesion was observed only in the -mg group in cases ( soft plaques, ulcerative plaques, p = . ). the patients on mg aspirin who continued smoking during the study showed significantly more progression as compared to the non-smokers in the -mg group ( plaques [ %] vs plaques [ %], p = . ). the results of our study indicate that aspirin treatment slows carotid plaque growth in a dose-dependent fashion, with a dose of mg daily more efficient than mg a day. gentamicin is frequently applied in neonatal patients for the treatment of severe infections. the use is associated with high incidences of oto-and nephrotoxicity. ototoxicity occurs in - . % of the treated patients. it can be avoided, if the trough levels remain under p.g/ml between the administered doses. if the nephrotoxic effect, which occurs in - % of the treated patients, can also be reduced under these circumstances is not yet clear. recently it has been suggested to administer gentamicin only once daily in order reach the desired trough levels. this applies especially to children, in which the terminal half-life can be prolonged to approximately hours. this dosage regimen assures a sufficient effectivity due to the marked postantibiotic effect, but should reduce the incidence of the toxic effects. in order to test this hypothesis we perform a case-control study with neonatal children treated with gentamicin. the cases of nephrotoxicity during the previous months are identified from the hospital records. for these patients the plasma level time curves are obtained with the use of a population based pharmacokinetics programme, which considers the measured plasma concentrations and the administered doses. plasma concentrations were measured routinely in the process of therapeutic drug monitoring. from these curves the duration of times with concentrations lower than p.g/ml during the last days prior to the diagnosis are estimated. for the cases matched controls were identified according to age, age of gestation and other factors affecting the renal function e.g. comedication. different analyses are carried out according to the definition of exposure. the results wiij be discussed with regards to the modification of the dosage schedule for gentamicin and the application of population kinetics in order to reduce the adverse effects. economic evaluation of drug therapy is an important task of clinical pharmacology. drug costs amount to a relatively small percentage of total health care costs. some drugs may help to avoid more expensive therapeutic procedures. drugs with small therapeutic ratio need e therapeutic drug monitoring (tdm) to optimize therapeutic results and to avoid side effects. tdm may enhance the costs of drug therapy and the question arises, whether the additional costs are justified from the economic point of view. the aim of the study was to investigate the effectiveness of theophylline tdm in chronic obstructive pulmonary disease. methods: outpatients suffering from chronic obstructive pu[monary disease and receiving theophylline as a basic medication were included in this study. during the one-year observation period the parameters of pulmonary function were examined four times ( st, rd, th and th month) by the same physician and theophylline trough levels were estimated to individualize the dosage. the clinical outcome was assessed by intraindividual comparison. results: impaired parameters of pulmonary function were improved. hospital admissions were reduced. during the one-year observation period out of patients had hospital admissions with totally days. under the conditions of theophylline therapy without tdm during the year before of them had hospital admissions with days of stay in hospital. discussion: the results agree with a previous study on inpatients. there it was observed, that parameters of pulmonary function significantly were improved due to tdm-supported theophy[line therapy in comparison to a group receiving theophylline without tdm. despite the additional costs thaopylline tdm contributes not only to improved clinical outcome but also to decreased costs for the care of patients with chronic obstructive pulmonary disease. in cancer treatment arterial blood flow reduction by embolisation followed by intra-artedal chemotherapy may be advantageous by achieving high and prolonged drug concentrations in the tumor and lower systemic drug exposure. pharmacokinetics of mitomycin-c (mmc) were investigated in colorectal cancer patients (body weight: - kg) with liver metastases undergoing chemoembolisation. after hepatic artery branch catheterization and microspheres injection (polyvinylalcobel, mg mmc dissolved in ml physiological naci solution were infused in - rain, followed by ethiblock application. serum mmc-concentrations were determined from peripheral venous blood samples by reverse-phase hplcsystem with ultraviolet detection (c -column, elution: phosphate-buffer ( . m, ph: . )) : methanol, v:v= : , nm). pharmacokinetic parameters were computed using non-compartmental or two-compartment model analysis assuming linear kinetics (top fit . software). results: on the average, maximum serum concentrations were observed at tmax = . rain following the beginning of mmc-infusion and amounted to cmax = ng/ml (range: - ng/ml). mean steady state distribution volume was . i/kg with a central compartment volume of . i/kg. concentration-time curves could be descnbed by a distributional phase (t / = . min) followed by a terminal elimination phase (]' / = . min). mean total clearance amounted to . ml/min.kg. the area under the concentration time curve was pg.m - .min. conclusion: pharmacokinetic parameters for mmc-disposition in this study do not substantially differ from those reported for similar mmc-amounts administered as intravenous bolus. systemic mmc exposure does not seem to be reduced following the embolisation procedure used compared to intravenous application. pseudocholinesterase-activity (pche) is an important determinant of the elimination kinetics of drugs like mivacudum. therefore, the time course of pche was investigated in patients undergoing cardiosurgery with cardiopulmonary bypass (cpb) in normo-or hypothermia. anesthesia was induced and maintained with midazolam, propofol and fentanyl using mivacurium as a muscle relaxant. by starting cpb, i of pdming fluid were added to circulation volume. from collected blood samples, pche (u/i) was determined photometrically according to eiiman ( )(butyrylthiocholine, dithiobisnitrobenzoate, nm, ph= . , oc) . total plasma protein concentration (g%) was measured according to lowry, and the specific pche was calculated as the quotient of pche and protein concentration (u/g). serum hepadn-activity was determined in patients (berichreme heparin test kit). moreover, the influence of hepadn and aprotinin, both contained in the cpb pdming solution, on pche in vitro was examined using fresh serum from healthy volunteers (n= ). results: the induction of anesthesia had no significant influence on pche or protein concentration (p> . ). cpb induced a significant decrease of pche (- %)(p< . ) and protein concentration (- %)(p< . ) and a less pronounced numedcal reduction the specific pche (- %)(p> . ). the reduction of pche and protein concentration was not significantly affected by ending cpb (p> . ), and the values remained low over the remaining operation time. there was no significant difference in pche, measured at °c in vitro, or protein concentration between the normothermic and hypothermic group (p> . ). furthermore, there was no correlation between serum hepadn-activity and pche reduction. pche in the plasma of healthy volunteers was not significantly affected by either hepadn up to u/ml or apretinin up to u/ml (p> . ). conclusion: ( ) the concentration of the antitumor antibiotic mitomycin c (mmc), used in ophtalmic surgery for its antiproliferative effects, was measured in the aqueous humor of glaucoma patients undergoing trabeculectomy. sponges soaked with mmc-solution ( ul of mmc-solution . mg/ml: rag) were applied intraoperatively under the scleral flap for rain. to ul of aqueous humor were drawn with a needle min following the end of topical mmc-treatment. samples were assayed for mmc using a reverse-phase hplc-system with ultraviolet detection (c -column, elution: phosphate-buffer ( . m, ph: . ):methanol, v:v = : , nm). swabs were extracted in phosphatebuffer ( . m, ph: . ) before hplc-analysis. external calibration was used for mmc quantitetion. quantitation limit was ng/ml. in all aqueous humor samples mmc-concentration was below ng/ml. mmc in the swabs amounted to % of the mmc amount applied. conclusion: after intraoperetive topical application, mmc concentration in the aqueous humor of patients is very low. the substantial loss of mmc from the swabs used for the topical mmc-treatment suggests ( ) rapid systemic absorption of mmc and/or ( ) a loss through irngation of the operative field following topical mmc-application. institut fur pharmakologie und * klinik for augenheilkunde, universitcit k n, gleuelerstrasse , k n al a due to runaway costs of the national health service which are reflected as well in growing expenditures for drugs at the university hospital of jena investigation of indication related drug administration patterns becomes more and more interesting. this holds especially true for intensive care units (itu's) which are determined by similar high costs for technical equipment as for drugs ( ) although any economical considerations seem to be questionable due to ethical reasons ( ). over a month period indication related drug administrations of surgical itu's of the university hospital jena have been recorded and analyzed by using a pc-notebook. total expenditures for all included patients add up to dm . . regarding these drugs and blood products which caused % of total costs in . the leading substances ( antithrombin ill, human albumin %, prothrembine complex, ...) represent % of total costs including blood products, antibiotics and ig m endched intravenous immunglobine. therefore the indication of particulary these drugs became mere interesting for further investigations. already during the study actual discussions with the treating medical staff have been made leading to new developed therapy recommendations. providing same high standard of medical treatment a remarkable cost saving of some drugs by more cdtical and purposeful use could already be achieved as a first result. however, the results of the study underline impressivly the benefit of such investigations for improvement of drug treatment. the simple replacement of expensive drugs ( e.g. prothrembine complex ) by higher quantities of cheaper ones of the same indication group ( e.g. fresh frozen plasma ( )) does not necessarily mean less expenditures in all cases but may cause unsiderable side effects. ( ketokonazole is known to decrease pituitary acth secretion in vitro and inhibits adrenal ll-hydroxylase activity. to work out the clinical significance of both effects analysis of episodic secretion of acth, cortisol (f) and ll-deoxycortisol (df) was performed in patients with cushing's syndrome (cs) requiring adrenostatic therapy. methods : ketokonazole was started in ii patients with cs ( acth-secreting pituitary adenomas [cd], adrenal adenoma [aa] ). in of them ( cd, aa) blood samples were obtained for hours at i min intervals ( samples/patient) before and again under treatment (mean dose i mg/d, > weeks). hormone levels were measured by ria and secretion patterns analysed by means of pulsar, cluster and desade. patients were investigated only once because treatment was stopped due to side effects. results : the we conclude that the observed % increase of plasma acth and the % decrease of f/df ratio demonstrate that inhibition of adrenal li -hydroxylase activity is the primary mode of action of ketoconzole in vivo. even at high doses acth and f secretion patterns could not be normalized. the improvement of pain and swelling conditions by means of drugs is an important method of achieving an enhanced perioperative quality of life in cases of dentoalveolar surgery. in prospective, randomised, double-blind studies the influence of various concentrations of local anaesthetics and accompanying analgesic and antioedematons drugs was investigated in the case of osteotoimes. all of the studies were carded out according to a standardised study procedure. a comparison of the local anaesthetics articaine % mad articaine % (study ) demonstrated the superior effect of articaiue % with respect to onset relief on pain, period of effectiveness and ischaemia. recordings of the cheek swelling in the remaining studies were made both sonographically and with tape measurement, while the documentation of the pain was carried out by means of visual analogue scales on the day of operation and on the first and third post-operative days. tile perioperative, exculsive administration of x mg dexamethasone (study ) resulted in a significant reduction in the swelling ( %) while the exclusive administration of x mg lbuprofen (study ) was accompained by a marked decrease in pain ( %) but no significant reduction of swelling in comparison to the placebo group. the combination of x mg ibuprofen und mg methylprednisolone (study ) yielded a decrease in pain of . % and a reduction in swelling of %. a cdmparison between a mono-drug ibuprofen and a combination drug ass/paracetamol (study ) resulted in no significant difference in the reduction of swelling and pain and therefore highlighted no advantages for the combined drug. a mono-drug should therefore be given priority as an analgesic. the combinatton of ibuprofen und methylprednisolone offers the greatest reduction in pain and swelling. using the results of the randomised studies, a phased plan for a patietu-orietued, anti-inflammatory therapy to accompany dento-alveolar surgery is presented. in a placebo controlled study patients with congestive heart failure (nyka class ii) were treated orally for seven days with i mg ibopamine t.i.d, i subjects had a normal renal function (mean inulin clearance (gfr) ± , ml/min), i patients suffered from chronic renal insufficiency (gfr ± , ml/min; x ± sem). pharmacokinetic parameters of epinine, the maximum plasma concentration, the time to reach maximum plasma concentration and the area under the curve from to hours were unaltered in impaired renal function when measured on the first or on the seventh treatment day. however plasma concentrations in both groups were significantly higher on the first treatment day than after one week of ibopamine administration. in this context antipyrine clearance as a parameter of oxidative liver metabolism which might have been induced by ibopamine revealed no differences between placebo and ibopamine values. in conclusion kinetic and dynamic behaviour of ibopamine was not altered by impaired renal function. human protein c (hpc) is a vitamin k-dependent in the liver produced glycoprotein with anticoagulant properties. when active protein c splits the coagulation factors va and vuia by means of limited proteolysis (kisiel et al ) . its concentration in normal plasma is - lag/m[ i-ipc's biological importance became evident when a congenital protein c deficiency, which results in difficult recurrent thromboembolic diseases was discovered (griffin eta/ ) . the recognition of a congenital hpc deficiency, as wall as the connection between acquired protein c deficiency and the appearance of thromboembolic complications by means of highly accurate and sensitive ascertained methods is therefore of great practical importance for the clinic. murine monoclonal antibodies (moabs) against hpc were formed. antibody producing hybridomas were tested by an ,,indirect elisa" against soluble antigens. the plates were coated with purified hpc up to ng/ al. the peroxydase-system was used to identify antibodies the antibodies were tested with the remaining vitamin k-dependent proteins for cross-reactivity, as well as with hpc deficiency plasma for disturbances by other plasma proteins. the above described experiment represents a sensitive and specific method for measuring the hpc concentration with moabs. assessment of local drug absorption differences ("absorption window") in the human gastrointestinal tract is relevant for the development of prolonged release preparations and for the prediction of possible absorption changes by modification of gastrointestinal motility. current methods are either invasive and expensive (catheterization of the intestinum, hf-capsule method) or do not deliver the drug to a precisely defined localization. we evaluated the delay of drug release from tablets coated with methacrylic acid copolymer dissolving at different ph values as an alternative method. three coated preparations of caffeine tablets (onset of drug release in in vitro tests at ph . , . and . ) and an uncoated tablet (control) were given to six healthy male volunteers in a randomized order. caffeine was used because of its rapid and complete absorption and good tolerability. blood samples were drawn up to h postdose (coating ph . up to h postdose), and caffeine concentrations were measured by hplc. auc, time to reach measurable caffeine concentrations (tia~), tr, ax, cmax and mean absorption time (mat) values for coated preparations were compared to the reference tablet (mean + sd of n= ): the relative bioavailibility for the coated preparations did not differ from the reference, suggesting complete release of caffeine. all coatings delayed caffeine absorption onset. the tlag for the ph . preparation suggests that release started immediately after the tablet had left the stomach. the mean delay of . h for the ph . coating was highly reproducible and should reflect small intestine release. the ph . coating delayed absorption to the highest extent, however the drug was probably released before the colon was reached. there is evidence that nitric oxide (no) plays a role in cardiovascular disease like hypertension, myocardial ischemia and septic cardiomyopath.y. no stimulates the guanylyl cyclase leading to an increase m cgmp content we investigated by immunoblotting the expression of the inducible nitric oxide synthase (inos) in left ventricular myocardium from failing human hearts due to idiopathic dilative cardiomyopathy (idc, n= ), ischemic cardiomyopathy (icm, n= ), beeker muscular dystrophy (n= ) and sepsis (sh, n= ) compared to non-failing human hearts (nf, n= ). cytokine-stimulated mouse macrophages were used as positive controls sds-polyacrylamide gel electrophoresis ( . %) was perfomed with homogenates of left ventricular myocardium and mouse macrophages respectively. proteins were detected by enhanced chemiluminescence using a mouse monoclnal antibody raised against inos. furthermore, we measured the cgmp content in these hearts by radioimmunoassy. a band at about kda was observed in two out of three hearts from patients with sepsis and in stimulated mouse macrophage~ no inos-protein expression was detected in either non-failing human hearts (n= ) or failing human hearts due to idc, ihd or bmd. in ventricular tissue from patients with sepsis cgmp content was increased to % ( + fmol/mg ww, n= ) compared to non-failing hearts ( % or + . fmol/mg ww, n= ). in left ventricular tissue tissue from patients with heart failure due to idc, ihd and bmd cgmp content did not differ from that in non-failing hearts. it is concluded that an enhanced inos protein expression may play a role in endotoxin shock, but is unlikely to be involved in the pathophysiology of end-stage heart failure due to idc, ihd and bmd. (supported by the dfg.) nitric oxide (no) has been shown to be a major messenger molecule regulating blood vessel dilatation, platelet aggregation and serving as central and peripheral neurotransmitter; furthermore no is a crucial mediator of macrophage cytotoxicity. no production can be assessed reliably by determination of its main metabolites nitrite and nitrate in serum, reflecting no synthesis at the time of sampling, or in h urine, reflecting daily no synthesis. farrell et ai. (ann rheum dis ; : ) recently reported elevated serum levels of nitrite in patients with rheumatoid arthritis (ra). we report here total body nitrate production and the effect of prednisolone in patients with ra. nitrate excretion in h urines of patients with ra as defined by the revised criteria of the american rheumatism association was measured by gas chromatography at times: first before start of a antiinflammatory therapy with prednisolone, when the patients had high inflammatory activity as indicated by mean crp serum concentrations of + sd mg/i and elevated esr with a mean of ]: after hour. secondly - weeks after start of prednisolone therapy in a dosage of . mg/kg body weight, when the patients showed clinical and biochemical improvement (crp + mg/i, p< . , esg + , p< . , two-tailed, paired t-test). for comparison h urines from healthy volunteers were obtained. before start of predniselone therapy the urinary nitrate excretion in patients with ra (mean + sd p.mol/mmol creatinine) was more than twofold higher (p< . , twoaailed unpaired t-test) than in healthy volunteers ( + ~tmol/mmol creatinine). the urinary nitrate excretion decreased significantly (p< . , two-tailed, paired t-test) to + i.tmol/mmol creatinine under therapy with prednisolone, when inflammatory activity was reduced considerably. despite the decrease the urinary nitrate excretion was still twc, fold higher (p< . , two-tailed, unpaired t-test) in patients with ra than in the control group. our data suggest that the endogenous no production is enhanced in patients with ra. furthermore the results indicate that this elevated no synthesis could be reduced in accordance with suppression of systemic inflammation by prednisolone therapy. but now as ever the physicians are entitled to prescribe drugs which have to prepare in a pharmacy for a particular patient. little information is available on the frequency and patterns of these prescriptions. we had occasion to analyse the prescriptions of drugs which were prepared in pharmacies in north thuringia (east germany) from october to december at the expense of a large health insurance company (allgemeine ortskrankenkasse). the selected pharmacies are loealised in cities. we found prescriptions of drugs made up in pharmacies among a total number of reviewed drug prescriptions. this is . % of the total. most of these prescriptions were performed by dermatologists ( . %), general practitioners ( . %), paediatrists ( . %) and otolaryngologists ( . %). according to this, the most frequently prescribed groups of drugs were dermatics enteric eoated tablets with nag and nag acetylsalicylic acid (asa) have been developed wluch should avoid the known gastrointestinal adverse events by a controlled drug release mainly in the duodenum after having passed the stomach. a -way cross-over study in healthy male subjects, aged from - years, was conducted to investigate the pharmacokinetics, bioavailability, safety, and tolerance of asa and its metabolites salicylic acid and salicylurie acid following enteric coated tablets in comparison with plain tablets. asa and its metabolites were determined by a sensitive, specific, and validated hplc method. pharmacokinetic parameters were determined by non-compartreental analysis. bioequivalence was assessed by % confidence intervals. following the admimstration of enteric coated tablets, a delayed absorption can be observed for both the mg dose and the rag dose. this is likely due to a delayed release of the active substance from the enteric-coated tablets in the small intestine arer gastric passage. considering the mean residence times (mrt), there is a difference of at least . h following the enteric coated tablets compared to the plain tablets for asa and the two metabolites measured• this difference represents the sum of residence time in the stomach plus the time needed to destroy the coating of the tablet when it left the stomach• in general, the maximum observed concentrations of both enteric coated formulations occurred - h post dose. the pharmacokinetics of a novel immunoglobulin g (lgg) preparation (bt , biotest, dreieich, frg) have been determined in healthy, male anti-hbs-negative volunteers. for this preparation only plasma from hiv-, hbv-and hcv-negative donors was used, the quality control for the product was in accordance with the ec-guideline for virus removal and inactivation procedures. each volunteer received a single, intravenous infusion of ml bt containing g igg and anti-hbs > , iu. anti-hbs was used as a simply measurable and representative marker for the igg. blood samples for determination of anti-hbs (ausab eia, abbott, frg) were drawn before and directly after the infusion, after , , , and hours, on day , , , , , , , , , and . additionally, total protein, igg, iga, igm and c /c complement were measured and blood hematology and clinical chemistry parameters determined. the phar~gacokinetic parameters of anti-hbs were calculated using the topfit ~" pc program assuming a -compartment model. pharmacoeconomic evaluations (pe) describe the relationship between a certain health care input (costs) for a defined treatment and the clinical outcome of patients measured in common natural units (e.g. blood pressure reduction in mmhg), quality of life (qol) gained, lifes saved or even in money saved due to the improvement in patients functional status. this implies that the efficacy of a treatment has been measured and proven in clinical trials. in addition, in order to transfer data obtained in clinical trials to the clinical setting, an epidemiological database for diseases and eventually drug utiiization may be required. the evaluation of the efficacy depends on the disease to be treated or prevented and the mode of treatment. for acute, e.g. infectious diseases, the endpoint can be defined easily by the cure rate, but for pe the time (length of hospital stay) and other factors (e.g. no. of dally drug administrations) have to be considered. in the case of chronic diseases, e.g. hypertension or hypercholesterolaemia, surrogate endpoints (blood pressure or serum cholesterol reduction) and information on side effects may be acceptable for the approval, but cannot be used for a meaningful pe. the latter should include the endpoints of the disease, i.e. cardiovascular events (requiring hospitalisation and additional treatment) and mortality. furthermore, the qol has to be measured and considered for chronic treatment. several questionaires have been developed to measure the overall qol or the health related qol. especially the latter may be a more useful tool to detect mad quantify the impact of a treatment on qol. combining the clinical endpoint mortality and qol by using qalys (quality-adjusted lifeyears) may be a useful tool to determine the value and costs of a given drug treatment but cannot be applied to all treatments under all circumstances. sorbitol was used as a model substance to investigate the dynamics of the initial distribution process following bolus intravenous injection of drugs. to avoid a priori assumptions on the existence of well-mixed compartments data analysis was based upon the concept of residence time density in a recirculatory system regarding the pulmonary and systemic circulation as subsystems. the inverse gaussian distribution was used as an empirical model for the transit time distribution of sorbitol across the subsystems, distribution kinetics was evaluated by the relative dispersion of transit (circulation) times. the distribution volumes calculated from the mean transit times were compared with the modelindependent estimate of the steady-state volume of distribution. kinetic data and estimates of cardiac output were obtained from patients after percutaneous transluminal coronary angioplasty. each received a single . g iv bolus dose of sorbitol. arterial blood samples were collected over hours. while the disposition curve could be well fitted by a tri-exponential function the results indicate that distribution kinetics is also influenced by the transit time through the lungs, in contrast to the assumption of a wellmixed plasma pool underlying compartmental modelling. a karit@ "bu£ter" is used traditionally in west afr%can manding colture as a cosmetic to protect the skin against the.sun. gas chromatography was used to analyze the ingredients of karit@ butter from guinea. we found % palmitic acid, % stearic acid, % oleic acid and % linoleic acid and . % of other fatty acids with higher chain lengths like arachidonio acid. some of these are essential fatty aclds (vitamine f). furthermore karit@ contains vitamine a and d as well as triterpene alcohols and phytosterines. an original extract was used to prepare a skin cream. this preparation was tested in volunteers ( women, men; age - y.). the cream contained at least % karit@, glycerol, emulsifiers and no preservative agent except for sorbic acid. of the volunteers very well tolerated the cream and thought it effective. the skin became more tender and elastic. good results were obtained when the volunteers suffered from very dry skin. two of them who were known to be allergic against the most available skin creams had no problems in using our karit cream. pure karit@ butter was used for four months to treat an african infant with neurodermitis. after this time the symptoms had markedly improved whereas previous therapy trials with other usual topical medicaments had been unsuccessful. these pre-studies had shown that dermatologic preparations containing karit# may be a good alternative in the treatment of therapyreslstent skin diseases and may in some cases be able to replace eorticoid treatment. ) and a low molecular weight heparin preparation (fragmin ~, iu/kg bodyweight s.c.) on coagulation and platelet activation in vivo by measuring specific coagulation activation peptides [prothrombin fragment + (f + ), thrombin antithrombin iii complex (tat), -thromboglobulin (~-tg)] in bleeding time blood (activated state) and in venous blood (basal state). in bleeding time blood, r-hirudin and the heparin preparations significantly inhibited the formation of both tat and f + . however, the inhibitory effect of r-hirudin on f + generation was short-lived and weaker compared to ufh and lmwh and the tat/f + ratio was significantly lower after r-hirudin than both ufh and lmwh. thus, in vivo when the coagulation system is in an activated state r-hirudin exerts its anticoagulant effects predominantly by inhibiting thrombin (lla), whereas ufh and lmwh are directed against both xa and ila. a different mode of action of ufh and lmwh was not detectable. in venous blood, r-hirudin caused a moderate reduction of tat formation and an increase (at hour) rather than decrease of f + generation. formation of tat and f + was suppressed at various time points following both ufh and lmwh. there was no difference in the tat/f + ratio after r-h[rudin and heparin. thus, a predominant effect of rhirudin on ila (as found in bleeding time blood) was not detectable in venous blood. in bleeding time blood, r-hirudin (but neither ufh nor lmwh) significantly inhibited ~-tg release. in contrast, both ufh and lmwh caused an increase of ~-tg hours after hepadn application. our observation of reduction of platelet function after r-hirudin compared to delayed platelet activation following ufh and lmwh suggests an advantage of r-h[rudin over heparin, especially in those clinical situations (such as arterial thromboembolism) where enhanced platelet activity has been shown to be of particular importance. the human cytochrome p isoform cyp a determines the level of a variety of drugs metabolized by the enzyme, including caffeine (ca) and theophylline (th). more than compounds are potential or proven inhibitors of this enzyme. some of them were reported to be substrates or inhibitors to cyp a in vitro, ethers caused pharmacokinetic interactions with drugs metabolised by cyp a . we characterized a series of these compounds with.respect to their effect on cyp a in human liver microsomes in relation to-published pharmacokinetic interactions in vivo. cyp a activity in vitro was measured as ca -demethylation at the high affinity site in human liver microsomes, using rain incubation at °c with - jm caffeine, an nadph generating system, and inhibitor concentrations covering . orders of magnitude. apparent kr values were estimated using nonlinear regression analysis. for inhibitory effects on cyp a activity in vivo, the absorbed oral dose causing % reduction in ca or th clearance (edso) was estimated from all published interaction studies using the emax model. %)i followed by disinfectants ( . %)r ointments ( . %) and solutions ( . %) were the most frequent drug forms %) or german ( . %). our results show that even now drugs prepared trend analysis of the expenses at the various departments may be a basis for a ratio-hal and economic use of the drug budget. total drug expenses amounted to mill. dm in . s milldm ( %) were used in surgical departments with intensive care units (icu) (general surgery, kardiovascular surgery, neurosurgery, gynecology, anaesthesiology) of wtfich % are needed by the icu and % in the operating rooms. surgical departments without scu but similar patient numbers (ophthalmology, ent, orthopedics and urology) get only % of the budget ( % needed for the operating rooms). the medical departments spent s mill.dm of which icu needs only % whereas the oncology (oncu) and antiinfective units uses more than %• similar relation could be seen in the child hospital ( . milldm, %) where % were spent for icu and % for oncu. the departments of dermatology and neurology get %, the depart-merits of radiology, nuclear medicine and radiation therapy only % of the budget. antiinfective drugs (antibiotics, antimycotics, virustatics) are most expensive ( % of budget) followed by drugs used for radiological procedures ( %) sncreasing the knowledge about the costs of medical items and the rational and economical use may stop the overproportional increase of the drug budget the mostly used : ) and a -fold higher efficiency than the r-form the elimination of the talinolol enantiomers was studied in healthy volunteers (age: - years, body weight: - kg) given a single oral dose ( mg) or an intravenous infusion ( rag) of the racemi c drug. three volunteers were phenotypically poor metabolisers and nine were extensive metabolisers of the debrisoquine-type of hydroxylation. the r-and senantiomers of talinolol were analysed in urine by a hplc method after enantioselective derivatisation. the concentrations of the enantiomers within every sampling period as well as the amounts of s-and r-enantiomer this corresponds to a s/r-ratio of , + , . the mean total amount (= s-+ r-enantiomer) eliminated was on average % &the administered dose. after oral administration _+ % of the dose were eliminated within h. the amounts of talinolol enantiomers recovered were equally (senantiomer: _+ gg the ratios of s-to r-concentrations at every sampling interval and of every volunteer were assessed between , and , (mean: , after infusion and , after oral administration, respectively) medizinische fakult~t carl gustav cams, teelmische university, t, fiedlerstr nitric oxide (no), synthesized by the inducib]e form of no synthase, has been implicated as an important mediator of-specific and non-specific immune response, little is known about the in vivo synthesis or no in inflammatory joint diseases. therefore we have studied the excretion of the major urinary metabolite of no, nitrate, in rats with adjuvant arthritis, a well established model of polyarthritis in addition we assessed the urinary excretion of cyclic gmp, which is known to serve as second messenger for the vascular effects of no, synthesized by the constitutive form of no synthase, affecting blood vessels, plate]et aggregation and neurotransmission, in h urines of male sprague daw]ey rats at day after induction of adjuvant arthritis we measured nitrate excretion by gas chromatography and cyclic gmp by radioimmunoassay. for contro] we determined the same parameters in h urines of non-arthritic rats of the same strain and age, we found a significant (p < , two-tailed, unpaired t-test), more than -fo]d increase of urinary nitrate excretion in arthritic rats (mean ± sd pmo]/mmol creatinine) as compared to non arthritic rats ( _+ izmot/mmo] creatinine). urinary cyclic gmp excretion was slightly, but not significant]y lower in arthritic rats ( ± nmol/mmol creatinine) than in controls ( ± nmo]/mmo] creatinine).there were no major differences in food or water intake which cou]d account for these results. the increased urinary nitrate excretion accompanied by normal cyclic gmp excretion suggests that no production by the inducible form of no synthase is enhanced in rats with adjuvant arthritis institute of c]inica] pharmacology~ hannover medical school, d- hannover, germany and *research center gr@nentha] gmbh, zieg]erstr , d- aachen, germany background: pge has been shown to be efficacious in the treatment of critical leg ischemia. despite of an almost complete first pass metabolism in the lung the clinical effects of intraarterial and intravenous pge do not differ significantly. in addition, it is not fully understood which of the various pharmacological actions of pge is the main factor; by most authors, however, it is thought to be the increase of cutaneous and muscular blood flow. by means of [ - ]-h -pet, we studied muscular blood flow (mbf) of the leg in patients with peripheral arterial disease comparing intraarterial and intravenous pge . patients and methods: patients ( f, m; mean age y) with pad were studied, ( atherosclerosis, thromboangiitis obliterans). at the first day, pg pge were infused intraarterially within minutes; pet scanning of the lower leg was performed at minutes , und . at the following day, pg pge were infused intravenously within hours; pet scanning was performed at minutes , , and . results: in the infused leg the increase of mbf caused by intraarterial pge averaged + % at minute and _ % at minute ; in the not infused leg there was no effect. the increase rate in the infused leg was highly variable but did not correlate with sex, age, disease or clinical outcome. for intravenous pge the change of mbf at any time averaged almost %. conclusion: unlike intraarterial pge , intravenous pge does not increase the muscular blood flow of the leg. a comparable clinical effect provided, increase of muscular blood flow may not be considered the main way of action of pge in critical leg ischemia. eslrogen(er) and progesterone(pr) receptor status as well as lymph node involvement are important factors in predicting prognosis and sensitivity to hormone and chemotherapy in patients with breast cancer. prognostic relevance of ps -protein, egfr and cathepsin d is currently under debate. especially ps and egfr expression appears to provide additional information regarding the responsiveness of the tumour tissue to tamoxifen. the aim of the present study was to investigate the relationships between these parameters and established prognostic factors in breast cancer. in a prospective study ps and cathepsin d were assayed immunoradiometricauy in the tumour cytosol of patients, egfr was measured by elisa. relating the level of these factors to the lymph node involvement, menopausal status as well as turnout size, no significant association could be established. jn our findings er and pr are significantly correlated with the expression of ps but none is correlated with the cathepsin d status. egfr was shown to be inversely correlated with the content of er. a significant association between cathepsin d and ps could be established in patients with early recurrence. at a median follow-up of - months, recurrence was more common in patients with tumours having negative status for ps , independent of receptor status. in conclusion, because of the relative independence on the er and pr status and other prognostic factors and the influence on the recurrence behaviour, demonslrated here, and their role in promoting tumour dissemination and changing hormone therapy sensitivity, all three factors represent markers of prognostic relevance.deparlancnts of clinical pharmacology l, nuclear medicine and surgery ,pharmacoeconomic studies, conducted either separately from or together with clinical trials are increasing in both number and meaning. in a period of limited health care budgets, political and medical decision makers alike run the risk of accepting the results of such studies without critical reflection. careful evaluation of those studies by state-of-the-art methods is one way out of the trap. another could be to refer to ethical considerations. the problem in this context is, that the discussion concerning ethical aspects of pharmacoeconomic research, at least in europe, is just in its beginning. therefore, no widely accepted standards are available. but they are essential to answer four main questions: . who should perfom a pharmacoeconomic study? . which objectives should be considered? . what kind of study should be performed (e. g. cost-effectiveness, cost-utility, cost-benefit analysis)? . which consequences will be drawn from the results?based on the case study-orientated "moral cost-benefit model" (r. wilson, sci. tech. human values : - , ) , a three-step decision and evaluation model is proposed to handle bioethical problems in pharmacoeconomic studies: . moral risk analysis . moral risk assessment . moral risk management. possible practical consequences for decision making in research policy, study design and assessment of results are discussed. hirudin is the most potent known natural inhibitor of thrombin and is presently gaining popularity as an anticoagulant since recombinant forms have become available. the aim of the present study was to compare platelet aggregation, sensitivity to prostaglandin e (pge ) and thromboxane a (txa ) release in r-hirudinized and heparinized blood. platelet aggregation was measured turbidimetrically using a dual channel aggregometer (labor, germany) in blood samples of healthy volunteers anticoagulated with r-hirndin w (behring) and hepatin ( gg/mi blood each). aggregation was induced by arachidonic acid (aa; . , . and . ram) and adp ( . lam). pge in concentrations , and ng/ml was used. plasma txb content was measured by gas chromatography/mass spectrometry. this study showed a significantly lower a.a-induced platelet aggregation in r-hirudinized plasma. three minutes after the aggregation induction by . mm aa the plasma txb concentration was ng/ml in blood anticoagulated with rhimdin and . ng/ml in heparin-anticoagulated blood. the extent of the adp-induced aggregation was nearly the same in rhimdinized and heparinized plasma. platelet sensitivity to pge was significantly higher in r-hirudinized blood. thus, aa-induced platelet aggregation is significantly lower and sensitivity to pgei higher in r-himdin-anticoagulated blood in comparison with beparin-anticoagulated blood.university of tartu, puusepa str. , tartu ee , estonia anaemia has been reported in renal transplant (ntx) recipients treated with azathioprine (aza) and angiotensin converting enzyme-inhibitors (ace-i). an abnormal aza metabolism with increased -thioguanine nucleotide (tgn) levels in erythrocytes is a possible cause of severe megaloblastic anaemia (lennard et al, br j clin pharmaco ). methods: ntx patients receiving aza ( , _+ , mg/kg/d), prednisolone ( , + , mg/kg/d) and enalapril (ena) ( , + , mg/kg/d) for more than months were studied prospectively. blood samples were taken before and h after administration of aza on visits during ena treatment and weeks after ena had been replaced by other antihypertensives (x). tgn in erythrocytes, -mercaptopurin (mp) and -thiouric acid (tua) in h post dose plasma (p.) und h urine (u.) samples were analyzed by hplc using a mercurial cellulose resin for selective absorption of the thiol compounds. pharmacodynamic variables were hemoglobin (hb), erythropoietin (epo) and creatinine clearance (ci ace~,lcholine plays an important role in regulating various functions in the airway's. in human lung less is known about regional differences in cholinergic innervation and about receptor-mediated r%m.flation of acetylcholine release. in the present study the tissue content of endogenous acetylcholine and the release of newly-synthesized [~h]acetylcholine were measured in human lung human tissue was obtained at thoracotomy from patients with lung cancer moreover, in isolated rat tracheae with intact extrinsic vagal innervation possible effects of g__-adrenoceptor agonists on evoked ph]acctylcholine release were studied. endogenous acetylcholine was measured by hplc with ec-detection; evoked ph]acetylcholme release was measured after a preceding incubation of the tissue with [~h]choline. huma n large (main bronchi) and small (subsegmental bronchi) airways contained similar amounts of acetylcholine ( pmol/ mg), whereas significantly less acetylcholine was found in lung parenchym ( pmol/ mg). release of [ h]acetylcholine ,,,,'as evoked in human bronchi by transmural electrical stimulation (four s trains at hz). oxotremorine, an agonist at muscarine receptors, inhibited evoked [~hiacetylcholine release indicating the existence of neuronal inhibitor ' receptors on pulmona~ parasympathetic neurones. scopolamine shifted the oxotremorine curve to the right suggesting a competitive interaction (pa value: : slope &the schild plot not different from unity) however, a rather sluggish schdd plot was obtained for pirenzepine. scopolamine but not pirenzepine enhanced evoked [ h]acetylcholine release. the present experiments indicate a dense cholinergic innervation in human bronchi; release of aceu, lcholine appears to be controlled by facilitatory and inhibitou' nmscarinc receptors. in isolated, mucosa-intact rat tracheae isoprenaline ( nm) inhibited [~h]acetylcholine release evoked by preganglionic nerve stimulation isoprenaline was ineffective in mucosa-denuded tracheae or in the presence of indomethacin thus, adrenoceptor agonists appear to inhibit acetylcholine release in the airways by the liberation of inhibitoiy prostanoids from the mucosa. the occurrence of the non-enzymatic reactions between glucose and structural proteins is well known (vlassara h et al. ( ) lab invest : - ) . the reaction between proteins and fructose (i.e. fmctation), however, can also occur. like glucose-protein adducts the fructose analognes are able to form so-called advanced glycation endproducts (age). the inhibition of early and advanced products of fmctation may be ilnportant for the prevention of diabetic late complications (mcpherson jd et al. ( ) biochemistry : - . we investigated the in vitro fmctation of human serum albumin (hsa) and its inhibition by selected drugs. hsa was fmctated by incubation with mmol/ fructose in . i mol/l phosphate buffer, ph= . .,at ° c for days. the rate of fmctation was measured by the following methods: -a colorimetric method based on deglycatien of glycated, proteins by hydrazine (kobayashi k et ai.( ) bioi pharm bull : - ), -affinity chromatography with aminophenyl-boronate-agarose, -fluorescence measurement for the delermination of age we used aminoguanidine, pcnicillamine, captopril and alpha-lipoic acid( mmol/ ) to study the inhibition of hsa fmctation. after three weeks incubation the formation of early glycation products was inhibited by aminogalanidine ( %) and captopril ( %) whereas penicillamine and alpha-lipoic acid showed minimal inhibition. aminognanidine inhibited the formation of age by %, penicillamine by %, alpha-lipoic acid by % and captopril by %. these results may suggest a potential use of the investigated drags in the prevention of the formation of protein-fructose addncts. key: cord- -vbzceozs authors: hu, zhi-hua; sheu, jiuh-biing; xiao, ling title: post-disaster evacuation and temporary resettlement considering panic and panic spread date: - - journal: transportation research part b: methodological doi: . /j.trb. . . sha: doc_id: cord_uid: vbzceozs abstract after a disaster, a huge number of homeless victims should be evacuated to temporary resettlement sites. however, because the number of temporary shelters is insufficient, as are shelter building capabilities, victims must be evacuated and resettled in batches. the perceived psychological penalty to victims may increase due to heightened panic when waiting for evacuation and resettlement, whereas psychological interventions can decrease the magnitude of this panic. based on the susceptible–infective-removal model, panic spread among homeless victims and other disaster-affected people is modeled, while considering the effects of psychological interventions on panic spread. a function is derived to compute the increase in the number of victims to be evacuated due to panic spread. a novel mixed-integer linear program is constructed for multi-step evacuation and temporary resettlement under minimization of panic-induced psychological penalty cost, psychological intervention cost, and costs associated with transportation and building shelters. the model is solved by aggregating objectives into a single objective by assigning weights to these objectives. with wenchuan county as the test case, the epicenter of the sichuan earthquake, the influence and the sensitivity of parameters, tradeoff among costs, and the effects of various functions of panic strength on psychological penalty and monetary costs are assessed using six experimental scenarios. analytical results reveal the complexity and managerial insights gained by applying the proposed method to post-disaster evacuation and temporary resettlement. worldwide. over the last decade, china, the united states, the philippines, india and indonesia, are the top five countries most frequently targeted by natural disasters (guha-sapir and below, ) . earthquakes, hurricanes, droughts, and floods are global challenges due to their unpredictability and potential scale of impact in terms of social, environmental, and economic costs. many natural disasters displace people. furthermore, natural disasters can cause psychological suffering, altering people's behaviors and decision-making, increasing rescue and relocation difficulties, and decreasing social stability and security (hu and sheu, ; rennemo et al., ) . although few direct evidences of the effects of panic and its spread on post-disaster evacuation and resettlement have been recorded in literature, the synthetic effects of psychological damage on victims have been identified. this study addresses the effects of panic and its spread on evacuation and temporary resettlement. evacuation is the allocation and transport of disaster victims from disaster sites to sites with temporary shelters, whereas temporary resettlement refers to the resettlement of disaster victims during the period from the moment a disaster occurs to their allocation to transitional or permanent houses (el-anwar and chen, ) . more than million people were evacuated to temporary shelters after the sichuan earthquake. about . million tents and tarpaulin shelters were produced and transported to the affected areas within three months after the disaster (lian, ) . many farmers and urban residents were persuaded to engage in reconstruction by subsuming their grief. about . % of the victims who were evacuated and temporarily resettled returned to their houses and reconstructed their lives and buildings (lian, ) . because panic was widespread, many people refused to return to their houses even through their houses were deemed safe and no aftershocks were observed. panic spread had affected those people in sichuan province. the psychological effects of panic on demand for evacuation and resettlement are examined in this study. shadow evacuation is representative of these effects. shadow evacuation, which is the evacuation of people outside the disaster area who ''shadow'' the evacuation of those within the target area, is considered in a group of studies on evidences of the psychological effects of evacuation, especially that of those experiencing panic and other emotions. in some areas, over-evacuation of people not threatened directly by a hazard can add to traffic and congestion within a network and hinder the transportation of evacuees who are directly threatened. lamb et al. ( ) identified the factors influencing shadow evacuation. dependent measures were used to assess the likelihood of a shadow evacuation and identify messages and presenter characteristics. despite a -year experience with shadow evacuations, its causes are not sufficiently understood. dash and gladwin ( ) argued that studies should examine the role of perceived risk of shadow evacuation. they examined evacuation as a social process that is affected by the complex relationship between compliance with an evacuation message and an individual's perception of their risk exposure. panic is a source of risk perceived by disaster victims, and panic spread is through typical social relations among victims. shadow evacuations also provide clues as to the number of people affected by psychological emotions, as typical negative effects on hurricane-induced evacuation operations. following hurricane andrew in , gladwin and peacock ( ) noted that % of residents outside the evacuation zone also evacuated, placing considerable additional stress on transportation networks. lindell and prater ( ) also observed shadow evacuations associated with hurricane events, often occurring from inland areas deemed safe. shadow evacuation overloaded the capacity of the transportation networks. dueñas-osorio et al. ( ) estimated risks of hurricane hazards and then compared them with risk perceptions of residents after hurricane ike's landfall in . comparison results show that shadow evacuation is partly attributable to risk overestimation. overestimation of damage risk from wind or water surges resulted in , shadow evacuees. in this work, those in a shadow evacuation are motivated by panic and panic spread. although homeless people are in urgent need of evacuation, the number of temporary shelters is often limited immediately after a disaster. additionally, temporary shelters (mainly tents and tarpaulin rooms) generally take three days to erect, and are costly and difficult to recycle when abandoned. therefore, how to evacuate and resettle victims effectively with a limited number of temporary shelters is a significant problem. uncertainty associated with a disaster and the possibility of secondary disasters typically causes panic as victims wait for evacuation and resettlement. when panic spreads among disaster-affected people, many people who should not be evacuated will request evacuation. panic and its spread among victims are practical problems which complicate post-disaster evacuation and resettlement processes. panic is the primary risk when a disaster occurs (fritz and marks, ; mawson, ) . the typical response to various threats and disasters is not to flee but to seek familiar persons and places; moreover, separation from what is familiar creates more stress than physical danger (mawson, ) . mawson also asserted that waiting for evacuation can induce panic among victims because those waiting often see others leaving for resettlement sites. although post-disaster panic is reported frequently, quantitative studies are few. people are prone to panic and it spreads easily after disasters, namely, panic is infectious. that panic can induce a series of non-adaptive crowd behaviors during evacuation, such as pushing and trampling, and trying to jump in the line of those to be evacuated first. these actions are responsible for many disaster-related injuries and deaths. price-smith et al. ( ) determined that, with severe acute respiratory syndrome (sars), which originated in guangdong, china, in , both infected and non-infected people panicked, and this panic had negative impacts (e.g., people fleeing, tourism declining, and trade slowing) on societies and their economies. this study characterized qualitatively the impact of panic spread, but did not consider the socio-psychological mechanism of panic a contagion. based on diffusion theory of general diseases and public opinion, the effects of panic and its spread on temporary resettlement are considered in this study. the panic degree while waiting for resettlement is quantified by this work, and panic spread typically increases the number of people who request resettlement and evacuation. to relieve panic and avoid mass incidents, governments organize mental health workers, including general mental health workers and mental health professionals (daly et al., ) , for post-disaster psychological treatment. mental health workers have become a part teams working with the disaster-affected populations worldwide. in this study, mental health workers are the main vehicle for psychological intervention. the effects of shelter preparation, psychological intervention, and logistics on evacuation and resettlement solutions are also examined. compared to studies of emergency management and decision, post-disaster resettlement, evacuation, psychological intervention, and the many other issues related to post-disaster relief and recovery, this study contributes to literature in the following ways. first, the evacuation and temporary resettlement are jointly considered as a multi-step decision problem. a novel post-disaster resettlement flow with five stages is applied to identify the stages of evacuation and temporary resettlement. second, the effects of panic and panic spread among victims on evacuation and temporary resettlement are considered. osuna ( ) showed that the psychological stress that accumulates while waiting is a marginally increasing function of wait times. therefore, this study formulates panic degree as a function of wait times for evacuation. further, panic spread is formulated using the susceptible-infective-removal (sir) model (anderson, ) . in this model, psychological intervention can reduce the degree of panic and panic spread. based on these formulations, the number of victims to be evacuated due to panic spread is derived. third, psychological penalty cost perceived by victims due to wait for evacuation, psychological intervention cost, cost of transporting victims from disaster sites to resettlement sites, and cost of building resettlement shelters are minimized in this formulation. these four costs are utilized to assess the efficacy of evacuation and resettlement solutions. the proposed models and formulated effects of panic and panic spread are characterized and analyzed based on estimated data from wenchuan county after the sichuan earthquake. via parameter sensitivity analysis and pareto analysis of the relationship between psychological penalty cost and monetary cost, the features of the proposed method are elucidated. the remainder of this paper is organized as follows. section briefly reviews relevant literature on temporary resettlement, psychological panic and resource allocation in disasters caused by infectious diseases. section introduces the problem of temporary resettlement impacted by panic-induced psychological panic and panic spread. section formulates a multi-objective optimization model for this problem. section lists the tasks required for acquiring data, and estimating parameter values. section gives numerical results for test scenarios, and findings based on these numerical results are summarized and discussed. section demonstrates the potential advantages and efficacy of the proposed method. this section has several parts. first, after introducing the behavioral and social features of evacuation, shadow evacuations are examined to explain the effects of panic and panic spread on demands for temporary shelters. second, psychological panic is further examined rationally by considering panic in post-disaster evacuation and resettlement processes. third, the principles of panic spread and base models are analyzed and applied to dynamic evacuation and resettlement. ( ) shadow evacuation evacuation research can be grouped into two main categories: behavioral and social science; and modeling and operations (tayfur and taaffe, ) . a fairly large body of research has focused on evacuation planning within behavioral and social science. in modeling and operations literature, researchers have focused mostly on the general population and used roadway infrastructure to move people away from a hazard. many of these researchers proposed operational policies for mass evacuations (daganzo and so, ; guo et al., ) , and studied routing and traffic problems during evacuation, whereas this study treats evacuation as a relief valve for psychological panic as perceived by disaster victims by considering the behavioral and social aspects of evacuation. as mentioned, shadow evacuation is the movement of evacuees who are not required to evacuate. numerous issues associated with shadow evacuation can affect the evacuation process. the shadow evacuation was a planning concern in the houston metropolitan region, which greatly restricted the movement of evacuees from high-risk areas (lamb et al., ) . in fact, this particular problem was underscored by conditions during and after hurricane rita. many shadow evacuees perceived that they were at high risk if they remained. although shadow evacuation has been examined to some degree as a part of the evacuation process, few studies have investigated shadow evacuation as a phenomenon separate from necessary evacuation. modeling techniques are similar to those outlined above. lamb et al. ( ) asserted that evacuation efficiency can be achieved in a variety of ways: minimizing the shadow evacuation and background traffic; sheltering victim in place; and phased (staged or sequenced) evacuation. reducing the magnitude of a shadow evacuation is likely best accomplished through clear communication and education. evacuation decisions are largely products of government bodies feeling that a particular population is at risk of harm. reducing background traffic also involves communication, typically urging people who do not need travel to cancel their trips. because evacuation demand considerations are not directly controlled by transportation agencies, they are seldom explored in transportation literature. this study considered dynamic evacuation demands that are affected by victims' psychology. as temporary shelters are for evacuated victims, shadow evacuations increase demands for shelters. this study makes a connection between post-disaster evacuation and temporary resettlement. temporary resettlement is a time-sensitive service that should be provided for evacuees. the temporary resettlement capacity and replenishment capability should be carefully considered when making post-disaster evacuation decisions. these considerations are based on the observation that a shadow evacuation may be induced by panic and panic spread. ( ) temporary resettlement the temporary resettlement period plays an important role in the economic and psychological recovery of a disasteraffected society (el-anwar and chen, ). the primary goal of emergency response efforts is to provide shelter and assistance to disaster victims as soon as possible (rawls and turnquist, ) . to restore normalcy and complete post-disaster reconstruction, temporary resettlement should proceed quickly. kellett ( ) argued that temporary resettlement not only provides shelters for victims, but also satisfies the functional and social demands of victims. however, temporary shelters have strong negative effects on homeless victims when they are forced to live in temporary shelters for prolonged periods, and often lead to significant social problems such as high unemployment rates, decreased quality of life, and increased crime rates (johnson, ) . therefore, those living in temporary resettlement sites should be transferred to transition resettlement houses and then resettled in permanent structures. el-anwar et al. ( ) proposed a multi-objective optimization model for assignments of large numbers of victims to temporary housing that minimizes social and economic disruption, temporary housing vulnerabilities, and adverse environmental impacts and public expenditures after a disaster. furthermore, el-anwar and chen ( ) determined the computational efficiency of the current socioeconomic model of the temporary resettlement problem, which is formulated as an integer linear programming model. moreover, to solve the model, an algorithm was developed based on the hungarian algorithm. johnson ( ) determined that erecting prefabricated temporary houses can minimize the negative effects associated with resettlement on disaster victims. although post-disaster housing and recovery have been well studied in literature, few researchers have formulated housing demands that are affected by shadow evacuation and the victims' psychology. although victim panic after a disaster is difficult to quantify, its effects on operations and management have been identified. although psychology literature is vast, for example, decision making under time pressure and safety concerns (see adam et al., ,dombroski et al., , few studies are related to evacuations. this may be due in part to that fact that it is often a non-quantitative topic. keating ( ) considered panic as a unique collective phenomenon when fear is the dominant psychological entity of a group. neria et al. ( ) argued that when first learning of an impending natural disaster, people will have negative psychological reactions, such as insecurity, anxiety, and/or fear. armfield ( ) argued that the danger levels associated with a disaster have a major impact on the severity and distribution of panic. some researchers have found indirect evidence of panic, even in the evacuation processes. first, once danger is recognized and people start responding, their information-processing and decision-making capabilities may be confounded by the mentally demanding circumstances, and those pressures are related to the perception of time pressure (adam et al., ) . second, egocentric behavior, a panic-related behavior, has been shown to be quite uncommon during an emergency (bohannon, ) . third, supported by a substantial number of empirical studies (e.g., dash and gladwin, ) , disaster victims do not automatically follow advice and orders from public officials, and tend to seek information, assess personal risk, and make their own evacuation decisions. according to adam et al. ( ) shadow evacuation is primarily motivated by people's perception of being at risk, and shadow evacuation may most commonly occur during mass evacuations, increasing pressure on infrastructure and hindering those who need to evacuate. on the other hand, the response rate for evacuation has a marked non-linear impact on evacuation traffic conditions and arrival patterns, and a higher response rate leads to more traffic on roads, results in traffic congestion. the socio-psychological and circumstantial factors markedly affect individuals' evacuation decisions. ( ) panic spread panic affects victims' evacuation decisions. panic-related emotions can be minimized or spread to others in a social network. one factor influencing evacuation decisions is the strength of a person's social network. based on the assumption that social cues are a causal factor, hasan and ukkusuri ( ) created a social contagion model to investigate the conditions for a cascade through a network of the decision to evacuate. they also considered the effects of social community mixing patterns, the first person to decide to evacuate, and the decisions neighbors made in previous time steps. their model used a mathematical approach and simulation to investigate these factors that bring about the desired behavior of evacuating people in an area based on social relationships. wang et al. ( ) created a qualitative simulation model of a large evacuation system while considering panic spread, and analyzed uncertainty factors that can affect panic spread during the evacuation process. a threshold model of social contagion was developed by hasan and ukkusuri ( ) . this model characterized social influence in decision about evacuation. based on these studies, this study considers the effects of panic spread among disaster victims on temporary resettlement. bi and ma ( ) and zhang et al. ( ) investigated panic psychology during resettlement of disaster victims. bi and ma assessed the effects of two resettlement modes, centralized and non-centralized resettlements, on psychological health, and noted that centralized resettlement mode improved mental health better than decentralized resettlement. zhang et al., who applied statistical methods to analyze factors affecting post-flood anxiety and psychological wellness, determined that timely and effective psychological intervention can minimize anxiety following a disaster. these two studies primarily analyzed the effects of resettlement modes, especially centralized resettlement, on the psychological health of victims; the effect of their psychology on the resettlement process did not get sufficient attention. this study differs from those by bi and ma, and zhang et al. in the following ways. first, they analyzed the effects of resettlement modes on panic and panic spread among victims, whereas this study considered the effects of panic and panic spread on evacuation and temporary resettlement decisions; multi-period evacuation and resettlement decisions are the main focus. second, they quantified the effects by questionnaire-based methods, which contributed to psychological research, whereas this study minimized psychological penalty or cost perceived by victims and three monetary costs by considering the effects of panic and panic spread. another stream of research involves resource allocation problems associated with the rapid spread of infectious diseases (brandeau et al., ; pietz et al., ; ren et al., ; mamani et al., ) . based on these studies, our study considers the following features. the resettlement problem while considering panic spread is formulated as a multi-step evacuation and resettlement problem. next, the time-varying strength of panic, and the time-varying costs and limits of shelters and mental health workers are modeled to study the effects of wait times on evacuation and resettlement solutions. finally, the panic spread in this study is affected by wait times and mental health workers. prevention, preparedness, response, and recovery are the four phases comprising the disaster management lifecycle (quarantelli, ) . quarantelli further divided the recovery period into four stages: instant settlement (within a couple of hours after a disaster); emergency settlement (within one or two days after a disaster); temporary resettlement (within a couple of weeks after a disaster); and permanent resettlement (within a couple of years after a disaster). according to practices in china (lian, ) , before permanent resettlement, a period of transitional resettlement is considered within a couple of months to years after a disaster based on the time needed to construct permanent houses. therefore, this study uses five-stage post-disaster resettlement processes ( fig. ) . after a disaster, emergency rescue resources are distributed and instant settlement sites are chosen and built immediately. shortly after rescue resources arrive, emergency settlement sites and facilities are usually built in safe spaces close to disaster sites. during this stage, young or healthy people take part in rescue operations at disaster sites, and some begin repairing their houses. for issues of water safety, sanitation, water and electricity supply, and other management-related factors, most homeless and injured victims should be evacuated to temporary resettlement sites gradually within about - days after the disaster. during these processes, panic and panic spread markedly affect victim psychology and demands for evacuation. to achieve efficient management of order and evacuation, demands of disaster-affected people should be met as soon as possible. as an indirect evidence of panic spread, some people will return to their homes and repair their houses after living in temporary resettlement shelters for a few days, partially because they were evacuated due to ''shadow evacuation''. this study focuses on the evacuation and temporary resettlement processes, which are enclosed by the grey box in fig. . shelters in the first three stages are typically tents and tarpaulin rooms, which cannot be deployed for extended periods due to issues related to safety and comfort. victims in temporary shelters wait for transition resettlement houses that they can live in for several years. shelters and/or houses for temporary, transitional and permanent resettlement are generally built in successive stages because of shortages in resources and capabilities. therefore, evacuation from one stage to the next stage is dynamic, and conducted step by step. after a disaster, a large number of victims must be evacuated in the temporary resettlement stage from disaster sites (including instant and emergency settlement sites) (denoted as set s that has jsj sites, and are indexed by s) to resettlement sites (denoted as d that has jdj sites, and are indexed by d). however, because resources are limited and lead times for producing and building shelters are long, not all homeless or injured victims can be evacuated immediately to temporary resettlement sites. due to these wait times for evacuation, the strengths of panic and panic spread among victims will rapidly increase. as a result, the number of panicked victims increases, as does the number of those who request evacuation and some young victims and those whose houses can be lived in victims evacuated from disasteraffected or emergency settlement sites to temporary resettlement sites that are relatively far from the disaster-affected sites affected by panic and panic spread, and focused on by this study fig. . five post-disaster settlement and resettlement stages. resettlement. indeed, some victims can remain at the emergency settlement sites close to disaster sites, support rescue processes, and repair their houses ( fig. ) . to examine the dynamic evacuation and resettlement processes, and the effects of wait times on the strengths of panic and panic spread, the temporary resettlement stage is divided into a set of discrete time steps (denoted as t that has jtj time steps, and are indexed by t). the spread of collective panic has a lifecycle comprised of initialization, development, outbreaks, and decline (she and sheng, ) , resembling that of infectious diseases. according to the sir model (anderson, ) , people are divided into three groups: infective, susceptible, and removed people. fig. shows relations among these three groups based on the sir model. initially, homeless victims whose houses have collapsed or are heavily damaged should be evacuated to temporary resettlement sites. however, temporary shelters are scarce and supplied gradually, such that victims are evacuated and gradually resettled at resettlement sites. because of the fear of secondary disasters, panic, especially that among the homeless, which are affected most by the disaster, will aggregate overtime. panic spreads from homeless victims to those that do not need to be evacuated. to keep the order in disaster areas and protect victims, most people affected by panic should be evacuated. ( ) ''infective people'' are victims who must be evacuated to resettlement sites at time t, denoted by x i t;s ; p i ;s denotes the initial number of homeless victims who must be evacuated and resettled. over time, x i t;s is increased by panic spread and decreased by evacuation. ( ) ''susceptible people'' are those who remain at a disaster site s at time, denoted by x s t;s . generally, the houses of susceptible people can still be used because these houses are not unsafe, or these people are healthy and young and can contribute to disaster relief. the initial number of these people is denoted by p s ;s . however, as disaster victims, their psychology is sensitive to the disaster and the willingness to be evacuated from disaster sites. some people in may be transferred to members in x i t;s . ( ) ''removed people'' are the victims evacuated from disaster site s to resettlement site d at time t, denoted as x r t;s;d . these three groups change dynamically as panic spreads (fig. ) . susceptible people become infective people when they are affected by infective people. the number of susceptible people who become infective people is the number of increase of people (nip) due to panic and panic spread at time t, denoted as x nip t;s . the infective people become removed people when they are evacuated to resettlement sites. notably, the nip value, or the transformation of susceptive people to infective people would be controlled by considering some eligibility criteria during on-the-spot operations. selection of these criteria affects the ratio of the nip to the entire group of susceptible people. as the number of infective people x i t;s increases, the probability of susceptible people x s t;s being ''infected'' increases, which increases x i t;s by x nip t;s . thus, x nip t;s is a function of the current number of infective people and susceptible people, denoted as , where x nip t;s is the nip added to infective people at the next time step x i tþ ;s . as b t,s represents panic spread strength, x nip t;s is proportional to b t;s , x i t;s , and x s t;s . based on the sir model, x nip t;s is then computed by eq. ( ). the panic spread strength (speed) is affected by internal and external factors. one internal factor is the panic degree of infective people. the panic degree directly affects the speed of panic spread. the speed of panic spread increases as panic degree increases. the panic degree at time t is denoted by a t . government intervention is an external factor. to maintain social stability, a government can provide psychological assistance for victims by dispatching mental health workers to the disaster areas. these workers can reduce the panic degree and slow panic spread. government intervention degree at disaster site s at time t is denoted by g t,s ; the number of mental health workers dispatched to site s at time t is denoted infective people (to be resettled) by e t,s ; and each mental health worker can provide psychological assistance to r disaster victims during one time step. obviously, g t,s is positively related to e t,s and r and negatively related to the total number of disaster victims, as in eq. ( ). panic spread strength (b t,s ) is a function of panic degree (a t ) and government intervention degree (g t,s ), as in eq. ( ). generally, b t,s is positively related to a t and negatively related to g t,s . the forms of b t,s (Á,Á) are examined in section . further, panic degree of infective people (a t ) is primarily affected by wait times for resettlement. thus, a t is a function of wait times for resettlement (osuna, ) . initially, panic degree of infective people is a constant. as wait times increase, panic degree increases, indicating that wait times and panic strength are positively correlated. moreover, when victims wait for excessively long times, psychological collapse may ensue. thus, panic degree (a t ) is computed by a function of wait times in the numerical studies (section ). the function (a t ) is defined in the data estimation section (section ). when one considers panic and panic spread among disaster victims, two factors that markedly affect temporary resettlement decisions are the nip, and the speed at which temporary shelters are erected at resettlement sites. the number of temporary shelters that can be erected at time t is denoted by z inc t;d . the number of erected temporary shelters directly determines the resettlement capacity, which is denoted by z a t;d . the initial resettlement capacity at site d is denoted by p a ;d . thus, to assess the efficacy of evacuation and resettlement solutions, the following costs are considered. panic-induced psychological penalty perceived by victims when waiting for evacuation is determined by panic degree (a t ) and number is a function of time and sensitive to time. to minimize panic of victims and maintain stability in disaster areas, local governments typically dispatch mental health workers to these areas. these psychological relief actions have allocation and training costs and emergency-related worker pay. the entire cost of one mental health worker serving for one time step at time t is denoted as c p t . three sets, a set of disaster sites, a set of resettlement sites, and a set of time steps, are involved, which are denoted by s, d and t, and indexed by s, d and t, respectively. two groups of known data for disaster site s at the initial time step are the initial number of infective people that should be evacuated to resettlement sites, p i ;s , and the initial number of susceptible people, p s ;s . the unit cost of building a shelter at time t is c inc t . transportation cost for one person from a disaster site s to a resettlement site d, is c t s;d Á p tc , where c t s;d is the distance from s to d, and p tc is the cost for a unit of distance (km). additionally, the cost of a mental health worker serving for a time step at time t is c p t , and the strength of panic spread at time t is denoted by a t . the first group of constraints characterizes the conditions at the initial and end time steps. at the initial time step, the number of infective people x i ;s À x nip ;s and the number of people who are evacuated from disaster areas to resettlement sites x r ;s equal the initial number of homeless disaster victims p i ;s , as defined by eq. ( ) . second, the number of people who should remain at the disaster site x s ;s þ x nip ;s equals the number of susceptible people at the initial time step p s ;s , as defined by eq. ( ) . third, the number of people who are evacuated to resettlement site d at time t is , as is defined by eq. ( ) . fourth, eq. ( ) sets the initial capacity of resettlement site d to p a ;d . the pre-establishment of an adequate capacity and amount of resources enables efficient response to a disaster (salmerón and apte, ). at the last time step (t = jtj), all homeless victims should be evacuated to resettlement sites, meaning that x i t;s and x nip t;s are both , namely, x i t;s þ x nip t;s ¼ for all t = jtj and s. further, by eq. ( ), when x i t;s is , x nip t;s is . therefore, as in eq. ( ), the number of infective people at the last time step (t = n t ) for each disaster site s must be . the second group of constraints comprises flow constraints. first, eqs. ( ) and ( ) the third group of constraints defines x r t;s (eq. ( )) and x r t;d (eq. ( )), and limits the number of the resettled people to the resettlement capacity (eq. ( ) ). furthermore, as a complementary constraint for eqs. ( ) and ( ) restricts the number of people served by mental health workers to the number of disaster-affected people. the fourth group of constraints handles the boundaries and integrities of decision variables. by eqs. ( ) and ( ), the boundaries of the shelter supply and the supply of mental health workers are set. additionally, by eq. ( ), the upper boundaries of the capacity of temporary resettlement sites are constrained. capacity should exceed the number of disaster victims. eq. ( ) indicates that homeless victims must be evacuated and resettled. all variables are non-negative numbers, as denoted by eq. ( ) s:t: eqs: ð Þ-ð Þ; ð Þ-ð Þ the earthquake that hit sichuan, china, on may , , caused numerous casualties and considerable property losses; that is, the earthquake killed , people, injured , , left , people missing, damaged , , houses, and toppled , , more (lian, ) . after the earthquake, about , , people had been relocated. more than , , tents or tarpaulin rooms were dispatched to the disaster sites. the earthquake's epicenter was in wenchuan county, which has a population of , , of which , ( . %) were killed (lian, ). the earthquake also brought psychological harm to residents. according to estimates by the psychology institute of chinese academy of science, more than , people suffered psychological problems. within five days after this earthquake, , people in wenchuan county were evacuated to temporary resettlement sites. after that, many victims were evacuated and resettled. the number of tents or tarpaulin rooms was not revealed in news reports on wenchuan county; neighboring guangdong province had built , movable houses within roughly one month after the disaster (kong, ) . because approximately % rural peasants would return to repair or rebuild their houses after staying at temporary shelters for a couple of weeks, we estimate that about , ( , Á /( À %)) victims, where a house can accommodate on average two victims, were evacuated to temporary resettlement sites after the earthquake. the number of evacuated people exceeds the number of damaged houses. panic and panic spread may account for this difference between number of damaged houses and number of evacuated people (lian, ). ( ) estimate the number of initial infective people p i ;s and susceptible people p s ;s the villages and towns in wenchuan county were disaster sites (with instant and emergency resettlement sites). based on the populations and geographic locations of these sites (fig. ) , two temporary resettlement sites were established in weizhou and yingxiu. according to data published by wenchuan government (http://www.wenchuan.gov.cn), the total number of people before the earthquake and the number of people killed by disaster at each village were summarized by statistical data. the total population at village after the disaster is denoted by p s . then, by using the data from lian ( ) , the ratio of damaged houses in village to the houses before the disaster is estimated, and denoted by. when reported data about the damage ratio differ, the average value is used as the estimation. the values of p i ;s and p s ;s can thus be estimated by p i ;s ¼ l s p s , and p s ;s ¼ p s À l s p s . table presents the statistical and estimated results (see fig. ). ( ) estimate transportation cost c t s;d the distance between a disaster site s and a resettlement site d c t s;d is estimated using the geographic information system (gis) ( table ). the unit transportation cost for one person for a kilometer (p tc ) is set to yuan. ( ) efficiency of a mental health worker (r) one psychological intervention group with one to three workers can serve for about victims per day (wo, ) . this study presumes that one mental health worker can serve roughly victims per day, r = . due to urgent demands for mental health workers and temporary shelters, and the marginally increasing degree of panic perceived by victims while waiting for evacuation and resettlement, three parameters (c p t ; c inc t and a t ) are defined as functions of wait times. reducing the time needed to dispatch mental health workers increases dispatching cost, subsidies, and training costs. the total cost of these types of costs for each worker is represented by c p t at time t. here, c p t is a linear decreasing function of time, and the highest cost is set to yuan, as in eq. ( ), where e p (t) is a monotone increasing function of time t. similarly, c inc t (the cost of increasing a unit of settlement capacity at time t) decreases as a function of time, as in eq. ( ), where e p (t) is also a monotone increasing function of time t. the space in a temporary resettlement shelter for each victim is set to five square meters; thus, resettlement cost per person is set to about yuan and the reference cost at safe times is set to yuan, as shown in eq. ( ). notably, for transition and permanent resettlement, the space allocated for each victim is larger than five square meters. the panic degree a t increases with wait times, and the initial and maximum degrees of panic are set to and , respectively, as in eq. ( ), where e a (t) is a monotone increasing function of time t. a t ¼ minf : Á e a ðtÞ; g to decide e p (t), e inc (t) and e a (t), three governmental officials, three scholars, and three disaster victims, which are all affected by the sichuan earthquake, are interviewed. three base forms, t, t , and e t are used and their figures with descriptions are shown to the interviewees. empirical results indicate that the answers are consistent and the analytical results are used to set the parameters: e t is extreme and too severe; the variances of c p t and c inc t can be described by linear relations (e.g., t), whereas e a (t) is more complex than both; and the forms based on t are used to describe the effects of panic on evacuation and resettlement solutions. a second interview was then conducted by using a curve figure (fig. ) in the questionnaire forms to determine e a (t). no distinct consistent answer was achieved for the question of choosing a property curve for e a (t). the interviewees thought the function values depended on many conditions. however, they believed that the degree affected by panic due to wait times was severe and generally managed well by the communities and local governments. therefore, very serious situations due to panic were rare. according to interviews and investigation results for the disaster areas, reference data for a t ; c p t and c inc t (table ) are basically linearly related to passed time. ( ) time-varying capacities for resettlement and mental health workers temporary resettlement capacities vary with time p zu t and are estimated by news reports (lian, ) . the values are determined by stocks of reserved disaster relief resources, and the capabilities of contracted manufacturers and suppliers. within several weeks after a disaster, many mental health workers p pu t dispatched are not specialized, coming from general hospitals, nearby schools and universities, and governmental and non-governmental organizations (ngos). these people are generally trained and can reach the disaster sites. however, because of the urgency and traveling costs, the number of people is limited. some reference data are set for p zu t and p pu t (table ) . ( ) strength of panic spread as defined in eq. ( ), panic spread strength (b t,s ) is a function of panic degree (a t ) and governmental intervention degree (g t,s ). generally, b t,s is positively related to g t;s and negatively related to g t,s . three typical forms of b t,s (a t ,g t,s ) for a given s are given in eqs. ( ) notably, g t,s = . t is used as the typical function representing governmental psychological intervention degrees that vary over time (fig. ) . however, set a t ¼ t (fig. (a) ), and set a t ¼ t (fig. (b) ). notably, a t Á ( À g t,s ) and a t /( + g t,s ) almost overlap, except for the first point (t = ) ( fig. (a) ). the differences between these two functions shown in fig. (b) are more apparent than those in fig. (a) in their slopes. the curves of a t /g t,s in fig. (a) are not representative. when interviewers are invited to select representative functions, a t Á ( À g t,s ) is comparably better than others when a t = t and a t = t . ( ) initial capacities of resettlement sites by default, p a ;d is set to . by these seven groups of estimates and analyses, the data of h in p (eq. ( )) are determined. for ease of reference, p represents the default settings for the parameters (eq. ( )). this section presents numerical studies, demonstrating the efficacy of the models whose primary known data are given in section . lingo (www.lindo.com) software was used to solve [m ] and [m ] . the purpose of the numerical experiments is to demonstrate application of the proposed methods for evacuation and temporary resettlement of people in wenchuan county after the sichuan earthquake. table summarizes the six experimental scenarios, corresponding purposes, and experimental steps. via these experiments, the proposed models are demonstrated; the effects of parameters on solutions and their sensitivities are analyzed; the consistencies of or conflicts between various costs are quantitatively studied. the analytical results can be used for decision-makings for post-disaster evacuation and temporary resettlement. performance of [m ] or [m ] is examined in each scenario. (fig. ) ( ) demonstrate the effects on f psych , f gint , f trans and f build , and the effects on f p and f c when they have equal weights ( ) use p as the base setting of parameters; ( ) set the weights in w to . ( ) use [m ] and use p as the base settings of parameters ( ) adjust weights (w p = , . , Á Á Á , , and w c = À w p ) to generate test cases and solve each [m ] ( ) draw and analyze the pareto fronts between f p and f c (fig. ) ( ) analyze the effects of the strength of psychological penalty on the solutions by altering e a (t) ( ) use [m ] and use p as the base settings of parameters ( ) set e a (t) to t,t / ,t / , t / and sequentially in the following experiments ( ) set w p = w c = . , and solve [m ] ( ) compare the experimental results of the five functions of (fig. ) based on the results (tables and ; figs. - ) for the six experiments (table ) , experimental results are summarized as follows. ( ) because their values will be changed when the parameters change. table presents values of min(f ⁄ ) and max (f ⁄ ) for the first and second experimental scenarios. ( ) experimental results by minimizing only f psych , f gint , f trans , f build or f c (fig. ) . ( ) by minimizing f psych , the evacuation and resettlement can be completed within the first days. the psychological penalty cost reaches its lowest ( ) after six days, and building cost remains high from day to day . the psychological intervention cost also reaches the allowed maximum within the first days. minimization of psychological penalty cost increases the costs of building shelters and psychological intervention. ( ) the minimum of f gint reaches while psychological penalty cost increases gradually during the first days. minimization of psychological intervention delays building shelters and victim evacuation. therefore, minimizing psychological intervention cost lowers psychological penalty costs and ignores the importance of evacuation and resettlement. ( ) minimizing f trans reduces the nip and the number of infective people who attempt to evacuate. therefore, the number of temporary shelters will also be minimized. psychological penalty cost also reaches its minimum on the sixth day. psychological intervention cost gradually increases during the first days and then declines to the following days. however, the reduction in this value does not minimize building cost. generally, the minimization of transportation cost is paid by the psychological intervention cost and building cost. ( ) the minimization of f build postpones building temporary shelters. before construction of many temporary shelters, psychological penalty cost and psychological intervention cost increase gradually. this increase in monetary costs is also reflected by the increase to the nip and number of mental health workers needed. ( ) minimizing psychological intervention cost has effect similar to those effects of minimizing f build . psychological penalty cost gradually increases over the first days. as building cost increases, psychological penalty cost decreases because more victims can be evacuated to temporary shelters. ( ) when weights of the costs returned by solving [m ] and [m ] are equal, the resulting solutions indicate that building costs have almost the same tendency during the first days, whereas the other three costs (psychological penalty cost, psychological intervention cost, and transportation cost) are distributed differently (fig. ) . when the four costs are equally weighted, psychological penalty cost is high while psychological intervention cost is minimized (fig. ( a) ). the curves of the four costs are explicitly distributed (fig. ( a) ). as indicated by the curves (fig. ( b) and ( b)), many mental health workers are needed during the period from day to day in the results of solving [m ] . when the sum of monetary costs (psychological penalty cost, psychological intervention cost, and transportation cost) is a single objective, the significance of psychological intervention is reflected by the fact that many mental health workers are deployed. however, when the four costs are weighted equally, building cost remains high while psychological penalty cost is also high. ( ) table presents sensitivity test results for eight groups of parameters (p i ;s ; p s ;s ; a; c p ( ) although psychological intervention cost is directly affected by c p t , it has no effect on other costs. ( ) notably, p pu t affects psychological intervention cost mostly, followed by psychological penalty cost. rationally, because the number of mental health workers is limited, sufficient workers in the early stages would be recruited when the recruit cost is high. when p pu t is increased markedly, the psychological intervention method will be used excessively, which can be seen with the large increase in psychological intervention cost, while its effects on other costs are minimal. ( ) decreasing p zu t affects the four costs significantly, especially psychological intervention cost. by decreasing p zu t , decision-makers will use the capacities as many as possible, such that capacities in the early days after a disaster with high building costs may be fully utilized. when p zu t is increased to the extent that resettlement sites can accommodate all disaster infective victims soon after the disaster, no worker is needed to treat the victims at the disaster sites. ( ) although c inc t directly affects building cost, it affects psychological intervention cost more, and affects psychological penalty cost consistently. the effect of c inc t on transportation cost is minor. ( ) transportation cost is markedly affected by c t s;d and also slightly affects psychological intervention cost; however, it does not affect the psychological penalty cost and building cost. ( ) fig. shows the effects of varying the initial stock of shelters at resettlement sites p a ;d on the four costs. notably, this study does not consider the cost for reserving the initial stock of shelters. with the increase in stock, the building cost decreases linearly; psychological intervention cost drops to almost when the stock increases to at each resettlement site; psychological penalty cost drops rapidly when stock reaches and then declines slowly; the varying curve of transportation cost reaches the minimum and then increases very slowly. ( ) by solving the [m ] with different weights for psychological penalty and monetary costs, pairs of the two costs are obtained. fig. shows the distribution of solutions and their pareto font. a decrease of % to psychological penalty cost can be achieved by increasing monetary cost by . %. ( ) fig. shows the effects of different psychological penalty costs associated with wait times on the solutions. here, a linear function (t) and four functions with different coefficients ( / , / , / and ) for squared wait times (t ) are considered and computational results for the four costs are compared. psychological intervention cost, transportation cost, and building cost increase almost linearly when e a (t) is formulated as t, t / , t / , t / and t . however, the increase in psychological intervention cost is fastest. psychological penalty cost increases nonlinearly when e a (t) is sequentially formulated as t, t / , t / , t / and t . when e a (t) is formulated as t / and t , the cost increases rapidly. the discussion of experimental results is generalized as follows. ( ) the costs of psychological penalty, psychological intervention, and transportation and building shelters may be correlated or conflict. first, minimizing psychological penalty cost is consistent the increased costs for building shelters and psychological intervention. second, when psychological intervention is excessively emphasized, the evacuation and resettlement process is delayed. third, although transportation cost is minor when compared to total monetary cost, it is representative of the number of victims affected by panic spread because it is computed based on the number of evacuated victims. fourth, building cost conflicts with psychological intervention cost and psychological penalty cost. further, psychological intervention cost and psychological penalty cost also conflict. fifth, minimization of transportation cost indicates that the number of victims influenced by panic spread is reduced, which in turn increases psychological intervention cost. finally, when psychological penalty cost and the entire monetary cost are considered separately, psychological intervention cost can be used to balance psychological penalty and psychological intervention costs. ( ) of all eight groups of parameters, p i ;s affects solutions most. all costs are related to the number of infective people and the increase in number of infective people. when the value of p i ;s is increased and resources are limited, the costs of psychological intervention and psychological penalty will increase markedly. the initial number of infective people and susceptible people affect the four costs greatly. therefore, these numbers should be reduced as much as possible. although the number of infective people cannot be reduced, some susceptible people can be persuaded to take part in the rescue processes. therefore, mental health workers can focus on infective people. the psychological pressure due to wait times (a) imposes great effects on costs related to victim psychology. ( ) psychological intervention may be overused when the number of available mental health workers is adequate. this overuse does not reduce psychological penalty cost significantly, and it also does not affect building and transportation costs significantly. ( ) expanding the limits of resources (e.g., mental health workers, reserved resettlement resources and replenishment capability for resettlements) can affect psychological intervention cost and psychological penalty cost considerably. and the expansion itself may have a high cost because the probability of disaster occurrences is in fact low. therefore, the increase in replenishment capability is practical for critical relief resources. ( ) although the unit cost of building a shelter affects building cost directly, it affects psychological intervention cost more, and also affects psychological penalty cost. therefore, controlling unit building cost for temporary shelters contributes markedly to cost reductions related to victim psychology. increasing the supply of shelters by emergency supply chain management and logistics is important to post-disaster psychological relief. ( ) in comparing to psychological intervention cost and building cost, the ratio of transportation cost to the overall monetary cost is small. although the ratios affect solutions, they can be neglected. however, when disasters such as earthquakes and landslides occur, road damage may increase transportation costs and evacuation difficulties. ( ) psychological intervention cost is the cost affected most by the initial resettlement capacity because infective people can be accommodated quickly after a disaster. then, the psychological penalty cost and building cost can be markedly reduced. therefore, improving the stock of temporary shelters or providing alternatives can reduce psychological and monetary costs significantly. ( ) by the pareto analytical tool, the tradeoffs between psychological penalty cost and monetary cost can be examined. in the experiments, a small increase to monetary cost ( . %) can cause psychological penalty cost to decrease significantly ( %). therefore, when the budget allows, psychological penalty perceived by victims can be minimized. by extending the proposed methods, the tradeoffs among four costs are revealed. ( ) the magnitude of psychological penalty cost induced by wait times directly affects total psychological penalty cost. however, this magnitude cannot be determined directly according to disaster type or other simple quantification methods. it is affected by the extent of damage, disaster severity, and post-disaster natural and social environments. therefore, minimizing the strength of panic and panic spread is important for evacuation and resettlement, and the recovery of disaster areas. psychological intervention and timely rescue are general ways to reduce the strength of panic and panic spread. this study addressed the post-disaster evacuation and temporary resettlement problem for victims affected by psychological penalty induced by panic and panic spread. psychological penalty due to panic is strengthened by wait times. victim panic and panic spread among victims may increase the number of people to be evacuated to temporary resettlement sites. this problem is formulated with following features. first, the considered period of evacuation and temporary resettlement is divided into time steps. second, panic degree is quantified as a function of wait time, and degree of governmental psychological intervention is determined by the number of dispatched mental health workers, infective people and susceptible people. third, based on the sir model of infectious disease spread, the increase in the number of people who are panicked and need to be evacuated to temporary resettlement sites in each time step is quantified. then, a multiobjective optimization model for evacuation and temporary resettlement was developed with the objectives of minimizing panic-induced psychological penalty, psychological intervention cost, transportation cost, and shelter building cost during evacuation and temporary resettlement. the model is solved by aggregating the objectives into a single objective by weights. the allocation of mental health workers to disaster sites, the shelter building and evacuation solutions are researched by solving the model and considering parameter sensitivities and tradeoffs among the four costs. the following experimental results are obtained. the costs of psychological penalty, psychological intervention, and transportation and building shelters may be affected by the number of victims that evolve from susceptible people to infective people, the number of dispatched mental health workers, timely evacuation, and building shelters. these factors also influence each other. indeed, among all parameters and factors, the initial number of homeless infective people imposes great effects on costs and solutions. psychological intervention apparently prevents the evolution of susceptible people to infective people. however, overuse of psychological intervention yields a big increase in monetary costs and a minor decrease in psychological penalty cost. therefore, timely and efficient evacuation and resettlement are very important. generally, psychological penalty and monetary costs conflict, and can be analyzed by the pareto analytical tool. experimentally, a small increase to monetary cost ( . %) can cause psychological penalty cost to decrease markedly ( %). this study used various functions to represent the strength of psychological penalty cost affected by wait times. the functions should be based on various conditions, such as extent of damage, disaster severity, and post-disaster natural and social environments. psychological intervention and timely rescue are general ways to reducing the strength of panic and panic spread. due to a lack of quantitative research on panic and panic spread during post-disaster relief and recovery, this study proposed a framework that considers the effects of panic and panic spread on post-disaster evacuation and temporary resettlement. therefore, quantification methods could be further researched and incorporate the new developments with practical evidence and theoretical research results in disaster psychology. next, this study formulated evacuation and temporary resettlement as a multi-step optimization model, where parameter uncertainties and conditions are assessed by sensitivity analysis. bell et al. ( ) asserted that the post-disaster transportation network may degrade due to disasters. in post-disaster scenarios, due to secondary disasters and disruptions, robust solutions for evacuation and resettlement should be researched. the evacuation routing problem is complex when one considers the roads adversely affected by disaster. third, this study focused on the temporary resettlement period and does not consider dependencies on other stages. moreover, pre-determined resettlement sites may affect the evacuation and resettlement solutions (rawls and turnquist, ) . incorporating the analytical results of this study, the entire post-disaster relief and recovery processes deserve further research under the consideration of psychological penalty as perceived by victims and rescuers. ( ) set s a set of n s disaster sites, denoted by s = { , , Á Á Á ,jsj} and indexed by s. d a set of n d resettlement sites, denoted by d = { , , Á Á Á ,jdj} and indexed by d. t a set of n t time steps, denoted by t = { , , Á Á Á , jtj} and indexed by t. ( ( ) other notations b t,s panic spread strength representing the speed of panic spread, which affects the nip x nip a review on travel behaviour modelling in dynamic traffic simulation models for evacuations discussion: the kermack-mckendrick epidemic threshold theorem cognitive vulnerability: a model of the etiology of fear depot location in degradable transport networks moderating effects of the community scenario on people's mental health after severe natural disaster: a contrastive analysis of two settlement models during the transition period of the wenchuan earthquake (in chinese) directing the herd: crowds and the science of evacuation resource allocation for control of infectious diseases in multiple independent populations: beyond costeffectiveness analysis managing evacuation networks disaster mental health workers responding to ground zero: one year later evacuation decision making and behavioral responses: individual and household predicting emergency evacuation and sheltering behavior: a structured analytical approach engineering-based hurricane risk estimates and comparison to perceived risks in storm-prone areas computing a displacement distance equivalent to optimize plans for post-disaster temporary housing projects maximizing the computational efficiency of temporary housing decision support following disasters optimizing large-scale temporary housing arrangements after natural disasters the norc studies of human behavior in disaster hurricane andrew: gender, ethnicity and the sociology of disasters route choice in pedestrian evacuation under conditions of good and zero visibility: experimental and simulation results a threshold model of social contagion process for evacuation decision making post-disaster debris reverses logistics management under psychological cost minimization impacts of prefabricated temporary housing after disasters: earthquakes in turkey the myth of panic residential mobility and consolidation processes in spontaneous settlements: the case of santa marta temporary housing construction supported by guangdong province was started at wenchuan county (in chinese) effect of authoritative information and message characteristics on evacuation and shadow evacuation in a simulated flood event wenchuan earthquake cases: emergency part (in chinese) critical behavioral assumptions in evacuation time estimate analysis for private vehicles: examples from hurricane research and planning a game-theoretic model of international influenza vaccination coordination understanding mass panic and other collective responses to threat and disaster post-traumatic stress disorder following disasters: a systematic review the psychological cost of waiting modeling and optimizing the public-health infrastructure for emergency response epidemic of fear: sars and the political economy of contagion in the pacific rim (chapter ) patterns of sheltering and housing in us disasters pre-positioning of emergency supplies for disaster response optimal resource allocation response to a smallpox outbreak a three-stage stochastic facility routing model for disaster response planning the sir epidemic mode-based analysis on the group behaviour under the situation of unconventional accident a model for allocating resources during hospital evacuations qualitative simulation of the panic spread in large-scale evacuation post-disaster psychological crisis research: a survey of the psychological crisis intervention after the sichuan earthquake anxiety and influencing factors of flood victims on pending arrange period the authors would like to thank the national science council of the republic of china, taiwan, for financially supporting this research under contracts nsc - -h- - -my and nsc - -h- - -my . moreover, this study is partially supported by the program for professor of special appointment (eastern scholar) at shanghai institutions of higher learning, shanghai recruitment program of global experts, the national nature science of china ( , , , ), the science foundation of ministry of education of china and shanghai ( , yz , , sg ), and the science and technology commission of shanghai ( zr , ). ( ) notably, p i ;s markedly affects solutions, especially when the values are increased. moreover, among the four costs, psychological intervention cost is affected most. when p i ;s increases by %, psychological intervention cost increases by roughly times. in addition to the psychological intervention cost, psychological penalty cost is affected considerably, especially when p i ;s decreases.( ) although p s ;s affects all four costs, the percentages of all variances for costs, expect for psychological intervention cost, are less than %. comparatively, the psychological intervention cost is affected by p s ;s more than the other costs. when p s ;s decreases by %, the psychological intervention cost will decrease by %.( ) psychological penalty cost and psychological intervention cost are markedly affected by a. the values of a also affect slightly the other two costs (transportation and shelter building costs). the values of a in this study are linearly related to time. sum of the weighted cost of f p and f c , or f psych , f gint , f trans and f build . e p (t) a function of time t, which determines the cost of a mental health worker at time t c p t . e c (t) a function of time t, which determines the cost of increasing a unit of resettlement capacity at time t c inc t . e a (t) a function of wait time, which determines panic degree at time t (a t ). key: cord- -hlyo fys authors: acher, alexandra w.; barrett, james r.; schwartz, patrick b.; stahl, chris; aiken, taylor; ronnekleiv-kelly, sean; minter, rebecca m.; leverson, glen; weber, sharon; abbott, daniel e. title: early vs late readmissions in pancreaticoduodenectomy patients: recognizing comprehensive episodic cost to help guide bundled payment plans and hospital resource allocation date: - - journal: j gastrointest surg doi: . /s - - - sha: doc_id: cord_uid: hlyo fys introduction: previous studies on readmission cost in pancreaticoduodenectomy patients use estimated cost data and do not delineate etiology or cost differences between early and late readmissions. we sought to identify relationships between postoperative complication type and readmission timing and cost in pancreaticoduodenectomy patients. methods: hospital cost data from date of discharge to postoperative day were merged with – nsqip data. early readmission was within days of surgery, and late readmission was to days from surgery. regression analyses for readmission controlled for patient comorbidities, complications, and surgeon. results: of patients included, ( %) were readmitted. the mean early and late readmission costs were $ , ± $ , and $ , ± $ , , respectively. early readmission was associated with index stay deep vein thrombosis (p < . ), delayed gastric emptying (p < . ), and grade b pancreatic fistula (p < . ). high-cost early readmission had long hospital stays or invasive procedures. common late readmission diagnoses were grade b pancreatic fistula requiring drainage (n = , %), failure to thrive (n = , %), and bowel obstruction requiring operation (n = , %). high-cost late readmissions were associated with chronic complications requiring reoperation. conclusion: early and late readmissions following pancreaticoduodenectomy differ in both etiology and cost. early readmission and cost are driven by common complications requiring percutaneous intervention while late readmission and cost are driven by chronic complications and reoperation. late readmissions are frequent and a significant source of resource utilization. negotiations of bundled care payment plans should account for significant late readmission resource utilization. electronic supplementary material: the online version of this article ( . /s - - - ) contains supplementary material, which is available to authorized users. introduction p a n c r e a s c a n c e r i s h i g h l y m o r b i d , a n d p o s tpancreaticoduodenectomy (pd) care is resource intense. the incidence of postoperative complications can be as high as %, , and the incidence of -and -day readmissions is - % and %, respectively. [ ] [ ] [ ] pancreatic fistula (pf) and delayed gastric emptying (dge) are highly associated with prolonged hospitalization and -day readmission. it is generally understood that postoperative complications drive readmission, which in turn increases the cost of care. however, the true cost of care in the perioperative period remains difficult to define. attempts to understand the true cost of cancer care in the us healthcare system are obscured by lack of access to actual costs incurred by hospitals and lack of centralized cost data. the best available cost studies report estimated costs, derived from medicare-based surveillance epidemiology and end results (seer) reimbursement data. [ ] [ ] [ ] these studies demonstrate that the highest estimated cost of pancreas cancer care is associated with treatment that includes surgery, which is more than double that of systemic therapy alone. [ ] [ ] [ ] these estimated cost derivations do not reflect true and specific costs of care delivery. in particular, they do not elucidate the actual cost of specific postoperative complications and readmissionassociated care, especially for readmission that occurs outside of the -day postoperative period. [ ] [ ] [ ] given that post-pd complications and readmission can continue to occur even months after surgery, the impact of this missing data has potentially profound implications on planning and payment/ reimbursement negotiations by hospital fiscal systems. to address this gap in the literature, we sought to delineate the relationship between different types of postoperative complications, the etiologies of early and late readmissions, and the true cost of early versus late readmissions in patients undergoing pd in an academic cancer center. this was a single institution retrospective study. hospital cost data from day of discharge to postoperative day were merged with nsqip data from day of discharge to postoperative day , from to . only patients with complete nsqip and cost data were included in the final analysis. total hospital cost was the sum of indirect and direct variables and fixed costs. direct fixed cost is static, and material costs are associated with running a hospital (buildings, equipment). direct variable costs are costs that depend on specific patient care (medications, procedures). indirect costs reflect hospital infrastructure (financial services, information technology). this cost data does not include physician professional fees. all cost data were generated directly from the electronic medical record system and reflect actual costs incurred by the hospital. high-cost care was defined as perpatient episodic cost within the top quartile of all cost data. early readmission was defined as occurring within days of surgery, and late readmission was defined as occurring between and days from surgery. nsqip definitions were used to categorize delayed gastric emptying (dge), and international study group on pancreatic fistula (isgpf) definitions were used to characterize pancreatic fistula (pf). this analysis focused on identifying risk factors for early and late readmissions and identifying readmission diagnoses associated with high cost. bivariate analysis of risk factors for early and late readmissions was based on preoperative patient factors, surgeon, and type of postoperative index stay complication. preoperative patient factors included in the analysis were age, gender, body mass index (bmi), preoperative albumin, american society of anesthesia physical classification status (asa), any history of hypertension, chronic obstructive pulmonary disease (copd), renal failure, preoperative weight loss greater than % body weight, and reduced functional status. index stay postoperative complications included in the analysis were wound infection, organ space infection, wound dehiscence, urinary tract infection (uti), clostridium difficile infection, pneumonia, pulmonary embolism, stroke, myocardial infarction, cardiac arrest, postoperative bleeding, deep vein thrombosis, dge, pf, and clavien-dindo score. there was a significant and consistent correlation between complication diagnosis and clavien-dindo score. for example, all utis presented as benign complications (clavien- dindo ) , whereas all postoperative pneumonia presented as severe complications (clavien-dindo b- ). therefore, to perform complication-specific analysis, complication diagnosis rather than severity was used in final regression analyses. all patients with index stay mortality were excluded from analysis. bivariate analyses included student's t test, chisquare test, fischer's exact test, and anova. multivariable logistic regression was conducted for early and late readmissions. only independent variables identified on bivariate analysis as significant risk factors for early or late readmission were included in the final multivariable regression analysis. all analyses were conducted with sas version . (sas institute, inc., cary nc, usa), and a p value < . was considered significant. of available patients, patients had an index stay mortality and were excluded; had complete nsqip and cost data and were included in the final analysis. patient demographics were similar for readmitted patients and non-readmitted patients aside from a higher prevalence of copd in readmitted patients (table ) . fifty-eight patients ( %) were readmitted: ( %) as early readmissions, ( %) as late readmissions, and ( %) as both early and late readmissions. the mean early readmission cost was $ , ± $ , , and the mean late readmission cost was $ , ± $ , per patient. forty-one patients were readmitted within days. on bivariate analysis, postoperative organ space infection (p < . ), deep vein thrombosis (dvt) (p = . ), uncontrolled grade b pf (p < . ), and dge (p < . ) were associated with early readmission. importantly, early readmission was not associated with any specific patient comorbidities or surgeon. multivariable analysis demonstrated that early readmission was associated with postoperative dvt (p = . ), dge (p < . ), and inadequately drained grade b pf (p < . ) but not postoperative organ space infection (p = . ) ( table ) . the most common early readmission diagnoses were dge (n = , % of early readmissions) and inadequately drained grade b pf (n = , % of early readmission) (fig. ) . on subset analysis, patients with pancreatic biochemical leaks or adequately drained grade b pfs did not demonstrate an increased risk of readmission compared to patients without pfs or postoperative complications (p > . across all analysis iterations). the mean early readmission cost per patient was $ , ± $ , . as depicted in fig. , the most common early readmission diagnoses were not always associated with the highest readmission costs; rather, high early readmission cost occurred with long hospital stays and/or an invasive procedure ( fig. ; table ). all patients readmitted with partial small bowel obstructions or gi bleeding were in the top cost quartile. in contrast, only % of patients readmitted with organ space infections, grade b pfs, or dvt treatment-related coagulopathy were in the top cost quartile. only % of patients readmitted with dge were in the top cost quartile (table ) . the most common late readmission diagnoses were inadequately drained grade b pf requiring intervention (n = , %), failure to thrive (n = , %), small bowel obstruction requiring operation (n = , %), and organ space infection requiring intervention (n = , %) (fig. ) . the mean late readmission cost per patient was $ , ± $ , . the high-cost late readmissions were related to time-dependent or chronic postoperative issues, some of which required reoperation: recurrent small bowel obstruction requiring reoperation, dge, pf-related enterocutaneous fistula, and organ space infection requiring invasive intervention (fig. , table ). there were patients who had both early and late readmissions. of these patients, patients had related early and late readmission diagnoses. four out of patients were in the top early readmission cost quartile, and out of were in the top late readmission cost quartile. this study is novel in that we quantify the true cost of early and late readmissions and demonstrate the persistent and currently unrecognized fiscal impacts of late readmission after pd. early readmission occurred secondary to the most common postoperative complications (dge and pf), but these complications did not always incur high costs. in contrast, late readmission occurred secondary to complications that required time to manifest into an intervenable issue (time-dependent complications: recurrent partial small bowel obstruction evolving into complete small bowel obstruction, failure to thrive, non-healing chronic wounds). late readmission cost was less varied within individual diagnoses, but when high costs occurred, they were exorbitant. much effort has been dedicated to predicting and preventing post-pd complciations, primarily to benefit patients but also to reduce cost of care. these efforts have centered on preoperative patient optimization, postoperative complication reduction, or improved transitions of care. despite these efforts, reducing common postoperative complications and readmissions has been difficult in pd patients. this suggests that some post-pd complications and readmissions may be inevitable. for example, preoperative cardiac disease and hypertension are patient-dependent risk factors associated with readmission. , prehabilitation programs designed to optimize cardiovascular comorbidities, however, have unclear impacts on postoperative morbidity and readmission. small randomized trials examining the impact of prehabilitation programs have not demonstrated reduction in perioperative morbidity or readmission while results from larger trials are pending. there has also been little progress in preventing common and often costly post-pd complications. meta-analysis and multi-institutional studies demonstrate that infectious complications, failure to thrive, dge, and complication severity drive readmission , , , and therefore represent cost saving opportunities. , despite this knowledge, there has been little progress in the preventing or improving treatment of the most common post-pd complications. for example, results from studies examining prevention and improved management of dge through operation type or postoperative erythromycin have conflicting results. [ ] [ ] [ ] [ ] [ ] [ ] [ ] similarly, studies examining operative strategies to prevent pf have failed to delineate evidence-based recommendations. , [ ] [ ] [ ] while the efficacy of somatostatin analogues to treat pf may be improving with the use of pasireotide, promising results have not been reproduced and pharmacologic intervention is costly. [ ] [ ] [ ] efforts have also been made to reduce post-pd readmissions through transitional care programs. although abbreviations: sbo small bowel obstruction, pf pancreatic fistula, dge delayed gastric emptying, dvt deep vein thrombosis, ecf enterocutaneous fistula, los length of stay these programs have shown promise in some surgical populations, their efficacy in pd patients is less clear. one retrospective study reviewed trends in postoperative morbidity and readmission over a -year period during which discharge coordination and patient education efforts were standardized; these data demonstrated a % reduction in morbidity and a % reduction in -day readmissions over this period. however, these results are inherently biased by study design and conflict with more recent results from a prospectively designed transitional care program. [ ] [ ] [ ] this study was designed to mitigate known clinical and patient-identified risk factors for readmission but failed to demonstrate a reduction in post-pd readmissions within days. [ ] [ ] [ ] these cumulative scenarios suggest that some post-pd complications may be inevitable, at least within current treatment paradigms. it is therefore important for hospital systems to acknowledge that post-pd complication-associated costs can remain significant for months after surgery. the cost trends presented in our study prompt further discussion of the potential for cost containment interventions despite the inevitability of post-pd complications. as apparent in figs. and , there is a wide range in cost for similar readmission diagnoses, which should be further examined for cost containment potential. for example, while postoperative pf and dge were associated with early readmission, they did not always incur high cost. readmission-associated cost of an inadequately drained grade b pf ranged from $ to $ , while that of dge ranged from $ to $ , . undoubtedly, severity of pf and dge is relevant to the cost of care; however, the range in cost could also reflect surgeon- specific management strategies and different thresholds for intervention or outpatient management. additionally, late readmission cost of small bowel obstruction requiring reoperation ranged from $ , to $ , . again, disease severity could explain the difference in cost but other factors should also be considered. the two highest cost patients had recurrent partial small bowel obstructions requiring multiple readmissions before definitive surgical management of complete small bowel obstructions. although the disease course of these patients may have been inevitable, their multiple readmissions before definitive management represent an opportunity for internal review and discussion of management optimization and cost containment strategies. alternatively, in the event that exorbitant cost is non-preventable, these cost data could be used to promote cost-containment standards of care across similar hospital systems. this may help internally and externally regulate pd-specific healthcare costs among hospitals caring for pd patients and provide data for negotiation of bundled payment plans. central to efforts to contain perioperative costs is an understanding that cost can be greatly influenced by not only the cumulative effect of relatively low-cost frequent complications but also the stand-alone effect of high-cost infrequent complications. patients with infrequent complications but exorbitant cost, or "super-users," represent between and % of the population but are responsible for - % of national healthcare expenditures. , researchers and policymakers have therefore debated whether mitigation of high-cost infrequent complications should be prioritized over that of lowcost frequent complications. our analysis demonstrates that within the pd population, both can have a profound impact on the fiscal stability of hospital systems. this is a retrospective small study; however, the trends in complications and readmissions align with findings from larger studies. it was not possible to decipher the breakdown of operative, supportive, or intervention-based costs within the readmission data which made it difficult to determine the cost distribution within a patient's hospital stay. only the cost of inpatient care was obtainable for this analysis; therefore, the total cost of care is underestimated due to lack of outpatient care costs. additionally, only readmissions to the index hospital were captured and the readmission incidence was therefore likely underestimated. early and late readmissions following pancreaticoduodenectomy vary in both etiology and cost. in our studied population, early readmission and cost were driven by common complications requiring intervention. late readmission is driven by timedependent complications, and cost is driven by reoperation and complex medical management requiring long hospital stays. as hospital systems and insurers work toward bundled payment plans for comprehensive episodes of care, it must be recognized that late readmissions for pd occur frequently and are a significant source of resource utilization. additionally, hospital systems should invest in comprehensive, longitudinal cost accounting systems to understand opportunities to prevent exorbitant cost within similar diagnoses. in the event that exorbitant cost is non-preventable, these cost data should be used to promote cost containment standards of care or goals. complications of pancreatic cancer resection readmission after major pancreatic resection: a necessary evil? indication for en bloc pancreatectomy with colectomy: when is it safe? readmission after pancreatic resection: causes and causality pattern pancreatic fistula and delayed gastric emptying are the highest-impact complications after whipple costs and trends in pancreatic cancer treatment the cost-effectiveness of neoadjuvant chemoradiation is superior to a surgery-first approach in the treatment of pancreatic head adenocarcinoma cost-effectiveness of treatment strategies for pancreatic head adenocarcinoma and potential opportunities for improvement distribution of variable vs fixed costs of hospital care acs nsqip operations manual: variables and definitions the update of the international study group (isgps) definition and grading of postoperative pancreatic fistula: years after. surgery classification of surgical complications: a new proposal with evaluation in a cohort of patients and results of a survey hospital readmission after pancreaticoduodenectomy: a systematic review and metaanalysis multivariable analysis of predictors of unplanned hospital readmission after pancreaticoduodenectomy: development of a validated risk score. hpb (oxford) prehabilitation in patients undergoing pancreaticoduodenectomy: a randomized controlled trial prehabilitation in pancreatic surgery -day readmission after pancreatic resection: a systematic review of the literature and meta-analysis factors influencing readmission after pancreaticoduodenectomy: a multi-institutional study of patients management and prevention of delayed gastric emptying after pancreaticoduodenectomy pancreaticoduodenectomy versus pylorus-preserving pancreaticoduodenectomy: the clinical impact no increase in delayed gastric emptying after pyloruspreserving pancreaticoduodenectomy: a meta-analysis of rcts low-dose erythromycin reduces delayed gastric emptying and improves gastric motility after billroth i pyloruspreserving pancreaticoduodenectomy erythromycin accelerates gastric emptying after pancreaticoduodenectomy. a prospective, randomized, placebocontrolled trial delayed gastric emptying: the state of the highest level of evidence clinical trials to reduce pancreatic fistula after pancreatic surgery-review of randomized controlled trials p a u s pancreaticogastrostomy reconstruction for the prevention of postoperative pancreatic fistula following pancreaticoduodenectomy. cochrane database syst rev fibrin sealants for the prevention of postoperative pancreatic fistula following pancreatic surgery pasireotide for postoperative pancreatic fistula pasireotide does not prevent postoperative pancreatic fistula: a prospective study prophylactic pasireotide administration following pancreatic resection reduces cost while improving outcomes transitional care interventions and hospital readmissions in surgical populations: a systematic review reducing readmissions after pancreatectomy: limiting complications and coordinating the care continuum adaptation and implementation of a transitional care protocol for patients undergoing complex abdominal surgery improving patient-centered transitional care after complex abdominal surgery adaptation of a transitional care protocol to reduce readmissino after complex abdominal surgery agency for healthcare research and quality. the high concentration of us health care expenditures superusers: drivers of health care resource utilization in the national trauma population is a strategy focused on super-utilizers equal to the task of health care system transformation? yes presentation at scientific meeting the abstract for this work was selected for quick shot presentation at the ssat/ddw conference publisher's note springer nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations authors' contributions alexandra w. acher md: concept design, data acquisition, analysis and interpretation, manuscript writing and critical revision, final approval, final accountability agreement.james r. conflict of interest the authors declare that they have no conflict of interest. key: cord- -g l s x authors: flyvbjerg, bent; budzier, alexander; lunn, daniel title: regression to the tail: why the olympics blow up date: - - journal: nan doi: . / x sha: doc_id: cord_uid: g l s x the olympic games are the largest, highest-profile, and most expensive megaevent hosted by cities and nations. average sports-related costs of hosting are $ . billion. non-sports-related costs are typically several times that. every olympics since has run over budget, at an average of percent in real terms, the highest overrun on record for any type of megaproject. the paper tests theoretical statistical distributions against empirical data for the costs of the games, in order to explain the cost risks faced by host cities and nations. it is documented, for the first time, that cost and cost overrun for the games follow a power-law distribution. olympic costs are subject to infinite mean and variance, with dire consequences for predictability and planning. we name this phenomenon"regression to the tail": it is only a matter of time until a new extreme event occurs, with an overrun larger than the largest so far, and thus more disruptive and less plannable. the generative mechanism for the olympic power law is identified as strong convexity prompted by six causal drivers: irreversibility, fixed deadlines, the blank check syndrome, tight coupling, long planning horizons, and an eternal beginner syndrome. the power law explains why the games are so difficult to plan and manage successfully, and why cities and nations should think twice before bidding to host. based on the power law, two heuristics are identified for better decision making on hosting. finally, the paper develops measures for good practice in planning and managing the games, including how to mitigate the extreme risks of the olympic power law. faced by cities and nations in deciding whether to host the games or not, and the implications of those risks. with the present paper we aim to fill this gap. first, we present our data for olympic costs and cost overruns since . second, we fit different theoretical statistical distributions to the data. finally, we interpret the best fits in terms of what they say about cost risk and how such risk may be mitigated and managed. annex contains more detail about previous academic research on cost and cost overrun at the games and how this relates to the present research. we measure olympic costs as actual outturn sports-related costs of hosting the games. non-sports-related, wider capital costs are not included, although they are typically substantial. cost overruns are measured as actual outturn cost in percent of estimated costs. for example, if a games was estimated to cost $ billion and actually cost $ billion, then these games incurred a cost overrun of percent. the baseline for measuring cost overrun is the cost estimate at the time of bidding for the games. all costs are measured in real terms, i.e., not including inflation. all games from to , for which data were available, are included, which is out of games for outturn cost and out of games for cost overrun. it is an interesting finding in its own right that for out of games valid and reliable data on cost overrun could not be found, despite trying hard. incredible as it may sound, for more than a third of games since no one seems to know what estimated and actual costs were. nevertheless, the dataset is the largest and most consistent of its kind, allowing statistical analysis. annex contains more detail about data and methodology. table shows actual outturn sports-related costs of the olympic games - together with the number of events and number of athletes in each games. [ table average cost for summer games - is $ . billion (median $ . billion). average cost for winter games over the same period is $ . billion (median $ . billion). the large difference between average and median cost for the winter games is mainly caused by the extreme cost of sochi , which at $ . billion cost more than all previous winter games combined. indeed, the sochi winter olympics are the most costly games ever, even when compared with the summer games. this is extraordinary, given that cost for the winter games is typically much lower than for the summer games, with the median cost for winter games being less than half the median cost for summer games. we will return to the importance of extreme values below as it turns out these are typical of the games to a degree where they define the particular cost risk profile of the games. average cost for all games games - . average cost for the five games held in the decade - is $ . billion (median $ . billion). it should again be remembered that wider capital costs (ocog indirect costs, see annex ) for urban, transportation, and business infrastructure are not included in these numbers and that such costs are typically substantial. figure shows the development of cost - . the trend lines indicate that the cost of the games have increased over time. conventional statistical analysis shows that the apparent increase is statistically overwhelmingly significant (p < . , winter olympics; p < . , summer olympics). however, due to the existence of extreme values in the data, documented below, results of conventional statistical analysis must be interpreted with caution. [ figure app. here] table [ table app. here] for cost per athlete, we found the winter games to be twice as costly as the summer games. the average cost per athlete is $ , for the summer games (median $ , ) and $ . million for the winter games (median $ , ). the highest cost per athlete in the summer games was found for london at $ . million. for the winter games, the highest cost per athlete was found for sochi at $ . million. the lowest cost per athlete in the summer games was found for tokyo at $ , , and in the winter games for innsbruck at $ , . figure shows the correlation of cost per athlete with time. we see a shift in trend from cost per athlete being generally higher for the summer than for the winter games until the mid 's, after which the winter games become more costly than the summer games, in terms of cost per athlete. we also see that cost per athlete was generally decreasing for the summer games from the mid- 's until the early noughties, after which cost per athlete has been increasing for both the summer and winter games, driven mainly by london and sochi . [ figure app. here] table shows percentage cost overrun in real terms (not including inflation) for the olympic games - . data on cost overrun were available for of the games - . statistical tests of the difference between bid budgets and final costs show this difference to be statistically overwhelmingly significant (v = , p < . ) . that is to say, cost overruns are statistically overwhelmingly manifest for the olympics. it should be mentioned that if the cost overruns had been calculated in nominal terms (including inflation) they would have been significantly larger. in this sense the numbers shown are conservative. [ table app. here] we find the following averages and medians for cost overrun in real terms: • summer games: average cost overrun is percent (median percent). • winter games: average cost overrun is percent (median percent). • all games: average cost overrun is percent (median percent). even though the difference between average cost overrun for the summer and winter games is relatively large at percentage points, the difference is statistically non-significant (non-parametric test, w = , p = . ). in statistical terms there is therefore no difference between cost overrun in the summer and winter games and the data may be pooled for statistical analyses, for instance in making more accurate reference class forecasts of budgets for future olympic games (flyvbjerg ) . we further see that: • of games ( percent) have cost overruns above percent. • of games ( percent) have cost overruns above percent. judging from these statistics it is clear that large risks of large cost overruns are inherent to the olympic games. in the next section we will explore further what type of risk is at play and what it entails for the decision to host the games. for the summer games, the largest cost overrun was found for montreal at percent. the smallest cost overrun for the summer games was found for beijing at two percent. for the winter games, the largest cost overrun was for lake placid at percent and the smallest for vancouver at percent. the vigilant reader may be skeptical that the lowest cost overrun of all games would be for beijing at two percent. china is known for its lack of reliability in economic reporting (koch-weser ). however, the total sports-related cost of $ . billion and the cost per athlete of $ , for the beijing games are higher than for the majority of other summer games (see tables and ). the reported costs are therefore deemed adequate for hosting the beijing games and we have no evidence that the official numbers have been manipulated and should be rejected. like other observers of economic data from china we therefore include the numbers, with the caveat that they are possibly less reliable than those from other nations, given the lack of transparency in chinese economic data. again, this means that our averages for cost overrun in the games are likely to be conservative. in sum, the financial and economic risk of hosting the olympic games holds policy and political consequences because: . all games, without exception, have cost overrun. for no other type of megaproject is this the case. for other capital investment types, typically - percent of investments come in on or under budget. for the olympics it is zero percent. it is worth considering this point carefully. a budget is typically established as the maximumor, alternatively, the expected -value to be spent on a capital investment. however, in the games the budget is more like a fictitious minimum that is consistently overspent. further, even more than in other megaprojects, each budget is established with a legal requirement for the host city and government to guarantee that they will cover the cost overruns of the games. our data suggest that this guarantee is akin to writing a blank check for the event, with certainty that the cost will be more than what has been quoted. we call this the "blank check syndrome." in practice, the bid budget is more of a down payment than it is a budget; further installments will follow, written on the blank check. . the olympics have the highest average cost overrun of any type of megaproject, at percent in real terms. to compare, flyvbjerg et al. ( ) found average cost overruns in major transportation investments of percent for roads, percent for large bridges and tunnels, and percent for rail; ansar et al. ( ) found percent overrun for megadams; and budzier and flyvbjerg ( ) percent for major it investments, all in real terms and measured in the same manner (see table ). the high cost overrun for the games may be related to the fixed deadline for delivery: the opening date cannot be moved. therefore, when problems arise there can be no trade-off between schedule and cost, as is common for other megaprojects. all that managers can do at the olympics is to allocate more money, which is what happens. this is the blank check syndrome, again. . the high average cost overrun for the games, combined with the existence of extreme values, should be cause for caution for anyone considering to host the games, and especially small or fragile economies with little capacity to absorb escalating costs and related debt. even a small risk of a + percent cost overrun on a multi-billion dollar investment should concern government officials and taxpayers when a guarantee to cover cost overrun is imposed, because such overrun may have fiscal implications for decades to come, as happened with montreal where it took years to pay off the debt incurred by the percent cost overrun on the summer games (vigor et al. : ) , and athens where olympic cost overruns and related debt exacerbated the - financial and economic crises, as mentioned above (flyvbjerg ) . [ table app. here] to identify more specifically the type of risk involved in bidding for the olympics, we fitted different theoretical statistical distributions to the data. we did this for data on cost overrun, bid cost, outturn cost, and cost per athlete, all presented above. here we focus on cost overrun, because this is the most critical variable for potential host cities and nations in deciding whether they can afford to host the games and would be likely to stay on budget or not. it should be mentioned, however, that results for the other variables are similar. we tried all possible theoretical frequency distributions and found that only fat-tailed distributions fit the data, i.e., distributions with many extreme values. thin-tailed distributions can be rejected, whereas fat-tailed ones cannot. power law and lognormal distributions were found to best fit the data (see figure ). [fig. app. here] power laws show potentially infinite variance and a volatile or nonexistent mean. as a consequence of the fattailedness of the data we cannot reliably establish the presence of time trends or compare means between subgroups of data, e.g., summer games v. winter games; the extreme values in the fat tail affect the results too much for this type of approach to be feasible. in essence, with fat tails the law of large numbers is not fast enough to give a reliable idea of the mean. this problem is exacerbated by the fact that each of the summer and winter games are held only every four years, so by the nature of their setup a large number of games are not and cannot be available for study. we examined in detail three models that fit the data on cost overrun particularly well, one lognormal and two power laws. first, the maximum likelihood estimator (mle) for the lognormal model gives μ = . and σ = . . while conventional thinking is that a lognormal distribution is thin-tailed, because it has all the moments (mean, variance, etc.), it actually behaves like a power law at a σ > . (taleb : ff.) , which is the case for our olympic costs dataset. the kurtosis for the lognormal model is . . the probability of costs being three times estimates is . percent. second, the maximum likelihood estimator for the first power law -the pareto , a simplified power law -finds λ = . (the minimum value in the sample) and the power law coefficient α = . . the advantage of the simplified power law is the estimation of a single parameter, α. finally, the second power law uses three parameters and for this model we find α ≈ . , which indicates thinner, but still fat, tails. we observe that < α < for both models, which means that the first moment (mean) is finite and can therefore be estimated, whereas the second moment (variance) is infinite and therefore cannot. if α were smaller than one the tail would be so fat that both mean and variance would be infinite and neither could be estimated. the controversy over the power-law fittings done by barabási et al. ( ) highlights that careful consideration must be observed when fitting and comparing fat-tailed distributions so as to not overstate the case (stouffer et al. , barabási et al. , shalizi , clauset ). to guard against the pitfall of overstatement, we tested how the power-law models compare with the lognormal model in fitting the data. for this, we used vuong's test, which is a likelihood ratio test for model selection using the kullback-leibler criteria. the null hypothesis is that both classes of distributions (power law and lognormal) are equally far from (or near to) the true distribution. the test statistic, r, is the ratio of the log-likelihoods of the data between the two competing models. the sign of r indicates which model is better. the size of r indicates the likelihood of the null hypothesis being true or false, with r going to ±infinity with probability if the null hypothesis is false, indicating that one type of distribution is closer to the true distribution than the other. using this test, we found r = - . with p = . , which favors the power law fit, however not at a level that allows us to reject the null. following clauset et al. ( ) we then used their iterative approach to search for values of λ that would improve the power law fit. we found λ = . and α = . . the mle fit of the lognormal distribution for the same λ gives lnn( . ; . ). vuong's likelihood ratio test now results in r = - . with p = . . at this higher starting point of the tail the model comparison can again not reject the null that either distribution (power law and lognormal) fits the data. we conclude that both the power law distribution and the lognormal distribution fit the data on olympic cost overrun and that both distributions are fat tailed. the findings translate into the following fast-andfrugal, practical heuristic for cities and nations trying to decide whether to host the games or not: heuristic no. : can we afford and accept a percent risk of a three-fold increase or higher in cost in real terms on the multi-billion-dollar expenditure for the olympics? if the answer to this question is yes, then proceed and become a host; if the answer is no, walk away. figure shows eight states of randomness, from zero ("degenerate") to infinite ("a £ "). thin-tailed distributions, with low randomness, are at the bottom of the figure. fat-tailed distributions, with high randomness, at the top. the eight states are somewhat similar to mandelbrot's ( ) classic categories of mild, slow, and wild randomness, but are statistically more detailed with regards to the implications for modelling risks. [ figure app. here] we saw above that olympic cost overruns follow fat-tailed distributions, specifically lognormal and power-law distributions with an alpha-value of . to . , i.e., smaller than and larger than . this is a highly significant finding to anyone who wants to understand how olympic costs work, and especially to anyone considering to host the games concerned about affordability and financial risks. comparing the established alpha-values with figure , we see that olympic cost overruns belong to the second-most extreme category in terms of randomness, here called "lévy-stable" after lévy processes in probability theory ( < a < ). distributions in this category are characterized by infinite variance and finite mean, whereas for the most extreme category (a £ ) variance and mean are both infinite, i.e., non-existent. with this finding we arrive at a basic empirical explanation of olympic cost blowouts that has not been uncovered before. such events are not just the unfortunate, happenstance incidents they appear to be, that are regrettable but will hopefully be avoided in the future, with more awareness and better luck. instead, olympic cost blowouts are systematic, ruled by a power law that will strike again and again, with more and more disastrous results. following power-law logic, it is just a matter of time until an event, here cost overrun, will occur that is even more extreme, with a larger cost overrun, than the most extreme event to date. we call this phenomenon "regression to the tail" and contrast it with "regression to the mean." sir francis galton coined the term regression to the mean -or "regression towards mediocrity," as he originally called it. it is now a widely used concept in statistics, describing how measurements of a sample mean will tend towards the population mean when done in sufficient numbers, although there may be large variations in individual measurements. galton illustrated his principle by the example that parents who are tall tend to have children who grow up to be shorter than their parents, closer to the mean of the population, and vice versa for short parents. in another example, pilots who performed well on recent flights tended to perform less well on later flights, closer to the mean of performance over many flights. this was not because the pilots' skills had deteriorated, but because their recent good performance was due not to an improvement of skills but to lucky combinations of random events. regression to the mean has been proven mathematically for many types of statistics and is highly useful for risk assessment and management, in, e.g., health, safety, insurance, casinos, and factories. but regression to the mean presupposes that a population mean exists. for some random events of great social consequence this is not the case. size-distributions of earthquakes, wars, floods, cybercrime, bankruptcies, and it procurement, e.g., have no population mean, or the mean is ill defined due to infinite variance. in other words, mean and/or variance do not exist. regression to the mean is a meaningless concept for such distributions, whereas regression to the tail is meaningful and consequential. regression to the tail applies to any distribution with non-vanishing probability density towards infinity. the frequency of new extremes and how much they exceed previous records is decisive for how fat tailed a distribution will be, that is, whether it will have infinite variance and mean. above a certain frequency and size of extremes, the mean increases with more events measured, with the mean eventually approaching infinity. in this case, regression to the mean means regression to infinity, i.e., a non-existent mean. deep disasters -e.g., earthquakes, tsunamis, pandemics, and wars -tend to follow this type of distribution. we call this phenomenon -extreme events recurring in the tail, with events more extreme than the most extreme so far -"the law of regression to the tail" (flyvbjerg forthcoming). we show that cost overrun for the olympic games follow this law. table compares the alpha-values for the olympic games with alpha-values for other events that follow power-law distributions. we see that with alpha-values between and , indicating infinite variance, olympic cost overruns fall in the same category as terrorist attacks, forest fires, floods, and bankruptcies in terms of fat-tailedness. [ table app. here] a further consequence of power-law behavior and regression to the tail, is that the sample mean cannot be trusted: even a single extreme value in the fat tails may significantly alter the mean. however, knowing that the power law distribution and the lognormal distribution fit the olympic cost data, we may use this knowledge to estimate the true (population) mean of the distributions, following taleb and cirillo ( ) . table shows the results. the sample mean, as we saw above, is . , equivalent to an average cost overrun of percent. the true mean, based on the lognormal distribution, is . , i.e., practically the same as the sample mean. based on the two pareto distributions, we find a true mean of . and . , respectively, i.e., substantially higher than the sample mean. the best estimate of the true mean of cost overrun in the olympics is therefore - percent. this finding gives us further detail to heuristic no. above and translates into a second fast-and-frugal heuristic that cities and nations may find useful when faced with deciding whether to host the games: heuristic no. : can we afford and accept an expected cost overrun in the range of - percent in real terms on the multi-billion-dollar expenditure for the olympics, with substantial risk of further overrun above this range? if the answer to this question is yes, then proceed and become a host; if the answer is no, walk away, or find effective ways to "cut the tail," i.e., reduce tail risk. [ table app. here] what is the mechanism that generates the power-law distribution for olympic cost and cost overrun? what are the implications? and would it be possible to mitigate the impacts of the power law? we address these questions below. at the most basic level, power laws are generated by non-linear amplification of random events, including infinite amplification (farmer and geanakoplos ) . in other words, when the value of an underlying variable changes, say the scope of the games, then the resulting output -say, cost -does not change linearly, but instead depends on the derivative. the change is accelerated, so to speak. geometrically, the relationship between variables is curved instead of linear. the degree of curvature is called convexity. the more curved, the more convexity, and the more amplification of random events. at the level of root causes, convexity is the primary mechanism that generates power laws, including the olympic cost power law. convexities are known to lead to serious financial fragilities (taleb ). convexity appears to be severe for olympic cost overruns, as evidenced by the alpha-values shown in table , documenting infinite variance. we find six reasons for this. first, hosting the games is a particularly difficult decision to reverse. this means that when scope and costs begin to escalate -as they have for every games on record -hosts generally do not have the option of walking away, as they do with most other investments, even should they think this would be the best course of action. in fact, denver is the only host city to ever have abandoned the games, in , after winning the bid to host them. instead, lock-in is near-absolute and hosts are forced to throw good money after bad, which is the type of behavior that leads to strong convexity in spending. second, not only is it difficult to reverse the decision to host the games, there is also no option to save on costs by trading off budget against schedule, because the timetable for the olympics is set in stone. the bid is won seven years in advance of the opening ceremony, the date of which was decided even earlier and cannot be moved. for other types of megaprojects, trading off budget against schedule is a common and often effective mechanism to dampen cost escalation. for the games, this mechanism is not available, again reinforcing convexity and power-law outcomes. third, there is the legally binding obligation, mentioned above, that the host must cover possible cost overruns. this means that the ioc has no incentive to curb cost overruns, but quite the opposite, as the ioc focuses on revenues, from which their profits derive, and some costs drive revenues and thus ioc profits. the host, on the other hand, has no choice but to spend more, whenever needed, whether they like it or not. this is the blank check syndrome mentioned above. the blank check is, in and of itself, an amplification mechanism that generates convexity and power law outcomes for olympic cost and cost overrun. things do not need to be like this. the ioc could choose to lead on costs instead of self-servingly focusing on revenues. but so far they do not. together, the first three points make for a clear case of strong convexity and go a long way in explaining why olympic costs are fat tailed. the way the games are set up, including their contracts and incentives, virtually guarantees that strong convexity and power-law outcomes ensue. but there is more. fourth, tight constraints on scope and quality in the delivery of investments are known to be an additional driver of convexity and power law outcomes (carlson and doyle ) . together with the immovable opening date for the games, mentioned above, a program scope that is rigorously defined by the many sports and events hosted, and design standards that are set in detail by the ioc and individual sports associations are examples of tight and non-negotiable constraints that set the olympics apart from more conventional megaprojects where trade-offs between budget, schedule, scope, and quality have wider margins. for the olympics such margins are zero or close to zero. for example, the delivery authority does not get to negotiate the standards for running the meter or for bob sleighing. they are givens, just like the opening date, as far as delivery is concerned. this means that staging the games can be conceived as a highly optimized system with budget uncertainty at the macro level, while constraints are exceptionally tight at the micro level. such systems have been demonstrated to be fragile to design flaws and perturbations and subject to convexity and power law outcomes. moreover, constraints and their effects are exacerbated by size, which the ioc and hosts should keep in mind as the olympics grow ever bigger (taleb : - ) . in conventional investment management, constraints are softened by the use of contingencies. so too with the olympics where bid budgets typically include reserves. for example, the bid for london included a . percent contingency. this proved sorely inadequate in the face of a percent cost overrun in real terms, i.e., an amplification times higher than the contingency allowed for. ten years later, with the bid for beijing , not much seemed to have changed. here the budget included a . percent contingency for unanticipated expenses, which, according to the ioc "is in line with the level of risk and the contingency for previous games" (ioc : ) . whether in line with previous practice or not, such contingencies are sorely inadequate, and table demonstrates they are not in line with the level of risk for previous games, as falsely claimed by the ioc. more than for any other type of megaproject, the contingency mechanism -which is crucial to effective delivery -fails for the olympics. when contingencies run out, as they have done for every olympics on record, such failure typically leads to what we call the "vicious circle of cost overrun," where top management gets distracted from delivery, because they are now forced to focus on pressing issues of negative media coverage about lack of funds, reputational damage, and fund raising aimed at closing the contingency gap. as a consequence, delivery suffers, which leads to further cost overruns, which lead to further distractions, etc. the vicious circle is caused in part by the exceptionally tight constraints that apply to delivery of the games and the high levels of contingency that are needed, but not available, to soften such constraints. again the consequence is higher convexity and power law outcomes. fifth, the longer the planning horizon the higher the variance of random variables and the more opportunities for random events to happen that may get amplified to power law outcomes (makridakis and taleb , farmer and geanakopolos ) . time is like a window. the longer the duration, the larger the window, and the greater the risk of a big, fat black swan flying through it, which tends to happen for the olympics, with its overincidence of extreme values documented above. some of these black swans may be deadly, in the sense that they kill plans altogether, as happened for the tokyo olympics when the covid- pandemic made it impossible to host the games in as planned. this resulted in billions of dollars of further cost overruns on top of billions already incurred, making the fat tails for olympic cost even fatter. by design, staging the olympics comes with a long planning horizon, specifically seven to eleven years. this is the length of an average business cycle in most nations. it should therefore come as no surprise that the price of labor and materials, inflation, interest rates, exchange rates, etc. may change significantly over a period this long and impact cost and cost overrun. cities and nations typically bid for the games when the economy is thriving, with the consequence that more often than not the business cycle has reversed to lower growth when the opening date arrives seven to eleven years later, as we saw in figure . furthermore, scope changes generated by random events will be the more likely to happen the longer the planning horizon, e.g., terrorist attacks that push up security standards and costs at the games, as has happened repeatedly. interestingly, the severity of terrorist attacks, measured by number of deaths, follow a power law distribution, just like olympic costs. this is an example of one power law (number of deaths from terrorist attacks) directly driving another power law (cost and cost overrun at the games). this is archetypical amplification and strong convexity at work. in addition to the risk of underestimating the true variability of forecasts, long planning horizons fundamentally change the nature of risk encountered. power laws arise when clustered volatility and correlated random variables exists in a time series, as has been observed for financial data (mandelbrot , gabaix ). such time series are dominated by random jumps instead of the smooth random walks often assumed by analysts and forecasters (mandelbrot ) . and random jumps -e.g., a sudden increase in the price of steel, a main ingredient of the olympics, or a jump in security costs triggered by a terrorist attack -lead to power law outcomes. the longer the planning horizon the more likely that random jumps will happen. existing forecasting techniques cannot deal with outcomes like these, if at all, beyond planning horizons of approximately one year (blair et al. , gabaix , beran , tetlock and gardner . in hydrology, mandelbrot and wallis ( ) called the rare instances of extreme precipitation that occur over the long term -often after long periods of low precipitation -the "noah effect." the coining of the term was inspired by the empirical observations of the fluctuations of the water level in the nile river by harold edwin hurst, but the effect works beyond hydrology. based on his observations, hurst increased the height of the planned aswan high dam far beyond conventional forecasts, and the dam was designed and built following his recommendations. hurst did not try to forecast the exact maximum level of water in the dam basin, which would have been in vain, and would likely have led to underestimation. instead he increased the height of the dam wall to take into account extreme values far beyond his observations. this is exactly what power laws dictate as the right course of action, because according to the power law it is only a matter of time until an event occurs that is more extreme than the most extreme event to date. unlike hurst, cost forecasters at the olympics fail to understand the power-law nature of the phenomenon at hand and therefore fail to increase their safety margins, with the consequence that they have underestimated costs for every olympics on record. sixth and finally, the problems described above are compounded by a phenomenon we call the eternal beginner syndrome. if, perversely, one wanted to make it as difficult as possible to deliver a megaproject to budget and schedule, then one would (a) make sure that those responsible for delivery had never delivered this type of project before, and (b) place the project in a location that had never seen such a project, or at least not for the past few decades so that any lessons previously learned would be either obsolete or forgotten. this, unfortunately, is a fairly accurate description of the situation for the olympics, as they move from nation to nation and city to city, forcing hosts into a role of "eternal beginners." it is also a further explanation of why the games hold the record for the highest cost overrun of any type of megaproject. inexperienced beginners are more prone than experienced experts to underestimate challenges and are less well suited in dealing with unexpected events when they happen. this means that such events spin out of control and amplify more easily for beginners than for experienced experts, which again contribute to convexity and the power-law outcomes we see in the data. a mistake made by eternal beginners is to assume that delivering the olympics is like delivering a scaled-up but otherwise conventional construction program. this emphatically is not the case. rio , for example, hosted sports with events in venues. tight coupling of the deliverables, fixed deadlines, and inadequate budget contingencies form a system that amplifies the impacts of random adverse events (dooley ) . the eternal beginner lacks experience in what this kind of scale and constraints mean in terms of increased delivery risks and therefore underestimates these. the lack of experience is aggravated by the fact that conventional approaches to estimating and containing risk do not work in this kind of system, as argued above. in sum, we find: (a) convexity is the root cause of the power-law nature and extreme randomness of cost and cost overrun for the olympics; (b) convexity is strong for the games, documented by alpha-values smaller than , indicating infinite variance; and (c) convexity at the games is driven by irreversibility, fixed time schedules, misaligned incentives, tight coupling, long planning horizons, and the eternal beginner syndrome. power-law fat tails -or extreme randomness -is the most challenging type of risk to manage, because it is maximal, unpredictable, and difficult to protect against via conventional methods. however, this is the type of risk that the olympics face in terms of cost, as shown above. what can be done to manage this risk intelligently? first, and most importantly, the ioc and potential hosts must understand the existence of fat tails as a matter of fact, i.e., that hosting the games is extremely risky business in terms of cost. if they do not understand the risk, and its particular power-law nature, they cannot hope to effectively protect themselves against it. today, there is some understanding of risk with the ioc and hosts, but nothing that reflects the real risks. instead of extreme randomness, the ioc assumes low randomness when it states that a . percent contingency for unanticipated expenses is in line with the level of risk for previous games, as we saw above. this number is glaringly insufficient when compared with actual cost overrun in the most recent games, which was percent for rio , percent for sochi , and percent for london , or when compared with the average cost overrun for all games since , which is percent. either the ioc is deluded about the real cost risk when it insists that a . percent contingency is sufficient, or the committee deliberately overlooks the uncomfortable facts. in either case, host cities and nations are misled, and as eternal beginners it is difficult for them to protect themselves against such misinformation. they understandably do not know what the real numbers are, because they have no experience in delivering the games. independent review of any cost and contingency forecast is therefore a must, including for estimates from the ioc. as said, hosts must understand the risk to be able to protect themselves against it. such understanding is therefore a necessary first step for mitigating cost risk at the games, and the ioc should be held accountable for misinforming hosts about the real risks under rules similar to those in corporate governance that make it unlawful for corporations to deliberately or recklessly misinform shareholders and investors. second, once the real risks are understood it becomes immediately clear that larger cost contingencies are needed for the games. reference class forecasting, based on kahneman and tversky's ( ) work on prediction under uncertainty, has been shown to produce the most reliable estimates of any forecasting method, mainly because it eliminates human bias and takes into account unknown unknowns (kahneman : , flyvbjerg , batselier and vanhoucke , chang et al. , horne . further de-biasing should be carried out, following kahneman et al. ( ) and flyvbjerg ( ) . based on the dataset presented above, reference class forecasting would produce a significantly more realistic estimate of the necessary cost contingencies for the olympics than the numbers put forward by the ioc. more realistic contingencies would have the additional advantage of softening the tight constraints identified above as a driver of power-law outcomes at the games, which in turn would help drive down cost blowouts and costs. in this manner, more realistic contingencies could help make a first clip in the fat tail of olympic costs. however, larger contingencies alone will not solve the problem. cost risks must also be actively managed down. third, the ioc should have skin in the game as regards cost, i.e., it should hold some of the cost risk that arise from staging the games (taleb and sandis ) . the ioc sets the agenda, defines the specs, and has ultimate decision making power over the games. nevertheless, the ioc holds none of the cost risk involved. as a result there is little alignment between incentives to keep costs down and making decisions about cost, which is one reason costs explode at the games, and will keep exploding. for any other type of megaproject such massively misaligned incentives would be unheard of. in order to change this state of affairs, we suggest the ioc is made to cover from its own funds minimum percent of any cost overrun for the games, to be paid on an annual basis as overruns happen. this would give the ioc the motivation it lacks today to effectively manage costs down and thus help reduce the blank check syndrome. cities and nations should refuse to host the games unless the ioc agrees to do this. lack of such agreement would be a clear sign that the ioc does not take cost control seriously. we further suggest that antitrust bodies take a look at the ioc, which today is an unregulated monopoly, and consider regulating it for better performance in accordance with antitrust law. fourth, anything that can be done to shorten the seven-year delivery phase for the games should be considered. the longer the delivery, the higher the risk, other things being equal. for many games, not much happens the first - years after winning the bid, which indicates that faster delivery would be possible, as it was before . faster delivery may be supported by a more standardized and modularized approach to delivery, without the need to reinvent the wheel at every games, and by using existing facilities as much as possible. here it is encouraging to see that the ioc has decided to consider "turnkey solutions" for ocogs in areas that require highly specific olympic expertise (ioc : ) . standardized turnkey solutions should be developed as far as possible to help hosts reduce costs. finally, a much more ambitious goal could be set for schedule and cost, to drive innovation at the games, for instance: "games delivered at half the cost, twice as fast, with zero cost overrun." we suggest that going forward the ioc adopts this slogan as one of its goals. that would show true ambition and willingness to innovate regarding cost control. today's budgets and schedules are so bloated that this goal would not be unrealistic for a professional and experienced delivery organization. unfortunately, as a monopoly that answers to no one the ioc is unlikely to innovate unless it is forced to do so from the outside. fifth, to directly tackle the eternal beginner syndrome proposals have been put forward to host the games in one or a few permanent locations -e.g., athens -or, alternatively, that two successive games should be given to the same host, so facilities could be used twice (short , baade and matheson ) . as a further variation on this theme, games could be spread geographically with different events going to different cities, but with each event having a more or less permanent home, say track and field in los angeles, tennis in london, equestrian events in hong kong, etc. this could be combined with a permanent and professional delivery authority, responsible for staging the games every time and accumulating experience in one place in order to secure effective learning and build what has been called a "high-reliability organization" for delivering the games, something that has been sorely missing so far in the history of the games (roberts ). finally, and perhaps most effectively, prospective host cities could mitigate their risk by simply walking away from the games. indeed, this has become a preferred strategy for many cities. over the past years the number of applicant and candidate cities have fallen drastically, from a dozen to a few (zimbalist : ; lauermann and vogelpohl ) . the exodus of candidate cities has been acutely embarrassing to the ioc and has caused reputational damage to the olympic brand. cities have explicitly and vocally cited extravagant costs and cost overruns as a main reason for exiting the bid process. with agenda , which is being touted by the ioc as protecting the interests of host cities, the ioc has made a first attempt to address the exodus, and to protect its brand and product (ioc (ioc , . but the initiative, though welcome, does not address the root causes and looks like too little too late. time will tell, with tokyo as the first test case, if it is held as planned in . our guess is that more reform will prove necessary to stop the exodus. as rightly observed by los angeles mayor eric garcetti, "most cities, unless you have a government that's willing to go into debt or pay the subsidy of what this costs, most cities will never say yes to the olympics again unless they find the right model," with the "right model" being defined by significantly lower costs (ford and drehs ) . garcetti said this before the covid- pandemic and its large extra debt burdens on governments. post-covid- , the appetite and ability for governments to go into further debt or pay a subsidy to finance the olympic games will likely be low and the pressure to keep costs down will likely be high. perhaps this can finally get the ioc to take olympic costs and cost overruns seriously and try to manage them down in collaboration with host cities. the novel explanations of olympic costs and cost risks uncovered in the present paper shows that the ioc and host cities would have every reason to do so and provides the evidence and guidelines for how to accomplish the task. interest in cost and cost overrun of the games has been high since the establishment of the modern olympics in . as long ago as baron pierre de coubertin, founder of the ioc and the modern games, referred to "the often exaggerated expenses incurred for the most recent olympiads" (coubertin, ) , and in jean drapeau, the mayor of montreal, infamously stated, "the montreal olympics can no more have a deficit, than a man can have a baby," which caused some peculiar cartoons in canadian media when the montreal games incurred a large deficit due to the biggest-ever olympic cost overrun, at percent (cbc ) . drapeau was wrong, and problems with cost and cost overrun are as prevalent today as they were in his time, and in coubertin's before him. despite substantial interest in the cost of the games, however, attempts to systematically evaluate such cost are few, perhaps because valid and reliable data that allow comparative analysis are difficult to come by (chappelet , kasimati , essex and chalkley , preuss , zimbalist , baade and matheson . instead, the attempts that exist have typically focused on a specific or a few games (brunet , bondonio and campaniello , müller and gaffney . another strain of research has focused on whether the games present a financially viable investment from the perspective of cost-benefit analysis (zimbalist : - ) . but what to measure when determining the costs and benefits of the games is open to debate and has varied widely between studies, again making it difficult to compare results across games and studies. in particular, legacy benefits described in the bid are often intangible, and as such pose a difficulty in ex-post evaluations. the benefits of increased tourism revenue, jobs created, or national pride are hugely varied and similarly difficult to quantify and compare. costs can also be hard to determine; for example, one could argue that if hotels in the host city have invested in renovations, and benefits of increased tourist revenues to those hotels are included in the analysis, then these costs should also be included in any accounting, but they rarely are. finally, the percentage of work that an employee in an outlying city spends on games-related work would be exceptionally difficult to estimate. preuss ( ) contains the perhaps most comprehensive multi-games economic analysis to date, looking at the final costs and revenues of the summer olympics from to . preuss finds that since every organizing committee of the olympic games (ocog), which leads the planning of the games in the host city, has produced a positive benefit as compared to cost, but only when investments are removed from ocog budgets. this restricts the analysis narrowly to only ocog activities, which typically represent a fairly small portion of the overall olympic cost and therefore, we argue, also denotes too limited a view for true cost-benefit analysis. further, other authors disagree with preuss' findings, and have suggested that the net economic benefits of the games are negligible at best, and are rarely offset by either revenue or increases in tourism and business (malfas, theodoraki, and houlihan ; billings and holladay ; von rekowsky ; goldblatt : - ) . furthermore, none of these studies have systematically compared projected cost to final cost, which is a problem, because evidence from other types of megaprojects show that cost overruns may, and often do, singlehandedly cause positive projected net benefits to become negative (flyvbjerg ; ansar et al. ) . taking the total body of knowledge into account, a recent study of the economics of the olympics, published in journal of economic perspectives, found that "the overwhelming conclusion is that in most cases the olympics are a money-losing proposition for host cities" (baade and matheson : ) . but there may be other legitimate reasons for hosting the olympics than mere money-making, for instance national pride or throwing the biggest party on the planet. whatever the reason for hosting the games, hosts should know the financial and economic risks involved, before making the decision. in sum, we find for previous academic research on cost and cost overrun for the olympic games: . earlier attempts to systematically evaluate cost and cost overrun in the games are few; . such attempts that exist are often focused on a specific games or are small-sample research; . earlier research on the cost of the games has focused on cost-benefit analysis, with debatable delimitations of costs and benefits making it difficult to systematically compare results across studies and games; . existing evidence indicate that benefits generally do not outweigh costs for host cities and nations. flyvbjerg and stewart ( ) is the first study to consistently compare cost and cost overrun for a large number of olympic games. that study took its inspiration in comparative research more broadly, looking at megaprojects, and used a method for measuring cost and cost overrun that is the international standard in this research field. the goal was to bring the same rigor and consistency to the study of olympic costs as that found in megaproject cost scholarship. previous research has established clearly that hosting the olympics is costly and financially risky. there is a gap in previous research, however, regarding the specific nature of the cost risks faced by cities and nations deciding whether to host the games or not. the present paper aims to fill this gap, by fitting different theoretical statistical distributions to olympic cost data, as described in the main text. olympic bidding for a specific games typically begins with a country's national olympic committee making a call for expressions of interest from prospective host cities eleven years before the games in question. interested cities then compete to become their country's favorite, which is decided nine years prior to the games. finally, among the favorite cities, which are known as "applicants," the ioc typically chooses three to five "candidate" cities, which enter the concluding competition to host the games, decided seven years before the games take place, with one city declared winner of the bid. in the competition to win, cities pitch their ideas to the ioc for how best to host the world's biggest sporting event and how to generate significant urban development in the process (andranovich, burbank, and heying ) . to demonstrate their ability to achieve these goals, bidding cities are required by the ioc to develop detailed plans in the form of so-called candidature files that are submitted to the ioc as part of the competition to host. the candidature files, or "bid books" as they are more commonly known, form part of the basis of the ioc's decision for the next host city. one of the requirements for the bid book is that it includes a budget that details the expected invest ment by the host, in addition to a budget for expected revenues (ioc ) . in their bid book, governments of candidate cities and nations are also required by the ioc to provide guarantees to "ensure the financing of all major capital infrastructure investments required to deliver the olympic games" and "cover a potential economic shortfall of the ocog [organizing committee of the olympic games]" (ibid: ). the candidature file is a legally binding agreement, which states to citizens, decision makers, and the ioc how much it will cost to host the games. as such the candidature file represents the baseline from which future cost and cost overrun should be measured. if cost overrun later turns out to be zero, then decision makers made a well-informed decision in the sense that what they were told the games would cost is what they actually ended up costing, so they had the correct information to make their decision. if cost overrun is significantly higher than zero, then the decision was misinformed in the sense that it was based on an unrealistically low estimate of cost. however, such measurement of cost against a consistent and relevant baseline is rarely done for olympic costs. new budgets are typically developed after the games were awarded, which are often very different to those presented at the bidding stage (jennings ) . these new budgets are then used as new baselines, rendering measurement of cost overrun inconsistent and misleading both within and between games. using later baselines typically makes cost overruns look smaller and this is a strong incentive for using them, as in the case for london mentioned in the introduction to the main text. new budgets continue to evolve over the course of the seven years of planning for the games, until the final actual cost is perhaps presented, often several years after the games' completion -if at all, as we will see. our objective was to measure cost and cost overrun for the games in a consistent and relevant manner. we therefore searched for valid and reliable bid book and outturn cost data for both summer and winter games, starting with the rome summer games and the squaw valley winter games, and continuing until the most recent winter and summer games. costs for hosting the games fall into the following three categories, established by the ioc: . operational costs incurred by the organizing committee for the purpose of "staging" the games. these include workforce, technology, transportation, administration, security, catering, ceremonies, and medical services. they are the variable costs of staging the games and are formally called "ocog costs" by the ioc. . direct capital costs incurred by the host city or country or private investors to build the competition venues, olympic village(s), international broadcast center, and media and press center, which are required to host the games. these are the direct capital costs of hosting the games and are formally called "non-ocog direct costs." . indirect capital costs, for instance for road, rail, or airport infrastructure, or for hotel upgrades or other business investment incurred in preparation for the games but not directly related to staging the games. these are wider capital costs and are formally called "non-ocog indirect costs." the first two items constitute the sports-related costs of the games and are covered in the present analysis. non-ocog indirect costs have been omitted, because ( ) data on such costs are rare, ( ) where data are available, their validity and reliability typically do not live up to the standards of academic research, and ( ) even where valid and reliable data exist, they are often less comparable across cities and nations than sports-related costs, because there is a much larger element of arbitrariness in what is included in indirect costs than in what is included in sports-related costs; plus many indirect costs cover expenditures that would have been incurred even without the games, although perhaps at a later time. it should be remembered, however, that the indirect costs are often higher than the direct costs. baade and matheson ( : ) found that for seven games for which they could obtain data for both sports infrastructure and general infrastructure, in all cases the cost of general infrastructure was higher than the cost of sports infrastructure, sometimes several times higher. for example, for barcelona , the cost of general infrastructure was eight times that of sports infrastructure; for vancouver five times higher. as developing nations increasingly bid for the games, costs for general infrastructure are bound to get even higher, because emerging economies typically have inadequate transportation, communications, energy, water, hospitality, and other infrastructure that must be upgraded before the games can be hosted (zimbalist : , fletcher and dowse ) . for measuring final outturn sports-related cost, data were available for out of the games - , or for percent of games. for measuring cost overrun, which involves comparing estimated bid cost with final outturn cost, data were available for of games. for the remaining games, valid and reliable data have not been reported that would make it possible to establish cost overrun for these games. this is an interesting research result in its own right, because it means -incredible as it may sound -that for more than a third of the games since no one seems to know what estimated and actual costs were. in addition to being a powerful indictment of olympic bidding and of the opacity of local olympic boosterism and ioc decision making, such ignorance hampers learning regarding how to develop more reliable budgets for the games. from a rational point of view, learning would appear to be a self-evident objective for billion-dollar events like the games, but often that is not the case. for some games, hiding costs and cost overruns seems to have been more important, for whatever reason. nevertheless, out of games is percent of all possible games for the years under consideration, which we deem sufficient for producing interesting and valid results. we measured costs in both nominal and real (adjusted for inflation) terms, and in both local currencies and us dollars. we followed international convention and made all comparisons across time and geographies in real terms, to ensure that like is compared with like. the dataset is the largest of its kind and is the first consistent dataset on olympic costs. further details on data and methodology are available in flyvbjerg and stewart ( ) . games not including road, rail, airport, hotel, and other infrastructure, which often cost more than the games themselves country events athletes cost, billion usd managers -even when faced with negative outcomes -keep rationalizing and funding their decisions rather than changing their course of action. this may happen, for instance, when managers think that a point of no return has been reached or that sunk costs are too high to opt out. as a consequence, funds keep being allocated to a project mainly because a lot of money was already allocated to it and no one wants to reverse the original decision. preferential attachment and escalation of commitment have been shown to lead to power-law outcomes (yule , farmer and geanakoplos , newman , gabaix , sleesman et al. ). spending at the olympics is as pure a case of preferential attachment and escalation of commitment that one can find, with complete lock in and a point of no return reached more than seven years before delivery. the more tight the budget for a games (measured as cost per athlete and cost per event in bid, with lower bid cost indicating a tighter budget) the higher the percentage cost overrun, although the correlation is not statistically significant, due to small sample sizes caused by the innate rarity of the games. before , games were delivered - years faster than this. several cities have hosted the games more than once, but over the past century only innsbruck ( innsbruck ( , revisiting some "established facts" in the field of management olympic cities: lessons learned from mega-event politics should we build more large dams? the actual costs of hydropower megaproject development big is fragile: an attempt at theorizing scale the state of self-organized criticality of the sun during the last three solar cycles going for the gold: the economics of the olympics self-organized criticality: an explanation of the /f noise the origin of bursts and heavy tails in human dynamics reply to comment on practical application and empirical evaluation of reference class forecasting for project management olympics and paralympics £ m under budget fitting long-memory models by generalized linear regression denver ' : the winter olympics and the politics of growth should cities go for the gold? the long-term impacts of hosting the olympics forecasting construction cost escalation developing expert political judgment: the impact of training and practice on judgmental accuracy in geopolitical forecasting tournaments from lake placid to salt lake city: the challenging growth of the olympic winter games since irrational exuberance plus indelible sniping yields delectable entertainment power-law distributions in empirical data on the frequency of severe terrorist events scaling behaviour in the number of criminal acts committed by individuals olympic review universality in solar flare and earthquake occurrence explaining complex organizational dynamics mega-sporting events in urban and regional policy: a history of the winter olympics power laws in finance and their implications for economic theory, mimeo power laws in economics and elsewhere sport mega-events, the non-west and the ethics of event hosting curbing optimism bias and strategic misrepresentation in planning: reference class forecasting in practice over budget, over time, over and over again: managing major projects quality control and due diligence in project management: getting decisions right by taking the outside view the fallacy of beneficial ignorance: a test of hirschman's hiding hand the law of regression to the tail: how to survive covid- , the climate crisis, and other disasters five things you should know about cost overrun digitization disasters: towards a theory of it investment risk underestimating costs in public works projects: error or lie olympic proportions: cost and cost overrun at the olympics - , working paper olympic bid : how mayors eric garcetti and anne hidalgo have changed the olympic bid game, espn power laws in economics and finance the relationship between corporate social responsibility and shareholder value: an empirical test of the risk management hypothesis the games: a global history of the olympics power law in firms bankruptcy the four 'knowns' of sports mega-events candidature procedure and questionnaire: games of the xxx olympiad in olympic agenda : + recommendations, lausanne: ioc report of the evaluation commission olympic agenda : midway report, lausanne: ioc why costs overrun: risk, optimism and uncertainty in budgeting for the london olympic games, construction management and economics thinking, fast and slow before you make that big decision intuitive prediction: biases and corrective procedures economic aspects of the summer olympics: a review of related research the reliability of china's economic data: an analysis of national output how greeks have adjusted to the forever crisis, the financial times fragile growth coalitions or powerful contestations? cancelled olympic bids in boston and hamburg heavy-tailed distribution of cyber-risks living in a world of low levels of predictability forest fires: an example of self-organized critical behavior the applicability of power-law frequency statistics to floods impacts of the olympic games as mega-events an informational theory of the statistical structure of language new methods in statistical economics after sochi : costs and impacts of russia's olympic games comparing the urban impacts of the fifa world cup and the olympic games from power laws, pareto distributions and zipf's law. contemporary physics olimpíada do rio fica sem custo total oficial, folha de s. paulo some characteristics of one type of high reliability organization gauss is not mocked cleaning up the big muddy: a metaanalytic review of the determinants of escalation of commitment exploring data to detect project problems comments on "the origin of bursts and heavy tails in human dynamics antifragile: how to live in a world we don't understand statistical consequences of fat tails: real world preasymptotics, epistemology, and applications, technical incerto collection on the shadow moments of apparently infinite-mean phenomena the skin in the game heuristic for protection against tail events superforecasting: the art and science of prediction after the gold rush: a sustainable olympics for london are the olympics a golden opportunity for investors? a mathematical theory of evolution, based on the conclusions of dr circus maximus: the economic gamble behind hosting the olympics and the world cup all dollar figures are usd at level, unless otherwise stated the paralympic games are not included here because they became fully integrated with the olympic games relatively recently and therefore do not compare across the period we study it should be mentioned that data on final outturn cost for the rio rio's cost accounting has been criticized in the press (nogueira ) for omitting some sports related cost and the true cost might be higher ) found that the mistakes are compounded by researchers typically being vague or not transparent about how so-called outliers are defined, identified, and removed, undermining the validity of results. removal of extreme values before analysis is like trying to understand how best to earthquake proof buildings without taking the biggest earthquakes into account-not a good idea. we did not remove any extreme values in the dataset, needless to say significance is here defined in the conventional manner we also did this fit for the winter and summer games separately and found no statistically significant difference between the two, wherefore their data should be, and are the true mean is sometimes also called the "shadow" mean no fully valid and reliable way exists for how to establish the existence of generative mechanisms for power law distributions. attempts at determining such mechanisms are therefore likely to contain elements of arbitrariness and speculation due to large cost overruns and worries about environmental impacts, citizens of colorado called into question the wisdom of hosting the games. a referendum was held, and voters rejected the games by a large margin. innsbruck, austria, stepped in to replace denver, hosting the winter olympics for a second time cost overrun is picked up by the host. the ioc makes its profit directly from the revenues, so any improvement in specs that leads to improved revenues is a positive for the ioc, irrespective of its cost. this lack of liability for costs for the ioc is deeply unhealthy the authors wish to thank nassim nicholas taleb for helping develop key ideas of the paper and for carrying out the statistical analysis involved in fitting the theoretical statistical distributions to the data. allison stewart helped collect the data for the games - and co-authored flyvbjerg and stewart ( ) , on which parts of the current paper builds. finally, the authors wish to thank the review team at the journal for highly useful comments and suggestions of improvements to an earlier version of the paper. key: cord- - rblzyry authors: hill, andrew; wang, junzheng; levi, jacob; heath, katie; fortunak, joseph title: minimum costs to manufacture new treatments for covid- date: - - journal: nan doi: nan sha: doc_id: cord_uid: rblzyry introduction: ‘repurposing’ existing drugs to treat covid- is vital to reducing mortality and controlling the pandemic. several promising drugs have been identified and are in various stages of clinical trials globally. if efficacy of these drugs is demonstrated, rapid, mass availability at an affordable cost would be essential to ensuring equity and access especially amongst low- and middle-income economies. methods: minimum costs of production were estimated from the costs of active pharmaceutical ingredients using established methodology, which had good predictive accuracy for medicines for hepatitis c and hiv amongst others. data were extracted from global export shipment records or analysis of the route of chemical synthesis. the estimated costs were compared with list prices from a range of countries where pricing data were available. results: minimum estimated costs of production were us $ . /day for remdesivir, $ . /day for favipiravir, $ . /day for hydroxychloroquine, $ . /day for chloroquine, $ . /day for azithromycin, $ . /day for lopinavir/ritonavir, $ . /day for sofosbuvir/daclatasvir and $ . /day for pirfenidone. costs of production ranged between $ . and $ per treatment course ( – days). current prices of these drugs were far higher than the costs of production, particularly in the us. conclusions: should repurposed drugs demonstrate efficacy against covid- , they could be manufactured profitably at very low costs, for much less than current list prices. estimations for the minimum production costs can strengthen price negotiations and help ensure affordable access to vital treatment for covid- at low prices globally. introduction as the sars-cov pandemic continues to grow, researchers worldwide are urgently looking for new treatments to prevent new infections, cure those already infected or lessen the severity of disease. an effective vaccine may not be widely available until late , even if trials are successful [ ] . there are clinical trials in progress to 'repurpose' drugs, normally indicated for other diseases, to treat covid- [ , ] . the shortened development timeline and reduced costs to this approach [ ] of using already existing compounds is particularly advantageous compared with new drug discovery in a pandemic situation, where time is of the essence. antiviral drugs include nucleotide analogue remdesivir, which was previously used experimentally but unsuccessfully against ebola [ ] [ ] [ ] [ ] , favipiravir, used to treat influenza [ ] , the hiv protease inhibitor lopinavir/ritonavir (lpv/r) [ , ] , the antimalarials chloroquine and hydroxychloroquine [ ] [ ] [ ] , and the directacting antivirals sofosbuvir and daclatasvir [ ] , which are all potential candidates. additionally, treatments to improve lung function and reduce inflammation, such as pirfenidone [ , ] and tocilizumab [ , ] , are being evaluated in clinical trials. most of the clinical trials reported so far are small pilot studies, often non-randomised, making interpretation of current evidence difficult. however, results from randomised trials of these repurposed treatments should become available from may onwards. if favourable results are shown from these new trials, there is the potential to rapidly upscale production of the most promising drugs. the safety profiles of these drugs have already been established from clinical trials for other diseases, so they could be rapidly deployed to treat covid- before vaccines become available. low-and middle-income countries will need access to these treatments at minimum prices to ensure all those in need can be treated. even in high-income countries, the burden of disease could be so great that access to drugs at minimum costs could also be necessary. the hiv epidemic has been controlled by mass treatment with antiretroviral drugs worldwide, at very low unit costs. large donor organisations such as the global fund for aids, tb and malaria (gfatm) and the president's emergency plan for aids relief (pepfar) order drugs to treat > million people with hiv, at prices close to the cost of production [ , ] . this system allows low-and middle-income countries to access high-quality drugs at affordable prices. the minimum costs of drug production can be estimated by calculating the cost of active pharmaceutical ingredients (api), which is combined with costs of excipients, formulation, packaging and a profit margin, to estimate the price of 'final finished product' (ffp) -the drug ready for use. there are established methods for these calculations, which have reliably predicted the minimum costs of drugs to cure hepatitis c [ , ] and other diseases [ , ] . the purpose of this analysis was to apply the same calculations to the new candidate treatments for covid- . the leading candidate drugs to treat covid- were selected based on recent reviews and analysis of ongoing clinical trials [ , , [ ] [ ] [ ] , [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] . the treatments selected were remdesivir, favipiravir, lopinavir/ritonavir, hydroxychloroquine, chloroquine, azithromycin, pirfenidone, tocilizumab and sofosbuvir/daclatasvir. all are currently being evaluated in randomised trials, with results expected between may and september . the methods used to estimate minimum costs of production have been described previously [ , ] . briefly, we analysed the costs of exports of api from india using the online tracking database panjiva [ ] , which shows details of all shipments of api with quantities and costs per kilogram. we used all available costing data for each drug api found on panjiva, excluding shipments < kg in size, alongside the lowest and highest % of results based on prices per kg. a % api loss during tableting process was factored into our calculations, and a conversion cost of us $ . per tablet was used. a multiplier based on api mass was applied to account for the price of excipients, which are additional substances needed to convert api into fpp. our estimated api costs presume that production is carried out by a generic provider of apis, where associated costs of capital investment, overhead and labour are not as high as for production by originator companies. this method was applied to small-molecule drugs only. in addition, the drug remdesivir, which is administered by iv infusion, was considered separately when estimating formulation costs. a profit margin of % and indian taxation of % on profit was added. these results were cross-checked with a second api database, pharmacompass [ ] , which displayed records up to only. panjiva was selected to provide real time, up-to-date shipment and cost data. the estimates assume a volume of kg of api for each compound. three drugs did not have available data on api production: remdesivir and favipiravir and tocilizumab. for the first two, api production costs were estimated based on published routes of chemical synthesis [ , , , ] . since remdesivir is administered by iv infusion, the costs of production were adjusted to include the cost of injections, according to an established method used for the world health organization (who) essential medicines list [ ] . it was not possible to track the cost of api for the monoclonal antibody tocilizumab; therefore, we tracked list prices in a range of countries, in particular developing economies as a proxy for minimum costs. the costs of regulatory filings and approvals are often significant add-ons to the initial use of drugs in any specific country. all drugs analysed in this study, except for remdesivir have been approved for treatment for some indications in all countries, but few are approved for treatment of covid- . remdesivir is an investigational drug without any prior regulatory approvals, but it has a known favourable safety profile after clinical trials against ebola. favipiravir has been approved for the treatment of influenza in japan since [ ] , and in china since march . it has also been approved for investigational use (china and italy) against covid- in march . we have not included the cost or timing associated with regulatory approvals for the use of these drugs. we are assuming in our analysis that the who and other influential regulatory agencies will cooperate to define a pathway for use of these drugs which does not include additional financial outlays or filing for marketing approvals. the minimum costs of production were then compared with published list prices for each drug in a range of countries -usa, uk, france, sweden, south africa, india, bangladesh, malaysia, brazil, turkey, pakistan, egypt and china [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] -to give a representative sample of prices in countries with different levels of economic development. for consistency, we selected a single data source per country to be used for all searches of drug prices within that country, based on the organisation of data and perceived reliability. not all drugs analysed in this study were available in our selected countries, and in some countries, online pharmacy sites were used because national databases were not available. where several prices were available in the same database, the lowest price was selected. available clinical trial data for each drug were collected from literature searches, clinicaltrials.gov and the chinese clinical trials registry. results from pilot studies were included, together with the planned clinical trial programmes and expected completion dates. the predicted costs of production, and the highest/lowest available list prices of all drugs analysed are shown in table , and chemical structures for all drugs are shown in figure . remdesivir remdesivir (formerly gs- ) has been evaluated for treatment of sars and ebola [ , ] . there are five randomised trials of remdesivir for sars-cov , with first results expected at the end of april . there is a -day course of treatment under evaluation. the dose of remdesivir is mg on the first day with mg per day thereafter. remdesivir is administered by iv infusion. one metric ton of remdesivir is sufficient api to manufacture , courses of treatment, without allowance for any losses during formulation. based on analysis of the published second-generation route of chemical synthesis and assumed overheads such as occupancy rate per hour and labour of an india-based generic producer, the cost of api was estimated to be $ /kg [ , ] for -kg batches produced without capital investment in dedicated remdesivir production facilities. the -day course of treatment would therefore cost $ . per person. after adjustment for the cost of formulation (and % losses projected during formulation), cost of vials, profit margin and tax, the estimated cost per treatment would be approximately $ per person as shown in figure a . daily cost is estimated to be $ . . as an iv infusion, there would be an additional cost for saline, estimated at $ per infusion [ ] , equipment such as syringe, sterile water for reconstitution and iv lines, as well as staff-time cost associated with the healthcare professional administrating the infusion. these additional non-drug components are likely to be more expensive than the estimated cost of the drug itself and are not included in our cost estimate. the main randomised trial of favipiravir evaluated up to days of treatment, vs another influenza drug, umifenovir, in patients. after days of treatment, the clinical recovery rate was % for favipiravir vs % for umifenovir (p< . ). recovery from fever was also faster for people treated with favipiravir (p< . ), but there was no difference between arms in auxiliary oxygen therapy or non-invasive mechanical ventilation rates [ ] . a second, non-randomised study in china evaluated days of treatment with either favipiravir or lpv/r. the median time to virus clearance was significantly shorter in the favipiravir group ( days) vs the lpv/r group ( days, p< . ) [ ] . favipiravir is dosed mg twice daily. a metric ton of favipiravir, therefore, would provide approximately , courses of treatment. the cost of api was estimated at about $ /kg, based [ , ] . this is a simple molecule to synthesise: several very basic steps involved are more akin to processes for manufacturing fine chemicals rather than pharmaceuticals. the structure is a -fluoro substituted -hydroxypyrazine- -carboxamide. based on a -day course of treatment, the cost of the api would be $ . . after adjustment for loss, formulation, packaging and profit margins, the estimated cost of production is $ per -day treatment cost (figure b ), or $ . per day. in late february , favipiravir was launched for sale in china for $ per treatment course, in an announcement by the shandong provincial public resource trading centre [ ] . the standard dose of lpv/r is / mg twice daily. lpv/r has been evaluated in two published studies of sars-cov infection [ ] . in these studies, there has been no difference in measures of efficacy between lpv/r (given for days) and control treatment. a systematic review by the who showed no clear evidence for the efficacy of lpv/r. however, there were very few clinical trials available for this evaluation at the time of the review [ ] . it is not clear whether this hiv protease inhibitor will work in the earlier stage of sars-cov infection, or if used in combination with other drugs. results from other clinical trials are expected between june and august . the estimated cost of production is $ . per treatment course, based on api cost of $ /kg for lopinavir and $ /kg for ritonavir. after adjustment for loss, formulation, packaging and profit margins, the estimated cost of production of the combined drug is $ per -day course, or $ . per day. searches in selected national databases and online pharmacies yielded a range of list prices between $ in the us, to $ in south africa per -day dose, as shown in figure a . additionally, lpv/r is also available through the global fund for a range of low-and middle-income countries, with a median cost of $ per -day course [ ] . in vitro, hydroxychloroquine is predicted to show superior activity against covid- compared with chloroquine [ ] . in a recent parallel-group, a randomised chinese study in wuhan of patients found treatment with hydroxychloroquine mg per day led to faster symptomatic improvement times in the treatment arm (n = ) after days in terms of temperature normalisation and cough remission, and a greater proportion of patients ( . % vs . %) with improved pneumonia [ ] . similarly, another chinese randomised trial in shanghai [ ] of patients also used mg per day in the treatment arm but found no statistically significant difference in clinical findings, symptomatic improvements or radiological improvements between the arms by day . the authors also highlighted the need for much larger, better powered trials to reach reliable conclusions. journal of virus eradication ; : - minimum costs to manufacture new treatments for covid- different dosing protocols are being used for hydroxychloroquine, including mg daily in the small, open-label, non-randomised french study by gautret et al. (n = ) that has suggested improved efficacy for hydroxychloroquine [ ] . for our analysis, we have therefore chosen mg daily, which was the most commonly used dosage. this is also the upper recommended dose by the british national formulary for existing indications [ ] . api cost-per-kilogram of hydroxychloroquine is $ /kg, with a -day dose-equivalent of api costing $ . . after adjustment for the cost of formulation, packaging and profit margin, the final cost would be $ per -day treatment course ($ . per day). globally, list prices range between $ per -day course in china and only $ per -day course in india ( figure b ). for chloroquine, the cost of api is $ /kg from panjiva data. a -day course equivalent of api would therefore cost $ . based on the equivalent dose of mg per day of chloroquine base. after adjustment for the cost of loss, formulation, packaging and profit margin, the final cost would be $ . per -day treatment course of chloroquine, equivalent to $ . per day. available list prices vary from $ in the us, to $ . per -day course in bangladesh, which is less than our estimated generic treatment cost. it is worth noting that the us price for chloroquine may be considered an outlier, given the next most expensive list price, found in the uk, was only $ per -day course (figure c ). azithromycin this macrolide antibiotic has been used as an adjunctive treatment for six patients in the small french pilot study of hydroxychloroquine by gautret et al. to prevent bacterial superinfection, with all six patients virologically cured by day six [ ] . however, this finding is contradicted by a small, open-label study (n = ) in paris by molina et al. [ ] , who found no strong viral clearance effect associated with hydroxychloroquine/azithromycin combination therapy. the cost of api derived from panjiva shipment data for azithromycin is $ /kg. a -day course equivalent of api at a dose of mg per day would therefore cost $ . . after adjustment for the cost of loss, formulation, packaging and profit margin, the final cost would be $ . per day or $ . per -day treatment course. list prices for azithromycin range between $ per -day course in the us, and $ per -day course in india and bangladesh (figure d ). sofosbuvir/daclatasvir combination treatment with sofosbuvir/daclatasvir, direct-acting antivirals normally used to treat hepatitis c, is being evaluated in iran for covid- patients with moderate to severe symptoms [ ] . the dosage of sofosbuvir/daclatasvir is / mg daily. api per kilogram was $ for sofosbuvir and $ for daclatasvir. fourteen-day dose-equivalent of api for the combined drug therefore costs $ . . after adjusting for the cost of loss, formulation, packaging and profit margin, the final cost would be $ per -day treatment course, or $ . per day. the cost of sofosbuvir/daclatasvir api has been falling significantly in recent years; earlier estimates in for a -week course of treatment were $ per patient, or $ per days [ ] and in , $ per -week course, or $ . per days [ ] . therefore, our new estimates represent a . -fold reduction in the minimum cost of production since . globally, -day course list prices range between $ , in the us and $ in india, or $ in neighbouring pakistan, as shown in figure e [ ] . there is a randomised trial of pirfenidone vs placebo in progress [ ] . there are patients with severe or critical sars-cov infection being evaluated in this clinical trial, with results expected in may . the dose being evaluated is mg three times daily for weeks. the cost of api from the panjiva database was $ /kg, representing a -week api cost of $ . after adjustment for costs of loss, formulation, packaging and profit margins, the minimum cost of treatment would be $ per person, or $ . per day. there is again a large variation between individual countries' list prices. pirfenidone is available in the us for $ for a -week course, but only $ in bangladesh and $ in india for a generic version (figure f) . however, even at $ per month, this is still higher than our api cost-based estimate. there are several large clinical trials of this monoclonal antibody in progress, for patients with late-stage disease [ , ] . as an iv infusion, doses are based on bodyweight ( mg/kg) with a maximum single dose of mg every hours. we therefore made the assumption of average bodyweight being kg, with a single dose of mg. no api data were available for tocilizumab; therefore, we were unable to estimate the minimum cost of production. list prices per mg single dose varied from $ in the us to $ in pakistan (figure g ). across several developing economies with available list prices -india, bangladesh, turkey, south africa, egypt and pakistan -the median was $ per dose. several tocilizumab biosimilars are currently under development [ , ] ; however, none has yet been approved and launched. the general experience so far of biosimilars has been that they offer health care systems the potential to lower costs significantly [ ] , with the uk alone expected to save up to gbp -gbp million per year through increased uptake of better-value biological medicines [ ] . this analysis shows that drugs to treat covid- could be manufactured for very low prices, between $ and $ per course. many of these drugs are already available as generics, at prices close to the cost of manufacture, in low-and middle-income countries. we do not yet know which of these drugs will show significant benefits. however, if promising results emerge from pivotal clinical trials, there is the potential to upscale generic production and provide treatment for millions of people at very low unit prices. there is an established mechanism to do this: donor organisations such as gfatm and pepfar already provide mass treatment of hiv, tb and malaria worldwide at prices close to the costs of production. the drugs in this analysis were not designed against the sars-cov virus; they were developed to treat other viruses or diseases. some, such as chloroquine, were developed in the s. most of the clinical trials of treatments have been funded by national health authorities and donor agencies rather than pharmaceutical companies. patients with covid- have risked their own health in these clinical trials, often with no clear benefits. companies should be encouraged to continue their research, with costs of clinical trials supported by public funding. since the start of the pandemic, the money spent by pharmaceutical companies on research and development of these drugs will be minimal, relative to funding from national health authorities. where pharmaceutical companies have donated drugs for clinical trials, there are already tax rebate systems in place that will recover the costs of the donated drugs. for mass production of these drugs, our analysis assumed a profit margin of % to companies manufacturing the drugs. this is similar to the pricing structure for hiv, tb and malaria, where generic drug companies still earn acceptable profits while mass producing these drugs at prices close to production costs. large-volume orders are needed to incentivise generic companies to manufacture drugs at low prices. other mechanisms exist to optimise drug manufacture. with pooled procurement, a set of countries can order drug supplies together, to take advantage of economies of scale. there can be volume-price guarantees to procure large amounts of drugs at fixed prices for a set number of years. prequalification of key companies by the who can be recognised by any country as an indicator of drug quality, including adherence to good manufacturing practice and the stability, or viability of the drug over its stated shelf life, alongside the bioequivalence of generic drugs vs the original branded versions. additionally, the costs of treatment may be higher if combinations of two or three drugs are needed. other infectious diseases such as hiv, hepatitis c or tb are best treated with two/threedrug combination treatments. drugs which have not shown efficacy against covid- as monotherapy should not necessarily be discarded: they might still contribute to the efficacy of two-or three-drug combination treatments. drugs that are not curative but lessen disease severity are also needed. these treatments could lessen the burden on healthcare systems, which could otherwise be overwhelmed by a lack of ventilators or other supportive services. when these drugs are repurposed to treat covid- , we will need to ensure a constant supply of drugs for the original indications, for example, pirfenidone for people with pulmonary fibrosis, or hydroxychloroquine for people with rheumatoid arthritis and systemic lupus erythematosus. the costs of these treatments will be higher if used for longerterm prevention, for example, in healthcare workers. randomised trials of chloroquine and hydroxychloroquine for prevention of sars-cov infection are in progress [ , ] and other candidate drugs could emerge for use as prophylactics. we highlight four limitations of our study, for consideration in future work. first, this analysis does not include all candidate drugs for covid- . there are drugs at earlier stages of development; a wide range of candidate drugs have been identified by machine learning models [ ] . these drugs may need further in vitro testing before being introduced into human studies. second, treatments like lpv/r and sofosbuvir/daclatasvir have only a small chance of showing significant benefits to patients in ongoing trials, given current evidence. third, for newer drugs such as remdesivir, favipiravir and pirfenidone, costs of production could continue to fall over time through economies of scale. this trend has been seen for drugs to treat hiv and hepatitis c. the cost of api for the hepatitis c drug daclatasvir fell by % in the years after initial launch, as more generic companies upscaled synthesis of the api with greater competition in the market. fourth, many drugs may have been given discounts from the list prices that we have identified for comparison in this analysis following in-country negotiations. even so, list prices can be over -fold higher than the predicted costs of production in some cases. we propose four main recommendations to ensure that any patient with covid- , in any country, can access the drugs they need: . treatments showing efficacy in well-powered clinical trials should be made available worldwide at prices close to the cost of manufacture. all the treatments being evaluated in clinical trials are very cheap to manufacture. clearly, the mass production of these drugs will need to be economically sustainable. treatments for hiv, tb and malaria are distributed worldwide by gfatm and pepfar, to treat millions of people at prices close the cost of manufacture. the prices paid allow generic companies to make acceptable profits. we should adopt a similar model of drug distribution for covid- . . there should be parallel manufacture by at least three different companies for each product, sourcing their api from different countries. in the early stages of the sars-cov epidemic, api production in china was severely disrupted because of quarantine of key workers and delays in transporting key raw materials between factories [ ] . india has suspended export of several key drugs because of anticipated local shortages. production of drugs in a range of countries will protect us from disruption or shortages in individual countries. . there should be no intellectual property barriers preventing mass production of these treatments worldwide. we need open 'technology transfer' so that the methods used to manufacture the key drugs can be shared with any country deciding to produce the drugs locally. . results and databases from all covid- clinical trials should be fully accessible so others can learn from them. to speed up access to these drugs, countries could rely on recognition of the review and approval of key treatments by regulatory authorities in the us or europe, or other stringent regulatory authorities. there may not be time for the normal times of regulatory review by all individual countries. in summary, repurposed drugs may be our only option to treat covid- for the next - months, until effective vaccines can be developed and manufactured at scale. if repurposed drugs do show efficacy against covid- , they could be manufactured at very low unit prices, in the range of $ to $ per treatment course. the system of mass production and distribution of drugs to treat hiv, tb and malaria via gfatm and pepfar could act as a blueprint for the treatment of sars-cov , to ensure access to effective drugs at low prices worldwide. responding to covid- -a once-in-a-century pandemic? therapeutic options for the novel coronavirus ( -ncov) predicting commercially available antiviral drugs that may act on the novel coronavirus ( -ncov), wuhan, china through a drug-target interaction deep learning model drug repurposing: progress, challenges and recommendations broad-spectrum antiviral gs- inhibits both epidemic and zoonotic coronaviruses methods for treating filoviridae. pubchem database: patent=us . available at: pubchem.ncbi.nlm.nih.gov/ patent/us discovery and synthesis of a phosphoramidate prodrug of a pyrrolo[ , -f][triazine- -amino] adenine c-nucleoside (gs- ) for the treatment of ebola and emerging viruses clinicaltrials.gov. severe -ncov remdesivir rct: identifier: nct . bethesda (md): us national library of medicine influenza virus polymerase inhibitors in clinical development systematic review of the efficacy and safety of antiretroviral drugs against sars, mers, or covid- : initial assessment a trial of lopinavir-ritonavir in adults hospitalized with severe covid- in vitro antiviral activity and projection of optimized dosing design of hydroxychloroquine for the treatment of severe acute respiratory syndrome coronavirus (sars-cov- ) drugs for treating protozoan infections a to ipca laboratories kumar a, vyas kd, singh d, nandavadekar s, bhise s, jadhav a. available at: patentscope.wipo.int/search/en/detail.jsf?docid=wo &tab=pctbiblio identifier: irct n . a prospective randomized controlled clinical trial comparing the therapeutic efficacy of sovodak (sofosbuvir/daclatasvir) with standard care in patients with moderate to severe coronavirus (covid- ) virus bollu rb, mandadapu vpk, indukuri vsk, chava s. . available at: patentscope.wipo.int/search/en/detail.jsf?docid=wo clinicaltrials.gov. a study to evaluate the efficacy and safety of pirfenidone with novel coronavirus infection: identifier: nct . bethesda (md): us national library of medicine favipiravir combined with tocilizumab for the treatment of coronavirus disease : identifier: nct . bethesda (md): us national library of medicine tocilizumab in covid- pneumonia (tocivid- ): identifier: nct . bethesda (md): us national library of medicine global fund for aids, tb and malaria. results report the united states president's emergency plan for aids relief. annual report to congress minimum costs for producing hepatitis c directacting antivirals for use in large-scale treatment access programs in developing countries rapid reductions in prices for generic sofosbuvir and daclatasvir to treat hepatitis c estimation of cost-based prices for injectable medicines in the who essential medicines list estimated costs of production and potential prices for the who essential medicines list efficacy of hydroxychloroquine in patients with covid- : results of a randomized clinical trial a pilot study of hydroxychloroquine in treatment of patients with common coronavirus disease- (covid- ) hydroxychloroquine and azithromycin as a treatment for covid- : results of an open-label, non-randomized clinical trial no evidence of rapid antiviral clearance or clinical benefit with the combination of hydroxychloroquine and azithromycin in patients with severe covid- infection breakthrough: chloroquine phosphate has shown apparent efficacy in treatment of covid- associated pneumonia in clinical studies favipiravir versus arbidol for covid- : a randomized clinical trial favipiravir treatment of covid- pneumonia initial clinical results released hydroxychloroquine post exposure prophylaxis for coronavirus disease (covid- ): identifier: nct . bethesda (md): us national library of medicine panjiva supply chain intelligence. available at: panjiva.com pharmacompass import-export database. available at: www.pharmacompass.com/ manufacturers-suppliers-exporters xiang h et al. a practical and step-economic route to favipiravir the complete synthesis of favipiravir from -aminopyrazine british national formulary. london: nice, . available at: bnf.nice.org.uk drug price information. available at: www.drugs.com/price-guide office of procurement, acquisition and logistics: pharmaceutical prices medicine price registry. south africa. available at: openup.org.za/tools/ mpr.html drug price lists. available at: portal.anvisa. gov.br/listas-de-precos french ministry of solidarity and health. base de données publique des médicaments. available at: base-donnees-publique.medicaments.gouv.fr online medicine index of bangladesh. available at: medex. com.bd pharmaceutical services programme. consumer price guide. available at: www.pharmacy.gov.my/v /en/apps/drug-price drug price of all the brand names. available at: www.medindia.net/ drug-price/index ilacabak turkey. available at: ilacabak.com egyption drug store. actemra mg/ ml (tocilizumab) vial for iv infusion. available at: egyptiandrugstore.com/index.php?route=product/product&product_ id= impact of fluid choice in systemic inflammatory response syndrome patients on hospital cost savings pharmaceutical favipiravir price announced: cny/box [chinese]. febuary. available at: med.sina.com/article_detail_ _ _ . html the global fund. price reference report - -present. available at: public. tableau.com/profile/the.global.fund#!/vizhome/pqrpricelist_english/pricelist treatment advocate tactics to expand access to antiviral therapy for hiv and viral hepatitis c in low-to high-income settings: making sure no one is left behind the road to elimination of hepatitis c: costs per cure fall to us$ per person. american association for the study of liver disease: the liver meeting mycenax biotech inc. lusinex (tocilizumab, mycenax in-house). available at: www.mycenax.com.tw/en/product.php?act=view&id= clinicaltrials.gov. comparative study of bat to roactemra® in rheumatoid arthritis patients with inadequate response to methotrexate: identifier: nct . bethesda (md): us national library of medicine of biosimilars in rheumatology: reshaping the healthcare environment commissioning framework for biological medicines (including biosimilar medicines indian pharma threatened by covid- shutdowns in china acknowledgements conflicts of interest ah received a consultancy payment from merck for a clinical trial review that is not connected with this project.funding ah received funding from the international treatment preparedness coalition (itpc) as part of the unitaid-supported project 'affordable medicines for developing countries'. journal of virus eradication ; : - minimum costs to manufacture new treatments for covid- key: cord- - d mb authors: hogg, william; gray, david; huston, patricia; zhang, wei title: the costs of preventing the spread of respiratory infection in family physician offices: a threshold analysis date: - - journal: bmc health serv res doi: . / - - - sha: doc_id: cord_uid: d mb background: influenza poses concerns about epidemic respiratory infection. interventions designed to prevent the spread of respiratory infection within family physician (fp) offices could potentially have a significant positive influence on the health of canadians. the main purpose of this paper is to estimate the explicit costs of such an intervention. methods: a cost analysis of a respiratory infection control was conducted. the costs were estimated from the perspective of provincial government. in addition, a threshold analysis was conducted to estimate a threshold value of the intervention's effectiveness that could generate potential savings in terms of averted health-care costs by the intervention that exceed the explicit costs. the informational requirements for these implicit costs savings are high, however. some of these elements, such as the cost of hospitalization in the event of contacting influenza, and the number of patients passing through the physicians' office, were readily available. other pertinent points of information, such as the proportion of infected people who require hospitalization, could be imported from the existing literature. we take an indirect approach to calculate a threshold value for the most uncertain piece of information, namely the reduction in the probability of the infection spreading as a direct result of the intervention, at which the intervention becomes worthwhile. results: the -week intervention costs amounted to a total of $ , . , or $ , . prorated according to the length of the flu season, or $ , . prorated for the entire calendar year. the variable costs that were incurred for this -week project amounted to approximately $ . per participating medical practice. the (fixed) training costs per practice were equivalent to $ . for the -week intervention, or $ . for -week flu season, or $ . for an entire one-year period. conclusion: based on our conservative estimates for the direct cost savings, there are indications that the outreach facilitation intervention program would be cost effective if it can achieve a reduction in the probability of infection on the order of . ( . , . ) percentage points. a facilitation intervention initiative tailored to the environment and needs of the family medical practice and walk-in clinics is of promise for improving respiratory infection control in the physicians' offices. there is a paucity of empirical evidence in the literature about actual intervention strategies to improve respiratory infection control practices and analyze the efficiency implications for health policy. prevention, especially within health care settings, has assumed paramount importance in the fight against respiratory infection. since influenza is typically transmitted by droplets and contact routes, there are precautions that can be taken to reduce its transmission [ ] [ ] [ ] . interventions designed to prevent the spread of respiratory infection within family physician (fp) offices could potentially have a significant positive influence on the health of canadians. while there are costs associated with the implementation of any intervention, the benefits stemming from the outcomes of such interventions have the potential to outweigh them. however, there are few evaluations of outreach facilitation that have studied the net costs of delivering interventions of this nature that exist in the literature. an exception is a study authored by hogg, baskerville, and lemelin [ ] , which consisted of a randomized, controlled, field trial of an intervention aimed at improving preventative care tailored to the needs of participating family practices. it demonstrated the effectiveness of a multi-faceted outreach facilitation in improving overall preventative care performance. it is the first analysis of cost consequences of an outreach facilitation intervention of which we are aware, and it indicated that the cost savings attributable to the reduction in inappropriate testing on the one hand, and increases in appropriate testing on the other hand, may outweigh all of the intervention costs. those authors argued that a costly intervention that achieves success may be preferred on a cost-benefit basis to a cheaper one that demonstrates very little or has no lasting effect. while based on an original and a very different application, this current study employs a similar approach to investigating the resource allocation implications of another type of outreach facilitation intervention that was designed to prevent the spread of respiratory infection within fps' offices. evidence from a systematic review has shown that influenza transmission occurs primarily by the droplet and short-distance contact routes [ ] . the best practices promoted by the intervention are the droplet and contact precautions, which are described presently. from a clinical perspective, improvement in adoptions of best practices prevents the respiratory virus transmission and therefore, is likely to reduce transmission rates. our particular case consists of an outreach facilitation intervention designed to improve respiratory infection control practices in community-based fp offices. it was conducted in the city of ottawa and delivered by five public health nurses. to our knowledge, it was the first facilitator-based intervention to promote respiratory infection control guidelines. although the intervention has been documented in detail elsewhere [ , ] , we provide a summary of the intervention and its outcomes in this paper. a total of family medicine practices participated in this pre-post intervention observational study, and all completed the study intervention. of the participating physicians, , or % of them, completed all or part of the pre-intervention questionnaire. the objective was to determine the effectiveness (in terms of compliance) of a short-term intervention to facilitate the incorporation of best practices in respiratory infection control in primary care offices. a mnemonic was developed for both the nurses and physicians to summarize the best practices by the acronym "masks" (mask for the patient with cough and a fever, alcohol gel hand sanitization, seating of potentially infectious patient apart from others, "kleen"-disinfection of hard surfaces and signage). the intervention commenced with the public health nurse facilitators providing the baseline audit feedback on the respiratory infection control practices in the participating family physicians' practice to physicians and to other practice staff. physicians were presented directly (and other staff either directly or indirectly through the physicians) with evidence-based best practices and a facilitative "tool kit". this tool kit contained colourful signage outlining best practices for respiratory infection control, signage demonstrating proper hand-washing techniques and use of alcohol-based gel, a reference list of major guidelines sources and web sites, four infection control articles, a box of procedural masks, wall-mounted alcohol gel dispensers with refills, alcohol gel pumps, and hospital-grade disinfectant wipes. during the five-week intervention with their assigned recruited practices, the facilitators worked independently. throughout the intervention the facilitators corresponded with the project team daily and attended scheduled weekly meetings to share information and strategies. in order to measure outcomes, four respiratory control activities for an ambulatory office were viewed as the primary indicators of effective respiratory infection control: ) signage posted in or about the waiting room; ) the receptionist giving masks to patients having a cough and/ or fever; ) instructing patients having a cough and/or fever to use alcohol gel to clean their hands; and ) requesting patients having a cough and/or fever to sit at least one meter away from others. professional nurse auditors were deployed once to obtain data before the intervention and once six weeks after the intervention. the auditor sat for an hour in the waiting room of the physicians' offices and noted the presence or absence of the four respiratory control activities listed just above. they also inquired as to how often potentially contaminated areas were cleaned with disinfectants, and if alcohol-based hand gels were used in examining rooms. the auditors were blinded to the outcome measures and aware only of data gathering requirements. in order to separate the intervention from the data collection, the physicians, office staff and facilitators were blinded from the outcomes and were not informed of the presence of the auditors. statistically significant differences between before and after the intervention were observed for all four of the primary outcome measures: . % ( % ci: . %- . %), . % ( . %- . %), . % ( . %- . %) and . % ( . %- . %), respectively. overall, the number of practices that applied all of the four audited primary prevention measures was . % ( %- . %) prior to the intervention and . % ( . %- . %) following the intervention (p < . ), demonstrating a ( . - . ) percentage point increase in the adoption rate of best practices. this study demonstrated that facilitation of a multi-faceted intervention by public health nurses successfully promoted best practices in respiratory infection control in primary care practices. however, it did not consider health-related outcomes before or after the intervention. we conducted a cost analysis of the respiratory infection control intervention. a standard cost-benefit analysis or cost-effectiveness analysis could not be conducted in this case due to the absence of information on health-related outcomes. as supportive information, we also attempted to evaluate a threshold value for the intervention's effectiveness that could justify the costs incurred by the intervention in terms of the potential cost savings. standard methodological approaches can be found in drummond et al. [ ] and muennig [ ] . we determined the explicit costs of the intervention from the perspective of the provincial government, which is responsible for financing health care in ontario. the potential cost savings for this intervention referred to the costs of medical care averted due to the improved respiratory infection control practices that reduce the probability of infection in the physicians' offices. these implicit cost savings can include the cost of health care provider visits by patients experiencing illness symptoms, the cost of medical tests and procedures, and the cost of hospitalizations that were avoided. the actual explicit costs of the intervention over weeks were gathered from the public health budget rationale ( ) for the inputs of labour, auditing services, supplies, facilitator travel, and honoraria that compensated the practices for the time diverted from normal activities. labour costs referred to the salaries and benefits of the five nurse facilitators and of the . full-time-equivalent project manager. the audit costs included the costs of the audit itself, involving feedback both before and after the intervention provided to the practices, as well as the traveling costs of the auditor. supply costs referred to the costs of the tool kits provided by the facilitators for each practice. an honorarium was paid to each fp practice site for the time it spent participating in the project. in addition to those variable costs, which vary directly with the number of practices that participate, it is important to include the fixed costs of the intervention, which consisted primarily of training the nurse facilitators. the investment in training generated returns extending well beyond the -week period of execution of the intervention. the amortization period for recovering the cost of training is much longer than this time frame for the initial intervention, as the skills obtained from training can be utilized again in subsequent years. the initial training cost should therefore be distributed across the estimated useful life of the investment item, taking into account the discount factor. we selected a discount factor of percent as recommended by other papers containing cost analyses [ , , ] . in addition to the discount rate, we have to select the length of the time period over which to amortize the training cost, which should be related to the life span of the training. as desai [ ] pointed out in his application to obstetrics, often the analyst must assume a length for the useful life span, but this is often initially unclear. existing research from the field of organizational behaviour indicates that the payoffs stemming from a one-time, up-front investment in employer-paid training for human resources intervention tend to decline after four years [ , ] . due to the fact that the facilitators received their training over a two-week period, we adopted a somewhat shorter amortization period for the cost of their training by assuming that it is valid for years. the cost of the training of the five nurse facilitators was amortized over a -year life span at a discount rate of % based on the training expenses that were initially incurred at the beginning of the intervention. in another scenario, we included the entire training cost into cost analysis instead of amortizing it, which would imply that the training has no value after the current season. results were also presented at discount rates of % and % in the mathematical summary which details the discounting process (see additional file ). the other costs listed in the public health budget rationale, such as recruiting participating practices, office assistance, and projection management were not included because these were costs incurred for this particular research pilot project rather than those of the intervention. those costs would not arise in the facilitation intervention implementation if it were to be adopted on a widespread basis. all of the direct costs were presented in micro detail for the -week period over which the intervention was executed, both in terms of total levels and on the basis of costs per practice. as such, the cost estimates that we generated should generalize to similar projects in other geographic areas that are on either larger or smaller scales. we have made some assumptions regarding how a facilitation program might be organized in order to deal with evaluating the costs of training the facilitators. our outreach facilitation program is most likely to be effective if delivered during or just before the peak season for respiratory infections (i.e., september, october, and november). hence our training activities would ideally be applied for months per year over years, generating a cumulative total of months of utilization. while the program would aim to introduce proper respiratory infection control practices to be followed all year round, the medical practitioners might be more interested just prior to the influenza season. therefore, although the training remains valid for years into the future, we envisaged that the program would be delivered during that -month period every year. we nevertheless also produced estimates based on the scenario for which the intervention is executed yearround. while the explicit costs of implementing this intervention can be assessed with accuracy, it is much more difficult to estimate the implicit cost savings because of the lack of information regarding a key event, namely the reduction in the probability of spread as a direct result of the intervention. we assume without solid evidence that improved infection control reduces the respiratory infection rate at physicians' offices, but we certainly do not know how by much the probability of infection changed after the intervention. in order to generate an accurate estimate of the total health-care costs averted by this intervention, one would require the following pieces of information: i) the incidence or frequencies of transmission at physicians' office, ii) the effect of the intervention in reducing those rates, iii) the probabilities of the various potential health outcomes that could arise given infection, and iv) the cost of the treatments associated with those outcomes. with the exception of item iv), these pieces of information were not available. drawing from several data sources in the literature, we therefore adopted an indirect approach to esti-mate the potential health-care costs that might be averted as a result of the intervention, and we attempted to make a case that the potential benefits were large relative to the explicit intervention costs. there are a range of treatments for different influenza patients according to the seriousness of the infections. the patient who is infected with influenza may rest at home, visit an emergency room, or be hospitalized. if the patient only needs care at home, he or she may request sick leave from his or her job. in such a case, cost arises from the patient's perspective or the societal perspective (from the lost output) but not from the ministry of health's perspective. another possibility is that a few patients die from influenza, but it is impossible to attach a precise value for the cost of death. therefore, we only took the intermediate events of outpatient visits and hospitalizations into account in estimating the avoided costs. the costs denominated in us dollars (as they were presented in some studies that we cited) were converted into canadian dollars by the current exchange rate [ ] , and costs from data in prior years were adjusted for inflation and denominated in constant dollars using appropriate component of the consumer price index [ ] where necessary. the underlying approach for the cost analysis of the intervention involves an efficacy rate, which is defined as the decrease in the probability of transmission that is attributable to the intervention. we could not evaluate this quantity, but we could evaluate the threshold value that would render the intervention beneficial, which was judged to be worthwhile if: cost savings -intervention cost > = . the cost savings attributable to the intervention were expressed as follows: (cost of hospitalization for a flu patient*number of flu cases avoided due to the intervention in the physician's office*proportion of the infected people who were hospitalized) + (cost of outpatient visit for a flu patient* number of flu cases avoided due to the intervention in the physician's office*proportion of the infected people who had an outpatient visit). note that infected individuals who were hospitalized or who had an out patient visit may or may not have passed through the fps' practices; there are other modes of infection besides transmission in these clinics. the second element in each of the terms in parentheses, which is a counterfactual, can be expressed as: the number of flu cases avoided in the physician's office due to the intervention = number of patients visiting the physician's office* (probability of contracting influenza in that office without the intervention -probability of contracting influenza in that office with the intervention). substituting that expression into the primary equation yields the following expression after a slight algebraic manipulation: (probability of contracting influenza in the office without the intervention -probability of contracting influenza in the office with the intervention) = intervention cost/ [number of patients visiting the physician's office*(cost of hospitalization*proportion of infected people who were hospitalized + cost of outpatient visit*proportion of infected people with an outpatient visit)]. a critical element for this calculation is transmission rates for influenza in settings such as physicians' offices. while there are articles in the literature dealing with the incidence of transmission of certain viruses within the general population, we were unable to find research pertaining to the incidence of transmission within physician offices or similar locations involving close contact with the public, such as waiting rooms, emergency rooms, and school busses. we searched for papers on medline, cinahl and embase by the key words "influenza or flu, and transmission or infection, and bus or waiting room or emergency room or emergency department or physician office", and we also asked for help from several experts in this area to search for the requisite information. we did locate some information regarding the incidence of transmission of influenza during airline flights. in our judgment, however, these figures are not reliable estimates of the rate of infection with and without the intervention that would occur in a fp's office. in light of that source of uncertainty, our approach was to calculate an estimated value for the left side of the above expression (i.e the reduction in the likelihood of infection) that represents a threshold value for the minimum efficacy of the intervention such that the potential cost savings of the intervention outweigh its costs. we solve that expression for the lowest possible value at which the net costs of the intervention would be negative. if the efficacy of the intervention is any lower than that value, its net costs would be greater than . table presents the number of hours of intervention work activity and the percentage of total hours spent at the medical practices by the facilitators. the total number of hours worked was ( days × facilitators × hours/day). in table , it should be noted that the time spent on "other" needs to be removed from the analysis, as that labour time was not allocated to the project. therefore, the total hours for the five public health nurse facilitators spent on the intervention should be the figures listed under the "total" label minus those listed under the "other" label, which worked out to a total of hours. on average, they spent approximately / hours at each practice for which they were responsible. given a yearly salary of one nurse facilitator of $ , (in the ottawa area) and an annual total of , hours worked in one year, the hourly wage rate of one nurse facilitator was $ . . this generated a labour cost per practice of $ . . in the -week intervention period, the labour costs (for time actually worked) for all five nurse facilitators combined amounted to $ , . ($ . × hrs). the total costs for the intervention are presented in table . the third column provides the data on the costs of the outreach facilitator intervention denominated in dollars on the basis of the -week period during which they actually worked. the fourth column contains similar data, except that all of the costs were estimated on the table were based on the assumption that the facilitators would work at the same pace for an entire year, and would thus visit approximately practices. the difference between these three scenarios consists of a pro-rating of all of the variable costs while holding the training costs fixed. the figures in the last column were exactly same as those in the third column except for the training cost. the training cost presented in the last column was not amortized over years, which accounts for the approximately threefold increase in the training costs coupled with no change in the other costs. the intervention costs amounted to a total of $ , . that was actually incurred over the -week intervention, $ , . per flu season, or $ , . per calendar year. in order to extrapolate these cost figures to other geographical areas, the distinction between the variable costs and the fixed costs plays an important role. the variable costs that were incurred for this -week project amount to $ , . , which is the sum of all of the costs listed in the third column of table the (fixed) training costs must be calculated in a different fashion, however. as explained in the mathematical summary (see additional file ), we calculated an annual value of $ , . for the training costs. this figure is equivalent to $ . for each participating practice. had these five facilitators worked for the entire -week flu season, the total training costs would still be $ , . , but many more practices could have been involved, thus lowering the per-practice training cost to $ . ($ , . / practices). if these same facilitators were to be assigned to this project on a year-round basis, the per-practice training costs would become one quarter of the prior figure, or $ . , because the nurses work times longer during the year. the first element that we obtained for the expression for averted costs was the number of patients that passed through the offices of the participating physicians, and were therefore at risk of becoming infected. in pre-intervention questionnaire, physicians were asked how many patients they typically see per half-day, from which we may estimate the number of patients visiting the physician offices during a -week period. physicians responded to the question, and the mean value was . . imputing this value to all of the physicians that were covered in our intervention, approximately , patients visited the participating physicians over weeks. *the training cost was amortized over years using % discount rate, and therefore the training cost for each calendar year was $ , . , as shown in the additional file . this training cost for five facilitators would be totally fixed for a -year period. even if these facilitators were to conduct this activity for the entire flu season, and thus serve more than practices, this cost would not change. costs for the -week intervention based on the actual -week length of the project. assume that intervention lasts months ( weeks) during flu season. all figures except the training cost were obtained by converting the -week totals (that apply to our particular intervention) listed in third column to weekly rates and then multiplying by . assume that intervention lasts one year ( weeks). all figures except the training cost were obtained by converting the -week totals (that apply to our particular intervention) listed in third column to weekly rates and then multiplying by . costs for the -week intervention based on the actual -week length of the project and the training cost was not amortized. the figures for the costs of treating influenza patients were drawn from several sources. hogg, baskerville and lemelin [ ] performed a cost savings analysis associated with administering influenza vaccine in the elderly. they obtained the estimated cost of an emergency room visit due to influenza from jacobs and hall, which was approximately cn $ . in or cn $ . in [ , ] . this cost was virtually identical to the costs for an outpatient visit reported in other studies [ , ] . we turn next to the proportion of people infected with influenza that ended up being hospitalized. an estimate of this proportion can be obtained by dividing the hospitalization rate among all subjects with influenza (regardless of where it was contracted) by the proportion of all subjects who become infected (regardless of where it was contracted). the latter quantity can be thought of as the illness or transmission rate of the influenza. in an analogous fashion, an estimate of this proportion of infected people who had an out-patient visit can be obtained by dividing the out-patient rate among all subjects with influenza by the proportion of all subjects who become infected. unfortunately, we could find no paper in the literature that provided values for the hospitalization rate or the outpatient visit rate given that a patient has influenza. in order to obtain rough estimates of these quantities, we borrowed heavily from the paper by nichol [ ] that dealt with vaccination against influenza. by a systematic literature review, the author obtained estimates of 'the hospitalization rate due to influenza and its complications', 'outpatient visit rate due to influenza and its complications', and 'the influenza (and its complications) illness rate' among healthy working adults aged between and years. nichol also derived from the monte carlo simulation the difference of the hospitalization rate (as well as the outpatient visit rate and the illness rate) for influenza and its complications between unvaccinated and vaccinated subjects. however, the influenza's complications were widely defined in nichol's paper. in our analyses, we focused on only influenza and pneumonia associated with influenza. therefore, by assuming that vaccination is % effective in preventing episodes of influenza (and pneumonia associated with influenza), we used the number of vaccinated individuals as a proxy for the number of non-infected subjects. in this respect, the three difference rates reported by nichol can be interpreted as each of these three incidence rates due to influenza only (and pneumonia associated only with influenza). therefore, we used these differences, . % ( % probability interval (pi): . %, . %), . % ( . %, . %), and . % ( . %, . %), as our estimates for 'the hospitalization rate due to influenza only', 'outpatient visit rate due to influenza only', and 'the influenza illness rate', respectively. when we inserted these values into the expression for the proportion of infected patients who ended up hospitalized, we obtained a value of . (pi: . , . ), and the value for the proportion of infected patients who had an outpatient visit was . (pi: . , . ). inserting all of the figures that we obtained above back to the primary expression for the cost savings, and combining that information with the value for the explicit costs of intervention, the efficacy of the intervention (probability of contracting influenza in the office without the intervention -probability of contracting influenza in the office with the intervention) was equal to . % (pi: . %, . %). the implication is that the threshold value for the efficacy at which the cost savings of the intervention barely outweigh the costs was . %. the goal would thus be to reduce the probability of infection occurring in fps' offices by at least . %. in addition, if we included the non-amortized training cost into analysis, the threshold value rose slightly to . %. the figures that entered into the calculations are presented in table . this paper has provided detailed information on the costs of an outreach facilitation initiative designed to prevent the spread of infectious diseases by promoting best practices in respiratory infection control in primary care practices. we have generated accurate estimates of the explicit costs of implementing such a program on a per-practice basis, which permits the extrapolation of these unit costs to other geographical domains. we have also provided some preliminary estimates of the potential cost savings to the health-care system. due to the lack of knowledge about the frequency of respiratory infection occurring at physicians' offices, particularly an estimate of the reduction in the probability of infection attributable to the intervention, we did not have enough evidence to evaluate precisely the benefits of the intervention. as an alternative approach, we undertook a threshold analysis to estimate a threshold value of the efficacy that could render the intervention cost saving. based on our conservative estimates referring to direct savings in the form of healthcare costs averted, there are indications that the outreach facilitation intervention program would result in cost savings if it could achieve a reduction in the probability of infection at the physician offices on the order of . percentage points. this implies that if we assume that there was a . % chance of contracting influenza in fp offices without intervention, to achieve the efficacy rate of . %, the probability of contracting influenza in fp offices with intervention would be . %, representing a large relative risk reduction in influenza transmission in fp offices. on the other hand, if we assume a higher probability of contracting influenza in fp offices without intervention, such as %, to achieve the targeted efficacy rate of . %, the probability of contracting influenza would be approximately . %, representing a smaller relative risk reduction in influenza transmission in fp offices. moreover, in addition to the direct cost savings to the health care system that may be realized, there are potential indirect cost savings associated with our intervention as well, such as the potential to avoid disastrous human loss and suffering caused by viruses such as the severe acute respiratory syndrome (sars). the scope of the influences of the infectious diseases such as sars and influenza extend far beyond the costs that were mentioned above, especially in the health-care and tourism sectors. the total costs in terms of lost production of the sars epidemic to toronto's economy had been estimated to be $ billion, and the estimate for the economic cost for all of canada was around $ . billion [ ] . within the health care sector, the indirect costs borne by non-sars patients were enormous. sars affected all health-care workers -especially those on the front line -and delayed "non-emergency surgeries" such as organ transplants and cancer radiation [ ] . according to ontario health minister tony clement, as of june , , sars had cost ontario's health-care system $ million, which was spent mostly on special supplies and added health-care workers needed to protect health-care workers, as well as on constructing specialized sars clinics and isolation rooms [ ] . this in turn had a huge impact on non-sars related health care system utilization, both due to diversion of resources as well as severe stress amongst the health care providers. for instance, a study comparing the periods before and during the sars outbreak in the gta and non-gta areas by woodward et al. found the greatest impact of sars on reduction in the utilization of inpatient and outpatient hospitalization, diagnostic testing, physician and emergency department visits, use of prescription drugs, intensive care bed availability, and cardiac care during april to may [ ] . avoiding such negative consequences implies that our intervention may also generate implicit or indirect cost savings. the -week intervention costs amounted to a total of $ , . . the results of the cost analysis suggest that the intervention can be cost saving because the . % point reduction of the probability of influenza at the physicians' offices appears to be a feasible target for the effectiveness of the studied intervention. a facilitation intervention tailored to the environment and needs of the family practice and walk-in clinics is of great promise for improving respiratory infection control in the physicians' offices. future research to conduct further economic evaluations of such an intervention based on adequate dataparticularly in relation to infection incidence rates and the ability to lower them -would aid in important public health policies and administrative decision-making on implementing preventive care guidelines. the author(s) declare that they have no competing interests. william hogg and patricia huston conceived the intervention, provided the data, participated in the study design and in critical revisions of the manuscript, and contributed to all other aspects of the study. david gray contributed to the cost evaluation study design, assisted with the calculations, and thoroughly revised the manuscript. wei zhang acquired the economic data, performed the cost evaluation analysis and interpretation, and drafted the first pass of the manuscript. all authors read and approved the final manuscript. transmission of influenza a in human beings promoting respiratory infection control practices in primary care: primary care/public health collaboration. accepted for publication in canadian family physician the inexact science of influenza prediction cost savings associated with improving appropriate and reducing inappropriate preventive care: cost-consequences analysis harnessing primary care capacity to enhance the public health response to respiratory epidemics. accepted for publication in canadian family physician methods for the economic evaluation of health care programmes oxford designing and conducting cost-effectiveness analyses in medicine and health care the cost of emergency obstetric care: concepts and issues assessing the economic impact of personnel programs on work-force productivity recouping training and development costs using preemployment agreements. employee responsibilities and rights journal daily foreign exchange rates look-up statistics canada: the consumer price index and major components for health care estimating the cost of outpatient hospital care cost benefit of influenza vaccination in healthy, working adults: an economic analysis based on the results of a clinical trial of trivalent live attenuated influenza virus vaccine the economic impact of pandemic influenza in the united states: priorities for intervention cost-benefit analysis of a strategy to vaccinate healthy working adults against influenza the economic impact of sars. cbc news online updated the toronto sars experience. division of infectious diseases utilization of ontario's health system during the sars outbreaks: an ices investigative report. institute for clinical evaluative sciences we are very thankful to dr. carmel martin, who contributed in the design and implementation of the facilitation study. we also thank three referees for their efforts and their constructive comments. the pre-publication history for this paper can be accessed here:http://www.biomedcentral.com/ - / / /pre pub mathematical summary -amortization of training costs. click here for file [http://www.biomedcentral.com/content/supplementary/ - - - -s .doc] key: cord- -p e authors: tan-torres edejer, tessa; hanssen, odd; mirelman, andrew; verboom, paul; lolong, glenn; watson, oliver john; boulanger, lucy linda; soucat, agnès title: projected health-care resource needs for an effective response to covid- in low-income and middle-income countries: a modelling study date: - - journal: lancet glob health doi: . /s - x( ) - sha: doc_id: cord_uid: p e background: since who declared the covid- pandemic a public health emergency of international concern, more than million cases have been reported, as of aug , . this study aimed to identify what the additional health-care costs of a strategic preparedness and response plan (sprp) would be if current transmission levels are maintained in a status quo scenario, or under scenarios where transmission is increased or decreased by %. methods: the number of covid- cases was projected for low-income and middle-income countries for each of the three scenarios for both -week and -week timeframes, starting from june , . an input-based approach was used to estimate the additional health-care costs associated with human resources, commodities, and capital inputs that would be accrued in implementing the sprp. findings: the total cost estimate for the covid- response in the status quo scenario was us$ · billion over weeks, at $ · per capita. for the decreased or increased transmission scenarios, the totals were $ · billion and $ · billion, respectively. costs would triple under the status quo and increased transmission scenarios at weeks. the costs of the decreased transmission scenario over weeks was equivalent to the cost of the status quo scenario at weeks. by percentage of the overall cost, case management ( %), maintaining essential services ( %), rapid response and case investigation ( %), and infection prevention and control ( %) were the main cost drivers. interpretation: the sizeable costs of a covid- response in the health sector will escalate, particularly if transmission increases. instituting early and comprehensive measures to limit the further spread of the virus will conserve resources and sustain the response. funding: who, and uk foreign commonwealth and development office. on jan , , who declared sars-cov- a public health emergency of international importance, later formally identified as covid- . the declaration advised the member states to prepare for containment and prevention of onward spread of the virus. after a week, cases were reported, % of which were in china and the rest in other countries. in response, who appealed for us$ million to support member states over a -month period, as they began implementing priority public health measures. the priority public health measures were outlined in the eight pillars of the strategic preparedness and response plan (sprp), and ranged from country coordination to clinical case management. as of july , , more than million cases of covid- , including more than deaths, had been reported globally. who explicitly expanded the scope of the sprp to include a ninth pillar on the maintenance of essential health services in acknowledgment that the pandemic was already straining the health system. who also released guidance on public health and social measures (phsm) to slow down the transmission of the virus. countries closed offices, schools, restaurants, places of worship, and banned large gatherings to restrict movement and to avoid further straining of the health system. epidemiological models have predicted that many more deaths and infections would have occurred if these measures were not implemented. however, the social and economic repercussions of the phsm are also beginning to emerge. the world bank has forecast global gdp will contract by · % in , on the assumption that measures will start to be lifted in the second half of the year. if the covid- pandemic persists, and movement restrictions are maintained or intensified, greater losses are predicted. this study aims to project the future costs of the strategic response and preparedness actions in the health sector to counter the covid- outbreak. given the uncertainty in the future course of the disease, estimates are provided in the short term, and separate scenarios are modelled where current measures to restrict movement are maintained, relaxed, or intensified. this study estimates the costs of implementing the nine pillars of the sprp in low-income and middle-income countries (appendix p ), accounting for · % of the total population in that group of countries. the nine pillars of the sprp and the key cost items in each pillar are presented in table . the study includes low-income countries, and the most populous lower-middle-income and upper-middleincome countries, and it excludes countries for which no gdp or epidemiological data were available. the costs were additional to what is currently known to exist, or to have been spent by the countries at the start of the analysis (june , ), and were estimated in the -week and -week periods after this date (ie, until july and sept , ). the costing was therefore synchronised chronologically to show the same time period in countries at different stages of the epidemic. all of the one-time and recurrent inputs that were expected to occur within these two time periods to prevent new cases, and to treat prevalent and incident cases, were costed. during this time, the course of the pandemic might change, depending on decisions taken by national leaders on either relaxing or intensifying phsm. in an attempt to capture this potential uncertainty, for each time period, three scenarios were analysed with the current measures to restrict movement, and facilitate physical and social distancing, being either maintained, relaxed, or intensified. only costs expected to be borne by the health sector were included, and costs related to any social mitigation interventions, such as cash or in-kind transfers, were excluded. an inputs-based approach was taken, where quantities of items related to each activity were multiplied by the unit price for each item. the interim guidance documents issued by who and consultations with experts from relevant technical programmes were the sources of the types and quantities for key items. the estimated number of cases of covid- were secondary data taken from the epidemiological model from imperial college (london, uk). this model was used because it provides publicly available estimates for a evidence before this study since jan , , when who labelled the covid- pandemic a public health emergency of international concern, countries have tried to limit its spread, instituting measures on physical distancing and restrictions on movement. with more than million cases reported, the world bank and other major financing institutions have projected an overall contraction of • % of global gross domestic product due to covid- in , with persisting effects in the years to come. this projection was made on the assumption that the restrictions will be lifted in the second half of ; however, the costs of the actions needed to respond to the pandemic, which could enable the lifting of these restrictions, have not been estimated for low-income and middle-income countries. from a different perspective, some costing work has been done on preparedness. in , after the ebola outbreak, the national academy of medicine launched the commission on a global health risk framework for the future. the commission estimated us$ • billion a year globally for pandemic preparedness versus an annualised expected loss from potential pandemics of more than $ billion. in december , the international working group on financing preparedness issued a report on investing in health security. based on a few country studies costing the multisectoral national action plans for health security, they estimated a cost of $ • - per person per year on preparedness. to our knowledge, this is the first study costing a strategic response to covid- , a public health emergency of international concern. considering the baseline preparedness of low-income and middle-income countries, and the limited resilience of their health systems, major investment will be needed to counter the virus. the result of the status quo scenario, a health-care cost total of us$ • billion or $ • per capita after weeks for low-income and middle-income countries, is not an insignificant cost, but reflects the constrained capacity in the countries facing a virus that has spread and established itself. some hope is offered by the scenario in which the public health and social measures are intensified, resulting in a decrease in transmission by %. however, the costs, when the restrictions are relaxed and transmission increases by %, escalated at weeks and further escalated at weeks. the results show which pillars of the strategic preparedness and response drive the costs. this study should inform governments, as they consider relaxing restrictions to jumpstart their economies. the arguments for investing in preparedness are strong, juxtaposed against the price tag for the response versus covid- , and coupled with the expected shock on the global economy. future work at the country level is needed to strategically identify the gaps in both preparedness and response against not only covid- , but also for other potential future pandemics. see online for appendix large number of low-income and middle-income countries. this susceptible, exposed, infected, and recovered or removed (seir) model was calibrated on confirmed deaths from the start of the covid- outbreak up to june , . imperial college runs the model regularly for all countries, except those where low levels of reported covid- deaths does not permit accurate modelling. for this costing exercise, countries without projected covid- epidemiology from the model were china, iran, tanzania, uganda, and zimbabwe. for these countries, a separate seir model, provided by imperial college as a script in the r programming language, was run by our research team using effective reproduction (rt) values taken from the centre for mathematical modelling and infectious diseases repository associated with the london school of hygiene & tropical medicine (london, uk). rt values are commonly described as the number of contacts that a case infects. the model projected cases for the weeks and weeks following june , , under three scenarios: status quo (maintain current transmission), an increase in transmission by %, and a decrease in transmission by %. the increased and decreased transmission scenarios work through changes in the rt and the level of mobility in the epidemiological model. as the projections are made based on the current state of the pandemic in each country, the results reflect a wide range of response strategies. we also report outputs of the epidemiology modelling at the start and the end of the period according to the oxford stringency index, which measures the level of covid- mitigation measures implemented at the country level. capital costs included within the resource needs for covid- response are intended for upgrading laboratories for diagnostic testing (pillar ), buying field hospitals to expand capacity for treating covid- patients (pillar ), and repurposing health facilities to enable them to cope with non-covid- patients who would otherwise have been treated in hospitals providing care to covid- patients (pillar ) to lift the supply side constraint of hospital and intensive care unit beds, procuring communications equipment (pillar ), and providing motorcycles for contact-tracing teams (pillar ; table ). another capital cost would be the provision of handwashing stations for hygiene (pillar ). in addition to capital costs, a series of one-time costs are included, such as the hiring of consultants to develop or adapt guidance documents, prepare online training courses, document plans, design communications materials, and other related duties. all these components would be scaled depending on the level of the epidemic and according to appropriate administrative scalars (eg, by the number of subnational administrative units or number of health facilities per country). the essential supplies forecasting tool version (esft ) was used to estimate the costs of key commodities and supplies as part of the covid- response. these commodities and supplies included personal protective equipment, single-use masks, diagnostic tests, supportive drugs (including dexamethasone), disposable supplies, and oxygen for hospitalised patients. to estimate the quantities of commodities needed for a country's covid- response, the esft combined the assumptions on the number of items related to each case with the number of cases, split by severity. only % of cases ( % severe and % critical) were assumed to need hospitalisation. the prices of each item, although found in the esft , were updated using international market prices. for diagnostics and testing, the esft assumed that all hospitalised covid- patients were tested, and that there was a targeted testing strategy, where % of all suspected cases were also tested. testing was constrained by a country's diagnostic capacity, as determined by the available diagnostic instruments and the number of laboratory technicians available to focus on covid- diagnostics and do the pcr-based tests. these supply-side constraints were lifted partly by the purchase of automated extraction platforms, expanding the working week for laboratories from days to days, and adding another h shift to laboratory operations. an assumption was made that a supply side constraint existed, and only a maximum of % of the existing health workers could be prioritised for the covid- response. they continued to receive their salaries, and these are not included in the costing. incentives, both financial and non-financial (eg, paid sick leave including time spent in quarantine; occupational risk insurance or life insurance; ensuring treatment for illness; provision of child or elder care support; or accommodation near the health facility, transport, or relocation allowance, or all three), estimated at % of the average monthly salary, were paid to all those working directly in the covid- response in health facilities. hazard pay at % of salary was paid to all those at increased risk, defined as those having close contact with a covid- -positive patient. to maintain essential health services, salaries were paid to new hires to replace half of the number of existing health workers prioritised for covid- response, on the assumption that % replacement was not needed because non-urgent consul tations and elective admissions are being postponed. the new hires were expected to come from the private sector, or from retirees or soon-to-be graduates. salaries were obtained from the who-choice salary database and were updated to us$. to capture the main uncertainty in the cost of the pandemic response that arises from the course of the pandemic itself and the policy responses of the governments, both increase and decrease in transmission of % were modelled. in addition, because providing incentives is a policy response that governments might choose to exercise, the costs are presented with (base case) and without the incentives. more details are available in the full technical documentation (appendix pp [ ] [ ] . resources from who (funding for consultants and salaries of staff) were used to produce the estimates in this paper. the authors (all from who except ow, who is funded by the uk foreign commonwealth and development office) were solely responsible for the design, conduct, analysis, and writing up of the study. the corresponding author had full access to the data and took the decision to submit for publication. at the start of the analysis on june , , seven countries had an rt of less than , two countries had an rt of or more, but most countries had an rt of - (table ) . across the rt categories, the median oxford stringency index, ranged from · to · (with representing the most stringent measures); the mean number of daily contacts, which is the number of personal interactions, ranged from • to · ; and the percentage of the population infected, which is an estimate of cumulative infections, ranged from • % to • %. at the end of weeks, the percentage infected was projected to increase in the status quo, particularly in those with an rt of · or greater, and much larger burdens were projected for the % increase in transmission scenario. under the % decrease in transmission scenario, only a slight increase in the percentage infected was projected. during the -week timeframe, a similar pattern emerged, and many more cases were projected in the status quo and % increased transmission scenarios, whereas in the % decreased transmission scenario, the case burden remained relatively stable, except in countries where the rt was or greater. the costs of the covid- response in low-income, lower-middle-income, and upper-middle-income countries after weeks and weeks under the different scenarios are shown in table . the total cost at this stage of the epidemic, if the status quo is maintained over weeks, is $ • billion with a per-capita cost of $ • . if more measures to facilitate physical and social distancing, and to restrict movement were applied, and countries' transmission was reduced by %, the -week resource requirements would be reduced to $ • billion ($ · per capita). with % increased transmission, under a scenario of relaxed restrictions, costs of $ • billion ($ · per capita) over the same -week period would be generated. in the -week projection, costs would more than tripled under the status quo and % increased transmission scenarios. the costs of the % decreased transmission scenario over weeks is equivalent to the cost of the status quo scenario at weeks. most of the costs would be accrued in the middle-income countries. the top ten countries (appendix p ) would account for % of the costs in the -month status quo scenario, and this pattern is stable across the different timeframes and scenarios. the dominance by a handful of countries is due to a combination of factors: larger populations, higher prices, and a more widespread epidemic. the distribution of the costs over the nine pillars are shown at weeks in table (data for weeks are provided in the technical documentation; appendix p ). under the status quo scenario, case management would account for around % of the costs, % would go to maintaining essential health services, and around % to investigation, surveillance, and rapid response. the building of handwashing stations, and procurement of personal protective equipment and cloth masks within pillar accounts for about % of the cost. these pillars would be the major cost drivers of implementing an effective covid- response. the pattern of the distribution of the costs is generally maintained under the % increased and decreased transmission scenarios, except for a decrease in the proportion of costs in investigation in the % decreased transmission scenario and an increase in the same costs under the % increased transmission scenario, compared with the status quo scenario. the costs by category for human resources, commodities, capital, and other costs at weeks and weeks for the status quo scenario are shown in table . at weeks, capital costs are nearly equivalent to human resources costs; however, at weeks, the costs of human resources becomes higher than all other categories, at % of the total cost. recurrent costs are primarily for human resources, and secondarily for commodities. costs for human resources are high, at $ · billion at weeks, and they are driven by salaries for newly hired staff and incentives. the cost of the status quo scenario would decrease to $ billion and $ billion at weeks and weeks, respectively, if incentives are not included. as of june , , the costs of the full, nine-pillar response to covid- in low-income and middleincome countries after weeks, on july , , were projected to be approximately $ billion, assuming that the rt was unchanged and the status quo continued. costs are estimated to be more than three times that amount after weeks on sept , , under a status quo scenario. the costs were projected to be greater at weeks and weeks if transmission values increased by %. this analysis shows that the cost of responding to a pandemic with % decreased transmission at weeks is coinci dentally equivalent to the cost at weeks under the status quo scenario. the per-capita cost of the response under the status quo scenario for weeks is $ • for low-income countries and $ • to $ • in lower-middle-income and upper-middle-income countries. for weeks, the costs per capita are $ • for low-income countries and about $ for middle-income countries. to put this in context, the health expenditure per capita in , for a whole year, in low-income countries was $ , and from $ to $ in lower-middle-income and upper-middleincome countries. the potentially huge opportunity costs within the health sector in not responding rapidly are clearly evident. the benefits of acting early and comprehensively, like in vietnam, are a clear lesson that can be drawn from this costing exercise. an early and rapid response will not only mitigate future covid- costs, but more importantly, it will be able to mitigate future covid- costs because of a lower number of covid- infections, and a corresponding lower number of deaths and long-term consequences among survivors. a strong pillar response on maintaining essential health services can also potentially decrease the number of deaths indirectly caused by covid- . social and economic disruptions can also be shortened. this analysis also shows the interconnectedness of the nine pillars of the covid- response. as the number of cases increases, the share of costs found in case management (in pillar seven) and in maintaining essential health services (in pillar nine) both increase. increases in the number of cases will also generate increased demand for personal protective equipment, hospitalisation and attendant costs, and contact tracing. however, it is important to note that, for preparedness, all countries must invest in more handwashing stations, and better risk communication and community engagement, even with low numbers of cases. the predicted resource needs for a full response for weeks continue to be onerous burdens for countries with a high expected number of cases. however, some of the resource requirements can be decreased by examining where efficiencies or cost savings can be made. the analysis described in this article has used international market prices that are readily obtainable for many commodities. however, some items can be locally produced, such as personal protective equipment (including gloves and cloth masks), some medicines, and single-use supplies. testing kits might be able to be produced at a lower price in countries with local manufacturing capacity, and good quality assurance and regulatory capacity. for human resources needed to respond to covid- and maintain essential health services, perhaps the current workforce is capable of providing enough surge capacity, and together with approaches such as telemedicine, task shifting, and quick upskilling through intensive training and supervision, there will be no need to replace the health workers directly engaged in the covid- response, and the large resource requirement this implies. however, the assumption of spare capacity within the health workforce in low-income and middleincome countries should be questioned. the health system response to covid- has been shown to have a negative impact on the delivery of other services, from immunisation to non-communicable diseases, with decreased coverage rates, substantiating the need to at least partially replace health workers prioritised to the covid- response. aside from hiring new health workers and paying salaries, hazard pay and incentives should be provided to workers in direct contact with patients diagnosed with covid- . countries might choose whether they will provide incentives, but hazard pay for arduous conditions is consistent with legally binding conventions of the international labor office. a more effective approach to reduce the costs will be to decrease the transmission of the virus and have fewer cases to respond to, from the implementation of interventions such as contact tracing and subsequent effective quarantine or isolation, washing stations. all these individual-level measures have been fully costed within this exercise, but their slowing of the transmission of the virus has not been taken into account, as each country's rt is fixed at the start for the period of analysis. as such, the true costs for countries would probably be lower than those estimated per scenario. this difference highlights the need for more dynamic and more frequent modelling and costing to get a more accurate estimate. the precision of the modelling used and the scope of the study have some limits. the first is that the costing is primarily driven by the epidemiological model used. running an epidemiological model and making projections for many countries is fraught with uncertainty, especially given the assumption that the rt remains fixed over the -week and -week timeframes. in this exercise, to cope with this uncertainty, scenarios with different transmission levels were projected to provide higher and lower bounds to the base case estimate. in terms of scope, this costing exercise did not include the isolation or quarantine costs of people with mild to moderate covid- , and their contacts who are unable to successfully isolate or quarantine themselves in their own homes, and where mass quarantine shelters or facilities would need to be set up. this could potentially be a large cost, but it is usually borne by local governments or ministries of social welfare. the use of international market prices, without freight, insurance, and import tariffs also underestimates the costs. however, countries have been known to allow time-bound, tariff-free entry for supplies and medicines for covid- . in addition, countries would have to bear costs of waste management of the covid- response, primarily for non-durable personal protective equipment, which are not included in our estimates, but could require significant amounts of resources. finally, these costs would change significantly once directly acting medicines or vaccines proven to be effective against covid- are produced and added to standard treatment or prevention protocols. in summary, the results of this study show the need to account for health systems in the context of health security. preparedness for health emergencies and disasters has been highlighted as a key component of the common goods for health that require explicit public investment to overcome market failures. these results emphasise that critical components of health systems essential to the surge capacity, which can deliver an effective response (eg, human resources and laboratories), need to be in place, and mechanisms for mobilisation need to exist for when an outbreak occurs. this study also shows that, when faced with a decision to adjust phsm, epidemiological modelling and costing of different scenarios based on different rt values, often reflecting various policy options, updated frequently and using good local data, can be informative. whatever the estimated costs of the response, it might be the case that this amount is not fully within the financial capacity of low-income and some middle-income countries. this gap in the resources can be partly filled by development partners and the private sector. to facilitate modelling, costing, and priority setting, who will be releasing a country level costing tool based on this exercise. it will be made available through the covid- partners platform, where countries and partners can interact in real time to prepare for and respond to the covid- pandemic. finally, this study highlights that while fully implementing a covid- response will entail significant resource needs, the impact of such an early and comprehensive response in limiting the spread of the virus will markedly reduce the resources needed to respond to a more widespread pandemic just a few weeks later. tt-te, oh, am, gl, llb, and as conceptualised the article. tt-te, oh, am, gl, and ojw reviewed articles, contacted experts, and collected data. tt-te, oh, am, pv, and ojw ran the analysis. tt-te, oh, and am wrote the first draft of the article and revised it based on feedback from co-authors. all authors reviewed and approved the article. we declare no competing interests. covid- as a public health emergency of international concern (pheic) under the ihr us$ million needed for new coronavirus preparedness and response global plan operational planning guidelines to support country preparedness and response director-general's opening remarks at the media briefing on covid- who. covid strategy update. overview of public health and social measures in the context of covid- oxford covid- government response tracker the effect of large-scale anti-contagion policies on the covid- pandemic global economic prospects the impact of covid- and strategies for mitigation and suppression in lowand middle-income countries projections of covid- epidemics in lmic countries covid- essential supplies forecasting tool clinical management of covid- : interim guidance emergency global supply chain system (covid- ) catalogue global health observatory data repository global health worker salary estimates: an econometric analysis of global earnings data global spending on health: a world in transition combating the covid- epidemic: experiences from vietnam excess mortality from the coronavirus pandemic (covid- ) updated estimates of the impact of covid- on global poverty the-use-of-masks-in-the-community-during-home-careand-in-healthcare-settings-in-the-context-of-the-novel-coronavirus-( -ncov)-outbreak new zealand eliminates covid- rapid assessment of service delivery for ncds during the covid- pandemic ilo nursing personnel convention no. effectiveness of isolation, testing, contact tracing, and physical distancing on reducing transmission of sars-cov- in different settings: a mathematical modelling study physical distancing, face masks, and eye protection to prevent person-to-person transmission of sars-cov- and covid- : a systematic review and meta-analysis de-escalation by reversing the escalation with a stronger synergistic package of contact tracing, quarantine, isolation and personal protection: feasibility of preventing a covid- rebound in ontario, canada, as a case study covid- : trade and trade-related measures who. global research on coronavirus disease (covid- ). financing common goods for health: core government functions in health emergency and disaster risk management covid- partners platform we acknowledge the modelling group from the mrc centre for global infectious disease analysis at imperial college london (azra ghani, patrick walker, and charlie whittaker) for providing up to date epidemiological projections. we also acknowledge justin graves, luke baertlein, zachary panos, and owen demke from the clinton health access initiative for their technical support with the use of essential supplies forecasting tool version , and technical inputs on diagnostic testing. finally, we acknowledge expert contributions from staff at various departments at who (juana paola bustamante, hong anh chu, giorgio cometto, bruce gordon, fiona gore, lisa hedman, sara hollis, teena kunjumen, ben lane, margaret montgomery, pryanka relan, teri reynolds, cris scotter, adriana velazquez, and lara vojnov). key: cord- -z vhxg authors: gardiner, fergus w.; de graaff, barbara; bishop, lara; campbell, julie a; mealing, susan; coleman, mathew title: mental health crises in rural and remote australia: an assessment of direct medical costs of air medical retrievals and the implications for the societal burden date: - - journal: air med j doi: . /j.amj. . . sha: doc_id: cord_uid: z vhxg objective: adequate mental health service provision in rural and remote australian communities is problematic because of the tyranny of distance. the royal flying doctor service provides air medical retrieval for people in rural and remote areas. the economic impact on both the royal flying doctor service and the public hospital system for mental health–related air medical retrievals is unknown. we aimed to estimate the direct medical costs associated with air medical retrievals and subsequent hospitalizations for mental and behavioral disorders for the calendar year. methods: all patients with a primary working diagnosis of international statistical classification of diseases and related health problems, th version, australian modification f to f (mental and behavioral disorders) who underwent an air medical retrieval were included in this cost analysis. international statistical classification of diseases and related health problems, th edition, australian modification codes were mapped to australian refined diagnosis related group codes, with hospital costs applied from the national hospital cost data collection ( / ). all costs are reported in australian dollars (auds). results: one hundred twenty-two primary evacuations and interhospital transfers occurred with an in-flight diagnosis of f to f , most commonly psychotic disorders, including schizophrenia and schizotypal disorders. the total direct medical costs were estimated to be aud $ , , . costs for primary evacuations accounted for % (aud $ , , ), with the majority of this associated with the subsequent hospital admission (aud $ , , ). similarly, the majority of the costs associated with interhospital transfers (total costs = aud $ , , ) were also related to hospital costs (aud $ , , ). conclusion: direct medical costs associated with air medical retrievals for people experiencing a mental health crisis are substantial. the majority of costs are associated with hospital admission and treatment; however, the indirect (loss of productivity) and intangible (quality of life) costs are likely to be far greater. demonstrated by suicide rates, which increase in line with the degree of remoteness, ranging from . per , persons in major cities to . per , in very remote settings. although the drivers of this unbalanced burden of mental health conditions are many and complex, of the key factors is the supply of mental health services. the majority of mental health professionals, measured as full-time equivalent (fte) per , population, are located in major cities. specifically, . fte psychiatrists per , persons are based in major cities compared with . per , in outer-regional settings, . per , in remote settings, and . per , in very remote areas of australia. similar trends are observed for mental health nurses ( . fte nurses/ , persons in major cities, . / , in remote setting, and . / , in very remote settings) and clinical psychologists ( . / , in major cities, . / , in remote settings, and . / , in very remote settings). in addition to these supply-side issues, demand for mental health services differs from that in major cities and many urban centers. although the overall prevalence of mental health conditions is similar across settings, people in rural and remote settings experience higher rates of substance use and acuity of mental health conditions along with the aforementioned rates of suicide. , furthermore, increased environmental challenges including drought, fires, and climate change, and the recent coronavirus pandemic, are placing greater pressure on the mental health of many rural and remote communities. [ ] [ ] [ ] [ ] in this context, the royal flying doctor service (rfds) provides air medical retrievals for australians living in rural and remote communities experiencing health crises, including acute mental health presentations. patients are typically transferred by aircraft to large metropolitan or inner regional public hospitals for urgent acute care. the economic costs of this approach, including those incurred by the rfds and public hospitals, has not been quantified. the primary aim of this article is to determine the annual air medical retrieval and in-patient hospital-direct medical costs associated with mental and behavioral disorders from a health payer perspective. a secondary aim includes determining the mismatch of the supply and the capacity of rural and remotely located mental health services with the demand of acute presentations, with retrieval signaling as a potential proxy for this unmet need. the rfds provides air medical, road ambulance, and primary health care to rural and remote areas of australia without traditional medical services, such as those associated with the medicare benefits schedule, a listing of the medicare services subsidized by the australian government. the focus of this article is on air medical retrievals for mental health crises from a health care payer perspective. , design and participants a partial economic evaluation was undertaken using routinely collected air medical data for patients diagnosed in flight with a mental and behavioral disorder (international statistical classification of diseases and related health problems, th edition, australian modification [icd- am], chapter v) between january , , and december , . participants included all rfds patients who underwent an air medical retrieval, including a primary evacuation and interhospital transfer, for mental and behavioral disorders within australia in . the majority of the rfds air medical retrievals are conducted in western australia, central australia, queensland, and new south wales, with limited air medical retrievals coming from tasmania and victoria. tasmania and victoria air medical services are mainly conducted by other services; however, the rfds in / conducted substantial road transportation in victoria and tasmania (n = , ). for the primary aim, data were collected and coded on each patient's in-flight working diagnosis using the icd- -am coding method. the in-flight primary working diagnosis was based on referral assessment information and an assessment of the current medical status by the in-flight medical team, which, in this patient group, mainly consisted of a senior medical officer and/or a senior flight nurse. the in-flight primary working diagnosis was then coded by trained administrative staff and cross-checked by of the authors (l.b. and f.g.). data were collected within flight on the patient's sex, age, and indigenous status. both paper-based and electronic methods were used in data collection. detailed patient histories were not routinely collected. all air medical retrieval patients with a primary working diagnosis of icd- -am chapter v codes f -f (mental and behavioral disorders) were included in the analysis. all other diagnoses were excluded from analysis. we defined separate types of air medical retrievals: ) primary evacuations of a patient and ) interhospital transfers that involve an rfds air medical evacuation from, typically, a small regional hospital to an inner regional or major city hospital. to determine the economic costs per primary evacuation, we collected the costs incurred by each rfds section and operation (loosely state based), including the queensland section, western operations, south eastern section, and central operations. this included determining the individual primary evacuation costs by rfds base from each section and operation. to protect patient and rfds base confidentiality (particularly for those bases conducting a small number of retrievals), these costs were then averaged. costs included engine hour and staffing by an rfds registered nurse (present on all flights) and rfds medical officer as required. the costs for interhospital transfers were based on $ , per engine hour, which includes an rfds registered nurse. for transfers in which an rfds medical officer was also required, an additional $ , per hour was added to the cost of each interhospital transfer. this formula is consistent with other published literature. to estimate inpatient admission costs, we mapped the icd- -am codes to australian refined diagnosis related group (ar-drg) codes provided in the national hospital cost data collection ( / ) ( table ) . this mapping was performed by of the authors (b.d.g.) and independently checked by of the authors (f.g. and m.c.). it is important to note that although there are > , icd- -am codes and > , ar-drg codes, the national hospital cost data collection only contains codes. cost data were then extracted from the / national hospital cost data collection for each relevant ar-drg ( table ) . the costs were then applied to each primary evacuation and interhospital transfer, respectively. for interhospital transfers, the costs associated for the first admission (ie, the hospital from which the patient was transferred from) were not included because no icd- -am or ar-drg data are collected for this. the total aggregate costs were estimated for both primary evacuations and interhospital transfers. to further understand where costs are incurred, the disaggregated "cost buckets" reported in the national hospital cost data collection were assessed. these disaggregated costs include ward medical, ward nursing, nonclinical salaries, pathology, imaging, allied pharmacy, critical care, operating room, emergency department, ward supplies, specialist procedure suites, prostheses, oncosts (eg, indirect salary costs such as superannuation), hotel and depreciation (ie, domestic services within the hospital that are not directly related to patient care), and emergency department (ed) to determine the secondary aim, we used the rfds service planning and operational tool (spot) to map service provision throughout australia. spot uses data from the australian bureau of statistics and data from health direct to derive geographic population estimates reflective of mental health services. primary evacuation statistical areas were defined according to the australian bureau of statistics statistical area level code. spot has been designed to help determine the geographic coverage of health care in australia. spot graphically represents population concentrations and health care services and calculates the proportion of the australian population who are covered by specific health care facilities (in this case, general mental health services) within a -minute drive time. to map the location of air medical retrievals for diagnoses associated with mental health, we used tableau mapping software (tableau software, salesforce company, seattle, wa united states of america). this study used descriptive statistics to summarize findings. cost data were extracted from the / national hospital cost data collection for each ar-drg derived from the icd- -am mapping exercise. costs associated with rfds air medical retrievals were applied to each ar-drg. these costs were then summed and multiplied by the number of retrievals per ar-drg. expenditure based on cost buckets was extracted from the / national hospital cost data collection. each item in the cost bucket was multiplied by the number of air medical transfers with the corresponding ar-drg. costs for each cost bucket were summed for all ar-drgs, allowing for the calculation of the proportion of the total expenditure associated with each cost bucket. all costs are reported in australian dollars. analyses were conducted in excel (microsoft, redmond, wa). in addition, cell sizes with or less patients are supressed for confidentiality. this project was deemed a low-risk quality assurance project by the rfds clinical and health services research committee, which provides oversight for rfds research projects, on march , . because this project involved routinely collected data, specific patient consent forms were not required. over the calendar year, the rfds conducted primary evacuations and interhospital transfers for patients with an inflight diagnosis associated with a mental and behavioral disorder (icd- -am chapter v f -f ). all of the primary evacuations and interhospital transfers were from remote and very remote areas to inner regional or metropolitan centers (fig. a) . the primary evacuation statistical areas level included alice springs (northern territory) ( . %, n = ), the far north (northern queensland) ( . %, n = ), gold fields (western australia) ( . %, n = ), and the kimberly ( . %, n = ). figure b provides an illustration of the supply of mental health services derived from spot. when looking at the general mental health service coverage within these areas reflective of population concentrations, the gold fields ( . %, n = , ) had the highest remote and very remote population level without coverage followed by alice springs ( . %, n = , ), kimberly ( . %, n = , ), and the far north ( . %, n = , ). for primary evacuations, the leading diagnoses were for the f to f group of psychotic disorders, including schizophrenia, schizotypal, delusional disorders, and other non−mood-related psychotic disorders ( . %, n = ). schizophrenia ( . %, n = ), acute and transient psychotic disorders ( . %, n = ), and unspecified nonorganic psychosis ( . %, n = ) were the most common diagnoses. one similar to primary evacuations, the majority of interhospital transfers were associated with f to f , schizophrenia, schizotypal, delusional disorders, and other non−mood-related psychotic disorders ( . %, n = ). one fifth of the transfers were for f to f , mood (affective) disorders ( . %, n = ); . % (n = ) were for f to f , mental and behavioral disorders caused by psychoactive substance use; and . % (n = ) for f to f , organic, including symptomatic, mental disorders. small numbers of interhospital transfers were reported for f to f , disorders of adult personality and behavior ( . %, n = ); f to f , neurotic, stress-related, and somatoform disorders ( . %, n = ); f , unspecified mental disorders ( . %, n = ); and f to f , behavioral syndromes associated with physiological disturbances and physical factors ( . %, n = ). five or less evacuations were reported for f to f , behavioral and emotional disorders with onset usually occurring in childhood and adolescence; f to f , disorders of psychological development; and f to f , mental retardation. the total costs for all mental and behavioral disorder air medical retrievals and subsequent hospital admissions in was $ , , . the cost associated with the primary evacuations was estimated to be $ , , . two thirds of this ($ , , , . %) was related to in-patient admissions (ar-drg costs) and the remaining $ , on air retrieval costs. reflecting the numbers of patients, the highest combined air retrieval and in-patient costs were for u a (schizophrenia disorders, major complexity; $ , ), u a (paranoia and acute psychotic disorders, major complexity; $ , ), and u a (major affective disorders, major complexity; $ , ). the average cost per primary evacuation patient was $ , . the total cost related to interhospital transfers was $ , , . similar to primary evacuations, the majority of these costs were related to in-patient admission costs ($ , , ) . reflecting the numbers of interhospital transfers, the greatest total costs were associated with u a (schizophrenia disorders, major complexity; $ , , ), u a (paranoia and acute psychotic disorders, major complexity; $ , , ), and u a (major affective disorders, major complexity; $ , , ). the average cost per interhospital transfer patient was $ , (see table for cost breakdowns for primary evacutations and interhospital transfers). almost one third of all hospital costs (for both primary evacuations and interhospital transfers) were associated with the ward nursing cost bucket ( . %) (fig. ) . these costs are associated with nursing care provided in general wards. the ward medical cost bucket, which includes both salaries and wages for medical officers, accounted for . % of hospital costs. other notable cost buckets were . % for ward supplies (costs for medical and surgical supplies, ward and clinical department overheads, and goods and services), . % for nonclinical salaries (other costs of service provision, predominantly wages for carers such as patient care assistants), . % for oncosts (eg, superannuation, fringe benefits tax, long service leave, worker's compensation, and recruitment), and . % for ed product (ie, the average cost per admitted ed patient). this is the first study to quantify the direct medical costs associated with air medical retrievals of patients experiencing mental health crises in australian rural and remote settings. we estimated that the annual direct medical cost associated with this was $ , , for . most of these costs were attributable to hospital costs, with over % of this expended on ward nursing staff, medical staff, and ward supplies. importantly, this total cost represents a substantial underestimate of both the health payer and the societal impacts of these acute mental health events in the rural and remote areas of australia. more specifically, in regard to the direct medical costs, the costs of the first hospital (or retrieval site) admission for the patients who received an interhospital transfer were not included. furthermore, the substantial indirect costs associated with air medical retrievals have not been assessed. these costs include those associated with the lost productivity of patients, their families, and caregivers. in addition, intangible costs, largely the suffering associated with a condition that can be captured using quality of life instruments, is likely to be substantial from both a physical and, importantly, from a psychosocial perspective and has not been assessed in this study. for patients, particularly for those experiencing a first episode of psychosis, inpatient admission in addition to air medical retrieval and dislocation from usual social supports and networks can be a traumatic experience, with a recent systematic review reporting % of firstepisode psychosis patients experienced symptoms of post-traumatic stress disorder. although the impact of this experience is partially understood with patients, less is known regarding the impact and costs for families and caregivers. further research is required to fully understand the true costs, including indirect (loss of productivity) and intangible costs (quality of life), associated with air medical retrievals for mental health care and the current level of service provision to rural and remote communities in location. ultimately, the implications of these health payer and societal costs need to be understood in the context of considered funding and capacity considerations for rural and remote mental health services and providers. tough longer-term health care policy decisions are required by governments and health planners through the prism of the known economic costs for air medical retrievals, albeit as an underestimate of the likely true costs identified in this study. additional attention and research are required to qualify the grossly under-researched costs to rural and remote communities of retrieving patients with mental and behavioral disorders out of their communities from a societal impact perspective. the severe lack of psychiatrists, mental health nurses, psychologists, and social workers in rural and remote areas requires structural change to the supply of a qualified and competent rural and remote workforce. the recent australian senate inquiry into the burden of mental health conditions on rural and remote communities recommends longer-term and more flexible funding and contract processes, in addition to working with professional colleges to improve support and training of workers in rural and remote communities, with the goal of supporting high-quality workers and services to remain in communities, thereby providing consistency of service provision. importantly, our article suggests that rfds air medical retrievals are a symptom of this unmet demand and also a potential proxy for the services and capacity that currently does not exist in rural and remote communities. to provide high-quality mental health services to rural and remote communities will require adequate resourcing. although the costs to governments will likely be substantial, we suggest that the societal benefits that will be gained from resourcing action to meet the unmet demand are likely to outweigh these costs. in other words, the poor mental health outcomes experienced by people in rural and remote communities deserves and demands action. furthermore, we do not understand the impact of high-acuity mental illnesses, the high rates of suicide and substance use on the indirect (productivity) and intangible (quality of life) costs for rural and remote communities, and the australian society more broadly. it is important to quantify these costs because they are likely to be substantial and offset much of the costs related to the provision of high-quality mental health services for rural and remote communities. a strength of this article is that it has provided annual direct medical costs associated with air medical retrievals in australia. these costs have not been quantified before and are important because they can be used to assess the total costs associated with the current approaches to mental health service provision in rural and remote communities. a limitation of this study is that to apply hospital costs from the national hospital cost data collection, we mapped icd- -am codes (of which there are > , ) to the ar-drg codes costed by the national hospital cost data collection. as such, the costs reported here are an estimate of the costs to governments through the public hospital system. a further limitation was that we were unable to access the icd- -am codes that were listed for the initial hospital admission before interhospital transfers; therefore, we were unable to apply these costs. in conclusion, the direct medical costs of air medical retrievals for mental and behavioral disorders in australia's most remote communities are substantial. the societal implications of these costs to the families and communities of these regions is unknown; nevertheless, the flow-on effects of these societal costs are likely to far exceed the direct medical costs. we challenge policy and decision makers to understand these societal implications for future health policy and planning of mental health services in australia's rural and remote communities. australian institute of health and welfare. rural and remote health. canberra, australia: australian institute of health and welfare time to end the drought in the bush the orange declaration on rural and remote mental health suicide and hospitalised self-harm in australia: trends and analysis australia's future health workforce − psychiatry. department of health australian institute of health and welfare. mental health services in australia: mental health workforce tables time to end the drought in the bush australian institute of health and welfare. national drug strategy household survey : detailed findings royal flying doctor service covid- activity and surge modelling in australia health care access, mental health, and preventive health: health priority survey findings for people in the bush suicide and mental health in rural, remote and metropolitan areas in australia drought-related stress among farmers: findings from the australian rural mental health study department of health. mbs online. australian government methods for the economic evaluation of health care programmes consolidated health economic evaluation reporting standards (cheers) statement international statistical classification of diseases and related health problems, th revision the medical and retrieval costs of road crashes in rural and remote northern queensland independent hospital pricing authority. national hospital cost data collection report, public sector, round (financial year - ). canberra: ihpa who we are volume -main structure and greater capital city statistical areas aeromedical retrievals of people for mental health care and the low level of clinical support in rural and remote australia the traumatic experience of first-episode psychosis: a systematic review and meta-analysis the australian senate: community affairs references committee. accessibility and quality of mental health services in rural and remote australia acknowledgmentin-flight patient data were extracted from the confidential royal flying doctor service data set, and cost data for australian refined diagnosis related groups codes were sourced from the publicly available national hospital cost data collection ( / ) website.*address for correspondence: fergus gardiner, phd medicine, royal flying doctor service, , level , - brisbane avenue, barton, act , e-mail address: fergus.gardiner@rfds.org.au (f.w. gardiner). key: cord- -gxv cbom authors: juneau, carl-etienne; pueyo, toma; bell, matt; gee, genevieve; potvin, louise title: evidence-based, cost-effective interventions to suppress the covid- pandemic: a rapid systematic review date: - - journal: nan doi: . / . . . sha: doc_id: cord_uid: gxv cbom background: countries vary in their response to the covid- pandemic. some emphasise social distancing, while others focus on other interventions. evidence on the effectiveness and cost-effectiveness of these interventions is urgently needed to guide public health policy and avoid unnecessary damage to the economy and other harms. we aimed to provide a comprehensive summary of the evidence on epidemic control, with a focus on cost-effectiveness. methods: medline ( to march week , ) and embase ( to march , ) were searched using a range of terms related to epidemic control. reviews, randomized trials, observational studies, and modelling studies were included. articles reporting on the effectiveness or cost-effectiveness of at least one intervention were included and grouped into higher-quality (randomized trials) and lower-quality evidence (other study designs). findings: we found , papers; were included. higher-quality evidence was only available to support the effectiveness of hand washing and face masks. modelling studies suggested that these measures are highly cost-effective. for other interventions, only evidence from observational and modelling studies was available. a cautious interpretation of this body of lower-quality evidence suggests that: ( ) the most cost-effective interventions are swift contact tracing and case isolation, surveillance networks, protective equipment for healthcare workers, and early vaccination (when available); ( ) home quarantines and stockpiling antivirals are less cost-effective; ( ) social distancing measures like workplace and school closures are effective but costly, making them the least cost-effective options; ( ) combinations are more cost-effective than single interventions; ( ) interventions are more cost-effective when adopted early and for severe viruses like sars-cov- . for h n influenza, contact tracing was estimated to be , times more cost-effective than school closures ($ , vs. $ , , per death prevented). conclusions: a cautious interpretation of this body of evidence suggests that for covid- : ( ) social distancing is effective but costly, especially when adopted late and ( ) adopting as early as possible a combination of interventions that includes hand washing, face masks, swift contact tracing and case isolation, and protective equipment for healthcare workers is likely to be the most cost-effective strategy. on march , , the world health organization (who) characterized covid- as a pandemic. since then, the virus has infected exponentially more men and women worldwide. as of april , the who reports , , confirmed cases and , deaths (who, ) . this is more than double the deaths reported on april (who, b), indicating that the burden of covid- is currently doubling approximately every days. countries have implemented a range of responses, and many are imposing nationwide school closures. these currently affect over . billion (almost %) of the world's students (who, b) . but closing school is costly-$ to $ billion for weeks in the us alone (lempel et al. )-and could lead to a greater number of deaths than they prevent by creating unintended downstream effects, such as child-care obligations and losses in health-care workforce capacity (bayham and fenichel, ) . other countries have focused on other interventions, such as contact tracing and case isolation. these have been estimated to be , times more cost-effective than school closures for h n influenza ($ , vs. $ , , per death prevented) (madhav et al. ) . as states in the us and around the world are faced with the challenge of balancing public health interventions with economic (and other) considerations, evidence on the effectiveness and cost-effectiveness of these interventions is urgently needed to guide policy and avoid unnecessary harm. in this rapid systematic review, we aimed to provide a comprehensive summary of the evidence on epidemic control and to identify cost-effective interventions in the context of covid- . we performed preliminary searches to locate review articles, devise our search strategy, and identify potential shortcomings in the literature. a systematic review analyzed studies, included only randomized trials, and concluded that the evidence was lacking for most non-pharmaceutical interventions (smith et al. ) . while we do not dispute this conclusion when looking only at randomized trials, we would argue that as urgent decisions of unknown cost-effectiveness are made in reaction to the covid- pandemic, some evidence, even of lower quality, is better than no evidence at all. therefore, we included a broad range of study designs in this review to provide a comprehensive summary of the peer-reviewed evidence. reviews (all types), randomized trials, all rights reserved. no reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medrxiv a license to display the preprint in perpetuity. the copyright holder for this preprint this version posted april , . . https://doi.org/ . observational studies, and modelling studies were included. articles reporting on the effectiveness or cost-effectiveness of at least one intervention were included. articles in english, french, spanish, and portuguese were included. studies of sexually transmitted infections (e.g. syphilis) and mosquito-borne diseases (e.g. dengue) were not included. abstracts, case reports, and conferences proceedings were also excluded. medline ( to april week , and embase ( ( to april , were then searched using the terms "non-pharmaceutical interventions", "outbreak control", "outbreak interventions", "epidemic control", "epidemic interventions", "pandemic control", and "pandemic interventions" (last search: april , ). screening of titles, abstracts, and full texts was carried out by a single investigator. reference lists and pubmed related articles of included studies were reviewed for additional studies. for this rapid systematic review, we followed prisma guidelines (moher et al. ), but we limited quality assessment to grouping studies into two categories: higher quality (randomized trials) and lower quality (other study designs). a total of , papers were found in medline and embase. removing duplicates left , . we retained based on title, based on abstract, and based on full text. we found additional studies via reference lists and pubmed related articles searches. therefore, a total of studies were included (efigure in the supplement). randomized trial evidence was only available for the effectiveness of hand washing and face masks (jefferson et al. ; smith et al. ; macintyre et al. ; saunders-hastings et al. ) . for other interventions, only lower-quality evidence was available (observational and modelling studies). treatment (cost-effective), antiviral prophylaxis (cost-effective), low efficiency vaccination (costeffective if timed before cases peak), high efficiency vaccination (cost-effective if timed before cases peak), stockpiling antiviral medicine (cost-effective for high-income countries), quarantining confirmed cases at home (cost-effective for viruses with a case fatality rate of %, not cost-effective for viruses with a case fatality rate of . %), self-isolation at home (costeffective with a case fatality rate of %, not cost-effective with a case fatality rate of . %), and school closure (not cost-effective). for h n influenza, contact tracing was estimated to be , times more cost-effective than school closures ($ , vs. $ , , per death prevented) (madhav et al. ). indeed, lempel et al. ( estimated that closing all schools in the us for weeks would cost $ to $ billion ( . - . % of gdp). other systematic reviews found that school closures did not contribute to the control of the sars epidemic in china, hong kong, and singapore and would prevent only - % of covid- deaths (viner et al. ) ; reduced the peak of epidemics by . % on average and were more effective when timed early (bin nafisah et al. ) ; would be most effective if they caused large reductions in contact, if transmissibility was low (e.g. a basic reproduction number < ), and if attack rates were higher in children than in adults (jackson et al. ) ; and appeared to be moderately effective in reducing the transmission of influenza and in delaying the peak of an epidemic, but were associated with very high costs (rashid et al. ) . differences in publication date, virus transmissibility, and study selection may explain the discrepancies among these reviews. contact tracing and case isolation was one of the most cost-effective interventions, based on h n data from hubei, china (wang et al. ) . pasquini-descomps et al. ( ) computed that it cost less than $ , per disability-adjusted life year. in a simulation study, hellewell et al. ( ) found that in most scenarios, highly effective contact tracing and case isolation would be enough to control a new outbreak of covid- within months. timing was important: with five initial cases, there was a greater than % chance of achieving control, even at modest contact-tracing levels. however, at initial cases, control was much less likely. similarly, any delay from symptom onset to isolation decreased the probability of control, highlighting the need for swift action. in another modelling study, chen et al. ( ) compared the effects of four interventions on the total attack rate and duration of a school influenza outbreak in changsha, china. case isolation was the most effective single intervention, and the addition of antiviral therapeutics, all rights reserved. no reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medrxiv a license to display the preprint in perpetuity. the copyright holder for this preprint this version posted april , . . https://doi.org/ . antiviral prophylaxis, vaccination prior to the outbreak, and school closure decreased the attack rate only slightly, and shortened outbreak duration by only days. saunders-hastings et al. ( ) carried out a systematic review and meta-analysis of personal protective measures to reduce pandemic influenza transmission. they did not find any study evaluating respiratory etiquette (e.g. covering the mouth during coughing). meta-analyses suggested that regular hand hygiene provided a significant protective effect (or = . ; % ci . - . ). face masks had a non-significant protective effect (or = . ; % ci . - . ) which became significant (or = . ; % ci . - . ) when randomized control trials and cohort studies were pooled with case-control studies (this also decreased heterogeneity). in an earlier systematic review, jefferson et al. ( ) also found a protective effect of masks. overall, they were the best performing intervention across populations, settings, and threats. similarly, in a narrative review, macintyre et al. ( ) drew on evidence from randomized community trials to conclude that face masks do provide protection against infection in various community settings, subject to compliance and early use. differences in publication date, search strategy, and study selection criteria may explain the discrepancies among these reviews. tracht et al. ( ) estimated savings of $ billion if % of the us population used masks in an unmitigated h n epidemic. for hand washing, townsend et al. ( ) estimated that a national behaviour change program in india would net $ . billion ( . - . ), a -fold return on investment. a similar program in china would net $ . billion ( . - . ), a -fold return on investment. preventive measures in hospitals include use of personal protective equipment for healthcare workers in direct contact with suspected patients. dan et al. ( ) estimated that this measure was cost-effective for h n ($ , per death prevented). however, adopting a wider set of measures (full personal protective equipment, restricting visitors, and cancelling elective procedures) was much less cost-effective ($ , , per death prevented). similarly, lee et al. ( ) found that increasing hand hygiene, use of protective apparel, and disinfection are the most cost-saving interventions to control a hospital outbreak of norovirus. suphanchaimat et al. ( ) found that influenza vaccination for prisoners in thailand was costeffective. the incremental cost-effectiveness ratio of vaccination (compared with routine outbreak all rights reserved. no reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medrxiv a license to display the preprint in perpetuity. the copyright holder for this preprint this version posted april , . . https://doi.org/ . control) was $ to $ per disability-adjusted life year. prosser et al. ( ) also found that h n vaccination in the us was cost-effective under many assumptions if initiated prior to the outbreak. incremental cost-effectiveness ratios ranged from $ , to $ , per quality-adjusted life year for persons aged months to years without high-risk conditions. the authors noted that all doses (two for some children, one for adults) should be delivered before the peak of a hypothetical influenza season. otherwise, vaccination may not be cost-effective. khazeni et al. ( ) also found that earlier vaccination is more cost-saving. if vaccine supplies are limited, lee et al. ( ) found that priority should be given to at-risk individuals, and to children within highrisk groups. in another systematic review of economic evaluations, pérez velasco et al. ( ) examined studies and found that combinations of pharmaceutical and non-pharmaceutical interventions were more cost-effective than vaccines and/or antivirals alone. reducing non-essential contacts, using pharmaceutical prophylaxis, and closing schools was the most cost-effective combination for all countries. however, quarantine for household contacts was not cost-effective, even in low-and middle-income countries. a modelling study by day et al. ( ) suggested that quarantine (of all individuals who have had contact with an infected individual) would be beneficial only when case isolation is ineffective, when there is significant asymptomatic transmission, and when the asymptomatic period is neither very long, nor very short. perlroth et al. ( ) estimated the health outcomes and costs of combinations of social distancing strategies and antiviral medication strategies. for a virus with a case fatality rate of % and a reproduction number of . or greater, school closure alone was the least cost-effective intervention and cost $ , per case averted. antiviral treatment ($ , ) , quarantine of infected individuals ($ , ), and adult and child social distancing ($ , ) had increasing levels of cost-effectiveness. however, combining interventions was more cost-effective, and the most cost-effective combination included adult and child social distancing, school closure, and antiviral treatment and prophylaxis ($ , per case). however, the same combination without school closure was more cost-effective for milder viruses (case fatality rate below %, reproduction number . or lower). if antivirals are not available, the combination of adult and child social distancing and school closure was most effective. all rights reserved. no reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medrxiv a license to display the preprint in perpetuity. the copyright holder for this preprint this version posted april , . . https://doi.org/ . . . . doi: medrxiv preprint halder et al. ( also aimed to determine the most cost-effective interventions for a pandemic similar to h n . they found that a combination of interventions was most cost-effective. this combination included treatment and household prophylaxis using antiviral drugs and limited duration school closure. if antiviral drugs are not available, limited duration school closure was significantly more cost-effective compared to continuous school closure. other social distancing strategies, such as reduced workplace attendance, were found to be costly due to productivity losses. closing school for to weeks without other interventions did not cost much more than doing nothing but gave a significant % to % reduction in cases "if optimally timed". the authors examined timing in another study, in which the effectiveness of school closures were analyzed for durations of , and weeks (halder et al. ) . they found that the most appropriate strategy depended on the virus' severity and transmissibility. for mild viruses, they concluded that individual school closures should begin once daily new cases reach to . for highly transmissible epidemics (reproduction number of or above), they concluded that long duration school closure should begin as soon as possible and be combined with other interventions. indeed, for such viruses, they found that school closure alone would be ineffective (~ % reduction in attack rate) and recommend "additional rigorous social distancing interventions." smith et al. ( ) carried out a systematic review of non-pharmaceutical interventions to reduce the transmission of influenza in adults. only randomized trials were included and studies met all selection criteria. the authors found that positive significant interventions included professional oral hygiene intervention in the elderly and hand washing, and noted that home quarantine may be useful, but required further assessment. jefferson et al. ( ) conducted a cochrane systematic review of physical interventions to interrupt or reduce the spread of respiratory viruses. they found that the highest quality randomized cluster trials suggested this could be achieved by hygienic measures such as handwashing, especially around younger children. they recommended that the following effective interventions be implemented, preferably in a combined fashion, to reduce transmission of viral respiratory disease: frequent handwashing with or without adjunct antiseptics; barrier measures all rights reserved. no reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medrxiv a license to display the preprint in perpetuity. the copyright holder for this preprint this version posted april , . . https://doi.org/ . such as gloves, gowns and masks with filtration apparatus; and suspicion diagnosis with isolation of likely cases. lee et al. ( ) carried out a systematic review of modelling studies quantifying the effectiveness of strategies for pandemic influenza response. they found that combinations of strategies increased the effectiveness of individual strategies and could reduce their potential negative impact. combinations delayed spread, reduced overall number of cases, and delayed and reduced peak attack rate more than individual strategies. pan et al. ( ) examined associations between public health interventions and the epidemiology of covid- in wuhan, china. traffic restrictions, cancellation of social gatherings, and home quarantines were associated with reduced transmission, but were not sufficient to prevent increases in confirmed cases. these were reduced and estimates of the effective reproduction number fell below only when additional interventions were implemented. those included hospital-based measures (designated hospitals and wards, use of personal protective equipment, increased testing capacity and accelerated reporting, and timely medical treatment) and community-based interventions (quarantine of confirmed and presumptive cases and of close contacts in designated facilities). influenza pandemic (isolation or quarantine, school closure, public gathering ban). they found that every city adopted at least one of these interventions, and that cities applied all three. the most common combination (school closure and public gathering bans) was implemented in cities ( %) for a median duration of weeks and was significantly associated with reductions in weekly excess death rate. cities that implemented interventions earlier had greater delays in reaching peak mortality (spearman r=− . , p< . ), lower peak mortality rates (spearman r= . , p=. ), and lower total mortality (spearman r= . , p=. ). there was a significant association between increased duration of interventions and a reduced total mortality burden (spearman r=− . , p=. ). all rights reserved. no reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medrxiv a license to display the preprint in perpetuity. the copyright holder for this preprint this version posted april , . . https://doi.org/ . ishola and phin ( ) reviewed the literature on mass gatherings. they found studies and cautiously concluded that there is some evidence to indicate that mass gatherings may be associated with an increased risk of influenza transmission. in a more recent systematic review, rainey et al. ( ) found that mass gathering-related respiratory disease outbreaks were relatively rare between and in the us. they concluded that this could suggest-perhaps surprisingly-low transmission at most types of gatherings, even during pandemics. similarly, in a us survey of state health departments and large local health departments, figueroa et al. ( ) found that outbreaks at mass gatherings were uncommon, even during the h n pandemic. in a modelling study, shi et al. ( ) found that mass gatherings that occur within days before the epidemic peak can result in a % relative increase in peak prevalence and total attack rate. conversely, they found that mass gatherings may have little effect when occurring more than days earlier or days later than the infection peak (when initial ro = . ). thus the timing of mass gatherings might explain the apparent lack of evidence in support of their ban. recently, zhao et al. ( ) china, and reduced international case importations by nearly % until mid february. modeling results also indicated that sustained % travel restrictions to and from china only modestly affect the epidemic trajectory unless combined with a % or higher reduction of transmission in the all rights reserved. no reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medrxiv a license to display the preprint in perpetuity. the copyright holder for this preprint this version posted april , . . https://doi.org/ . community. travel limitations may be more effective when neighbouring countries fail to implement adequate outbreak control efforts (bwire et al. ) . this rapid systematic review aimed to provide a comprehensive summary of the evidence on pandemic control, with a focus on cost-effective interventions in the context of covid- . randomized trial evidence was only available to support the effectiveness of hand washing and face masks. modelling studies suggested that these measures are highly cost-effective. for other interventions, only evidence from observational and modelling studies was available. this lowerquality evidence suggests that overall, when timed appropriately, the following interventions are likely to be highly cost-effective: contact tracing and case isolation, protective equipment for healthcare workers, and vaccination prior to the outbreak (when available). surveillance networks and protective equipment for healthcare workers are also likely to be cost-effective. home quarantine for confirmed cases and stockpiling antivirals appear less cost-effective. the least costeffective interventions appear to be social distancing measures like workplace and school closures. the evidence suggests that these are more cost-effective for severe viruses like sars-cov- , and when timed early in the outbreak. vaccination past the peak of infections and long-term school closures late in the outbreak appear less cost-effective, underscoring the importance of timing. (march ). they point out that the high rates in italy may be due to differences in population age, definition of covid- -related deaths, and testing strategies. based on the above, our assessment of the cost-effectiveness of interventions is based on an estimate of to % case fatality rate for all rights reserved. no reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medrxiv a license to display the preprint in perpetuity. the copyright holder for this preprint this version posted april , . . https://doi.org/ . covid- . the cost-effectiveness of interventions also depends on their timing. taking this into account, we propose a -stage framework for pandemic control interventions, adapted from madhav et al. ( ) . this framework is illustrated in figure . stages are shown from left to right, and interventions are shown from top (most cost-effective) to bottom (least cost-effective). figure shows the cost-effectiveness of interventions to control covid- by stage of the pandemic. according to this framework, surveillance networks are highly cost-effective, should be established before the pandemic starts (stage ), and maintained through stages and . vaccination, when available, should occur before the pandemic, or as early as possible. antivirals can be stockpiled cost-effectively in high-income countries. as the pandemic starts (stage ), early contact tracing and case isolation is the most cost-effective intervention. it may be sufficient to contain the outbreak. if the outbreak is not contained, hand hygiene, face masks, and protective equipment for healthcare workers are all highly cost-effective. if these measures are not sufficient, home quarantines, social distancing, and school closures are all effective, albeit increasingly costly measures. assuming a to % case fatality rate for covid- , these measures are likely to be cost-effective nonetheless, especially if implemented early. as covid- spreads (stage ), and especially past the peak, the costliest interventions can be replaced cost-effectively by a all rights reserved. no reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medrxiv a license to display the preprint in perpetuity. the copyright holder for this preprint this version posted april , . . https://doi.org/ . combination of interventions centered on swift contact tracing and case isolation. once antivirals are available, they can also replace the costlier interventions cost-effectively. this rapid systematic review has a number of limitations. first, randomized trial evidence was not available for most epidemic control interventions. consequently, we included observational and modelling studies. their results should be interpreted with caution. still, as covid- forces urgent decision-making, we submit that some evidence is better than none. second, because of time constraints, our search was limited to two databases (medline and embase). only one investigator screened titles, abstracts, and the full text of papers. we performed limited quality assessment and we did not examine risk of bias. third, most of the studies we have reviewed focused on h n and other viruses, not sars-cov- . fourth, estimates of covid- case fatality rate are subject to substantial uncertainties. we assessed the cost-effectiveness of interventions based on estimates of to %. these are likely to change as more data emerge. should the true rate be higher, all interventions would be more cost-effective. conversely, should it be lower, costly interventions such as extended school closures may not be cost-effective. hand washing and face masks were the only measures supported by higher-quality evidence. other interventions were supported by lower-quality evidence. in the context of covid- , a cautious interpretation suggests that ( ) social distancing is effective but costly, especially when adopted late and ( ) adopting as early as possible a combination of interventions that includes hand washing, face masks, swift contact tracing and case isolation, and protective equipment for healthcare workers is likely to be the most cost-effective strategy. all rights reserved. no reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medrxiv a license to display the preprint in perpetuity. the copyright holder for this preprint this version posted april , . . https://doi.org/ . funding lp holds the canada research chair in community approaches and health inequalities (crc - ). this funding source had no role in the design, conduct, or reporting of the study. cej, tp, mb, and lp declare no conflict of interest. gg holds a contractual position with the millar group (provider of personal protective equipment) and executive roles at panacea health solutions and angular momentum (providers of diabetes and corporate wellness programs). cej, tp, and lp designed the study. cej, gg, and mb searched and analyzed the literature. cej and tp interpreted the findings. cej wrote the first draft. all authors revised drafts and approved the final manuscript. all rights reserved. no reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medrxiv a license to display the preprint in perpetuity. the copyright holder for this preprint this version posted april , . all rights reserved. no reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medrxiv a license to display the preprint in perpetuity. the copyright holder for this preprint this version posted april , . . https://doi.org/ . step searches results pandemic control.mp. [mp=ti, ab, hw, tn, ot, dm, mf, dv, kw, fx, dq, nm, kf, ox, px, rx, ui, sy] pandemic interventions.mp. [mp=ti, ab, hw, tn, ot, dm, mf, dv, kw, fx, dq, nm, kf, ox, px, rx, ui, sy] non-pharmaceutical interventions.mp. [mp=ti, ab, hw, tn, ot, dm, mf, dv, kw, fx, dq, nm, kf, ox, px, rx, ui, sy] outbreak control.mp. [mp=ti, ab, hw, tn, ot, dm, mf, dv, kw, fx, dq, nm, kf, ox, px, rx, ui, sy] epidemic control.mp. [mp=ti, ab, hw, tn, ot, dm, mf, dv, kw, fx, dq, nm, kf, ox, px, rx, ui, sy] epidemic interventions.mp. [mp=ti, ab, hw, tn, ot, dm, mf, dv, kw, fx, dq, nm, kf, ox, px, rx, ui, sy] outbreak interventions.mp. [mp=ti, ab, hw, tn, ot, dm, mf, dv, kw, fx, dq, nm, kf, ox, px, rx, ui, sy] or or or or or or remove duplicates from all rights reserved. no reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medrxiv a license to display the preprint in perpetuity. the copyright holder for this preprint this version posted april , . for more information, visit www.prisma-statement.org. records identified through database searching (n = , ) additional studies identified through other sources (n = ) records after duplicates removed (n = , ) records screened (n = ) full-text articles assessed for eligibility (n = ) full-text articles excluded (n = ) all rights reserved. no reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medrxiv a license to display the preprint in perpetuity. the copyright holder for this preprint this version posted april , . . https://doi.org/ . grade evidence to decision (etd) frameworks: a systematic and transparent approach to making well informed healthcare choices. : introduction impact of school closures for covid- on the us health-care workforce and net mortality: a modelling study. the lancet public health school closure during novel influenza: a systematic review cross-border cholera outbreaks in sub-saharan africa, the mystery behind the silent illness: what needs to be done? simulation of key interventions for seasonal influenza outbreak control at school in changsha the effects of school closures on influenza outbreaks and pandemics: systematic review of simulation studies physical interventions to interrupt or reduce the spread of respiratory viruses. cochrane database syst rev effectiveness and cost-effectiveness of vaccination against pandemic influenza (h n ) a computer simulation of vaccine prioritization, allocation, and rationing during the h n influenza pandemic. vaccine combination strategies for pandemic influenza response -a systematic review of mathematical modeling studies nonpharmaceutical interventions implemented by us cities during the - influenza pandemic pandemics: risks, impacts, and mitigation in disease control priorities: improving health and reducing poverty preferred reporting items for systematic reviews and meta-analyses: the prisma statement case-fatality rate and characteristics of patients dying in relation to covid- in italy association of public health interventions with the epidemiology of the covid- outbreak in wuhan, china health outcomes and costs of community mitigation strategies for an influenza pandemic in the united states the use of mathematical modeling studies for evidence synthesis and guideline development: a glossary costeffectiveness of pandemic influenza a(h n ) vaccination in the united states mass gatherings and respiratory disease outbreaks in the united states -should we be worried? results from a systematic literature review and analysis of the national outbreak reporting system evidence compendium and advice on social distancing and other related measures for response to an influenza pandemic sars-cov- : fear versus data estimating the infection and case fatality ratio for coronavirus disease (covid- ) using age-adjusted data from the outbreak on the diamond princess cruise ship nonpharmaceutical measures for pandemic influenza in nonhealthcare settings-international travel-related measures. emerg infect dis introduction of mobile health tools to support ebola surveillance and contact tracing in guinea. glob health sci pract effectiveness of personal protective measures in reducing pandemic influenza transmission: a systematic review and meta-analysis the impact of mass gatherings and holiday traveling on the course of an influenza pandemic: a computational model use of non-pharmaceutical interventions to reduce the transmission of influenza in adults: a systematic review cost effectiveness and budget impact analyses of influenza vaccination for prisoners in thailand: an application of system dynamic modelling costs of diarrhoea and acute respiratory infection attributable to not handwashing: the cases of india and china economic analysis of the use of facemasks during pandemic (h n ) school closure and management practices during coronavirus outbreaks including covid- : a rapid systematic review. lancet child adolesc health is a mass prevention and control program for pandemic (h n ) good value for money? evidence from the chinese experience estimating clinical severity of covid- from the transmission dynamics in wuhan, china key: cord- -gn tsz authors: higginson, irene j.; yi, deokhee; johnston, bridget m.; ryan, karen; mcquillan, regina; selman, lucy; pantilat, stephen z.; daveson, barbara a.; morrison, r. sean; normand, charles title: associations between informal care costs, care quality, carer rewards, burden and subsequent grief: the international, access, rights and empowerment mortality follow-back study of the last months of life (iare i study) date: - - journal: bmc med doi: . /s - - - sha: doc_id: cord_uid: gn tsz background: at the end of life, formal care costs are high. informal care (ic) costs, and their effects on outcomes, are not known. this study aimed to determine the ic costs for older adults in the last months of life, and their relationships with outcomes, adjusting for care quality. methods: mortality follow-back postal survey. setting: palliative care services in england (london), ireland (dublin) and the usa (new york, san francisco). participants: informal carers (icrs) of decedents who had received palliative care. data: icrs reported hours and activities, care quality, positive aspects and burdens of caregiving, and completed the texas revised inventory of grief (trig). analysis: all costs (formal, informal) were calculated by multiplying reported hours of activities by country-specific costs for that activity. ic costs used country-specific shadow prices, e.g. average hourly wages and unit costs for nursing care. multivariable logistic regression analysis explored the association of potential explanatory variables, including ic costs and care quality, on three outcomes: positive aspects and burdens of caregiving, and subsequent grief. results: we received completed surveys, from london, dublin, new york and san francisco. most respondents were women ( %); average age was years. on average, patients received – h per week from icrs for ‘being on call’, – h for icrs being with them, – h for personal care, – h for household tasks, – h for medical procedures and – h for appointments. mean (sd) ic costs were as follows: usa $ , ( , ), england $ , ( , ) and ireland $ , ( , ). ic costs accounted for % of total (formal plus informal) costs. higher ic costs were associated with less grief and more positive perspectives of caregiving. poor home care was associated with greater caregiver burden. conclusions: costs to informal carers are larger than those to formal care services for people in the last three months of life. if well supported icrs can play a role in providing care, and this can be done without detriment to them, providing that they are helped. improving community palliative care and informal carer support should be a focus for future investment. in palliative care, those important to the patient, such as family members and informal carers (hereafter called 'informal carers', icrs) are part of the unit of care. icrs often provide high levels of demanding care and support willingly, because they see this as part of their relationship with the patient and are ambivalent to considering their own needs [ ] . while some icrs report positive outcomes such as closer relationships with others, greater appreciation of life, increased empathy and positive self-view, many can experience anxiety, depression, decline in quality of life and/or post-traumatic stress disorder [ ] . patient symptoms, such as breathlessness, fatigue or cognitive impairments, and advanced illness can increase caregiver burden and/or reduce caregiver rewards [ ] [ ] [ ] [ ] . however, this research is usually based on small, single setting studies [ ] [ ] [ ] [ ] . formal care costs in the last year of life account for between and % of health care costs [ ] [ ] [ ] [ ] , but to the best of our knowledge, informal care costs have never been compared internationally. their costs are much less recognised or understood [ ] . systematic review evidence indicates that living with relatives and/ or extended family support doubles or more (up to . times) the odds of patients being able to be cared for and to die at home, something for which many people wish [ ] . the activities of icrs likely save states billions in health and social care spending each year [ ] . home palliative care is thought to be cost-effective [ , ] . but what are the costs for icrs? dying at home when preferred can often require informal caregivers to be given time off work [ ] , which is unpaid. the relationship of ic costs with the outcomes for carers, such as burden, or subsequent grief are not known. therefore, as part of the international, access, rights, and empowerment (iare i) study of palliative care in three countries, we aimed to determine and compare the informal care (ic) costs and their associations with selfreported caregiver burden, rewards and subsequent caregiver grief, taking account of care quality, as reported by icrs. we conducted a mortality follow-back postal survey of key informants (normally relatives and informal carers) of decedents identified by palliative care services in participating hospitals. reporting follows strobe [ ] and morecare statements [ ] . see declarations for ethical approvals. further details are provided elsewhere [ ] . three countries included are in the top (of countries) of the economist intelligence unit quality of death index; rankings (scores) are as follows: england − ( . ), ireland − ( . ) and the usa − ( . ) [ ] . this index ranks the quality of palliative and end of life care across the world according to predetermined criteria, national data and interviews. the countries had different health care systems (england: national health service; ireland: national health insurance; usa: private health system, palliative care covered by most insurance agencies and medicare and medicaid [ ] ) and philanthropy supporting hospice and palliative care [ ] . participating palliative care services in london (england), dublin (ireland), new york and san francisco (usa) were as follows: established hospital palliative care consulting teams in all countries, a hospital-based community outreach team in london and an inpatient palliative care ward in new york. details of the participating services are found elsewhere [ , , ] . we identified patients aged ≥ years who had accessed (≥ contact) a participating palliative care team and died - months prior to the survey date. their next of kin, as indicated in clinical records, was sent study information and a postal questionnaire from their clinical service (following data-protection regulations), with a pre-paid envelope addressed to the research team. the next of kin was asked to complete the questionnaire or pass it to the most appropriate individual who was close to the patient for completion. all data were analysed anonymously. consenting respondents returned a self-completed questionnaire, pre-piloted in all countries. respondents reported demographic data including socio-economic status, living arrangements and relationship to patient and the patient's illnesses. this was supplemented by patient record data on age, diagnosis and co-morbidities. in addition, icrs reported health and social care services used by patients in the last months of life. informal time spent caring was counted with the client service receipt inventory (csri) [ , ] , by asking respondents to document all ic time spent by family and friends as well as the respondents during the last months of the patient's life. six questions covered a wide range of possible physical, social, emotional and other caring activities, including time spent 'on call' and being available for the patient (additional file : table s ). answers were given as categories of hours per week: less than h, - h, - h, - h, or more hours and all the time. quality of care of the last place of stay (e.g. hospital, home) in the last months was rated using likert scales from (very poor) to (excellent). carer burden and positive aspects of caregiving (pac) at the time when patients died was measured according to the zarit [ , ] and a set of eight questions derived from previous studies, respectively [ , , ] . subsequent grief was assessed using the revised texas revised inventory of grief (trig). this measures the intensity of grief after the death of a close person and has two scales: trig i (past behaviours when patient died, eight items) and trig ii (present feelings, referred to as subsequent grief in this study, items) [ ] [ ] [ ] . hours of ic spent per week for each item were converted as the middle point of the given range: . for less than h, for - h and , , and for the rest. to determine the ic costs for each patient, we multiplied the number of hours of care with country-specific shadow prices such as average hourly wages and unit costs for nursing care. all costs were translated into usd ($) for comparison, using the purchasing power parity (ppp) index. we checked the summary statistics and plotted the informal cost distributions of all patients in each country for illustrative purposes. formal care costs were extracted from an earlier analysis on this dataset and are presented to aid interpretation. these were calculated by multiplying the quantity of specific services used according to the csri with corresponding country-specific unit costs [ ] . we described provision and hours of ic by country. we explored the distribution of formal health and social care costs (in $ ) and ic costs (in $ ) and calculated the proportion of ic costs in the total societal costs. we also described the ic costs by carers' relationship to the patients. after examining the distribution of subsequent grief, carer burden and pac by country and carer's relationship to the patient, we plotted the univariate relationship between these variables and ic costs. we examined the factors associated with subsequent grief, carer burden and pac using multiple regression analysis. we selected explanatory variables based on previous literature reviews, meta-analysis and theoretical considerations [ , , ] . these included age, gender, patient's cause of death (cancer or not), carer's relationship to patient, a religious faith of carer, carer's feeling about household financial status, carer's quality rating with care at hospital or home and informal and formal care costs. country fixed effects were also included in the models. we used complete cases only. we calculated the sample size based on being able to detect a difference in the mean ic costs between countries, with % of power and α = . ( . with bonferroni correction for two pairs of comparison), which would require individuals in each country. we received completed surveys: ( . %) of delivered surveys in london, / ( . %) in dublin, / ( . %) in new york and / ( . %) in san francisco. missing values were infrequent ( - % of variables), if any, and scattered with no patterns. most respondents were women ( %) and average age was years (table ). . % of all respondents were daughter of patients, followed by wife or female partner ( . %), husband or male partner ( . %) and son ( . %). in ireland, . % of respondents had a religious belief, which is higher than in the uk ( . %) and the usa ( . %). about / were living comfortably or doing alright regarding the household income of themselves. more carers were in paid employment in the usa ( . %) than in the uk ( . %) or ireland ( . %). subsequent grief was the summation of items of trig, with mean scores of . (sd . ) and . (sd . ). mean scores of zarit measuring carer's burden were . (sd . ). positive aspects of caregiving (pac) were measured using items and its mean score was . (sd . ). the average age of those who had died was years, with similar numbers of women and men (table ) . patients had on average - icrs; % had an icr living with them; % had cancer as a cause of death. health and social care costs were on average $ , (sd $ , ), highest in the usa ($ , ) and lowest in the uk ($ , ). the most common care giving activities in all countries were spending time with the patient ( - % of patients received this) and 'being on call' ( - % of patients received this), i.e. being there to watch for problems at least a few hours per week ( table ) . household tasks were provided to - % of patients. more than half of the patients were helped with personal care and medical procedures. on average, patients received - h of ic per week from friends or family for personal care, - h for medical procedure, - h for appointments and - h for household tasks. friends and/or family spent - h per week on being on call and - h with patients. (fig. , table , additional file : figure s and figure s ). removing the being on call element of ic cost estimates reduced the costs, although the distributions remained unchanged (fig. , table , additional file : figure s and figure s ). ic costs varied less than did formal care costs between countries (table ) . ic costs were not associated with total formal care costs (pearson's r = − . ), nor with hospice/palliative care costs (pearson's r = . ). ic costs were higher for husband/male partners ($ , ) and wife/female partners ($ , ) than those of daughters/sons or other relatives/others (additional file : table s ). carer burden, positive aspects of caregiving (pac), subsequent grief and informal care costs subsequent grief, carer burden and pac were near normally distributed and were similar across the three countries (additional file : figure s ). wives, husbands and daughters reported higher subsequent trig grief score than other relatives, implying lesser distress (additional file : table s ). burden felt by carers was slightly higher among daughters and sons than others. positive aspects of caregiving did not differ by the relationship to patient much except the male relatives. in all three countries, there was a consistent pattern that subsequent grief was positively associated with ic costs, implying that more time spent on caring for patients was associated with lesser distress felt afterwards (fig. ) . as for the univariate analysis, higher ic costs were associated with higher trig scores for subsequent grief, indicative of less grief (models and , table ). however, the coefficients were small and the effects varied between individuals. for older patients, the carer's subsequent grief score was lower, i.e. the carer grieved more. higher ic costs were also associated with higher scores for pac (models and , table ). higher carer's satisfaction with home care was associated with higher pac. ic costs were not associated with caregiver burden. satisfaction with home care was negatively associated with caregiver burden, i.e. carers felt more burdened when they felt care provided for patients was not satisfactory (model , table ). carers with a religious faith felt less burdened. daughters felt more burdened, compared to wife/female partner of patients. this is the first international study of icrs of older people in their last months of life. we found that ic costs were high ($ , ), representing % of total societal costs. even when the elements of being on call are removed, ic costs still account for % of total care costs. those carers who reported higher ic costs, and/or more hours of informal care, had lesser subsequent grief and reported more positive aspects of caregiving, without a negative effect on caregiver burden. quality of care, as reported by the carer, was an important mediator; poorer experiences with home care were associated with more caregiver burden and fewer positive aspects of caregiving. the inverse relationship between ic costs and subsequent grief surprised us. it appears that providing more hours of ic to patients protected the carer during more hours of ic also led to a more positive feelings about caregiving. it may be that providing support protected icrs from guilt in later bereavement. it may also be that more hours of ic support were provided by larger families and groups, and so the icrs were not required to do so much individually, and possibly also gained from mutual support from other family members and friends. other possible explanations include the following: that icrs providing more support were more prepared for the death and ensured that the person they cared for did not feel burden to others; both factors possibly protect against complicated grief [ ] . this finding needs more study, widening the usual approach of considering 'single' patient-family dyads. however, it is a promising development, as it suggests that icrs can and often do want to be part of the caring team-and that this can be done without being harmful to them; they just need help to allow them to do this well. this finding should also inform the development of caregiver support interventions [ ] , which need to provide support across changes in setting [ ] . our finding did not support other meta-analyses that higher number of hours spent caregiving led to higher caregiver burden [ ] , and we suggest that this difference is because other studies did not account for variations in care quality and other potential confounders. poor formal home care quality was associated with poorer icr outcomes, in terms of greater burden and fewer rewards. poor quality of end-of-life care has been associated also with complicated grief in other population-based research [ ] . earlier analysis across these countries found that poor home care also was associated with high formal care costs [ ] . in contrast, palliative care services had high quality, but were little used, accounting for only - % of formal care costs [ ] . taken together, these findings suggest that improving community palliative care may improve care value, the care experience for patients and icrs; increase ic rewards; and reduce ic burden and formal care costs. treatments and interventions are ever more intricate, especially in the face of a multimorbid, older person, who is approaching the end of life [ , ] . coordination between settings and between the diversity of care interventions and treatments, communication and the response of staff in the face of clinical uncertainty collectively are vital to improve care experience, yet are often lacking [ , ] . as in the closely related field of patient safely, consideration of the whole patient journey is vital [ ] . our finding is particularly important at a time when carers are being asked to do more, due to self-and family isolation for older people as a result of the covid- pandemic [ ] , including potentially to administer medicines [ ] . interestingly, ic costs were quite similar between our countries, in contrast to formal end of life care costs which varied much more [ ] . we found the contribution of icrs in all three countries was similar in terms of hours of care provided and types of support given. most studies of icrs have considered the impacts on individual caregivers, rather than the needs of patients overall. a uk survey of bereaved cancer carers ( % response) found that respondents reported a median of h of caregiving each week [ ] . however, a societal perspective to supporting end of life care requires that the contribution of all caregivers be considered [ ] [ ] [ ] , as in our study. we observed that patients received higher levels of support, which is identified in earlier studies. patients were provided with - h of ic per week for personal care, - h for medical procedures, - h for appointments and - h for household tasks. icrs also spent - h per week on being on call and - h being with patients. using the societal perspective taken, our data suggest that ic costs at the end of life (usually based on allocating a minimum wage to caregiver's activities) account for more than half of total care costs. out of pocket payment for medications and private health insurance are not included in this analysis, and so may slightly underestimate care costs. however, we do not believe that this substantially alters our findings. payment for prescription medicines varies between countries so would limit international comparison. this descriptive data is an important contribution, because end of life care lags behind much of health care in economic appraisal [ ] . we were able to identify some tasks that would likely need to be performed at a relatively fixed times of day (such as medical procedures, personal care, appointments), and others (such as household tasks) that may be adjustable and performed at other times or even on different days, reducing the 'time-bound' opportunity costs [ ] . we were surprised by the amount of time spent by icrs on call, this was the most common activity. it may be that the uncertainties encountered in end of life care [ ] [ ] [ ] [ ] mean that icrs felt that someone in the family was on call most of the time. this will require a flexibility by employers to allow icrs to be able to respond to unpredictable needs. it also highlights the need for effective out-of-hours palliative care, in all settings, to support not only patients but caregivers, who are supporting patients throughout much of the week, and, at least in our study, anticipate problems outside 'normal' - working hours. we were also surprised by the lower numbers of people with dementia in our study. however, this was a sample of people who were recruited from primarily hospital-based palliative care services, where people in late stage dementia may have limited access. dementia may also have been under-represented in our data on primary diagnosis and is sometimes missed. this warrants further study, and we have planned the international, access, rights, and empowerment ii (iare ii), to study older people with symptoms and frailty who are not receiving specialist palliative care. we took a societal approach to costing, including ic costs, which places greater recognition on the role of icrs. we were able to collect the same data across our different countries, making the finding of similar patterns, in contrast to formal care costs which varied more between countries [ ] , more noteworthy and widely generalisable. we also had data from four major cities, all in the top ranks of the global power city index: rankings are as follows: london , dublin , new york and san francisco [ ] , cities are becoming the norm for many societies, and so our focus on cities makes our findings highly relevant to care for the future. we had a response similar to or better than similar mortality follow-back surveys [ , ] . we focussed on the last months of life, when it is known that formal care costs increase especially [ ] [ ] [ ] [ ] . our results are based on responses from bereaved carers or next of kin; thus, we do not know the informal care provided to patients who do not have such carers but might have had support from neighbours, friends or relatives missed by our survey. icrs may have different perspectives from patients and may have recall bias about the care provided, although the time window used in our mortality follow-back survey, - months after bereavement, is usually considered optimal [ , ] . our respondents were identified by the specialist palliative care teams in participating hospitals which provided palliative care services for the bereaved carers. icrs of patients who did not have access to palliative care or bereavement care may have different experiences, possibly worse. more than % of patients and icrs reported having a religious belief, which is higher than might be expected. this may be due to sample or measurement bias. we do not know whether or how icrs were practising beliefs. thus, the results suggesting that icrs with a religious belief felt less burdened should be treated with caution. we were not allowed under the ethics approval to collect data on non-responders, so we were not able to compare the characteristics of responders and non-responders. there were some differences between cities in their response rates, but the similarities between countries in informal care activities do not suggest that this altered our overall findings or conclusions. we asked about total ic activities involving all members of the family and friends, but the assessments of grief relate to the main icrs, and for practical reasons, we were not able to study grief among all those involved in caregiving. the data are crosssectional, which limits the basis for establishing causality: we cannot positively determine that increased ic costs protected against grief, nor that poor quality care resulted in greater ic burden or fewer rewards. however, our findings meet many of the bradford-hill criteria for supporting causal relationships, such as consistency with other literature and across settings, temporality, plausibility and coherence [ ] . further, in this complex situation of end of life care, it is impossible to understand any causal chain perfectly, i.e. know every factor that could be considered a cause [ , ] . even if the chain were clear, it would not be clear how best to change the outcome, as the interventions are by definition complex [ ] . our data provide insights into how to improve care value at the end of life, which is profoundly needed and can also help with the appropriate modelling of complex interventions [ ] . our data may also help with the development of robust business cases for palliative care [ ] . the contribution of icrs is considerable, accounting for around % of total care costs. these costs are similar across countries. training and support interventions for icrs should target the wide range of activities that they undertake. increased informal care hours and costs, can lead to more rewards and lesser subsequent grief. therefore icrs, including family and friends and beyond one main informal carer, are central at the end of life and should be considered in all interventions. our finding of an association between poor care quality and poorer icrs outcomes, including greater burden and fewer rewards, suggests an urgent need to improve care quality, through the better integration and support for dedicated community palliative care services, and support people across the whole journey of care. improving community palliative care may improve care value, the care experience for patients and icrs; increase ic rewards; and reduce ic burden and formal care costs and should be a focus for investment, including and importantly during the covid- pandemic. supplementary information accompanies this paper at https://doi.org/ . /s - - - . additional file : table s . informal care (ic) costs in the last three months of life categorized by the relationship of carers to patients. figure s . distribution of subsequent grief, carer burden and positive aspects of caregiving and by country. table s . subsequent grief, carer burden and positive feeling of care by carers' relationship to patients. figure s . histogram of costs of informal care (ic) provided for older patients in their last three months of life, with and without 'time being on call' in three countries. figure s . cumulative distribution of costs of informal care (ic) provided for older patients in their last three months of life, with and without 'time being on call' in three countries. we are grateful for the main support for this programme from cicely saunders international (csi, uk charity) and the atlantic philanthropies. additional support was provided by the nihr south london clahrc and higginson's nihr senior investigator award. a component of this research was supported by the national institute for health research collaboration for leadership in applied health research south london (nihr clahrc south london), now recommissioned as nihr applied research collaboration south london. the views expressed in this publication are those of the author(s) and not necessarily those of the nihr or the department of health and social care. the funding sources had no role in the design of this study, its execution, analyses, interpretation of the data or decision to submit results. the anonymised datasets supporting the conclusions of this manuscript are available upon request to the corresponding authors and buildcare team. if participants were agreeable to being approached, a researcher fully explained the study to them, provided an information sheet and gained written informed consent. adults unable to give informed consent or deemed too ill to complete any part of the interview were excluded. not applicable. we declare no conflicts of or competing interests. author details working with ambivalence: informal caregivers of patients at the end of life informal caregivers in cancer: roles, burden, and support (pdq(r)): health professional version. bethesda: pdq supportive and palliative care editorial board a prospective analysis on fatigue and experienced burden in informal caregivers of cancer patients during cancer treatment in the palliative phase living with breathlessness: a survey of caregivers of breathless patients with lung cancer or heart failure costs and burden associated with loss of labor productivity in informal caregivers of people with dementia: results from spain care participation and burden among informal caregivers of older adults with care needs and associations with dementia end-of-life medical spending in last twelve months of life is lower than previously reported exploring the cost of care at the end of life medicare cost at end of life the myth regarding the high cost of end-of-life care bringing the economic cost of informal caregiving into focus factors influencing death at home in terminally ill patients with cancer: systematic review the hidden costs of caring. london: marie curie uk is homebased palliative care cost-effective? an economic evaluation of the palliative care extended packages at home (peach) pilot effectiveness and cost-effectiveness of home palliative care services for adults with advanced illness and their caregivers is dying in hospital better than home in incurable cancer and what factors influence this? a population-based study the strengthening the reporting of observational studies in epidemiology (strobe) statement: guidelines for reporting observational studies evaluating complex interventions in end of life care: the morecare statement on good practice generated by a synthesis of transparent expert consultations and systematic reviews drivers of care costs and quality in the last months of life among older people receiving palliative care: a multinational mortality follow-back survey across england, ireland and the united states -quality-deathindex?redirect=true: the economist intelligence unit five types of oecd healthcare systems: empirical results of a deductive classification research priorities in health economics and funding for palliative care: views of an international think tank social and clinical determinants of preferences and their achievement at the end of life: prospective cohort study of older adults receiving palliative care in three countries variations in the quality and costs of end-of-life care, preferences and palliative outcomes for cancer patients by place of death: the qualycare study short-form zarit caregiver burden interviews were valid in advanced conditions the zarit burden interview: a new short version and screening version positive aspects of caregiving: contributions of the reach project to the development of new measures for alzheimer's caregiving validity and reliability of the positive aspects of caregiving (pac) scale and development of its shorter version (s-pac) among family caregivers of older adults psychometric evaluation of the texas revised inventory of grief in a sample of bereaved family caregivers factorial validity of the texas revised inventory of grief-present scale among bereaved older adults development of the texas inventory of grief caregiver burden: a clinical review principles of confounder selection factors associated with possible complicated grief and major depressive disorders caregiver bereavement outcome: relationship with hospice at home, satisfaction with care, and home death from 'conductor' to 'second fiddle': older adult care recipients' perspectives on transitions in family caring at hospital admission association between quality of end-of-life care and possible complicated grief among bereaved family members a framework for complexity in palliative care: a qualitative study with patients, family carers and professionals understanding the experiences of patients with advanced multimorbidity to improve care: serial interviews with patients and carers coordination of care for individuals with advanced progressive conditions: a multi-site ethnographic and serial interview study improving primary palliative care in scotland: lessons from a mixed methods study safer healthcare the family caregiving crisis meets an actual pandemic administration of end-of-life drugs by family caregivers during covid- pandemic the contributions of family care-givers at end of life: a national postbereavement census survey of cancer carers' hours of care and expenditures an overview of methods and applications to value informal care in economic evaluations of healthcare variations in the cost of formal and informal health care for patients with advanced chronic disease and refractory breathlessness: a cross-sectional secondary analysis costs of formal and informal care in the last year of life for patients in receipt of specialist palliative care an economic view on the current state of the economics of palliative and end-of-life care time-bound opportunity costs of informal care: consequences for access to professional care, caregiver support, and labour supply estimates the value of uncertainty in critical illness? an ethnographic study of patterns and conflicts in care and decision-making trajectories patient and carer experiences of clinical uncertainty and deterioration, in the face of limited reversibility: a comparative observational study of the amber care bundle the challenges of uncertainty and interprofessional collaboration in palliative care for non-cancer patients in the community: a systematic review of views from patients, carers and health-care professionals uncertainty in illness. an integrative review tokyo: mori memorial foundation home care in the last year of life: family member perceptions of unmet need associated with last place of care the environment and disease: association or causation? causal inference in public health evaluating public health interventions: . causal inference for time-invariant interventions the business case for palliative care: translating research into program development in the u.s this international, access, rights, and empowerment (iare) study was part of buildcare, an international research programme supported by cicely saunders international ( higginson, normand, meier and morrison were grant leads. daveson, normand, ryan, mcquillan, morrison, selman and pantilat were study pis for the sites of london, dublin, new york and san francisco; selman and johnston coordinated data collection in different sites. yi led the analysis. higginson drafted the manuscript, with significant contributions from yi, normand, johnston and ryan. higginson and yi had full access to all the data in the study and take responsibility for the integrity of the data and analysis. all authors contributed to the development of study aims, integration of the interpretation across the buildcare study and analysis plan and provided critical revision of the manuscript for important intellectual content. all authors read and approved the final manuscript. springer nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. key: cord- -rrsgl authors: beutels, philippe; scuffham, paul a; macintyre, c raina title: funding of drugs: do vaccines warrant a different approach? date: - - journal: the lancet infectious diseases doi: . /s - ( ) - sha: doc_id: cord_uid: rrsgl summary vaccines have features that require special consideration when assessing their cost-effectiveness. these features are related to herd immunity, quality-of-life losses in young children, parental care and work loss, time preference, uncertainty, eradication, macroeconomics, and tiered pricing. advisory committees on public funding for vaccines, or for pharmaceuticals in general, should be knowledgable about these special features. we discuss key issues and difficulties in decision making for vaccines against rotavirus, human papillomavirus, varicella-zoster virus, influenza virus, and streptococcus pneumoniae. we argue that guidelines for economic evaluation should be reconsidered generally to recommend ( ) modelling options for the assessment of interventions against infectious diseases; ( ) a wider perspective to account for impacts on third parties, if relevant; ( ) a wider scope of costs than health-care system costs alone, if appropriate; and ( ) alternative discounting techniques to explore social time preference over long periods. in many high-income countries, public funding of preventive vaccines is assessed based on the same criteria as the funding of curative pharmaceutical drugs. such routine drug assessment processes consider evidence on quality, safety, effi cacy, and cost-eff ectiveness. because of the increase in the number of diff erent vaccines available and advances in the science behind decision making, we have drawn on existing literature and practices to develop the arguments around potential disparities with other pharmaceuticals when assessing vaccines for public funding. these arguments revolve around vaccine-specifi c features of herd immunity and eradication, which are not evident in pharmaceuticals, and features for which the eff ects of quality-of-life losses in very young children, parental care and work loss, time preference, macroeconomics, and uncertainty substantially infl uence cost-eff ectiveness estimates. vaccines may increasingly be judged as unacceptable if these features are not acknowledged. we also illustrate these points for fi ve specifi c vaccines that are currently under consideration for widespread use in high-income countries. we use the term "cost-eff ectiveness" in a broad sense throughout this article, encompassing cost-utility and cost-benefi t analysis, although there are technical diff erences. in , australia was the fi rst country to make evidence on cost-eff ectiveness a mandatory part of funding decisions of drugs. the australian pharmaceutical benefi ts advisory committee is a rigorous and well-run system for evaluating drugs for acute care, chronic disease, palliation, and more recently vaccines. many other countries have adopted a similar philosophy towards cost-eff ectiveness considerations for funding pharmaceuticals (eg, belgium, finland, norway, canada [ontario] , portugal, sweden, netherlands, uk, and usa [some organisations]), but they deal with preventive public-health measures, such as mass vaccination, in diff erent ways. some countries have specifi c advisory groups to make funding recommendations on vaccinations (eg, uk joint committee on vaccination and immunisation, us advisory committee on immunization practices). often, cost-eff ectiveness evidence for vaccines is assessed in the same manner as for any drug. nevertheless, as we discuss below, vaccination has special features that make it particularly challenging to assess. furthermore, vaccination constitutes one of the largest preventive health programmes around the world, and increasing pressures on health-care budgets are as much a challenge for the use of vaccines as for other drugs. vaccines provide primary prevention of future morbidity and mortality. thus, unlike secondary prevention interventions, such as statins for cholesterol lowering, vaccines are targeted before, or in the initial stages of, the recipient's potential risk exposure. additionally, the recipient may or may not benefi t on an individual basis. vaccination may even harm some recipients through vaccine-associated adverse events (panel); for example, - % of varicella-zoster virus (vzv) vaccine recipients report a localised rash. the individual perception of risks of disease and risks of adverse events drives the demand panel: why many vaccines require a diff erent approach • primary prevention in healthy people, but with possibility of adverse events • unvaccinated or poorly vaccinated people may experience benefi cial or, more rarely, detrimental impact from herd immunity • many vaccines prevent short-lived illness in very young children, causing extra family care and work loss, for which evaluation methods lack credibility and acceptability • the cost-eff ectiveness of many vaccines is highly sensitive to the choice of discount method • some infections are eradicable • some emerging infections (eg, sars, pandemic infl uenza) would have a major macroeconomic impact that goes beyond lost productivity of sick people sars=severe acute respiratory syndrome. for vaccines, and may dominate the infl uence of other factors, such as price. the need to show protective effi cacy beyond the typical duration of clinical trials generally aff ects the assessment of vaccines more than therapeutic pharmaceuticals, primarily because the endpoints may not be immediate. in fact, the clinical endpoints might not show clinical effi cacy at the time of trial reporting because the numbers required can be extremely large. clinical endpoints of mortality or hospital admissions might require follow-up of thousands to millions of participants over as long as several decades. as such, some vaccines have been funded on the basis of immunogenicity data or intermediate endpoints alone (eg, meningococcal c conjugate vaccine and human papillomavirus [hpv] vaccine in several countries). , vaccination not only protects vaccine recipients, but reduces exposure of unvaccinated people to infection through herd immunity. herd immunity, in addition to lowering the incidence of infection in the unvaccinated, is well known to lead to an increased average age at infection. vaccination is therefore not always entirely benefi cial to public health because some childhood infections are more severe if contracted in adolescence or adulthood. furthermore, vaccination itself may modify vaccine eff ectiveness over time because of factors such as strain replacement and cross reactivity. some of these indirect eff ects improve the cost-eff ectiveness (eg, non-exposure of most of the unvaccinated, cross reactivity), whereas others may reduce the costeff ectiveness (eg, shift in the average age of infection, serotype replacement). for most vaccination programmes, the sum of these eff ects substantially improves cost-eff ectiveness, but sometimes the reverse may be true. , convincing evidence for the extent of herd immunity, and the duration of immunity, may only come from widespread use in another country, not from clinical trials. for example, the population impact of vaccinations against vzv and streptococcus pneumoniae in the usa are of major interest to other countries. appropriately parameterised dynamic transmission models could also provide credible estimates of herd-immunity eff ects. lieu and colleagues were the fi rst to estimate the costeff ectiveness of a vaccine based on dynamic model simulations. such models, which take into account the above indirect eff ects, are gradually becoming more widespread, but are not yet part of the traditional toolbox of epidemiologists or health economists. all these features add to the uncertainty under which vaccine funding decisions are made, as opposed to those of other drugs. for whatever the reason some people decline vaccination for their child, they may trade the uncertain value of direct protection for the certainty of avoiding the risk of vaccine-associated adverse events and the cost of vaccination, while potentially counting on a "free ride" from herd immunity induced by others being vaccinated. the risk perceptions driving this trade-off are distorted as a result of imperfect information. reductions in vaccine-preventable disease make people believe that their child's risk of disease has decreased. however, their risk is highly dependent on historical and future rates of exposure and vaccination in the rest of the population and can quickly rebound when uptake declines. , therefore, government intervention in the form of subsidies or public funding is required to ensure that vaccine uptake remains high enough to guarantee benefi cial herd immunity. the uk's recent struggle with the measles, mumps, and rubella vaccine uptake illustrates this point. for other pharmaceuticals, this kind of trade-off is not even conceivable. potential global eradication is another feature that sets some vaccines apart. for example, polio has been eliminated in high-income and middle-income countries. the risk of acquiring paralytic polio from the live oral polio vaccine is thus particularly sensitive to public scrutiny. however, replacing the oral vaccine with the risk-free inactivated polio vaccine is far more expensive, and would be judged unacceptable if cost-eff ectiveness were the only criterion under consideration. nevertheless, until polio is eradicated globally, vaccination must continue or polio will again become endemic, as shown by occasional outbreaks in unvaccinated communities. although not usually quantifi ed in cost-eff ectiveness analysis, , the prospect of eradication and concerns over the public's perception about the entire vaccination programme has led to the replacement of oral vaccine by inactivated vaccine in nearly all highincome countries. some infections have the capacity to aff ect not only patients and their direct contacts (ie, their family, health-care provider, employer) in terms of economic costs and medical eff ects, but they may also aff ect health-care use, and expectations and behaviour of consumers and investors. for instance, pandemic infl uenza is likely to lead to capacity problems within the health-care system, aff ecting the timely treatment of patients with infl uenza in addition to those with unrelated illnesses. additionally, it would have a macroeconomic impact that goes beyond lost productivity to employers of sick patients, because virtually everyone-employers, con sumers, and investors-would adapt their intentions under its perceived threat. , the latter was also shown in countries aff ected by the outbreak of severe acute respiratory syndrome. finally, affl uent countries pay much higher prices than poorer countries. this system of tiered pricing is not unique to vaccines, but might be most relevant for new vaccines (eg, rotavirus, pneumococcal, and hpv) and medications (eg, highly active antiretroviral therapy) with great lifesaving potential in poor countries. some economists argue that market prices set for high-income countries need to be much higher to suffi ciently stimulate personal view innovation through market mechanisms, rather than rely on publicly funded research. conversely, if a vaccine is added to a low-income country's national programme, it is likely to become cheaper for high-income countries through price discrimination mechanisms. clearly, decision making becomes more complex if such moral or opportunistic considerations are thought to be important. there are some methodological aspects to which the cost-eff ectiveness of vaccines is particularly sensitive. first, the defi nition of the analytical viewpoint is crucial. guidelines for economic evaluation, as used by most advisory committees, generally focus on direct health-care costs and do not consider indirect costs to society (eg, the value of lost productive and leisure time from illness or caregiving). these indirect costs can be very large for infectious diseases that aff ect virtually the entire population, even for generally benign illness. for example, the cost-eff ectiveness of childhood vzv vaccination is unlikely to be thought acceptable from the health-care budget perspective, but is possibly cost-saving from a societal perspective. , , second, the use of quality-adjusted life-years is widely advocated as the best measure currently available for valuing health states. however, standardised quality-oflife estimates for short-term diseases in young children are virtually non-existent, and the appropriate methods to measure them are subject to debate. [ ] [ ] [ ] additionally, the impact of a child's illness on the quality of life of caregivers can be substantial, just as it is for lifethreatening and severe chronic diseases in adults (eg, cancer). however, such indirect quality-of-life losses are typically not accounted for. these impacts have the potential to change decisions, for instance on rotavirus vaccine. finally, the peace of mind off ered through the reassurance of vaccine protection is a quality-of-life improvement of prevention programmes that is routinely ignored in economic evaluation. a third issue is the impact that discounting has in accounting for time preference. discounting is a technique that aims to put costs and benefi ts occurring at diff erent timepoints on the same basis of comparison. discounting scales down future events, such that, the further into the future they occur or the higher the discount rate, the less important they are to a decision maker in the present. in health economics, there is continued debate about whether the discount rate for health outcomes should be lower than or equal to that for costs. , for curative therapies, most benefi ts accrue immediately or shortly after the intervention is initiated, and the cost-eff ectiveness of these interventions is therefore largely independent of these methodological disagreements on discounting. conversely, the costeff ectiveness of most prevention programmes is highly sensitive to discounting because of the long time spans over which benefi ts accrue. a slight decrease in discount rate-eg, from % to %-could change the costeff ectiveness of vaccination from unacceptable to attractive. country-specifi c recommend ations on discount rates vary to the extent that a vaccine could be deemed cost eff ective in one country and cost-ineff ective in another for this reason alone (table ). in the standard discount procedure, as recommended in all guidelines known to us, the discount rate is constant, implying that preferences between outcomes are held constant through time and depend only on the length of the time interval between them. one can argue that discounting at a constant rate exaggerates the importance we give for the present over the future. [ ] [ ] [ ] this assertion is backed by psychological empirical evidence, which suggests that the diff erence between equidistant outcomes is thought less important the further into the future the outcomes occur. so-called "slow" discounting procedures could be used for cases in which the discount rate decreases and falls close to zero for the more distant future (eg, · % for years - , · % for years - , % thereafter), thus yielding a higher present value of benefi ts. , additionally, time preference may exist only to the time until risk exposure, and not the time until health consequences from risk exposure arise (eg, cervical cancer is the health consequence of a much earlier exposure to hpv). adjustment of the discount procedure to account for these aspects is not current practice, but would substantially improve the estimated cost-eff ectiveness of prevention versus cure. , currently, policy makers are presented with very wide cost-eff ectiveness ranges for preventive public-health actions when sensitivity to discounting is illustrated to them. in , baumol noted the "sorry spectacle" that economists provided through their diverging understandings on this subject, and his assertion that "little help is provided to the decision maker who is confronted with such an enormous range of finally, the equity impact of vaccination is far less predictable than for most drugs. generally, the less healthy or less wealthy are those least likely to be vaccinated, and thus more likely to experience the eff ect of herd immunity from other people receiving vaccination. as shown for measles in bangladesh, this eff ect is often equitable, but the reverse may also occur for poorly executed vaccination programmes. the redistribution eff ects on health and wealth are thus less straightforward in the prediction of decisions on vaccination compared with those used for therapeutic medicines. the fi rst generation of vaccines, such as measles, pertussis, and polio vaccines, were against serious childhood diseases that were common worldwide. little analysis was done before their introduction because their benefi ts were obvious and their costs were low in an era when there was less pressure on the health-care budget. new vaccines are much more expensive and often aimed at less common or less serious diseases, particularly in wealthy countries. thus, whether these vaccines are worth introducing is less clear. we will explain key aspects of the cost-eff ectiveness of current vaccines, while focusing on high-income countries. rotavirus is the commonest cause of dehydrating gastroenteritis in the world and accounts for most gastroenteritis hospital admissions in children under years of age. deaths are infrequent because of good medical care in high-income countries (eg, about three deaths per year in the uk). a challenge to the evaluation of both current oral rotavirus vaccines is the estimation of the part of the gastroenteritis disease burden specifi cally attributable to rotavirus, as well as assessing the extent to which these vaccines would invoke herd immunity. in high-income countries, the main benefi t of rotavirus vaccines is the prevention of parental care and productivity losses in virtually all households with infants or toddlers. however, as we have outlined, gains in quality-adjusted life-years in such young children and their parents, as well as parental care and work loss, are not standard features in cost-eff ectiveness analyses. given the current price setting (€ - per fully vaccinated child) and the recommended schedule for these vaccines (two doses rotarix [glaxosmithkline]; three doses rotateq [merck]), they are unlikely to be judged as cost eff ective unless these so-called "soft" benefi ts are also included. , [ ] [ ] [ ] but if they are, why should they not also be considered for all other health-care interventions, thus potentially reshuffl ing the comparison between all health-care programmes (including the other vaccines discussed here)? table describes potential consequences of including soft costs and benefi ts at various levels of government decision making. hpv vaccines are eff ective against the two hpv serotypes associated with most cervical cancers, and one of these vaccines also protects against two of the serotypes that cause genital warts. eff ectiveness against cervical cancer would have to be modelled based on the premise that hpv infection is a necessary condition for cervical cancer to develop, although often only decades later. the cost- programmes that prevent disease, with a proportionately larger aggregated impact on the quality of life and productivity of patients and/or their families, become more cost eff ective compared with other vaccination programmes ·· hpv=human papillomavirus. *costs and benefi ts arising to parties generally not considered relevant in guidelines for economic evaluation of pharmaceuticals for which public funding is sought. these third parties can consist of people not receiving the intervention, parents of patients, employers of patients, and employers in general. †cost-benefi t analyses do not routinely inform other sector decisions in many countries (eg, education, transport infrastructure, military, etc.). politics may dominate rational decision rules in other sectors more than in health care. ‡produced by glaxosmithkline. §produced by merck. ¶produced by wyeth. personal view eff ectiveness of hpv vaccines depends heavily on the choice of the discounting approach used. , furthermore, mathematical models for hpv vaccination ideally have to build in complexities related to herd-immunity eff ects from vaccinating cohorts of girls only and boys additionally, the optimum frequency of cervical cancer screening, and type-specifi c progressive infection and replacement, all over long time periods, which makes this a very complex programme to assess properly. , however, a more simple approach, based on static models, could give insights on the basic question: should we vaccinate girls before their sexual debut? such models would underestimate the benefi ts of hpv vaccination, and therefore would only be helpful for policy if they resulted in favourable cost-eff ectiveness ratios. the static models that have been published so far have tended to be favourable. , , policy makers could therefore quickly decide about vaccinating a limited number of cohorts before their sexual debut, and have reasonably confi dent cost-eff ectiveness evidence to support this decision. however, they cannot rely on such analyses to decide on more complicated aspects of the programme, such as the breadth of the programme in girls and boys. in view of the high costs of this programme (€ - per fully vaccinated individual), the uncertainty surrounding these more complicated decisions could unnecessarily postpone policy on the more basic issue. vzv childhood vaccination prevents chickenpox in vaccinated children and is likely to protect these vaccinees against shingles later in life. since chickenpox infects virtually all children by age years, the accumulated societal savings, including avoided parental care and productivity losses, are likely to be greater than the costs of vaccination at a price of € - per fully vaccinated person. however, childhood vzv vaccination increases the occurrence of shingles in adults and this may be such that it counteracts these societal savings and leads to adverse health eff ects. a further complication is that with single-dose infant vaccination many teenage breakthrough cases can still be expected, but the addition of a second dose to prevent this would make it a much less cost-eff ective programme. modifi ed vzv vaccine in adults was recently shown to prevent shingles, and was shown by static models to be cost eff ective. , finally, vaccination of susceptible pre-adolescents is an alternative strategy that has consistently been shown to be cost eff ective to the health-care budget, and is thus independent of the wider societal perspective. however, it is not advocated by public-health specialists, because it would only prevent a small part of all chickenpox disease, albeit the most severe proportion. , , clearly, the simultaneous modelling of all these strategies and considerations requires complex models and data from various sources to establish eff ectiveness. empirical studies alone cannot answer all these questions. infant infl uenza vaccination may be a cost-eff ective way of preventing seasonal infl uenza and pneumonia in young children directly and the elderly indirectly through herd immunity. , however, vaccinating a child partially to save a grandparent from experiencing serious illness does not only raise concerns over intergenerational equity, but also the eff ectiveness of such an approach could only be shown if put into practice on a large scale, or by applying an appropriately parameterised model. seasonal variations in incidence, severity of disease, and vaccine effi cacy are complicating factors that contribute to uncertainty. furthermore, preparing for pandemic infl uenza demands very large investments, and this can only be shown to be worthwhile by modelling. a policyrelevant approach to modelling the cost-eff ectiveness of pandemic infl uenza vaccination would entail considering macroeconomic impacts across sectors and across countries. clearly, deciding on the best options to prevent and control infl uenza requires an analytical framework and applied modelling work that substantially digresses from usual drug assessments. the currently available seven-valent pneumococcal conjugate vaccine (pcv ), which costs about € - per vaccinated child, is eff ective against invasive and non-invasive disease caused by seven serotypes of s pneumoniae. because of its high price, in the short term the cost-eff ectiveness of pcv depends in most high-income countries on the inclusion of positive herd-immunity eff ects in adults, which were observed after year of widespread use in the usa. if the long-term eff ect of its widespread use, consisting of a mix of herd immunity, serotype replacement, antibiotic resistance, and cross reactivity, remains benefi cial and if the cheaper three-dose schedule confers near-equivalent protection to the original four-dose schedule, pcv vaccination programmes are judged to be cost eff ective in high-income countries. to budget for this vaccine, european policy makers should accept imputations from herd-immunity eff ects observed in other countries in the short term as well as uncertainties with both positive and negative impacts of the programme in the longer term. advisory processes on drug funding can be generally eff ective at selecting which pharmaceuticals, and which subgroups of patients, should be subsidised to make the most of scarce health-care resources. vaccines are diff erent and more complex than most drugs assessed by such processes for the reasons we have outlined. this implies that such processes should be more fl exible in accepting the best available quantifi ed evidence of the unique features of vaccination programmes, and that decision makers and their advisers should be aware of these features if they cannot be quantifi ed. the best personal view available evidence depends on the type of infection and vaccine, and the time of its consideration. guidelines for economic assessment of pharmaceuticals dictate the approach to use to make such analyses acceptable for a country's decision makers. since economic evaluation is not an exact science, such guidelines are made on the basis of compromises between the people designing them and therefore can be changed (table ) . economic evaluation requires quantifi cation of the eff ects of interventions, as well as valuing these eff ects. in terms of quantifying the eff ects of vaccination, governments should adapt their guidelines to specify modelling options for the assessment of interventions against infectious diseases. this should enable submitters and drug-reimbursement committees to better understand which models are acceptable (or unacceptable) under which circumstances. crucially, drug-reimbursement committees must be represented by the required expertise to properly understand and evaluate complex vaccine models. in terms of valuing the eff ects of vaccination, we do not plead for a special case, but for a level playing fi eld. that is, we argue that not all aspects of ill health and time preference are currently captured by recommended techniques for economic evaluation, and that this may disadvantage the cost-eff ectiveness of interventions against diseases in children relative to interventions against diseases in adults, and prevention relative to cure. therefore, guidelines should also be adapted in general terms to allow for ( ) a wider perspective to account for eff ects on third parties, if these are aff ected substantially by specifi c interventions (eg, parents experiencing a quality-of-life impact through the illness of their child); ( ) a wider scope of costs to be included, if appropriate, than health-care system costs alone (eg, irrecoverable losses caused by modifi ed behaviour when faced with a large public-health threat); and ( ) alternative discounting techniques to deal with social time preference over long time periods. large uncertainties about the value and distribution of particular variables imply that timely vaccine decisions may need to be taken with more uncertainty than decisions on other drugs. this should not deter the widespread use of new safe and effi cacious vaccines, if-all things considered-these are unlikely to be judged cost-ineff ective relative to other interventions. furthermore, other criteria, including the programme's acceptability, feasibility, budget, and equity impact, are also important. a who guide for the standardisation of economic evaluations of immunisation programmes, which will become shortly available for public use, could be used as a starting point for governments to adapt their guidelines with respect to some of the issues mentioned here. economic evaluation of vaccination vaccine adverse events: separating myth from reality economic epidemiology and infectious disease the role of economic evaluation in vaccine decision making: focus on meningococcal group c conjugate vaccine human papillomavirus vaccination-reasons for caution the seroepidemiology of human papillomavirus infection in australia herd immunity: history, theory, practice impact of model, methodological, and parameter uncertainty in the economic analysis of vaccination programs evaluating the cost-eff ectiveness of vaccination programmes: a dynamic perspective eff ectiveness of sevenvalent pneumococcal conjugate vaccine against invasive pneumococcal disease: a matched case-control study cost-eff ectiveness of a routine varicella vaccination program for us children theoretical epidemiologic and morbidity eff ects of routine varicella immunization of preschool children in the united states impact of anti-vaccine movements on pertussis control: the untold story eradication versus control for poliomyelitis: an economic analysis compulsory vaccination and conscientious or philosophical exemptions: past, present, and future improving uptake of mmr vaccine cost-eff ectiveness analysis of changing from live oral poliovirus vaccine to inactivated poliovirus vaccine in australia poliomyelitis outbreak in an unvaccinated community in the netherlands, - cost-eff ectiveness of incorporating inactivated poliovirus vaccine into the routine childhood immunization schedule world wide experience with inactivated poliovirus vaccine partially wrong? partial equilibrium and the economic analysis of public health emergencies of international concern precautionary behavior in response to perceived threat of pandemic infl uenza the economic impact of sars: how does the reality match the predictions? tiered pricing of vaccines: a win-win-win situation, not a subsidy who benefi ts from new medical technologies? estimates of consumer and producer surpluses for hiv/aids drugs is g putting profi ts before the world's poorest children? economic evaluations of varicella vaccination programmes: review of the literature comments on the prosser et al approach to value disease reduction in children quality-adjusted life-years lack quality in pediatric care: a critical review of published cost-utility studies in child health preferences and willingness to pay for health states prevented by pneumococcal conjugate vaccine eff ect of pneumococcal vaccination on quality of life in children with recurrent acute otitis media: a randomized, controlled trial cost-eff ectiveness of rotavirus vaccination: exploring caregiver(s) and "no medical care" disease impact in belgium methodological issues and new developments in the economic evaluation of vaccines need for diff erential discounting of costs and health eff ects in cost eff ectiveness analyses discounting and cost-eff ectiveness in nice-stepping back to sort out a confusion recommendations of the panel on cost-eff ectiveness in health and medicine guidelines for the pharmaceutical industry on preparation of submissions to the pharmaceutical benefi ts advisory committee (pbac): including major submissions involving economic analyses. canberra: commonwealth department of health and ageing guide to the methods of technology appraisal. london: national institute for health and clinical excellence dutch health insurance board. guidelines for pharmacoeconomic research canadian agency for drugs and technologies in health. guidelines for the economic evaluation of health technologies: canada, rd edn. ottawa: canadian agency for drugs and technologies in health prescription for pharmacoeconomic analysis national institute for clinical excellence. guidance for manufacturers and sponsors pharmac responds to richard milne on discounting health benefi ts and costs a prescription for pharmacoeconomic analysis time preference, the discounted utility model and health anomalies in intertemporal choice: evidence and an interpretation the social rate of discount and the optimal rate of investment saving future lives. a comparison of three discounting models proportional discounting of future costs and benefi ts valuing prevention through economic evaluation: some considerations regarding the choice of discount model for health eff ects with focus on infectious diseases social rate of discount measles vaccination improves the equity of health outcomes: evidence from bangladesh epidemiology of rubella and congenital rubella syndrome in greece estimating the number of deaths with rotavirus as a cause in england and wales evaluating rotavirus vaccination in england and wales. part ii. the potential cost-eff ectiveness of vaccination the cost-eff ectiveness of rotavirus vaccination in australia cost-eff ectiveness and potential impact of rotavirus vaccination in the united states cost-eff ectiveness analyses of human papillomavirus vaccination the potential cost-eff ectiveness of prophylactic human papillomavirus vaccines in canada the epidemiological and economic impact of a quadrivalent human papillomavirus vaccine ( / / / ) in the uk modeling human papillomavirus vaccine eff ectiveness: quantifying the impact of parameter uncertainty varicella vaccination in england and wales: cost-utility analysis a vaccine to prevent herpes zoster and postherpetic neuralgia in older adults the epidemiology of herpes zoster and potential cost-eff ectiveness of vaccination in england and wales an economic evaluation of varicella vaccination in italian adolescents infl uenza vaccine eff ectiveness in healthy -to -month-old children during the - season interdisciplinary epidemiologic and economic research needed to support a universal childhood infl uenza vaccination policy school-based infl uenza vaccination program reduces infl uenza-related outcomes among household members optimal allocation of pandemic infl uenza vaccine depends on age, risk and timing convincing or confusing? economic evaluations of childhood pneumococcal conjugate vaccination-a review the global burden of disease assessments-who is responsible? pb acknowledges funding from simulation models for infectious disease processes (simid), a strategic basic research project funded by the institute for the promotion of innovation by science and technology in flanders (project number ). we thank the anonymous referees for their helpful comments. key: cord- - y b authors: madanoglu, melih title: state-of-the-art cost of capital in hospitality strategic management date: - - journal: handbook of hospitality strategic management doi: . /b - - - - . - sha: doc_id: cord_uid: y b nan making well-informed and effective capital investment decisions lies at the heart of any successful business organization. however, prior to investing in a project, an executive/manager should make three key estimates to ensure the viability of a business project: economic useful life of the asset, future cash flows that the project will generate, and the discount rate that properly accounts for the time value of the capital invested and compensates the investors for the risk they bear by investing in that project ( olsen et al. , ) . although the first two items are fairly challenging to estimate, the last one is even more challenging. in their book related to cost of capital, ogier et al. ( ) provided an excellent example which i would like to use to provide a practical introduction to this chapter. i take the liberty to modify the story in accordance with the needs of this chapter. imagine yourself at the edge of a river where your goal is to pass the river getting minimally wet in the least possible time. before making your move you need to turn to a local inhabitant who knows which stepping stones are safe, what the velocity and the viscosity of the water are, what the turning moments are, and what the probability of loose stones on the stream bed is. this situation is similar to the world of today's business investments. that is, executives need to make informed decisions about their investments and find out the minimum acceptable rate of return their shareholders expect as a compensation for the risks investors undertake. in addition, when an investment consists of both debt and equity, then the executives need to estimate the total cost of capital employed in this project to be able to pay their debt holders. this chapter intends to serve as a field guide or handbook of the cost of capital estimation for hospitality executives and practitioners. however, before getting into the practical aspects of cost of capital, some relevant concepts will be discussed from a theoretical perspective to better understand the background of this important topic. prior to getting into the core of the subject of estimating cost of capital, it is useful to define what risk is and describe the role it plays in investment decisions. in the hospitality field, risk is often defined as the variation in returns (probable outcomes) over the life of an investment project ( choi, ; olsen et al. , ) . the concept of risk is at the foundation of every firm as it seeks to compete in its business environment. financial theory states that shareholders face two types of risk: systematic and unsystematic. the examples of systematic risk could be changes in monetary and fiscal policies, the cost of energy, tax laws, and the demographics of the marketplace. finance scholars refer to the variability of a firm's stock returns that moves in unison with these macroeconomic influences as systematic, or stockholder, risk ( lubatkin and chatterjee, ) . stated differently, the level of a firm's systematic risk is determined by the degree of uncertainty associated with general economic forces and the responsiveness, or sensitivity, of a firm's returns to those forces ( helfat and teece, ) . in other words, these types of risk are external to the company and are outside of its control. however, a loss of a major customer as a result of its bankruptcy represents one source of unsystematic, or firmspecific risk (idiosyncratic or stakeholder risk). other sources of unsystematic risk include the death of a high-ranking executive, a fire at a production facility, and the sudden obsolescence of a critical product technology ( lubatkin and chatterjee, ) . unsystematic risk is a type of risk that can be eliminated by an individual investor by investing his/her funds in multiple companies ' stocks. the same rule may not be applied by company executives, since the success of a single project determines their tenure within their firms. the traditional financial theory looks at investment in securities from a portfolio perspective by assuming that investors are risk-averse and can eliminate the unsystematic risks (variance) associated with investing in any particular firm by holding a diversified portfolio of stocks ( markowitz, ( markowitz, , . markowitz pioneered the application of decision theory to investments by contending that portfolio optimization is characterized by a trade-off of the reward (expected return) of that individual security against portfolio risk. since the key aspect to that theory is the notion that a security's risk is the contribution to portfolio risk, rather than its own risk, it presumes that the only risks that matter to investors are those that are systematically associated with market-wide variance in returns ( lubatkin and schulze, ; rosenberg, ) . investors, it argues, should only be concerned about the impact that an alternative investment might have on the risk-return properties of their portfolio. however, the capital asset pricing model (capm) ( lintner, ; sharpe, ) (to be discussed in detail later) does not explicitly explain what criteria investors should use to select the alternative investments and how they should assess the risk features of these investments. moreover, the capm assumes that because investors can eliminate the risks they do not wish to bear, at relatively low costs to them, through diversification and other financial strategies, there is little need, therefore, for managers to engage in risk-management activities ( lubatkin and schulze, ) . in contrast, the field of strategic management is based on the premise that to gain competitive advantage, firms must make strategic, or hard-to-reverse, investments in competitive methods (portfolios of products and services) that create value for their shareholders, employees, and customers in ways that rivals will have difficulty imitating ( olsen et al ., ) . these investments enable the firms to protect their earnings from competitive pressure, and allow firms to increase the level of their future cash flow, while simultaneously reducing the uncertainty associated with them. the management of firmspecific risk lies at the heart of strategic management theories ( bettis, ; lubatkin and schulze, ) , and, from this perspective, management must work hard at avoiding investments that create additional levels of risk for the firm. bettis ( ) further affirms that the capm's emphasis on the equilibration of returns across firms (i.e., systematic risk) relegates to a secondary role strategy's central concern with managerial actions that seek to delay the calibration of returns (i.e., unsystematic risks). thus, the claim that systematic risk is paramount to the firm is undermined by the two arguable assumptions from portfolio theory: stockholders are fully diversified, and the capital markets operate without such imperfections as transaction costs and taxes. some stockholders, however, are not fully diversified, particularly the corporate managers, who have heavily invested, both financially and personally, in a single company ( vancil, ) . also, transaction costs, such as brokerage fees, act as a minor impediment, inhibiting other stockholders from completely eliminating unsystematic risk ( constantinides, ) . finally, taxes make all stockholders somewhat concerned with unsystematic risk (amit and wernerfelt, ; hayn, ) because interest on debt financing is tax deductible, thereby allowing firms to pass a portion of the cost of capital from their stockholders to the government. thus, firms can create value for their stockholders, within limits, by financing investments with debt rather than equity (kaplan, ; smith, ) . the limits are determined in part by the amount a firm is allowed to borrow and the terms of such debt, both of which are contingent upon the unsystematic variation in the firm's income streams. lubatkin and chatterjee ( ) contend that the debt markets favour firms with low unsystematic risk because they are less likely to default on their loans (this is particularly the case of the hospitality industry firms). in summary, the discussion of partially diversified stockholders, transaction costs, and leverage suggests that some stockholders may be concerned with unsystematic risk and factor it along with market risk to determine the value of a firm's stock (amit and wernerfelt, ; aron, ; lubatkin and schulze, ; marshall et al. , ) . cost of capital is defined as the rate of return a firm must earn on its investment projects in order to maintain its market value and continue attracting needed funds for its operations ( fields and kwansa, ; gitman, ) . consequently, a firm adds shareholder wealth when it undertakes the projects that generate a return higher than the cost of capital of the project. cost of capital is an anchor in firm valuation, project valuation, and capital investment decisions. cost of capital is generally referred to as weighted average cost of capital (wacc): where e is the market value of equity, d the market value of debt (and thus v ϭ e ϩ d ), t c the corporate tax rate, r e the cost of equity, and r d the cost of debt ( copeland et al. , ) . both of these items ( r d and r e ) are difficult to estimate and require some careful deliberations. the cost of debt is relatively simpler to calculate when a hypothetical firm issues bonds that are rated by the major bond-rating agencies such as standard & poor's and moody's. thus, these ratings may be used as a guide in computing the cost of debt. in addition, an investor may use the bond's yield to maturity or the rate of return that is in congruence with the rating of a bond. averaging the interest rates of long-term obligations of a firm is another method to calculate the cost of debt. the cost of debt estimation becomes difficult when a given firm has no bonds and no outstanding long-term debt. the cost of equity is difficult to estimate in its own right. first, cost of equity is generally estimated using historical data, which may be confounded by business cycles and abnormal • • • events affecting firm stock returns (e.g., fire in a hotel property) and industry returns (e.g., the terrorism events of september ). second, although several methods were developed in the last years, there is not one single method that produces consistent and reliable estimates. last, a hypothetical executive/ entrepreneur will face greater challenges as he/she needs to estimate the required rate of a single restaurant/hotel unit. the next section covers some of the common methods that are used by practitioners in the fields of financial and strategic management. cost of equity can be defined as the rate of return a firm must deliver to its shareholders who have foregone other investment opportunities and elected to invest in this particular company. however, cost of equity is a complex concept because firms do not promise paying a certain level of dividends and delivering a certain level of stock returns. thus, since there is no contractual agreement between the shareholders and the firm, the expected rate of return on invested equity is extremely challenging to estimate. fortunately, there are some models that can help us in tackling this challenging task. the next section will cover the major cost of equity models that gained prominence among practitioners and researchers in the last four decades. one of the early forward-looking methodologies is the dividend growth model (dgm) originally developed by gordon ( ) . it offers a very parsimonious method for estimating discount rate and thus accounts for risk. the dividend growth approach to cost of equity states that where, k e is the cost of common equity, dps the projected dividend per share, p the current market price per share, and g the projected dividend growth rate. the model assumes that over time, successful reinvestment of the value received by retained earnings will lead to growth and growing dividends. the approach suffers from oversimplification because firms vary greatly in their rate of dividend payout ( helfert, ) . this is due to the fact that common stockholders are the residual owners of all earnings not reserved for other obligations, and dividends paid are usually only a portion of the earnings accruing to common shares. the other major difficulty in applying this model lies in determining the specific dividend growth rate, which is based on future performance tempered by past experience. another key issue is that the model becomes unusable when a firm is not a dividend payer. the capm ( lintner, ; sharpe, ) is based on the assumption of a positive risk-return trade-off and asserts that the expected return of an asset is determined by three variables: β (a function of the stock's responsiveness to the overall movements in the market), the risk-free rate of return, and the expected market return ( fama and french, ) . the model assumes that investors are risk-averse and, when choosing among portfolios, they are only concerned about the mean and variance of their one-period investment return. this argument is, in essence, the cornerstone of the capm. the model can be stated as where, r m is the market return of stocks and securities, r f the risk-free rate, β the coefficient that measures the covariance of the risky asset with the market portfolio, and e ( r i ) the expected return of i stock. although the capm is touted for its relatively simple application, several other studies ( lakonishok and shapiro, ; reinganum, ) present evidence that the positive relationship between β and returns could not be demonstrated for the period of . particularly over the last two decades, even stronger evidence has been developed against the capm by fama and french ( , , and roll and ross ( ) . these researchers challenged the model by contending that it is difficult to find the right proxy for the market portfolio and that capm does not appear to accurately reflect the firm size in the cost of equity calculation, and that not all systematic risk factors are reflected in returns of the market portfolio. from the strategic management perspective, business executives face the following issues. implicit to the capm is the recommendation that managers should focus on managing their firm's overall market risk by focusing on β or the firm's • • • systematic risk and not be concerned with what strategists may focus on: firm-specific (unsystematic) risk. chatterjee et al. ( ) claim that herein lie two dilemmas: first, decreasing β requires managers to reduce investors ' exposure to macroeconomic uncertainties at a cost lower than what investors could transact on their own by diversifying their own portfolio; and second, to downplay the importance of firm-specific risk that not only is contrary to the strategic management field but also tempts corporate bankruptcy ( bettis, ) . therefore, an executive of a given company has to take into account the total risk of the project because, unlike investors holding stocks of multiple companies, the executive may not be able to diversify the risk of his/her company's investment by investing in multiple projects. another prominent cost of equity model is the arbitrage pricing theory (apt) developed by ross ( ) . the model states that actors other than β affect the systematic risk. the apt is based on the assumption that there are some major macroeconomic factors that influence security returns. the apt states that no matter how thoroughly investors diversify, they cannot avoid these factors. thus, investors will " price " these factors precisely because they are sources of risk that cannot be diversified away. that is, they will demand compensation in terms of expected return for holding securities exposed to these risks ( goetzmann, ) . although the model does not explicitly specify the risk factors, the apt depicts a world with many possible sources of risk and uncertainty, instead of seeking for equilibrium in which all investors hold the same portfolio. more formally, the apt is based on the assumption that there are some major macroeconomic factors that influence security returns. the apt states that no matter how thoroughly investors diversify, they cannot avoid these factors. thus, investors will " price " these factors precisely because they are the sources of risk that cannot be diversified away. that is, they will demand compensation in terms of expected return for holding securities exposed to these risks. just like the capm, this exposure is measured by a factor β ( goetzmann, ) . chen et al. ( ) managed to identify five macroeconomic factors that, in their view, explain the expected asset returns: the industrial production index, which is a measure of state of the economy based on the actual physical output; the shortterm interest rate, measured by the difference between the yield on treasury bills (tb) and the consumer price index (cpi); short-term inflation, measured by unexpected changes in cpi; long-term inflation, measured as the difference between the yield to maturity on long-and short-term u.s. government bonds; and default risk, measured by the difference between the yield to maturity on aaa-and baa-rated long-term corporate bonds (chen et al ., ; copeland et al. , ) . the apt describes a world in which investors behave intelligently by diversifying, but they may choose their own systematic profile of risk and return by selecting a portfolio with its own peculiar array of β s. the apt allows a world where occasional mispricings occur. investors constantly seek information about these mispricings and exploit them as they find them. in other words, the apt somewhat realistically reflects the world in which we live ( goetzmann, ) . although the apt provides the benefits explained above, these benefits come with some drawbacks. the apt demands that investors perceive the risk sources, and that they reasonably estimate factor sensitivities. in fact, even professionals and academics are yet to agree on the identity of the risk factors, and the more β s they have to estimate, the more statistical noise they have to put up with. last, this model does not offer much guidance to business executives as it focuses primarily on investors. one of the major proponents of the capm fama and french ( ) found that the relationship between average returns and β was flat and there was a strong size effect on stock returns. as a result, they developed a model that has gained popularity in recent years among the scholars and practitioners in the hospitality industry. the fama-french (ff) model is a multifactor model that argues that factors other than the movement of the market and the risk-free rate impact security prices. the ff is a multiple regression model that incorporates both size and financial distress in the regression equation. the ff model is typically stated as where β is the coefficient that measures the covariance of the risky asset with the market portfolio, r m the market return, r f , the risk-free rate, s the slope coefficient, and small minus big (smb) the difference between the returns on portfolios of small and big company stocks (below or above the nyse median), h the slope coefficient, and high minus low (hml) the difference between the returns on portfolios of high-and low-be/me (book equity/market equity) stocks (above and below the . and . fractiles of be/me) ( fama and french, ) . the size factor is denoted as smb premium where size is measured by market capitalization. smb is the average return on three small portfolios minus the average return on three big portfolios as described by fama and french ( ) . hml is the average return on two value portfolios minus the average return on two growth portfolios ( fama and french, ) . high be/me (value) stocks are associated with distress that produces persistently low earnings on book equity which result in low stock prices. in practice, the ff model shows that investors holding stocks of small capitalization companies and firms with high bookto-market value ratios ( annin, ) need to be compensated for the additional risk they are bearing. the size argument is supported by barad ( ) who reports that small stocks have outperformed their larger counterparts by an average of . % over the last years . however, fama and french ( ) find that the book-to-market factor (hml) produces an average premium of . % per month ( t ϭ . ) for the - period, which, in the authors ' view, is large both in practical and statistical terms. the starting point for selecting the best method for the estimation of the cost of equity can be achieved by reviewing the relevant studies undertaken in the fields of hospitality and tourism. fields and kwansa ( ) conducted the first study that directly looked into the cost of equity and suggested the use of pureplay technique for estimating the cost of equity for the divisions of a diversified firm. later, several studies investigated how macroeconomic variables affect security returns in the hospitality industry (hotels and restaurants). the first study was conducted by barrows and naka ( ) . their study encompassed the -year period between and and employed five factors that were slightly different than the five factors of chen et al . ( ) . barrows and naka postulated that the return of the stocks is a function of the following five factors: where einf is the expected inflation, m the money supply, conn the domestic consumption, term the term structure of interest rates, and ip the industrial production. the results revealed that none of the macroeconomic factors was significant in explaining the variance of u.s. hotel stocks at . level and the factors accounted for the . % of the variance in the lodging stocks. however, einf, m , and conn had significant effect on the variation of the stock returns in the u.s. restaurant industry. in terms of the signs of the β coefficients einf had a negative whereas m and conn had a positive relationship with the restaurant stock returns . the postulated model explained % of the variance in the restaurant stocks. the authors cautioned that the results should be interpreted with care due to the small sample size of both restaurant and hotel portfolios, which were represented by five and three stocks, respectively. the second study was undertaken by chen et al. ( ) who used hotel stocks listed on taiwan stock exchange. the macroeconomic variables included in their study were ip, cpi, unemployment rate (uep), money supply (m ), -year government bond yield (lgb), and -month tb rate. these variables were used in the following way: cpi was utilized to estimate einf, and lgb, and tb were used for the computation of the yield spread (spd). based on the six time-series data the authors arrived at the common five macroeconomic variables which were predominantly used in the literature, namely, ip (change in ip), einf uep (change in unemployment rate), m (change in money supply), and spd (rate of the yield spread). these five variables explained merely % of the variation in hotel stock returns while only two of these variables were significant at the . level (m and uep). the regression coefficient of change in money supply had a positive relationship with hotel stock returns, whereas the relationship between change in uep and lodging returns was negative. in madanoglu and olsen ( ) proposed a conceptual framework that called for the inclusion of some of the intangible variables into the cost of equity estimation in the lodging industry. some of these variables were human capital, brand, technology, and safety and security. it is common knowledge that these variables were relevant for the lodging industry; however, there exists no time-series data to include them in the cost of equity estimations. publicly traded multinational lodging companies tend to differ on some key points regarding how assets are treated on their balance sheets. many of these companies do not actually own assets and produce their future cash flows from management contracts or franchise agreements. in many cases, they may also lease hotels or restaurants and the leases do not appear on their balance sheets. instead, these firms hold an equity position in a different company that holds these leases. therefore, it is almost unfeasible to properly assess the book value of the hospitality firms, which confounds the application of the ff model. sheel ( ) was the first researcher in the hospitality industry to point out that capm does not seem to meet the industry needs and called for further research into industry-specific factors. in the mainstream financial economics, downe ( ) argued that in a world of increasing returns, risk cannot be considered a function of only systematic factors, and thus β . he pointed out that the position of the firm in the industry, as well as the nature of the industry itself become a risk factor. thus, firms with a dominant position in the industry that succeed to adapt to the complexities of the business environment, will have a different risk profile than their competitors. this argument is particularly well fitting in the context of the hospitality industry where companies such as mcdonald's and marriott may demonstrate a different risk profile based on their market share in their segments. as for ff factors, professionals in the lodging industry are sceptical about such measures as the book-to-market value ratio (hml). some hospitality industry experts argue that hml is an inappropriate measure for the industry and attribute it to the fact that the difference between the firms whose value is captured by the assets they own and the firms whose value is derived from their intangible assets is not as distinct as in some manufacturing firms. while jagannathan and wang's study ( ) added a human capital variable to their cost of equity capital model, it measured human capital effects from the macroeconomic perspective as opposed to a micro level where most hotel firms operate. in other words, the overall labour index may not properly reflect the state of the human capital in the hospitality industry. as fama and french ( ) stated, their work (ff model) leaves many open questions. the most important missing piece of the puzzle is that fama and french ( ) have not shown how the size and book-to-market factors in security returns are driven by the stochastic behaviour of firm earnings. this implies that it is not yet known how firm fundamentals such as profitability or growth produce common variation in returns associated with size and be/me factors and this variation is not captured by the market return itself. these authors further query whether specific fundamentals can be identified as state variables (variables that describe variation in the investment opportunity set) and these variables are independent of the market and carry a different premium than general market risk. this question is of utmost importance for lodging industry executives who are aiming to identify the major drivers of their companies ' stock returns in their effort to create value for their stockholders. in their current state, the cost of equity models are far from satisfying the needs of the hospitality industry. as fama and french ( ) pointed out, the cost of equity estimates yielded by these models are distressingly imprecise. standard errors of more than % per year were typical when the capm and ff models were used to estimate industry costs of equity in their study ( fama and french, ) . they stated that large standard errors are driven primarily by uncertainty about true factor risk premiums. since the hospitality industry is really the aggregate of individual units that all have their own unique business environments and return on equity structures, this means that the standard errors, and thus, cost of equity capital on a per-company, single-unit (a hotel property or a restaurant) basis, or for a new project will be even more imprecise. therefore, the risk determinants of cost of equity and risk factor loadings for individual operating units will be even more difficult to estimate. thus, it is very important to consider the purpose for which the cost of equity is estimated (e.g., a single project, business division, or an entire corporation). particularly, in the case of single project cost of equity estimations there might be several factors that need to be considered before arriving at the proper discount rate of the project. these factors might be location of the project, local/regional competition, political risk, credit risk, and other risk idiosyncratic to a given project. consequently, as ogier et al. ( ) suggest when estimating a cost of equity for a given project the risk of the project will be much more important than the risk level of the corporation making the investment. in other words, when marriott corporation makes a capital investment decision in nairobi, kenya, the marriott corporation executives will be much more concerned with the risks surrounding that project. unlike cost of equity, cost of debt does not require the use of sophisticated theoretical models. rather, cost of debt is simply the rate at which a given company can borrow capital from a lender (e.g., bank) or the rate at which the aforementioned company can issue bonds. some experts caution that the • • • promised and the expected yields of debt are two different concepts. in other words, when a firm makes contracted debt payments on time it meets " the promised yield " to its lender. however, in reality, there is always a possibility for default and thus the difference between the promised yield and the probability for default equals the expected yield. the expected yield can be regarded as true cost of debt since it is more realistic. although many textbooks calculate the cost of debt as promised yield, it should be noted that expected yield is more meaningful since it includes not only the systematic risk of the market but also the firm-specific risk of a given firm. another challenge for calculating the cost of debt might occur when a firm uses multiple debt instruments (e.g., bank loans, commercial papers, bonds). in this case, it may be fruitful to average the rate of these instruments based on their weight in the debt portfolio. however, an easier and more simplistic approach would be to use the " generic long-term debt " rate which can be calculated from the current rate of a company's bond or current rate at which the company can borrow a longterm loan ( ogier et al ., ) . last, to estimate the cost of debt, the issue of tax shield should be given a close consideration. for instance, although the majority of the finance textbooks use or % as an average for corporate tax rate in the united states, it is common occurrence to observe companies whose effective corporate tax rate is often lower than the statutory rate. here, an executive should assess the situation and decide whether the effective tax rate trend is expected to continue to be below the statutory corporate tax rate in the long term. if that is the case, then he/she should use the effective tax rate in calculating the cost of debt. however, if a low effective tax rate is a short-term occurrence, then a given firm should use the statutory corporate tax rate instead ( ogier et al ., ) . hospitality industry is part of the overall service sector and is dependent on human capital in order to maintain and grow its operations. in an increasingly competitive environment, the human factor becomes one of the keys in creating sustainable competitive advantage. therefore, murphy ( ) stated that the hospitality industry should learn to view its employees from a new paradigm that human capital is a strategic intangible asset (knowledge, experience, skills, etc.). this implies that, like other assets, it is an important determinant of firm value. however, studies have concluded that " the research of human resources expenditures " is in its infancy and is seriously hampered by the absence of publicly disclosed corporate data on human resources ( lev, ) . caroll and sikich ( ) argued that keeping track of at least a -year history of labour costs would serve to identify the dollar value of " premium " labour-related costs, which could be thought of as all labour/benefit costs above federally mandated minimum wage. other techniques proposed by the authors were ( ) to design a scoring system that illustrates productivity versus both baseline and premium labour/benefit costs by departments, and ( ) to establish metrics to determine a productivity level for guest experience standards, facilities standards, and targeted revenue improvements on a department-by-department basis. bloxham ( ) advocated adjustments to certain human resource expenditures to capitalize them over the time of the investment. in that approach, one-time human resources costs are amortized and capitalized in the value creation equation in an effort to demonstrate that human capital investments go beyond being a cost item in the firms ' operations. these costs can include recruiting, interviewing, and hiring costs; one-time hiring bonuses and relocation expenses; and training costs. the costs are capitalized and amortized over the average employee tenure with the company. in this case, if employee turnover is high, these costs would be amortized over a shorter time period (thus the costs will be higher), whereas the longer tenure of the workforce will enable the firm to spread the costs over a longer period of time. kalafut and low ( ) reported that in a study of the airline industry conducted by cap gemini ernst & young's center for business innovation (cbi), the employee category was the single greatest value driver that had an impact on the firm's market value. the employee factor had a positive correlation of . with the firm value. thus, kalafut and low ( ) conclude that in the aggregate, quality and the talent of the workforce, quality of labour management relations, and diversity are critically important in the value creation process of the airline companies. the arguments above can be justified on the grounds that higher-quality human resources decrease labour turnover and increase employee productivity. this results in better organizational performance that results in stabilization of cash flows which in turn decreases the uncertainty of firms ' stock returns. therefore, one would expect that hospitality firms that have institutionalized quality human resource management practices would achieve a more realistic cost of equity estimates that reflect the lower risk associated with these practices. although definitions of the concept of brand differ across the professional and trade literature, the underlying notion is that of a distinctive name with which the customer has a higher level of awareness and a willingness to pay a higher-thanotherwise average price or make a higher-than-otherwise purchase frequency ( barth et al ., ) . a brand is the product or service of a particular supplier which is differentiated by its name and perceived expectations on the part of the consumer. brands are important and valuable because they provide a " certainty " as to future cash flows ( murphy, ) . however, since the task of estimating brand value is yet an improbable one, its value is not specifically reflected on the company's balance sheet. yet, the lodging industry has made much of the importance of the value of the brand but has not been able to unequivocally substantiate the role of the brand in reducing the variance in firm cash flows, and thus contributing to lower cost of capital for the firm. srivastava et al. ( ) provided an analytical example of how successful market-based assets (the term authors use in lieu of intangibles) lower costs by building superior relationships with customers, enable firms to attain price premiums, and generate competitive barriers (via customer loyalty and switching costs). all these factors lead to the conclusion that a strong brand reduces the uncertainty pertaining to the future cash flows which in turn decreases the required return by the investors for the risk they bear by investing in a particular firm. in attempts to value the brand in the manufacturing industries, the use of the following methods has been cited by murphy ( ) : • valuation based on the aggregate cost of all marketing, advertising, and research and development expenditures devoted to the brand over a stipulated period. • valuation based on premium pricing of a branded product over a non-branded product. • valuation at market value. • valuation based on various consumer-related factors such as esteem, recognition, or awareness. • valuation based on future earning potential discounted to present-day value. in further analysis, the investigators rejected these methods because, if indeed, brand values were the function of its cost of development, then failed brands would be attributed high values. in addition, brand valuation based solely on the consumer esteem or awareness factor would bear no relationship to commercial reality ( murphy, ) . in an effort to link the firm's security returns with brand value, simon and sullivan ( ) proposed a technique to estimate the firm's brand equity based on its value. this was done by estimating the cost of tangible assets and then subtracting it from the market capitalization of the firm to obtain the value of intangible assets. as a second step, the researchers tried to break down the intangible assets into brand value and nonbrand value components. the authors utilized the aaker and jacobson ( ) equitrend brand quality measure to evaluate the quality of major brands. they examined associations between measures of brand quality and stock returns and reported that the relationship is positive. according to murphy ( ) , the only logical and consistent way to develop a multiple for brand profit was through the brand strength concept. brand strength is a composite of six weighted factors: leadership, stability, market, trend, support, and protection. the brand is scored on each of these factors according to different weightings and the resultant total known as " brand strength score. " a further addition to the brand strength concept came from prasad and dev ( ) who developed a hypothetical brand equity index via customer ratings of the brand using five key brand attributes in two sets of indicators-brand performance and brand awareness. brand performance was measured by overall satisfaction with the product or service, return intent, price-value perception, and brand preference, while brand awareness was measured as top-of-mind brand recall. olsen ( ) proposed brand-related value drivers specific to the lodging industry such as brand dilution and brand sincerity ratio. brand dilution is related to the question of how many new corporate sub-brands must be introduced in order to maintain growth, whereas, brand duration deals with what percentage of hotels in the portfolio currently meet the brand standards or promise. as a result, it is argued that hospitality companies that possess higher-brand strength will be able to achieve a lower cost of equity capital. according to connolly ( ) , one of the greatest issues plaguing the advancement of technology in the hospitality industry is the difficulty of calculating return on investment. until recently, most technology investment decisions have been considered using a support or utility mentality that stems from a manufacturing paradigm. current policies rely more on faith than on a rational business assessment. as a result, the hotel industry is perceived to be lagging behind the rival industries in the use of technology ( sangster, ) . in part, this is attributed to the fragmented nature of the hotel business itself; however, it is also believed to be closely related to hoteliers ' lack of experience and understanding in technology investments ( sangster, ) . connolly further argued that " today's financial models are inadequate for estimating the financial benefits for most of the technology projects under consideration. while the hospitality industry has disciplined models and sufficient history to determine the financial gains or success of opening a new property in a given city, it lacks the same rigorous models and historical data for technology, especially since each technology projects are unique. although this problem is not specific to the hospitality industry, it is particularly problematic since the industry tends to be technologically conservative and unwilling to adopt new technology applications based on the promises of their long-term merits especially if it cannot quantify the results and calculate a defined payback period. when uncertainty surrounds the investment, when the timing of the cash flows is unpredictable, and when the investment is perceived as risky, owners and investors will most likely channel their investment capital to projects with more certain returns and minimal risk. thus, under this thinking, technology will always take a back seat to other organizational priorities and initiatives. efforts must be made to change this thinking and to develop financial models that can accurately predict and capture the financial benefits derived from technology ( connolly, ; p. iii) . " although there are no hard and fast rules to facilitate the valuations of technology investments, it is common knowledge that technology is transforming the way business is conducted in the lodging industry. particularly the surge in internet usage in the early years of the new millennium brought about the issue of capacity control for hotel room inventory holders. therefore, firms that are more adaptive to utilize technology to market and sell their perishable product (hotel rooms) may accomplish a lower variation in their future cash flows, since they are able to retain greater control over pricing. the author would like to acknowledge the fact that the body of literature does not offer a direct causal relationship between the cost of equity capital and the technology utilization. however, based on the arguments discussed above, the author contends that firms that invest in technology wisely may achieve a higher average daily rate or revpar in their properties which in turn will lead to a decrease in the variance in firm's cash flows. thus, better utilization of information technology can possibly reduce the uncertainty surrounding the future earnings of the firm. as a result, capital markets will assign a lower risk premium to hospitality firms that successfully utilize and deploy technology into their operations. guest safety and security topics in the lodging industry can vary from building safety codes and bacterial contamination of hotel whirlpools to restaurant food safety and hotel crime statistics ( olsen and merna, ) . the need for greater commitment to safety and security for the hospitality industry became evident in after the san francisco earthquake and hurricane hugo occurred ( olsen and merna, ) . the culmination of these events and all the other events sparked an effort by the hotel industry to manage the risk and liability related to guest safety and security. ray ellis, the director of risk management and operations in the american hotel & motel association (at that time in ), contended that after the end of the gulf war the benefits of increased security for the industry go far beyond intangibles such as peace of mind ( jesitus, ) . ellis stressed that improved safety and security will significantly decrease the insurance premiums of the properties, and thus enable the companies to have more resources to invest in their operations. although ellis said that chances of terrorist attacks on the united states post gulf war were fairly remote, he warned that the hotels, particularly those serving international markets, be most wary of arson and bomb threats. the international hotel and restaurant association in identified safety and security as one of the major forces driving change in the global hospitality industry ( olsen, ) . with the destruction of the world trade center in , and subsequent terrorist attacks in bali and kenya, it is clear that force has emerged now as a major risk factor for all tourismrelated enterprises. in february , the federal bureau of investigation (fbi) alerted its law enforcement partners that " soft targets, " such as hotels, can be subject to terrorist attacks ( arena et al. , ) . this report simply reaffirms the argument proposed by olsen ( olsen ( , that lodging properties which are situated in an area exposed to terrorist attacks, should factor that risk into their cost of capital estimates. therefore, lodging property executives should apply this risk factor into their future capital investment decisions. in addition, outbreaks related to food-borne diseases, infectious bacteria occurrences on cruise ships, increased crime, and the growing threats of human immunodeficiency virus (hiv), and other viral infections such as severe acute respiratory syndrome (sars) have created a significant challenge for hospitality managers worldwide. these must be considered as important risk variables that will no doubt have an impact on the estimates of cost of capital. although the factors mentioned above are critical in estimating the cost of capital of a given project, there are no methods that can quantify these factors and apply them to the cost of equity models. however, executives are advised to consider these industryspecific risk factors before making a capital investment decision. the models covered thus far do not provide any guidance for estimating the cost of equity in a global setting or multinational projects. in order to fill this void, academics and practitioners developed adjustment models to account for differences in cost of equity among markets in developing and emerging countries. the adjustment models are primarily concerned with whether the emerging markets are segmented or integrated with the world markets. that is, in a completely segmented market, assets will be priced based on local market return. the local expected return is a product of the local β times and the local market risk premium (mrp) ( bekaert and harvey, ) . bekaert and harvey ( ) developed a modified model after researching emerging markets for the pre- and post- periods and reported that the correlation of the emerging markets with the morgan stanley capital international (msci) world index increased noticeably. for instance, turkey is one of the countries whose market correlation with msci world index increased from less than . to more than . . based on this, turkey may be considered an integrated capital market where the expected return is determined by the β with respect to the world market portfolio multiplied by the world risk premium. this is the core argument of the bekaert-harvey mixture model ( bekaert and harvey, ) . in cases when integrated markets assumption does not apply, investment banks and business advisory firms use a method called " the sovereign spread model (goldman model). " this is conducted by regressing an individual stock against the standard & poor's stock price index returns to obtain the risk premium. then, an additional " factor " is added which is called the " sovereign spread " (ss). this spread between respective country's lgb for bonds denominated in u.s. dollars and the u.s. treasury bond yield is " added in. " the bond spread serves as a tool to increase an " unreasonably low " country risk premium ( harvey, ) . this section offers a practical example for managers to estimate the wacc of their projects. in addition, this section breaks down the wacc into its respective components in order to assist executives in the capital investment decisions. the major components of the wacc estimations are a firm's stock return, market return, risk-free rate, regression coefficients ( β , s , and h ), smb, hml and equity market risk premium (emrp) (which is r m Ϫ r f ), capital structure (proportion of debt and equity), corporate tax rate, and cost of borrowed debt. if you are an executive of a company that is not publicly traded, you have two options to estimate the cost of equity. you can either use the industry average for cost of equity or locate two or three comparable firms that compete in the same line of business and estimate their cost of equity. however, even if you are an executive of a large restaurant corporation that is traded publicly, it is still recommended that you estimate the cost of equity for the entire restaurant industry because the standard error of regression coefficients for a single firm is fairly high, which decreases the reliability of these coefficients. my past research experience has showed me that at times using a single firm may create a situation in which cost of equity cannot be even estimated. more often than not, i obtained distressing results when running a regression for small-or medium-size hospitality firms. as a result, in the practical example, i will estimate the restaurant industry's cost of equity. since the cost of equity calculation process may be a fairly complex process for someone who is not familiar with data analysis, i will offer a step-by-step procedure, which should better clarify this process: step : obtaining a -year monthly stock return for your company/industry and the market • • • ideally, you need years of monthly stock return data for your firm and the -year market return. the issue of selecting the best index of all traded assets in the world is a very challenging and sometimes a controversial issue. based on seminal • • • studies in financial management, the market index that yields most reliable results in the united states is center of research in security prices value weight (crspvw) index housed at the university of chicago. both your company's stock and market return should be used as excess return (i.e., return less risk-free rate which is -month tb rate) in order to measure the cost of equity in real units (i.e., after accounting for inflation). for reasons mentioned before, i will be estimating the u.s. restaurant industry's cost of equity and leave the decision to restaurant industry executives to adjust this value to their specific projects at hand. in order to be able to observe the accuracy of cost of equity models, we estimate the restaurant industry cost of equity by using the capm and ff model. the observation period of this example is between and . the reason for not selecting a longer observation period is that the values of β and other variables become unstable over extended periods. the sample is developed from the nation's restaurant news (nrn) index, which entails restaurant firms. in cases when executives are not familiar with building stock portfolios, they can alternatively use monthly returns of hospitality indices for lodging and restaurant industries from data providers such as yahoo! finance, wall street journal, or industry publications such as nrn. step : estimating β and fama-french factor coefficients • • • the capm's β can be computed by regressing excess stock return of a firm over the excess market return. the monthly returns for ff factors (smb and hml) can be retrieved from eventus database housed in the wharton school at the university of pennsylvania or from kenneth french's website at dartmouth college. by regressing monthly smb and hml returns on market returns you can obtain " s " and " h " coefficients that can later be inserted into the equation to estimate the cost of equity. in our practical example, the results indicate that the ff model explains more than half ( . %) of the variation in the returns of the nrn index. in addition, the ff model results in a significant r change over the capm, which showed that the two ff variables (smb and hml) explained some extra variance over and above the capm which accounted for . % of the variation in the restaurant industry stock returns. the analysis at the variable level indicates that the market index variable ( β ) and the hml are significant at . level (see table . ). however, the smb was not significant at the . level, which means that the size factor does not affect the restaurant industry stock returns while controlling for β and hml. in practice, this means that restaurant industry portfolio behaves as a large company stock, and therefore there is no size premium when considering the overall cost of equity for the restaurant industry. it should be remembered that if you are an executive of a small restaurant company there is a high possibility that your stock returns will have a size premium. step : the risk-free rate, market, size and distress premiums • • • there are certain rules of thumb that executives should be aware of before inserting the regression coefficients into the cost of equity calculation. first, it should be pointed out that there are two risk-free rates ( r f ) in the capm and ff models. the first r f is used in order to demonstrate the level of risk-free rate that a firm needs to exceed to compensate its investors for the risk they undertake. the second r f should ideally match the life of an asset. in other words, if the asset in this project is expected to last at least years, then a given investor/executive should use a -year government bond as its risk-free rate to obtain the mrp ( r m Ϫ r f ). another important issue is calculating market, size and distress premiums. executives/investors may often face challenges when the -year mrp (which equals r m Ϫ r f ) is negative or extremely low, or when size premium (smb) and distress premium (hml) figures are negative. in these cases, i would recommend that executives/investors use the longterm equity premium ( r m Ϫ r f ) figure of % ( siegel, ) , - , - , - , and so on) until and verified that in all instances smb, and hml premiums were positive. step : solving cost of equity equation • • • since the market index (vwcrsp) has a very low return ( . %) for the -year period, i will use the long-term equity premium of % ( siegel, ) . next, by using the obtained regression coefficients in table . , the regression equations provide the following results: as it can be seen from the results above, the restaurant industry cost of equity is considerably higher when estimated by using the ff model. in basic terms, this means that a hypothetical investor will expect a return of % from the u.s. restaurant industry in order to invest his/her funds in the u.s. restaurant portfolio. however, if a restaurant executive believes that % is a fairly high rate of return and his/her restaurant company does not have the same risk profile as the overall u.s. restaurant industry, he/she may elect to use the average of the capm and ff estimates, which is around %. next, a restaurant executive may adjust the rate of his/ her firm's project by considering whether the project will be riskier than the restaurant industry's expected return. here one should consider factors such as competition, life of the project, and the events that may have an impact on the risk of the project by influencing forces driving change in firm's external (e.g., economic, political, technological) and internal (e.g., industry, local) environment. the next step in estimating the cost of capital is to estimate the cost of debt. unlike cost of equity, cost of debt does not require consideration of the average cost of debt for the hospitality industry. this is because in simple terms, cost of debt denotes an interest rate at which a given company can borrow. therefore, a given company can calculate the cost of debt for a given project in a relatively simple manner. the situation is little more complex in cases when a corporation has multiple projects to invest in and has to estimate its corporate cost of debt. this is because some of the projects may be expansion projects that are already financed by loans obtained in the past. consequently, executives need to average out the interest rate of the outstanding debt related to this project and also consider the interest rate at which the company can borrow new funds. in this particular example, we will assume that a hypothetical company plans to issue bonds which mature in years and will also secure a -year loan to finance a portion of the project. in this scenario, we assume that both the bond issuance and the loan will have equal contribution to the funding of the project (e.g., % each). let us assume that the hypothetical company in this example issues -year bonds whose expected yield-to-maturity is %. this rate is assumed based on the present bond rating of this company. we also assume that the rate of a -year bank loan is % and the corporate tax rate %. thus, the cost of debt can be calculated as follows: before entering the values from previous sections we assume that the current project will be financed with % equity and % debt. we use the average cost of equity estimate ( . %) and the cost of debt ( . %) we obtained before. consequently, the weighted cost of capital for this project can be calculated as follows: . % it should be noted that the executive of this hypothetical firm needs to make adjustments to this project if the project carries any specific risk such as political risk, divisional risk (if the firm has multiple divisions), risk of early termination, stiff competition, and so on. this section considers a case when the cost of equity needs to be estimated for an international project. here i use a hypothetical scenario where a thai investor plans to make a hotel in line with the suggestions made by annin ( ) , and barad and mcdowell ( ) , a minimum of months ' stock market trading is the criterion for a hospitality firm to be included in the turkish tourism index. in addition, crspwv index is used as a market portfolio index for the united states. this is in congruence with the previous seminal studies related to asset pricing models ( fama and french, , jaganathan and wang, ) . however, imkb ulusal index is utilized as a market portfolio for turkey. β is computed by regressing excess return of the four seasons and turkish tourism index over the excess market return; therefore, both variables are analysed in real units (e.g., after subtracting inflation). excess market return (mrp) for the united states is computed by subtracting -month tb rate from the monthly vwcrsp index return. the mrp for turkey is calculated by subtracting the turkish government's tb from the monthly ise ulusal index return. the data for the five apt variables are obtained from global insight database. the apt variables are calculated as in chen et al. ( ) . einf is estimated following the method of fama and gibbons ( ) . country risk premium is adapted from aswath damodaran at new york university. damodaran ( ) explains the estimation procedure as " to estimate the long term country risk premium, i start with the country rating (from moody's: www.moodys.com ) and estimate the default spread for that rating (us corporate and country bonds) over the treasury bond rate. this becomes a measure of the added country risk premium for that country. i add this default spread to the historical risk premium for a mature equity market (estimated from us historical data) to estimate the total risk premium. " both direct and indirect approaches are used to estimate the expected return (indirect and direct) of an investment. in this method, i first compute the expected rate of return for the u.s. stock (in this case four seasons) by using the average estimates for the capm and apt. then i adjust for country risks of turkey and thailand based on moody's country risk ratings as reported by damodaran ( ) . this method assumes that the turkish stock market is integrated and thus using the u.s. market indices to estimate the cost of equity for four seasons is equivalent to using ulusal market index for the turkish tourism portfolio. first, i run a regression of the monthly returns of four seasons over the crspvw return for the - period. the results show that the β for four seasons is . . next, the -year annualized return for the crsp was calculated in order to estimate the mrp. the -year historical return for crsp was . %. the riskfree rate for the - period was . %. as a result, the cost of equity estimate based on the capm for four seasons is as follows: . % ϭ ϩ ϫ Ϫ ϭ in an effort to have less biased estimates, i also use the five apt variables (chen et al. , ) to calculate the expected return for four seasons. the results reveal that, among the five apt variables, only the default risk variable (upr) is significant at the . level. however, it is not feasible to use this variable to estimate the expected return because the regression coefficient for upr is a negative number. as a result, the four seasons is likely to have a negative expected return based on the apt. as a consequence, i elect not to use the apt results in the final stage of the direct approach, since the results of the apt are in conflict with the contemporary financial theories. therefore, i use the capm's estimate of . % and adjust this estimate with the country risk of turkey and thailand. according to damodaran ( ) , the historical risk premium for the united states is . %. turkey's country risk premium is . % above the united states value and that for thailand is . % above the risk premium for the united states. this denotes that turkey's country risk premium is . % over that of thailand. these figures result in an expected return of . % ( . ϩ . %) for the thai entrepreneur who is undertaking an equity investment in a hotel in turkey. in the direct approach, i estimate the nominal required rate of return for the portfolio of turkish tourism and hospitality stocks. as a next step, i adjust for the sovereign spreads of turkey and thailand as it is assumed that the thai investor will repatriate the returns from an investment to his/her home country. in this method, i regress the monthly return of the turkish tourism index over the return of the ise. the β for the tourism index was merely . . the -year average for the risk-free rate (turkish government's tb) for the - period was . %. the annualized return of the market index (ise) for the - period was . %. the expected return for the tourism portfolio was calculated by applying the capm and it provided the following results: . % ϭ ϩ ϫ Ϫ ϭ ϩ ϭ the next step entails the addition of the sovereign spread between thailand and turkey to arrive at the estimate for the cost of equity capital for the thai investor. the sovereign spreads are obtained from fuentes and godoy ( ) . the spread for turkey was . % and that of thailand . %. based on these figures, the cost of equity for the direct approach was . % ( . ϩ . %). as it can be seen from both the examples of cost of equity estimation (the united states and international), the expected returns (costs of equity) varied widely. in the example of united states, the use of the capm resulted in a cost of equity that was fairly low (less than %). it is worth asking, would a given investor invest in a u.s. restaurant portfolio of stocks for less than % a year? the answer would probably be " no. " however, if one elects to use ff as its main cost of equity model then the possibility of obtaining more relevant results is likely to increase. as it can be seen in this example, the cost of equity by using the ff model yielded a fairly logical return which far exceeds the historical equity premium for the united states. for the international example, one of the main reasons for the stark difference in cost of equity estimates using the two approaches (direct and indirect) is the high historical inflation in turkey. this is demonstrated by the gap in the tb rates for this country ( . % for and . % for ). hence, if a hypothetical investor elects to use the " going-rate ( . %) in then the new expected return for the turkish tourism portfolio would be at least twice lower than the original estimate of . %. another challenge in the direct approach for international cost of equity estimations is the low β estimate for the turkish tourism portfolio ( . ). does this mean that the tourism portfolio is five times less risky than the overall ise index? what if the real risk of tourism stocks is twice higher than that of the market? (this is quite likely as the β for four seasons in the united states was . .) if that is the case, then the thai investor needs to require a rate of return that is more than % in thai currency. how can the investor hedge his investments against the large swings in the cost of equity estimates? as the results indicated thus far, cost of equity estimations for hospitality investments in emerging and developed markets are beset with uncertainty. the main shortcomings stem from the challenge of applying the seminal models such as the capm, ff, and the apt. the second set of challenges arises when countries such as turkey tend to have high historical rates of inflation but now are entering a more stabilized period of fiscal reforms. thus, should an investor use the historical data or try to forecast the future interest rates in turkey? although the practical examples provided some answers to these questions, few more questions are left for future research. hence, i suggest two interim solutions for this cost of equity conundrum in the emerging markets: ( ) the investors and academics should either solely focus on future cash flows of the project, or ( ) use simulations such as monte carlo in order to create multiple scenarios that approximate the investment realities of the emerging markets. otherwise, the expected return remains to be a " gut feeling " estimate for foreign investors in emerging markets. the financial information content of perceived quality why do firms reduce business risk? fama-french and small company cost of equity calculations preparations for possible attacks gear up: new flight restrictions planned around washington ability, moral hazard, firm size, and diversification technical analysis of the size premium. business valuation alert capturing industry risk in a buildup model use of macroeconomic variables to evaluate selected hospitality stock returns in the u brand values and capital market valuation research in emerging markets finance: looking into the future modern financial theory, corporate strategy and public policy: three conundrums economic value management: applications and techniques what is your irr on human capital? towards a strategic theory of risk premium: moving beyond capm the impact of macroeconomic and non-macroeconomic forces on hotel stock returns economic forces and the stock market the restaurant industry, business cycles, strategies, financial practices, economic indicators, and forecasting. unpublished dissertation understanding information technology investment decision making in the context of hotel global distribution systems: a multiple-case study . unpublished dissertation capital market equilibrium with transaction costs valuation: measuring and managing the value of companies country default spreads and risk premiums increasing returns: a theoretical explanation for the demise of beta the cross section of expected stock returns common risk factors in the returns on stocks and bonds size and book-to-market factors in earnings and returns industry costs of equity risk, return and equilibrium: empirical tests a comparison of inflation forecasts analysis of pure play technique in the hospitality industry sovereign spreads in emerging markets: a principal components analysis principles of managerial finance introduction to investment theory (hyper textbook). retrieved the investment, financing, and valuation of the corporation the theory and practice of corporate finance: evidence from the field twelve ways to calculate the international cost of capital tax attributes as determinants of shareholder gains in corporate acquisitions vertical integration and risk reduction techniques of financial analysis: a guide to value creation conditional capm and cross section of expected returns valuation in emerging markets march . safety and security: risk management, threat of terrorism top hoteliers ' concerns in the value creation index: quantifying intangible value the effects of management buyouts on operations and value systematic risk, total risk and size as determinants of stock market returns intangibles: management, measurement, and reporting the valuation of risk assets and the selection of risky investments in stock portfolios and capital budgets extending modern portfolio theory into the domain of corporate diversification: does it apply? risk, strategy, and finance: unifying two world views (editorial) . long range planning cost of equity conundrum in the lodging industry: a conceptual framework portfolio selection portfolio selection: efficient diversification of investments incentives for diversification and the structure of the conglomerate firm assessing the value of brands a proposed structure for obtaining human resource intangible value in restaurant organizations using economic value added the real cost of capital: a business field guide to better financial decisions into the new millennium: the iha white paper on the global hospitality industry: events shaping the future of the industry global hotel finance--the future . one-day program co-sponsored by hong kong shanghai bank corporation, deloitte & touche consulting group and richard ellis international property consultants leading hospitality into the age of excellence: competition and vision in the multinational hotel industry march . trends in safety & security . hotel and motel management strategic management in the hospitality industry managing brand equity: a customer-centric framework for assessing performance model estimates financial impact of guest satisfaction efforts cost of capital: estimation and applications a new empirical perspective on the capm on the cross-sectional relation between expected returns and betas the capital asset pricing model and the market model persuasive evidence of market inefficiency the arbitrage theory of capital asset pricing technology: the importance of technology in the hotel industry capital asset prices: a theory of market equilibrium under conditions of risk an empirical analysis of anomalies in the relationship between earnings ' yield and returns of common stocks: the case of lodging and hotel firms stocks for the long run the measurement and determinants of brand equity: a financial approach corporate ownership structure and performance: the case of management buyouts marketbased assets and shareholder value: a framework for analysis passing the baton: managing the process of ceo succession key: cord- -e mr kqb authors: adler, nicole; gellman, aaron title: strategies for managing risk in a changing aviation environment date: - - journal: journal of air transport management doi: . /j.jairtraman. . . sha: doc_id: cord_uid: e mr kqb abstract given the increasing volatility in the economic performance of airlines, partially reflecting the dynamics of demand for air transport and the fixed costs associated with the industry, all stakeholders need to consider appropriate strategies for better managing the risks. many risks were identified in the literature previously, some even decades ago, however most have yet to be satisfactorily addressed. urgency is growing. removal of the remaining barriers to competition at all levels, congestion management, open skies policies across continents, computer-centric air traffic management systems and increased research and development into the processes and technology needed to reduce environmental externalities remain among the top challenges for the next decade. the aviation industry is entering a new era in part due to two major issues. the first issue involves the increasing interest in the perceived environmental damage caused by transportation in general and by aviation in particular. the second issue involves the impact of multiple exogenous shocks such as the financial meltdown of as a result of which the aggregate airline industry profits of the past seventy years, which were admittedly marginal, were completely wiped out. fig. presents the data drawn from the air transport association ( ). the variability of the exogenous shocks on airline demand levels has been increasing at a rapid pace hence the need to develop strategies for all stakeholders in the aviation sector. a major risk to the sustainability of the aviation system is that legal principles rather than economic rationality will prevail such that competition and good managerial leadership are swamped by market distortions. understanding the markets, removing barriers to both entry and exit and encouraging competition on all links of the aviation sector leads to innovation and internalization of the inherent risks of volatile demand, economic cycles and climate change. deregulation in the airline sector led to the development of a new breed of carriers that has in turn increased consumer surplus. corporatization and privatization of airports led to a substantial increase in alternative revenue streams at airports which improved both producer and consumer surplus. on the other hand, distortionary subsidies given to airframe manufacturers led to the development of aircraft that are not financially viable, such as concorde and the a (gellman et al., ) . in this article we discuss potentially fruitful strategies that may aid the airlines, airports, airframe and engine manufacturers and their first tier suppliers as well as those bodies governing the industry. these strategies need to provide a cushion whereby companies can reasonably handle the risk of fuel price instability, the introduction of carbon cap and trade regulation, the need to finance airport infrastructure, air traffic management systems, aircraft and other assets, the competitive inequalities drawing from subsidies across the globe at various levels of the supply chain and the effects of increasing ad-hoc consumer protection laws. the industry is dynamic and in returned to growth. pro-active strategies are needed to ensure that further growth is viable in an economically, politically and environmentally sustainable manner since the alternative will involve regulation and a reduction in overall social welfare and mobility. most airlines provide a scheduled service over which supply and demand must be carefully balanced, especially in light of the exogenous shocks that have substantially impacted demand in the short to medium term such as the explosion of the dot-com bubble in , the security implications of september th , the severe acute respiratory syndrome outbreak in and the united states housing price bubble of that led to the current recession felt in many parts of the world. the effects of these downturns will continue to be felt at airlines that fail to adopt a plan to replenish, upgrade and perhaps increase their fleet in order to account for the longer term, underlying growth pattern that is likely to transpire over time. good management would appear to be one of the most important elements of building and maintaining a successful airline and prudent aircraft purchasing decisions are at the epicenter of this approach (tretheway and waters, ; government accountability office, ) . furthermore, management must consider direct risks to the supply side, including for example the future price of fuel as well as the pricing and/or regulation of environmental externalities such as global greenhouse gas emissions, local air pollutants and noise. this section first discusses the issues of managing a heterogeneous customer base and the life cycle of the airline market in section . , the issues of achieving profitability in section . , the approaches to handling competition in section . and the remaining supply side strategies in section . . aviation is often treated as a discretionary service in comparison to other forms of transport such as daily trips to work, which leads to volatility and seasonality of demand. however airlines do provide mobility which is unique in longer haul markets and spans heavily business oriented destinations (e.g. belgium and shanghai), almost purely touristic hotspots (e.g. hawaii and las palmas) with the majority of origin-destination pairs a mix of the two to varying degrees. overall growth in demand has been decidedly positive over the longer term in line with the different stages of maturity of the industry around the globe and the respective income levels. business travel demand appears to be shrinking which is a process that began as far back as (mason, ) and has continued as a result of the current financial crisis, with companies searching for alternative forms of communication or at the very least, economy class tickets (cobb, ) . consequently, airlines need to encourage business passengers to move to the front of the cabin by maintaining frequency where reasonable, improving frequent flyer programs and attracting long term corporate travel agent agreements. the standard scheduled carriers have lost some business demand to the business jet market, although this is obviously limited to the extremely time constrained with a substantial willingness-to-pay (mason, ) . private aircraft and related traffic have so far avoided most of the security regulations that the legacy and low cost carriers must handle, which contributes a reasonable amount of additional time to a trip particularly in the shorter haul markets. leisure travelers choose holiday purchases given their discretionary income levels which have been reduced since . this passenger type is the most price sensitive, which has encouraged airlines to unbundle their product, providing the airlines with the ability to further price discriminate whilst arguably allowing passengers greater choice (brons et al., ; clemons et al., ; bilotkach, ) . airlines must utilize their existing staffing levels and fleet of aircraft at least in the short term, which has led to a heavy reliance on revenue management technology. the heart of the airline business lies in attracting the two consumer types, namely the business passenger interested in high levels of frequency and less so the airfare as compared to the leisure passenger who places much greater emphasis on fares (proussaloglou and koppelman, ; adler, ; adler et al., c) . ignoring one type at the expense of the other would appear to be extremely perilous. despite the high margins on business travel, a scheduled airline model catering specifically to this type of consumer does not appear to be viable, see for example maxjet, eos and silverjet, pure business class airlines serving transatlantic routes, all of which filed for bankruptcy in . one of their major issues were the problems of connectivity, as none of the airlines developed a web of interline or codeshare services which is so important to beyond or behind gateway travel (holloway, ) . charter carriers serving the pure leisure market also appear to be a waning business model as the low cost scheduled carriers take their place in maturing airline markets (gillen, ) . for scheduled service, the high frequency demanded by business consumers can only be served if the remainder of the aircraft is filled with a sufficient number of passengers willing to at least cover the marginal cost of the seats. relatively high frequency ensures a disproportionately higher market share (swan, ; belobaba, ) which is only worthwhile if the yield at the very least covers the average costs of the flight, including the cost of capital. airlines in the more mature, standardized markets achieve competitive advantage through lower costs. it may also be true that on longer flights (more than five hours), passengers are more willing to pay for additional comfort which would permit the differentiation strategy to survive and prosper. strategies also need to match the life cycle of the market in which they exist. until now, airlines have placed extreme emphasis on maintaining or increasing market share rather than profit potential and origin-destination yields. it would appear that the american domestic market, currently the largest aviation market in the world, has achieved a level of maturation such that market growth is flattening out. whilst the european union is moving towards saturation, the south american, far east and intercontinental markets are all a long way from maturation. furthermore, the african and middle eastern markets have yet to begin their exponential growth rates (swelbar and belobaba, ). consequently, low cost strategies in the united states and european union domestic markets appear to be the most profitable strategy given the current market life cycle, whereas the differentiated strategy would appear to be more profitable on the intercontinental routes and in regions that have yet to develop their markets more fully. it is extremely important for airlines to analyze the markets not as short-run revenue maximizers rather as long-run profit maximizers, in which case the reasonably substantial fixed costs would be covered such that a normal return on capital could be achieved. gillen ( ) argues that the legacy carriers focus on profitability at the network level rather than individual links which has lead to managerial myopia, excessive network size and severe price discounting. tretheway ( ) argues that the low cost carrier pricing policy differs subtly but importantly from that of the legacy carrier revenue maximization procedure. whilst the low cost carriers require all flights to fully cover allocated costs thus ignoring the issue of transfer passengers, the legacy carriers separate the decision making apparatus such that in the first stage, capacity choices are made and in the second stage, yield management systems maximize revenue given the first stage decisions. this separation in decision-making reduces the pricing policies to short term decisions which has resulted in declining yields and a failure to cover the capital costs needed to replenish a fleet. proussaloglou and koppelman ( ) analyze air carrier demand and demonstrate that new carriers with limited frequent flyer programs must provide substantially lower airfares or a superior level of service in order to compete effectively with incumbents. however, the recent erosion of the gates required to ensure successful revenue management models has left the legacy carriers with a reduction in fare classes, for example as a result of the disaggregation of return fares into single unidirectional tickets that has occurred due to low cost carrier policies (cobb, ) . following porter's competitive strategy approach ( ), we argue that the likely market outcome that would permit airlines to achieve long run profitability suggests that low cost carriers should serve the domestic or regional markets whereas legacy carriers should continue with their differentiated approach on the intercontinental, longer distance routes. this would permit the legacy carriers to reduce the variety of aircraft currently required to serve greatly differing stage lengths, in turn reducing maintenance and training costs and increasing the productivity of the remaining fleet. codesharing across the two business models would be a logical next step and although low cost carriers have not generally participated in interlining or codesharing, examples do exist such as virgin blue and united ( e ) then delta (from onwards) and westjet with southwest for a short period and cathay pacific (from may ). another important set of strategies available to airline managers to better manage risk include choices with regard to interlining, codesharing, joining an alliance or merging with complementary partners or rivals, subject to government anti-trust regulation. interlining became a feature of the airline landscape as a result of the chicago conference held in which permitted an airline to sell a single ticket to a consumer despite the fact that the origin and destination were not directly connected by the carrier, rather passengers would need to change both planes and airlines on the single itinerary. this was advantageous to the consumer who would not need to carry baggage at the connection and was organized between the airlines through the international air transport association (iata). the iata conferences organized the airlines, enabling them to reach pricing decisions per region and to subsequently share interline revenues according to the geographical distance each carrier provided per itinerary. codesharing first appeared in international markets in (gellman research associates, ) . collaboration between airlines was at first designed in order to offer the international passenger a "seamless" travel experience by minimizing some of the inconveniences of traditional interline itineraries. benefits to consumers of codeshares over interline itineraries include agreements on standardized levels of service, access to airport lounges and frequent flyer programs. for the suppliers, codeshares based on block space or free sale agreements encourage the airlines to consider the issue of double marginalization but also lead to closer associations and a softening of competition, such that the agreements are a somewhat double edged sword. the transportation research board ( ) noted that % of global alliances include provisions for codesharing, % include provisions relating to sharing of frequent flyer programs and % also include agreements to share facilities such as catering, training, maintenance and aircraft purchasing. the web of codeshares that form the basis of an alliance help airlines to better handle risk, permitting a reduction in capacity during bear markets and faster response to unexpected short-term changes in demand. gillen ( ) argues that along with the development of hub-and-spoke systems, domestic feeds have contributed to the development of international alliances in which one airline feeds another hence utilizing the capacity of both to increase service and pricing. codesharing began as a pure marketing exercise but has now become an important element for both suppliers and customers. the supplier offers a greater network span and enjoys economies of scope and density. consumers avoid the issue of double marginalization that arises when required to purchase two or more tickets from different vendors, enjoy reductions in schedule delay and reduce complications arising from delays particularly on the first leg of an itinerary. adler and hanany ( ) demonstrate that consumer welfare on thin origindestination markets is higher with code-sharing airlines than purely competing carriers. consequently, codesharing increases the level of service provided to the consumer. aviation should develop into an industry in which reasonable levels of profit are achievable throughout the economic cycle. under the current regulatory regime, cross-border mergers are not permitted since foreign ownership rights are curtailed to varying degrees, except in the australasian domestic markets. however, as demonstrated in adler and smilowitz ( ) , airlines would always prefer to merge based on economic considerations, drawing from improved cost efficiency and subsequently higher profits. indeed international gateway choice would change were mergers to be permitted. adler and hanany ( ) also demonstrate this point but purely from the demand side perspective whilst the cost advantages are ignored. consumer preferences for higher frequencies and home carrier bias permit airlines to achieve their highest profits under mergers although to some extent at the expense of consumer surplus. airline competition may not always be acting on a fair playing ground which is a sign of supply side risk. airlines in the middle east, including emirates, etihad and qatar, have a growing presence in the aviation markets and enjoy a business environment to which other airlines do not have access. according to o'connell ( ) , emirates enjoys zero corporate tax under the united arab emirate's laws, extremely low airport charges at its dubai hub since the chairman of the airline is also minister in charge of civil aviation governing the airport, an uncongested hub that reduces fuel costs, low labor costs and a labor force that is not permitted to join a union or strike. altogether, this contributes to an estimated % cost advantage over british airways and a % advantage over air france/klm (o'connell, ) . were the middle east aviation market to develop alongside regional stability and liberalization, adler and hashai ( ) predict that cairo and tehran are likely to develop regional hubs with istanbul and riyadh emerging along with the prosperity of the region based on geographic and demographic considerations. current transport investments also suggest that the dubai region is succeeding in its attempt to develop a major hub system connecting the continents of north america, europe, africa, the far east and australasia via the middle east. the growing lack of trained pilots is another issue of note to both airlines and aircraft manufacturers. as the number of unmanned aerial vehicles grows globally (the economist, ), fewer fighter pilots are being trained, leaving an insufficient number to subsequently enter the civilian industry once their military careers are completed. embraer has announced that within the coming decade it plans to build a single pilot certified aircraft (flightglobal, ) and it is likely that pilotless cargo aircraft will be in use within this timeframe as well. we predict that pilotless passenger aircraft are likely to enter the skies within two decades, once the next generation of computer-centric air traffic management systems and avionics enter the market. in the meantime, the burden to push for increased funding of pilot training appears to lie on the shoulders of the pilots association and trade associations, such as the air transport association and regional airline association. finally the climate change debate is gradually pushing all sectors of society to measure, manage and subsequently reduce their carbon footprint. the aviation sector is slowly feeling this pressure too with new zealand and the european union at the vanguard of this process. the pressure on aviation has more to do with the prominence of air travel in society today than with the real contribution of aviation to global warming, since trucking and cars are a far more important contributor. new zealand introduced an emissions trading scheme (ets) in that extends only to domestic flights and can be applied to either the petroleum supplier or the airline. the new zealand government intends to reduce carbon emissions to levels. scheelhaase et al. ( ) discuss the likely impact of the european union (e.u.) emissions trading scheme currently expected to begin implementation in january , which is to be applied to both domestic and international flights. scheelhaase et al. argue that the e.u.-ets will probably provide a competitive advantage to non-e.u. carriers whose short-haul, less environmentally efficient flights are not within the e.u. jurisdiction. forsyth ( ) argues the opposite by suggesting that the free permits would provide a financial advantage to those receiving them, although the impact is not expected to be substantial. the question then remains as to whether other regions of the world will follow suit and set up emission trading schemes or introduce carbon taxes in order to internalize the environmental externalities. in addition, various individual airports have gradually introduced night flight curfews and noise charges as well as local air pollution charges covering both nitrogen oxide and hydrocarbon (scheelhaase, ) over the past decade. governments need to decide whether they are interested in dampening demand to reduce global warming or push for innovation such that each flight pollutes at lower levels hence permitting "green growth". if the latter has a greater priority, then subsidizing research and development in this area is a necessary and currently underutilized component. finally, it would probably be extremely beneficial to the various players in the aviation supply chain were the economic instruments chosen, whether restrictions, charges or taxes, to be applied equally across the globe and in a harmonized manner. airports have been changing as a result of privatization and corporatization, the deregulation of airline markets regionally and inter-continentally and the development of the low cost carrier model which demands different services from the secondary airports that they generally serve (deneufville, ) . airports in many parts of the world are no longer viewed as public utilities rather as private enterprises aiming to maximize shareholder value and profits from a fixed facility (adler et al., b) . the trend to privatize airports began in the united kingdom in with the flotation of the british airports authority, a company that owned and managed seven airports, three of which were located in london. the recent forced sale of gatwick airport has the intended aim of encouraging competition among the airports of london. within the london catchment area, baa now owns and runs heathrow and stansted, global infrastructure partners owns and runs gatwick and london city whilst luton is owned by the local council and run by a private company. as airports have required infrastructure investments beyond the budgets of local and federal governments, the airports have gradually been privatized in europe, south america, south africa, asia and australasia. perhaps surprisingly, airports in the united states are owned either at the state or local authority level and are operated by divisions of municipal governments or airport authorities. however many of the sub-processes at american airports are managed by private companies and a mere e % of the employees on the airport site are directly employed by the government authority (deneufville, ) . until the s, much of the investment in airport infrastructure drew from the airport improvement program, a federal aviation authority based fund. the fund has gradually reduced in importance, particularly at the larger hub airports, and has been replaced with direct passenger facility fees and the issuing of bonds often underwritten by the relevant hubbing airline (odoni, ) . whilst many airports remain natural or locational monopolists, for example in small countries with little to no domestic traffic, others operate in competitive markets as a result of the deregulation of both the airlines and airports (starkie, ) . tretheway and kincaid ( ) define airport competition to include local demand located in overlapping catchment areas e.g. multi-airport cities, connecting traffic served by hubs, cargo traffic, alternative modes and destinations. barrett ( ) argues airport competition is a new element of european aviation as a direct result of liberalization, whereby airports within one hour ground surface access are in direct competition for their respective catchment area, as occurs in multiple cases in france, germany and the united kingdom. hooper ( ) argues that governments in asia may rely on competition to impose a significant degree of discipline on airport managerial behavior. adler and liebert ( ) demonstrate that competition for connecting passengers and/or over catchment areas appears to be sufficient to encourage cost efficiency independent of ownership form or economic regulation. however, apart from australia and new zealand, airports around the world remain price regulated. according to fu et al. ( ) , the light handed regulatory approach of australasia in which price monitoring replaced formal regulation has not been successful, mostly due to the lack of competition inherent in a system with large distances between airports. consequently, it would appear that competition is sufficient to ensure that airports are cost efficient but without it, independent of ownership form, some form of economic regulation is necessary. such regulation would reduce the likelihood of litigation as has occurred on multiple occasions in australia with virgin blue, currently the second largest australian airline. adler and liebert ( ) also demonstrate that privatized airports operating in a competitive environment may still require economic regulation in order to avoid excessive pricing in comparison to their unregulated, public counterparts operating in a similar environment. strategies for airport managers therefore need to account for ownership form. in section . we discuss strategies for the shorter term timeframe and in section . , we discuss size and pricing policies relevant to the longer term issues identified. in the short term, airport managers may be interested in maximizing variable factor productivity, given a fixed airport capacity. this is particularly true for privatized airports and those who are price capped under an inflationary less efficiency formulation which permits the airport to retain productivity gains beyond the minimum level required by the regulator. variable factor productivity includes labor, supplies and materials and outsourcing costs and quantities, given passenger and cargo throughput, air traffic movements and non-aeronautical revenues. shorter term decision making includes searching for a balance between in-house production and outsourcing activities. partial analyses of subprocesses such as baggage handling and passenger flow through terminals may also help managers to highlight bottlenecks in the system. benchmarking good practice is crucial to effective management and public disclosure requirements, an approach adopted in britain and australia (hooper, ) , is an important missing link in encouraging productive efficiency. a uniform system of airport accounts similar to that of the international civil aviation organization (icao) airline reporting practices would be helpful to both airport managers and regulators alike. indeed, there are no generally accepted accounting practices even for airports within a single country which means that the capital input mix cannot be analyzed. the academic literature contains many potentially useful methodologies for benchmarking processes, such as stochastic frontier analysis (oum et al., ; and data envelopment analysis (sarkis and talluri, ; adler et al., b) which could be applied were comparable data to be made available. transparency in data collection would also encourage analyses of dynamic efficiency which is extremely important in an industry with lumpy and large fixed costs. in the medium term, uncongested airports with low capacity utilization need to reduce their asset base and/or increase their customer base. to attract greater output, either in terms of passengers or cargo, may require offering lower charges for new destinations served for the first couple of years of service or unbundling the airport services, thus permitting airlines to choose varying levels of service according to their desires. congested airports require different managerial policies including expanding capacity at the margin wherever bottlenecks are identified and incentivizing airlines to use off-peak slots through pricing. the icao governs the rules for landing fees on all international flights and requires that charges do not exceed the full cost including a return on capital which is needed to provide the facilities and services. a revenue neutral congestion pricing policy would remain within the guidelines of the icao and may result in negative prices for off-peak air traffic movements but this should improve capacity utilization without being discriminatory. alternatively, larger planes could attract price reductions which again would provide incentives for airlines to maximize capacity utilization in line with social welfare optimization. additional medium term strategies include actively identifying ground access improvement opportunities, such as high or higher speed train service, or improved road access which may widen an airport's catchment area. the longer term issues are the most difficult to solve since they generally require capacity expansion or reduction, both of which are very difficult to undertake. barriers to expansion include political interests, noise and environmental restrictions, the time and expense involved in receiving planning permission, not in my backyard syndrome and the lack of active management interest, likely to be more relevant at public airport authorities. in addition, there are sufficient examples of airports who undertook the risk and expense of expansion only to be underutilized afterwards, such as the city of dayton that decided to build a hub at the behest of u.s. air which then drastically reduced its services. american airlines behaved similarly at raleigh and nashville and, after acquiring reno, left san jose airport in the lurch to a large degree. in order for an airport to be cost efficient, it is necessary to utilize resources carefully, which generally leads to congestion and the need to deal with this issue fairly with respect to passengers, airlines and the environment. the toughest issue for airport managers is the lack of signals inherent in a system whereby congestion and delay are not priced. the lack of congestion pricing incentivizes airlines to increase frequency and reduce aircraft size even during peak periods. indeed, the trend in airplane size in the united states has been on the decline since because smaller aircraft achieve shorter turn-around times hence higher utilization, consumers value higher frequency which is reflected in airfares, smaller aircraft produce marginally lower levels of noise which is relevant at hub airports with aggregate noise constraints and congestion pricing which is missing from the equation (swan, ) . without peak pricing in the united states or scarcity pricing in europe under the slot allocation system, from where do the signals come to expand or define optimal capacity levels? as levine wrote in , the existing pricing system fails to guide investment so as to achieve the appropriate mix and level of output with a minimum investment of resources and the same could be said today. congestion pricing and the direct valuation of slots would appear to be strictly preferable to the current system of rationing defined in the form of slot allocation regulation in europe and department of transport brokerage in the united states (johnson and savage, ) . one could argue that were congestion fees collected for the transparent purpose of building or expanding specific bottlenecks at an airport, such charges would indeed be in line with the icao policy mandate. slot allocation policies exist to ensure that delays in air transport are not excessive and appear to be effective when comparing american and european delay outcomes (forsyth, ) . indeed, the lack of slot allocations at american airports has led to the development of a ground delay program operated by air traffic management through the federal aviation administration (faa). however, the bartering involved with this system prevents new entrants from entering congested airports hence provides an economic advantage to legacy carriers. adler et al. ( a) discuss the slot allocation issues in the greater tokyo region which permit the producers to extract surplus from consumers, to the extent that an aggressive low cost carrier is not capable of increasing competition either domestically or regionally. czerny et al. ( ) summarize much research that promotes the use of auctions as an alternative form of scarce resource allocation, however it is rather unlikely that the incumbent airlines would readily agree (see sentance, for an incumbent airline's response). the lack of clear legal ownership with respect to landing rights is an issue that needs to be solved in order to allow airports to efficiently match supply with demand. permitting slots to become a tradable asset would substantially improve the capacity allocation issue although regulation would still be necessary in order to ensure that airports are not reregulating the airline sector. whilst slot allocation is not an issue in the united states where a first come, first served policy exists, gate allocation acts as a barrier to entry instead (dresner et al., ) . gate allocations in the united states are often accompanied by a e year lease contract in order to allow airports to issue bonds that fund the expansion. despite deneufville's ( ) argument that the collaborative approach in the united states has led to a better airport system than other areas of the world, controlling access to busy airports acts as a barrier to entry for airlines, which severely curtails competition and the positive impacts of deregulation. over time, many airframe manufacturers merged, exited or failed to the point that two major markets remain; large airframe and regional jet manufacturers. the large airframe market currently consists of two firms, the european airbus and american boeing companies. the duopolists have chosen to compete headon, with each firm producing a range of aircraft in direct competition, such as the a and boeing - , the a and b and later variants of the b , and the smaller a with the b . to some extent the b is also in competition with the a over certain routes. for example, in the american-japanese market, the a may well serve the jfk-narita hub-to-hub market given the level of congestion at both airports whereas the b may serve the jfk-nagoya or newark-nagoya market as a way of avoiding at least one major hub and providing improved service to passengers through a direct itinerary. in the regional jet market, brazilian embraer and canadian bombardier are the two major players but they may be competing with manufacturers located in russia, japan and china shortly. small airframe development has benefitted from subsidies to customers in the form of low interest loans from their respective governments in order to support development of aircraft of up to seats, despite and world trade organization (wto) rulings that this should not continue. recently bombardier, which is subsidized by the canadian government, announced the development of their c series which will ultimately accommodate seats. in an unusual move, airbus and boeing joined forces and jointly argued before the wto that such financial subsidies should be limited to seat capacities, if not stopped entirely. however, both the japanese and chinese governments provide subsidies to companies developing aircraft components within their respective borders that encouraged outsourcing by both airbus and boeing. another form of subsidy occurs when new aircraft require a change in the capabilities of airports and the cost is borne by the airports rather than the relevant airframe manufacturer. in the 's, mcdonnell douglas began producing the dc - but the conditions for sale were that the new york airports could accommodate the aircraft, which required strengthening the taxiways and widening the runways. the new york airport authority argued that the costs involved were prohibitive and the mcdonnell douglas company, after reducing the costs through a radical redesign, paid for the changes necessary. multiple airports are currently under expansion in order to accommodate the a , but these costs are being borne by the airports, which represents a distortion in the airframe market. clearly, subsidies are unlikely to disappear despite wto rulings and it would appear that the more appropriate policy would be to encourage discussions and reach agreements across countries in order to limit the imbalance such distortions create. an example of the results of such discussions includes the e.u.-u.s. agreement that calls for a critical project appraisal before permitting any subsidization of the research and development of airframes. the agreement called for the repayment of direct government support over a period of years beginning from the date that the first state aid was received. however, as argued in gellman et al. ( ) , such an appraisal of the a was never undertaken and had this been the case, it is unlikely that the aircraft would have been produced. hence, it is insufficient to reach such agreements unless a legal entity exists that can uphold the clauses therein. other expensive inputs such as the engines and avionics are manufactured by various companies located in europe, south america and north east asia. in the parts market, under current american regulation, the original equipment manufacturer controls the supply of parts for aircraft still under production. alternative producers do not receive faa approval and their parts are tagged with the negative connotation of 'bogus' parts. a similar situation occurs with engine parts but in this market, alternative producers have tried to receive approval from the faa on the basis of 'functional equivalence'. to date, functional equivalence has not been approved and the spare parts market is limited, ensuring high mark-ups which inflate airline input costs. since the american policy with respect to the parts approval process is emulated globally, this issue crosses borders. we would argue that if a comprehensive functional equivalence test can be developed and the testing was undertaken by an independent agency, providing approval for these parts would break the current stranglehold in this first tier market. another major risk to the aviation sector is the continuing fluctuations in the price of oil. it is unlikely that a battery powered aircraft engine will be developed in the near future due to issues with the weight and size of the batteries available under current technological capabilities. consequently, aviation is likely to continue to be dependent on oil for the foreseeable future. two types of government action may be helpful in this regard. first, it would appear to be important to begin regulating oil speculation in order to prevent oil upside spikes that caused the massive changes in the price of oil inputs mid . second, were the united states, united kingdom, france, germany and japan to agree, it would be possible to break the stranglehold of the opec cartel on current oil prices. the current price of jet fuel has little connection to the cost of production. the relevant governments could restrict oil imports if prices were deemed unacceptably high. independently, these governments could subsidize research and development into new, cleaner technologies that would encourage universities and the private sector to explore ways of reducing greenhouse gas emissions. current promising avenues include the use of lithium aluminum or composite materials to reduce the weight of the aircraft and the development of alternative fuels, such as bio-fuels which reduce carbon dioxide based on the full life cycle approach. government funding, such as the european union's clean sky joint technology initiative, appears to be necessary at this point in time due to the high risk involved in this research. it is not yet clear whether camelina or algae have the potential to be grown in sufficient quantities to serve the market for bio-fuels without displacing land needed for food production. finally, operational research and development could encourage air traffic management systems to search for greener routings and manufacturers to further improve aerodynamics and engine efficiency. noise remains a major issue, particularly in regions with high density populations such as europe and asia but also at out of the busiest airports in the united states (girvin, ). there are examples of airports for whom capacity restrictions are defined by noise regulation rather than their physical capabilities such as schiphol. brueckner and girvin ( ) argue that continuing to limit cumulative noise at airports or equivalently, to charge a noise tax, pressures stakeholders to attempt to mitigate the issue hence maximize social welfare. swan ( ) argues that the use of smaller airplanes is preferable with respect to their noise output than an equivalent number of seats on larger aircraft. clarke ( ) calls for automated air traffic management procedures which would improve noise abatement measures beyond the impact of improvements in individual aircraft. clearly research and development needs to consider all elements of the aviation sector. two initiatives are currently being funded including nasa's 'quiet aircraft technology' program financed by the american government and the silent aircraft initiative undertaken at the cambridge-mit institute together with industrial partners, mainly funded by the british government. due to the trade-offs between reductions in local air pollution, noise in the vicinity of the airport catchment area and global greenhouse gas emissions affecting climate change, one of the major tasks of the new decade will be to strike the correct balance. in this section, we discuss the risks that exist within each of the links of the aviation industry and the potential strategies available to regulators to counteract the issues. we discuss the on-going process of deregulation of the airline markets in section . , the conditions under which airport regulation continues to be a necessity in a gradually privatized and corporatized airport industry in section . and the issues arising as a result of the changes in ownership form of the air traffic control sector in section . . over the history of the aviation industry, both airlines and airports have been heavily regulated and subsidized. in the united states, airlines have always been in private hands but until deregulation in , the civil aeronautics board chose the carriers to serve specific markets and their respective airfares. after deregulation, american carriers were free to fly wherever they chose in domestic markets but international services remain regulated according to reciprocal bi-lateral agreements. the american government has gradually opened the skies by encouraging multilaterals which led to the horizontal open skies agreement with the european union in , effective as of . however, american airlines are still protected through the standard chapter bankruptcy proceedings under which airlines restructure their debt and operations but continue to serve their markets (button, ) . whilst chapter proceedings are not specific to the aviation sector, the impact of this law is to produce an effective barrier to free exit from the market. in the european union, most airlines were defined as flag carriers up to deregulation in the third package of in which airline subsidies, which had been quite substantial up until that point, were no longer deemed acceptable. whilst there remain a few state owned airlines, such as olympic and tap, the majority of carriers are now in private hands. the european union and individual countries have permitted airlines to fail, for example sabena and swissair, however other airlines continue to survive due to either protectionist international bilateral agreements or subsidies, as has occurred in the cases of olympic and alitalia. the domestic chinese airline market has been gradually deregulated with china eastern airlines listed on three stock exchanges in , marking the beginning of the process. in there was a wave of airline consolidations resulting in the emergence of three large airline groups; air china, china eastern and china southern with major hubs in beijing, shanghai and guangzhou respectively (zhang and round, ) . however, the chinese skies remain relatively closed as the government continues to protect chinese airlines from foreign competition. southeast asian liberalization permitted a wave of new entrants in the early 's although many did not survive the regional economic crisis of (hooper, ) . the world trade organization has placed on their website a geographical tool that demonstrates the level of openness of bi-lateral agreements and awards each country a weighted air liberalization index score based on the level of air freedoms permitted, ownership restrictions, pricing and carrier designations. new zealand and australia receive relatively high scores, the united states is somewhat lower and china's score is close to the bottom of the scale currently. in order to protect airlines on the grounds of security considerations and potential job losses, the united states currently limits all foreign ownership of american carriers to % of the voting shares and at least two-thirds of the board as well as the chair must be american nationals. the european union limits foreign ownership to % of the airline's shares. a second open skies u.s.-e.u. agreement, signed in june but still requiring ratification on both sides of the atlantic, aims to loosen airline ownership and control restrictions reciprocally but as yet the details have not been revealed. in the chinese government began to permit foreign investment in chinese airlines of up to % of registered capital, which has since been increased to %, although foreign owners may not purchase more than % of the voting stock (zhang and round, ) . similar restrictions exist in south america, africa and asia. tretheway ( ) calls for the elimination of foreign ownership restrictions of air carriers and the permission for mergers across borders, arguing that national security benefits do not exceed the economic inefficiencies arising from the prevention of cross-border consolidation. the failure to permit consolidation is likely to result in either further bankruptcies or bailouts. new zealand was the first to remove foreign ownership restrictions on domestic carriers and australia followed suit in . indeed a multilateral open skies agreement (maliat) was signed in between brunei, chile, malaysia, new zealand and the united states in which the nationality clause was replaced with "the principal place of business and effective control" (hsu and chang, ) . as a result of the existing ownership restrictions, airlines currently unable to merge across borders have chosen to develop strategic alliances through the development of a web of codeshares which pools risk and increases network access. it would appear that codeshares have positive benefits for both consumers and producers alike even on parallel links and anti-trust immunity should only be necessary on thin routes (adler and hanany, ) . furthermore, bilateral agreements between two countries appear to be the worst of all worlds, limiting frequency and hiking prices at the expense of consumer surplus (gillen et al., ; adler and hanany, ) . therefore, the most important strategy from the regulators perspective should be to open up the skies through multi-laterals. cabotage, defined as the eighth and ninth freedoms of the air, would be another way to circumvent the archaic ownership rules. conservatism has ruled to date, for example the association of southeast asian nations (asean) have discussed opening the skies regionally for over a decade but still appear to be a long distance from achieving this goal (tan, ) , although the maliat agreement has shown that this is a distinct possibility. deregulation of the airline industry has served to highlight the importance of ongoing ex-post application of normal anti-trust law. to protect the lower prices and higher frequencies that strongly support the argument that the aviation market is better off without regulation (kahn, ) , it is equally important to protect the premise on which competitive markets develop. free entry and exit are the cornerstones of such a policy and prevent market distortions and inefficiencies. however, it would appear that both tenets are ignored in different geographical corners of the world. free entry only occurs if there are neither bi-laterals protecting designated carriers nor restrictions on the freedom to land and take-off at the airport level. within the far east and european union, almost all airports are slot controlled and many are highly congested, both of which present serious barriers to entry. within the united states, slot controlled airports no longer exist, however gate constraints due to high utilization or exclusive use designations are proving to be real barriers to entry (dresner et al., ) . in order to support revenue bond financing of facilities, many of the larger airport operators have required airline tenants to lease gates and counter space for a period of up to thirty years and in some instances, dominant airline carriers have built their own terminals and subsequently retain complete control whether fully utilized or not (cohen, ) . consequently, independent investment in airport gates, restrictions on minimum aircraft sizes during peaks and congestion or scarcity pricing are important policies to be considered. needless to say, the academic literature has discussed replacing the weight based landing charges with peak pricing for the last forty years but so far to no avail. levine ( ) and carlin and park ( ) were among the first to discuss this issue. daniel ( ) developed a bottleneck model and applied it to minneapolis-st. paul airport, arguing that by spreading the peak, the airport could increase air traffic movements by as much as %. a series of papers by brueckner ( brueckner ( , and brueckner and van dender ( ) argued that at least some of the congestion is internalized by hub airlines, namely that which it imposes on itself, however this does not remove the need for peak pricing nor the need to ensure access for potential new entrants. morrison and winston ( ) argue that second-best, atomistic congestion charges would improve social welfare and significantly reduce delays at congested airports in the united states even if internalized congestion is essentially charged twice. schank ( ) argued that peak pricing has so far been unsuccessful, citing three attempts at boston logan, the port authority of new york and new jersey and the british airports authority. his main line of reasoning suggests that implementation is only acceptable and likely to stand in subsequent litigation if the airlines removed from the peak timeslots have the ability to move to an alternative, efficient time, which the american carriers flying into london in the early morning successfully argued was not the case in the subsequent court proceedings, or to alternative airports, which was not available in boston. as starkie ( ) noted, most airports are not necessarily congested rather demand is peaked over the course of a day which is currently not managed efficiently through the weight-based charges but is the current basis for deciding on the need to expand. free exit is the other single most important strategy for governments to consider. ensuring that no company is 'too big to fail' is equally applicable to the airline industry. if chapter and subsidies or bailouts permit airlines to survive rather than be liquidated, the creativity and strong managerial skills that were engendered in this market apparatus will fail. it is important to permit failure and bankruptcy in order to ensure that the best survive and profit with as few market distortions as possible. reregulating the airline industry is a perennial discussion that has been highlighted once again at the initiative of oberstar and others in the united states congress recently (lowy, ) . a government accountability office report to congress in argues that such a move would likely reverse consumer benefits without saving airline pensions, such as those lost during the bankruptcy proceedings at united and us airways in . the report argues that the reduction in prices and increase in flight frequency and competition which have benefited consumers to varying degrees would be derailed by reregulation. poole and butler ( ) argue that the serious problems remaining in the aviation sector draw from the fact that although airlines were deregulated in the united states, neither the airports nor the air traffic management systems followed the same path which has led to serious distortions in the market. tretheway and waters ( ) argue that neither the civil aeronautics board nor price cap regulation would provide the stability that the political leadership is attempting to encourage. if the main aim of the politicians is to increase the levels of competition in an increasingly concentrated market, dresner et al. ( ) suggest that the construction of new gates, alternative provisions that permit gate access to new entrants during peak periods, specifying minimum aircraft size provisions during peak periods and/or peak load pricing policies may be sufficient to increase competition in congested corridors. winston ( ) argues that the use of reregulation to avoid 'destructive' competition draws from the traditional but flawed theory of regulation which assumes that perfectly informed social welfare maximizers are either managing the regulation or running the regulated firms. it is argued that the airline industry appears to oscillate between periods of excessive concentration and destructive competition. the regulator needs to help the industry to find a happy medium in which neither extreme occurs. there is sufficient anecdotal evidence that airlines use hubs, gate access and frequent flyer programs as barriers to entrance, yet the hub-spoke system allows airlines to be cost efficient and serve markets that otherwise would not be served. hubs are likely to continue for the foreseeable future because half the origin-destination traffic in the world is in markets too small to be served directly (swan, ) . however, as opposed to the discussions of excessive concentration being held in the united states congress currently, swan ( ) points out that the united states airline industry has not consolidated over the period of to according to the herfindahl index, despite numerous mergers and bankruptcies. winston ( ) argues that deregulation in multiple industries, including that of airlines, has proven to be positive for consumers, labor and producers, although not necessarily on an equal basis even within a group. consequently, the question remaining for the regulator is how to protect the advantages of deregulation whilst maintaining reasonable levels of competition in city pair markets. removing the remaining barriers to entry and exit, including the independent investment in gates and pricing of slots, will help further the impact so far achieved. finally, consumer protection rules need to be carefully balanced in order to ensure reasonable levels of service and behavior only where producers have been shown to be derelict. examples of such laws include the three hour tarmac rule that passed through congress in . this rule has increased the likelihood of canceling flights due to the maximal $ , fine per passenger were the travelers to be forced to remain onboard the aircraft whilst waiting on the tarmac for longer than the legal limit. in , the european court of justice ruled that passengers on flights delayed for more than three hours are entitled to compensation from airlines as is true for passengers on canceled flights. this begs the question as to whether these consumer rights in fact protect or harm passengers and whether there is a better way to handle congestion. we would argue that the issue of congestion and delay is better served through pricing appropriately rather than court cases or ad-hoc government restrictions imposed after a public outcry through the popular media. the aim of airport regulation is to ensure that airports do not abuse monopoly power, to incentivize airport managers to achieve productive efficiency and to provide the correct signals in the marketplace that would encourage appropriate utilization of the fixed facility. it would appear that all of these issues have yet to be resolved satisfactorily and will be discussed respectively. niemeier ( ) argues that ex-ante regulation should be limited to activities with natural monopoly characteristics. based on the premise that airports enjoy locational monopoly power, economic regulation has been undertaken in various forms ranging from cost based principles or rate of return regulation to incentive based structures. in europe, prices are capped by the relevant civil aviation authority or department of transport, generally for a period of five years, after which a new review is undertaken. the price caps are frequently based on a value that changes according to inflation, for example the retail or consumer price index, less a pre-specified level of efficiency (rpi-x). an airport that achieves levels of efficiency greater than x will reap the cost reductions at least until the next review. asymmetric information between the regulator and airport owners ensures that the review process is both time-consuming and relatively expensive but necessary where competition does not exist. furthermore, privatized airports working under competitive conditions still may require regulation in order to prevent excessive pricing relative to their public counterparts serving under similar market conditions (adler and liebert, ). an additional complication concerns the question of whether the regulation is based on a single or dual till computation because airports produce two revenue streams. on the aeronautical side airlines are charged per landing, based on maximum take-off weight, as well as a seat based fee. the non-aeronautical revenue stream draws from the terminal side in the form of concessions, car parking fees and rents from the development of airport land. niemeier ( ) argues that single till regulation, which constrains overall airport profitability, may represent a first best solution for unconstrained airports provided non-aviation rents are sufficiently high. at the london airports price caps are set per airport and specify the upper level the airports may charge for their aeronautical services, however within this calculation the british civil aviation authority takes into account the revenues that the airport realizes from the commercial side of the business, which represents a single till approach. if the british government was concerned with levels of congestion, this approach is clearly inappropriate (jones et al., ) . according to averch and johnson ( ) , if a company is prevented from fully exploiting monopoly power, there is a clear incentive to cross subsidize competitive offerings from those that are regulated. according to kahn ( ) this is precisely what occurs at a single till, regulated airport and the solution is to sever the link between the revenues and costs associated with the airside from the revenues attainable on the commercial side. in the united states, airports are viewed as notfor-profit, public utilities and their pricing mechanism is based on cost recovery using a residual, compensatory or hybrid cost pricing approach. consequently, this system does not require price regulation which appears to be advantageous. however, airports who do achieve profitability must then reinvest the revenues into the airport whether necessary or not. the residual cost approach that is more likely to arise at a hub in effect restricts airports to the equivalent of a single-till regulatory system which appears to be less appropriate for congested airports. jones et al. ( ) argue that all airport services should be regulated because the airports enjoy monopoly presence in many markets including terminal side car parking services as well as airline related services. reductions in the costs of services applicable to consumers directly, such as car parking, would stop the cross subsidization from commercial to airside activities and the consequent transfer of consumer surplus to the producers. fu et al. ( ) argue that airports enjoy substantial market power due to low price elasticity on the aeronautical side which may be moderated by the vertical relationship between the airport and hubbing airline. in summation, dual-till regulation is preferable to the single till form at congested airports both in terms of encouraging productive efficiency and ensuring sufficient investment in infrastructure (oum et al., ) . starkie ( ) argues that rpi-x price cap regulation encourages productive efficiency provided the airport acts as a profit maximizer rather than monopolist, however the same style of regulation also encourages excessive investment as defined in the averch johnson ( ) effect. consequently, we would argue that dual till economic regulation is preferable with separate price caps on aviation and commercial services, restricted to only those activities over which airports enjoy monopolistic rents. another important issue for regulators concerns the need to ensure optimal capital investment in an industry with large fixed costs. the current pricing policies at airports do not provide the signals necessary to evaluate the need for capacity expansion or reduction. barrett ( ) argues that there is no reason to assume that privatized airports are more likely to under-invest in infrastructure rather that this is more likely to occur under monopolistic regulatory conditions that restrict output below competitive levels, as indicated by the level of congestion that occurred under the traditional organization of airports prior to liberalization. however, basso ( ) argues that social welfare maximizing public airports subject to a budget constraint are strictly preferable to unregulated profit maximizing private airports because the latter would overcharge for congestion leading to excessive traffic contractions. martin and socorro ( ) argue that a private, congested airport does not require price regulation provided the regulator ensures an appropriate capacity investment under which private and public objectives coincide. cost plus regulation leads to over investment in either capacity or quality which leads to an unnecessarily expensive airport due to the spiraling regulated asset base cost issue. since governments are frequently interested in stimulating economic activity, incentives may exist that encourage over investment (forsyth, ) . whilst cost based regulation may lead to over investment, incentive based regulation may lead to under investment in which case the regulator then needs to consider an investment incentive mechanism as a counter balance. swelbar and belobaba ( ) argue that the lack of infrastructure capacity at airports and air navigation service provision enroute are two of the most critical issues facing international and national air services today. odoni ( ) argues that airport access is becoming the new form of market regulation that distorts the competitive outcome so sought after by many countries around the world. one of the major issues with regulation and optimal investment in airports lies in the mismatch between regulated price caps which are normally set every five years and the lifetime of an investment which may be closer to fifty. privatized airports will be willing to invest only if they are reasonably sure that they will cover their investment costs. carrier-served airports in the united states are defined as not-for-profits which allows them to receive infrastructure grants through the airport improvement program but as the funds are drying up, taxes on passengers, i.e. the passenger facility charge added to airfares, and bond issues cover the remaining costs. consequently, irrespective of airport ownership, the timing of capacity expansion will always be an issue unless the pricing policies change, permitting the market to signal the need for expansion through congestion and/or slot pricing mechanisms. air traffic management is another link in the system that requires change in order to prevent further restrictions on airline service. air traffic management is generally supplied by government entities although there are a few notable exceptions such as navcanada, where this control is now in the hands of a not-forprofit agency and nats, a public-private partnership in the united kingdom. weakly-led civil aviation authorities who prefer a quiet life rather than ensuring an efficient, highly utilized system have led to a mismatch between supply and demand. the dual role leads to limitations in the system that ensures neither efficiency nor productivity. around the world, air traffic management needs more rapid deployment of proven technologies and to become computer-centric rather than human-centric as is currently true. it is equally important that the individual links within the air traffic management system are understood and their respective capacities analyzed in order to set priorities for research and development to be directed specifically at the bottlenecks. prior to deregulation of the airlines, many questioned whether profit oriented companies would serve the public as safely as under the regulated era. the same is true for air traffic management. there is a visible trend towards privatization and corporatization of the air navigation service providers around the world, however the ability to introduce competition in this market is clearly suspect. separate companies, whether not-for-profit or economically regulated privatized concerns, appear to have reduced some of the inefficiency that existed previously (mcdougall and roberts, ). the departments of transport or civil aviation authorities could then promote their rightful positions as safety regulators, at arms length from the service providers. adam smith's ( ) treatise argues that competition enhances economic welfare whereas monopoly power, for example in the form of labor association restrictions or government regulation, detracts from rational pricing. multiple domestic airline markets have been deregulated over the past years, however international routes are still associated with restrictive bi-laterals for the most part, the maliat and u.s.-e.u. open skies pact being among the first to remove such restrictions. there would appear to be ample evidence of the success of deregulation in the form of business model innovation and increased consumer surplus, hence the global policy emulation. however, the volatility of demand seriously impacts the airline industry pushing the players between two extremes, excess concentration and destructive competition, which requires regulators worldwide to continue their vigilance. first, government oversight in the market should be restricted to the protection of competition rather than the protection of competitors such that no firm is too big to fail. second, in order to protect the positive impact of airline deregulation, it is necessary to remove the remaining barriers to free entry and exit including bilateral agreements between nations, restrictive slot and gate allocations that grant preferred status to incumbent airlines and the foreign ownership restrictions and controls that prevent mergers across borders. in summation, the risks to society and possible solutions are summarized in table . consequently, pricing congestion or scarcity, noise and emissions are far superior to the system of government restrictions that are currently applied to solve the bottlenecks in the aviation supply chain. pricing provides the signals necessary to identify and subsequently search for solutions to constraints based on demand rather than ad-hoc short-term solutions. specifically, one of the major limitations to the prosperity of air travel today is the ongoing regulatory regime that restricts and controls the airport and air traffic management capacities. separation of powers is necessary in both arenas in order to prevent either elements from reregulating airlines. the airports, whether private corporations or public entities, need to be separated from political pressures defining slot or gate allocations and the civil aviation authorities, who set the air traffic management levels, need to be separated from the body that operates the system. market distortions, limitations and inefficiencies will thus be removed. these strategies will enable the airlines, airframe and engine manufacturers and airports to better respond to demand and reduce the risks inherent in the existing system. for fruitful discussions that led to the development of this paper. nicole would also like to thank the recanati foundation for partial support of this work. the effect of competition on the choice of an optimal network in a liberalized aviation market with an application to western europe trading capacity and capacity trade-offs. working paper at the center for transportation studies effect of open skies in the middle eastern region the impact of ownership form, competition and regulation on airport efficiency. working paper at the center for transportation studies air transport liberalization and airport slot allocation: the case of the northeast asian transport market benchmarking airports from a managerial perspective. working paper at the center for transportation studies high-speed rail and air transport competition: game engineering as tool for cost-benefit analysis hub-and-spoke network alliances and mergers: pricelocation competition in the airline industry annual results: world airlines behavior of the firm under regulatory constraint airport competition in the deregulated european aviation market airport deregulation: effects on pricing and capacity overview of airline economics, markets and demand reputation, search cost and airfares price elasticities of demand for passenger air travel: a meta-analysis airport congestion pricing when carriers have market power internalization of airport congestion: a network analysis airport noise regulation, airline service quality and social welfare atomistic congestion tolls at concentrated airports? seeking a unified view in the internalization debate low cost airlines: a failed business model? marginal cost pricing of airport runway capacity the role of advanced air traffic management in reducing the impact of aircraft noise and enabling aviation growth price dispersion and differentiation in online travel: an empirical investigation today's airlines should adopt a low-cost strategy: can this popular idea be supported by the facts? the antitrust implications of airline deregulation airport slots: international experiences and options for reform congestion pricing and capacity of large hub airports: a bottleneck model with stochastic queues airport barriers to entry in the u embraer reveals vision for single-pilot airliners the impacts of emerging aviation trends on airport infrastructure the impact of climate change policy on competition in the air transport industry airport-airline vertical relationships, their effects and regulatory policy implications airline business models and networks: regulation, competition and evolution in aviation markets measuring the economic effects of bilateral liberalization air transport aircraft noise-abatement and mitigation strategies a study of international airline code sharing straight and level: practical airline economics privatization of airports in asia the environment for southeast asia's new and evolving airlines the influences of airline ownership rules on aviation policies and carriers' strategies departure delays, the pricing of congestion and expansion proposals at chicago o'hare airport the economics of airport slots deregulatory schizophrenia airline deregulation: a mixed bag, but a clear success nevertheless landing fees and the airport congestion problem reregulation of airlines being pondered. the washington times on commercializing air traffic control: have the reforms worked? canadian public administration a stochastic frontier analysis to estimate the relative efficiency of spanish airports a new era for airport regulators through capacity investments airframe manufacturers: which has the better view of the future? observations of fundamental changes in the demand for aviation services another look at airport congestion pricing regulation of airports: the case of hamburg airport e a view from the perspective of regional policy the changing dynamics of the arab gulf based airlines and an investigation into the strategies that are making emirates into a global challenger the international institutional and regulatory environment ownership forms matter for airport efficiency: a stochastic frontier investigation of worldwide airports alternative forms of economic regulation and their efficiency implications for airports airline deregulation: the unfinished revolution competitive strategy: techniques for analyzing industries and competitors air carrier demand: an analysis of market share determinants performance based clustering for benchmarking of us airports solving airside airport congestion: why peak runway pricing is not working the inclusion of aviation into the eu emission trading scheme: impacts on competition between european and non-european network airlines local emission charges e a new economic instrument at german airports airport slot auctions: desirable or feasible? utilities policy , e . smith, a., . the wealth of nations airport regulation and competition aviation markets: studies in competition and regulatory reform misunderstandings about airline growth critical issues and prospects for the global airline industry the asean multilateral agreement on air service: en route to open skies entry and competition in the u.s. airline industry: issues and opportunities distortions of airline revenues: why the network airline business model is broken competition between airports: occurrence and strategy reregulation of the airline industry: could price-cap regulation play a role economic deregulation: days of reckoning for microeconomists china's airline deregulation since and the driving forces behind the airline consolidations the authors would like to sincerely thank the organizers and participants of the hamburg aviation conference of february key: cord- - ftephy authors: aldridge, arnie p.; barbosa, carolina; barocas, joshua a.; bush, joshua l.; chhatwal, jagpreet; harlow, kristin j.; hyder, ayaz; linas, benjamin p.; mccollister, kathryn e.; morgan, jake r.; murphy, sean m.; savitzky, caroline; schackman, bruce r.; seiber, eric e.; e starbird, laura; villani, jennifer; zarkin, gary a. title: health economic design for evaluating cost, cost-effectiveness and simulation analyses in the healing communities study date: - - journal: drug alcohol depend doi: . /j.drugalcdep. . sha: doc_id: cord_uid: ftephy background: the healing communities study (hcs) is designed to implement and evaluate the communities that heal (cth) intervention, a conceptually driven framework to assist communities in selecting and adopting evidence-based practices to reduce opioid overdose deaths. the goal of the hcs is to produce generalizable information for policy makers and community stakeholders seeking to implement cth or a similar community intervention. to support this objective, one aim of the hcs is a health economics study (hes), the results of which will inform decisions around fiscal feasibility and sustainability relevant to other community settings. methods: the hes is integrated into the hcs design: an unblinded, multisite, parallel arm, cluster randomized, wait list–controlled trial of the cth intervention implemented in communities in four u.s. states: kentucky, massachusetts, new york, and ohio. the objectives of the hes are to estimate the economic costs to communities of implementing and sustaining cth; estimate broader societal costs associated with cth; estimate the cost-effectiveness of cth for overdose deaths avoided; and use simulation modeling to evaluate the short- and long-term health and economic impact of cth, including future overdose deaths avoided and quality-adjusted life years saved, and to develop a simulation policy tool for communities that seek to implement cth or a similar community intervention. discussion: the hcs offers an unprecedented opportunity to conduct health economics research on solutions to the opioid crisis and to increase understanding of the impact and value of complex, community-level interventions. the u.s. opioid crisis persists with nearly , deaths attributed to opioid overdose in (wilson et al., ) . opioid misuse and opioid use disorder (oud) have multiple and long-lasting economic impacts on individuals, families, communities, and society (florence et j o u r n a l p r e -p r o o f al., ; inocencio et al., ; leslie et al., ; roland et al., ; scavette, ; segel et al., ) . despite the demonstrated efficacy of existing evidence-based practices (ebps) to support treatment and recovery from oud, only a small proportion of individuals with oud are identified as needing treatment, and less than % actually receive recommended services (samhsa, ; wu et al., ) . reasons for underutilization of these services include lack of screening for oud by health care and legal systems, insufficient treatment capacity especially for medications for opioid use disorder (moud) and documentation of oud, lack of access and awareness among individuals with oud about treatment options, and stigma surrounding the use of moud (braithwaite and nolan, ; jones et al., ; mclean and kavanaugh, ) . additional challenges include limited uptake of overdose prevention approaches such as utilization of community-based naloxone distribution (meisenberg et al., ) . the healing communities study (hcs) is a four-year, multi-site, parallel group, cluster randomized wait-list controlled trial testing the impact of the communities that heal (cth) intervention on reducing opioid overdose deaths in disproportionately affected communities across four states--kentucky, ohio, new york and massachusetts. the hcs will assess the effectiveness of the cth intervention, a stepwise community change process that seeks to mobilize hcs communities to implement ebps in a range of settings, including behavioral health, health care, and criminal justice systems (oesterle et al., ) . cth has three components: ( ) a community engagement to facilitate data-driven selection and implementation (martinez et al., in press) ; ( ) the opioid overdose reduction continuum of care approach (orcca) (winhusen et al., in press) ; and ( ) communication campaigns to reduce stigma and raise awareness and demand for ebps (lefebvre et al., in press) . the orcca itself comprises multiple options within three areas: ( ) overdose education and naloxone distribution; ( ) j o u r n a l p r e -p r o o f effective delivery of moud; and ( ) safer opioid prescribing and dispensing. more detail on the cth components and orcca ebps are provided in walsh et al. (in press) and winhusen et al. (in press) . the hcs is designed to produce generalizable information for policy makers and community stakeholders seeking to implement cth or a similar community intervention. to support this objective, one aim of the hcs is focused on health economics analysis and simulation modeling to gain information about fiscal feasibility and sustainability that may be relevant to other states and settings. this paper describes the design for the hcs health economics study (hes) . the objectives of the hcs hes are to . estimate the economic costs to communities of implementing and sustaining cth; . estimate the broader societal costs associated with cth; . estimate the cost-effectiveness of cth compared with standard practice, measured as cost per overdose death avoided; use simulation modeling to evaluate the short-and long-term health and economic impact of cth, including future overdose deaths avoided and qualityadjusted life years (qalys) saved; and . develop a simulation policy tool for communities that want to implement cth or a similar community intervention. health and medicine (neumann et al., ) and incorporates methods from previous health economics studies of interventions to reduce oud and its consequences, including definitions of the study's perspectives, approaches to micro-costing, and best practices for cost and costeffectiveness analyses. there are few published health economics studies of complex, community-driven interventions for substance use disorders. the most relevant are two benefit-cost analyses j o u r n a l p r e -p r o o f (kuklinski et al., ; kuklinski et al., ) of the communities that care model (oesterle et al., ) from which the cth was adapted. these analyses used a micro-costing approach to estimate the resources consumed by communities to implement the intervention, including the process for engaging community coalitions, selecting and implementing a menu of ebps, and ongoing training and technical assistance. economic benefits were estimated for a sample of adolescent participants and compared with the costs, ultimately demonstrating a positive economic value (i.e., positive net benefits) for the intervention implementation. we will follow a similar approach to costing the multilevel/multisystem intervention in the hcs. we will estimate the resources needed to engage community coalitions, to support communities selecting and implementing ebps, and to conduct ongoing training and technical assistance in service to those ebps. our planned analyses, however, differ in important ways. first, our primary goal is to estimate the cost-effectiveness of cth with respect to opioid overdose deaths and related outcomes rather than to compare monetized benefits with intervention costs to measure economic value. second, rather than measuring efficacy using outcomes of a cohort of identified individuals who are tracked over time, we will rely on population-level outcomes for the study communities that will be measured in the hcs. the individual orcca ebp components of cth have a substantial literature assessing their costs and cost-effectiveness. for instance, the cost-effectiveness of moud has an established evidence base across a variety of settings and populations, particularly for methadone maintenance therapy and buprenorphine (barocas et al., ; busch et al., ; dunlap et al., ; gisev et al., ; krebs et al., ; murphy et al., ; murphy and polsky, ) . additionally, programs providing naloxone to high-risk individuals and first responders for reducing opioid overdose mortality have also been found cost-effective (coffin and sullivan, j o u r n a l p r e -p r o o f ; townsend et al., ) . notably, there is a lack of literature on the cost-effectiveness of safer opioid prescribing and dispensing programs. simulation modeling has become more common in economic evaluation of interventions to address opioid misuse (barbosa et al., ; briggs et al., ) . specifically, it draws information from multiple sources to integrate information on the trajectory of oud and associated complications with evidence of the clinical and economic impact of public health strategies to inform the best responses to the opioid crisis. simulation modeling can augment clinical studies by projecting clinical and economic outcomes over long time horizons and can explore outcomes for populations that differ from the ones that participate in clinical studies. simulation models can also enhance cost-effectiveness analyses conducted alongside clinical trials by improving the measurement of uncertainty around estimates of economic value, examining causal factors, and characterizing alternative scenarios to inform policy (buxton et al., ) . models with longer-term time horizons can capture the multiplicity of outcomes characteristic of oud, which often has periods of relapse and recovery accompanied by several comorbid conditions (barbosa et al., ; nosyk, ) . the hcs presents an ideal platform for developing simulation models through collaboration among research sites and for engaging communities during model development so that model results can be most useful to decision makers. the hcs offers an unprecedented opportunity to conduct health economics research on solutions to the opioid crisis-solutions that combine a community-level approach with combinations of proven ebps and communication campaigns. it provides a significant opportunity to answer questions about cth costs and to evaluate its cost-effectiveness at reducing opioid overdose fatalities. additionally, the potential for understanding policies that j o u r n a l p r e -p r o o f may lead to sustainable changes in opioid use disorder through simulation modeling provides an opportunity for broad impact. the methods that will be employed in the hcs hes are described below, including a brief overview of the main study design, cost data collection and analysis methods, costeffectiveness analysis methods, and simulation modeling methods. the protocol (pro ) was approved by a single (advarra) institutional review board for all sites on october , . the hcs is an unblinded, multisite, parallel arm, cluster randomized, wait list-controlled trial of the cth intervention implemented in communities in four u.s. states: kentucky, massachusetts, new york, and ohio. wave communities (n = ) will implement cth for years. during this time, the wait-listed wave communities (n = ) will not receive any intervention. in year of the cth intervention, wave communities will begin to implement cth for months. the full effect of cth on wave outcomes is expected to occur after the intervention is fully implemented (by the end of year of the cth intervention.) thus, the primary analysis will compare opioid overdose deaths between wave and wave communities during year of the cth intervention. communities are defined as towns, cities, or counties; more detail on communities is provided in walsh et al. in in press. the goal of estimating startup and ongoing cth implementation costs is to inform other communities outside of hcs about the resources and other investments required to implement cth, supporting both replicability and sustainability. importantly, hcs is investigating the effectiveness, cost, and cost-effectiveness of the cth process and its resulting intervention j o u r n a l p r e -p r o o f priorities and implementation in the real world. hcs does not seek to test the effectiveness or cost-effectiveness of the specific evidence-based interventions selected and implemented in combination by each community. as a result, hcs economic results will inform decision makers about what they can expect in terms of cost and cost-effectiveness before initiating cth, rather than providing generalizable cost or cost-effectiveness estimates for each ebi. we have adopted a micro-costing approach (drummond et al., ; glick et al., ; zarkin et al., ) that first identifies activities required to implement the intervention and then identifies resources required to perform those activities. two perspectives guide our study: ( ) because cost-effectiveness analyses in public health and medicine frequently address decisions made in a health care setting for which substantial costs will accrue to health care payers and providers, the health care sector perspective is prioritized as it is most relevant to those decision-j o u r n a l p r e -p r o o f making stakeholders. hcs is a community-based intervention; therefore, we chose to substitute the community perspective for the health care perspective to reflect costs incurred that are relevant to community decision makers. appendix table illustrates the relevant components for calculating cth costs from the community and societal perspectives, or impact inventory. cth costs to the community arise from two components: ( ) startup investments required to launch the cth in study communities and ( ) labor and non-labor cost inputs for both startup and ongoing implementation costs will be assigned monetary values by multiplying a resource quantity (e.g., an hour of time, a square foot of space, a software license) by an appropriate unit price (e.g., a wage per hour or rental/lease value). ongoing implementation costs for wave communities will generally be those incurred in year and year of the cth intervention, and for wave communities these will be costs incurred in year of the cth intervention. cth is expected to influence resources used by communities as oud prevention and care services are increased or enhanced and individuals receive more services from providers and other organizations implementing ebps from the orcca. the populations tracked for resource use cost estimation will be those targeted by the ebps: individuals with an oud diagnosis or who are receiving oud treatment, or individuals who receive prescription opioids for pain. costs associated with these resources are not included in the community perspective costs. most of the measures of utilization of these services will be derived from secondary data sources (see table ). we will estimate the resource utilization costs for each community by multiplying resource units by relevant estimated unit costs. table lists the resources, measure characteristics, resource data sources, and unit cost data sources for cth. the majority of implementation resources are collected from three sources: ) surveys of intervention participants and key informants regarding the time spent on intervention activities and other non-labor costs; ) hcs study data including, for example, numbers of staff hired for intervention roles by job title; and ) financial documentation, such as invoices for purchased services and materials. to avoid "double counting" resources, the surveys and data analyses separate out costs associated with conducting research from those required to implement the cth. in addition to primary data collection, we will also employ administrative data collection for analysis. we will use state prescription monitoring program databases to obtain quantities of buprenorphine prescribed for treatment of oud and prescription opioids prescribed for pain relief. we will use medicaid claims data to quantify changes in moud, behavioral health treatment, and other health care service utilization. medicaid claims data will be linked with state departments of corrections data to identify the number of linkages to moud treatment among individuals released from jail. from this, we will assign costs to these linkages to care activities. to improve generalizability, we will explore supplementing medicaid data with other sources such as massachusetts' all payers claims databases (apcds). dispensing. similar to the planned analysis of the hcs primary and secondary outcomes (slavova et al., in press) , data collected from each of these sources will be attributed to specific communities based on the locations of individual residences, of providers, or the service or event specific to the intervention time periods described above. wage costs will be based primarily on publicly available data from the bureau of labor care organizations to explore budget impact for these payers. the cost of naloxone kits will be based on two sources. estimates of allowable charges for naloxone in iqvia-provided dispensed prescription data will be used as a cost for pharmacy naloxone kit distribution. for all healthcare units and medication, we will consider alternative price estimates to reflect different perspectives and uncertainty about costs incurred. other published costs or micro-costing (primary data collection and analysis of costs conducted by the hcs team and focused on specific intervention components) estimates will be considered for naloxone distributed by community partners and administered by first responders. finally, some of the orcca ebps implemented in specific communities may use unique j o u r n a l p r e -p r o o f approaches or resources that may not have readily available unit cost estimates in the literature. for example, academic detailing for moud prescribing is being adapted specifically for use in the hcs. when appropriate, micro-costing studies of these interventions will be conducted. we anticipate several types of missing data and nonresponse in primary data collected for the hcs hes. instruments may not be administered during some periods of time (e.g., early in the study for some community coalitions), and data will need to be either collected retrospectively or imputed using appropriate proxies. eligible respondents may not respond at all or may choose not to provide answers to certain questions. we will reduce the amount of missing data and non-response by using multiple modes of data collection (e.g., telephone, email, and inperson interviews) and leveraging collaborative relationships between hcs research staff and community stakeholders. in addition, we will have administrative data to support imputation for most participants. for example, we will have records of who attended coalition meetings and how long those meetings last even if a participant did not complete a coalition meeting survey. finally, we will use model-based imputation (including multiple imputation when appropriate) and sensitivity analyses to account for missing data, sampling variability, and other sources of uncertainty (briggs et al., ; dunlap et al., ; faria et al., ; michalowsky et al., ) . these methods will be conducted within the broader cost-effectiveness analysis described below. our cost-effectiveness methodology follows best practices as described in the literature (neumann et al., ) . following the design of the hcs study, we will calculate incremental effectiveness as the estimated difference in opioid overdose mortality in wave compared with wave communities in year of the cth intervention. during this period of the cth j o u r n a l p r e -p r o o f intervention, orcca ebps will be fully implemented in wave communities, and no hcs interventions will have begun in wave communities. by study design, wave communities do not have any cth implementation costs during the phase where they are in the wait list condition. wave communities are assumed to have zero relevant costs before the beginning of the study. therefore, the incremental cost of community implementation is the estimated total cost needed to implement cth in wave communities in year and year of the cth intervention. the incremental cost of other resource utilization attributable to cth is the estimated difference in costs between wave and wave communities occurring in year of the cth intervention after controlling for potential confounders that were not accounted for in the randomization process (walsh et al., in press) , such as communities' resource utilization in the year before year of the cth intervention or pre-existing infrastructure that can be used to implement cth ebps. such differences may influence which ebps are adopted by each community and the resources needed to implement them. the hcs is systematically collecting data to characterize communities' pre-cth assets and infrastructure and is tracking decisions around ebp selection ebps (knudsen et al., in press) . additionally, information on, funding related to ebps and infrastructure from federal, state, and community sources is being collected, as well as funding received directly from the hcs. all of these factors will be used to evaluate the extent to which selection or other biases were not eliminated through randomization. to the extent that these factors appear salient, we will construct analytic variables that we will include as controls in our cost models. we will estimate the incremental cost-effectiveness ratios (icers) between waves and , defined as the ratio of incremental costs to the difference in opioid overdose deaths, and that j o u r n a l p r e -p r o o f represents the incremental cost per additional opioid overdose death averted. icers will be estimated for both the community and the societal perspectives. outcomes may be influenced by factors not balanced by randomization. in addition to model adjustments described above related to costs, we will incorporate adjustments used for the primary and secondary outcomes analyses in our cost-effectiveness analyses (slavova et al., in press ). we will use monte carlo, nonparametric bootstrapping (e.g., dunlap et al., ) , or parametric methods (e.g., murphy et al., ) to characterize joint parameter uncertainty around our cost and icer estimates (e.g., adjusted standard errors, confidence intervals). the methods will account for missing data and measurement error when data are observed and will be used jointly with multiple imputation and other sensitivity analyses to provide a comprehensive set of cost and cost-effectiveness results with well-characterized uncertainty. we will also consider alternative values of key parameters or assumptions (parameter uncertainty) in sensitivity analyses. these alternative analyses range in complexity from simply including "high" and "low" alternative value scenarios to sampling explicitly from specified probability distributions of possible ranges for cost and effectiveness. to evaluate the cost-effectiveness results, we will assess how stakeholder willingness to pay (wtp) affects the results. cost-effectiveness acceptability curves (ceacs) will be an important tool for exploring the probability that cth is cost-effective compared with no cth intervention over a range of stakeholder wtp values of cost per opioid overdose death avoided. ceacs incorporate the joint variability of the cost and outcome estimates and show the probability that an intervention is the cost-effective choice as a function of the policy maker's wtp over a range of values (e.g., $ , to $ , wtp per opioid overdose death avoided) (see neumann et al. [ ] and murphy et al. [ ] for qaly wtp examples). j o u r n a l p r e -p r o o f the hcs hes will use simulation modeling to evaluate the short-and long-term health and economic impacts of the cth intervention and to develop a policy tool for communities that want to implement cth. simulation modeling will be used to extend the cost-effectiveness analysis described above to alternative community scenarios and longer time horizons. models will provide a "lifetime" time horizon and enable us to estimate the cost-effectiveness of the hcs intervention on a cost per qaly gained basis. the second panel (neumann et al., ) recommends using qalys as an outcome measure in cea, which are years of life saved adjusted by the quality of those years. qalys are useful because they combine mortality and morbidity into a single metric, reflect societal preferences for the value assigned to each year of life, and can be used as a standard measure of health gains across diverse treatments and settings (neumann and cohen, ) . cost per qaly estimates will allow us to compare the economic value of hcs with other community-based interventions. the hcs hes will benefit from simulation models that will be developed through a collaborative approach among modelers representing each research site and the data coordinating center. several modeling approaches will be used, including agent-based modeling (abm), microsimulation, system dynamics, and dynamic compartmental models (neumann et al., ) . throughout the model building and estimation processes, modelers will share progress and compare key model outputs, enabling model cross-validation. although models will be built independently, they will share parameterization approaches and be subject to the scrutiny of other modelers, thus improving the face validity and internal validity of each model (eddy et al., ) . developing simulation models will also benefit from continuous engagement with j o u r n a l p r e -p r o o f community stakeholders, so modelers can ensure the most up-to-date inputs are used and that the outcomes are of greatest interest to decision makers. all models will simulate the trajectory of opioid use disorder by modeling transition of people to different stages of opioid use including opioid misuse and oud, remission, relapse, fatal and nonfatal overdoses, and/or death (chen et al., ; mclellan et al., ) . this structure enables ebps adopted in each community to potentially alter individuals' trajectories, which will subsequently affect their lifetime costs, mortality, and quality of life. models will use two types of parameters, those that are context-specific to communities and those that can be applied more generally across communities-parameters that characterize details of oud epidemiology and treatment seeking behaviors are context specific, while parameters about the pharmacologic efficacy of treatment and quality of life with oud are generalizable. context-specific parameters include population size, prevalence and incidence of opioid misuse, prevalence of oud, opioid overdose deaths, other causes of death, moud admissions, and moud retention. general parameters include moud efficacy and the relative risk of death both during and immediately following moud treatment. other general parameters are utility weights representing quality of life with oud or opioid use. these inputs will be drawn from national or publicly available data sources (e.g., cdc) and published literature. we will account for parameter uncertainty as described in section . . key model outputs will include temporal trends in population health outcomes such as fatal and nonfatal opioid overdoses, number of individuals misusing opioids, number of individuals with oud, number receiving and maintaining use of moud, and naloxone coverage. additionally, models will estimate the long-term impact of ebps on costs, opioid overdoses, life years gained, and qalys. although hcs will not examine the effectiveness of j o u r n a l p r e -p r o o f individual ebps, the models could account for synergies among multiple practices to explore the relative impact of different intervention combinations beyond the short-term period of the ebp interventions implemented at the research sites. the models can thus inform optimal resource allocation at the community level to achieve a targeted reduction in opioid overdose deaths. one of the goals of the hcs hes is to inform cth implementation decisions by communities not participating in the hcs. we will develop the opioid policy simulator which will be an interactive online translational tool for policy makers and non-hcs communities to use as they plan their approaches to reducing opioid overdoses. one of the hcs modeling teams has previously developed similar web-based tools for hepatitis c infection (chhatwal et al., ) and for covid- (www.covid sim.org). the inputs and outputs of simulation models will feed the simulator through a "metamodeling" approach (ferreiro-cabello et al., ) , and the simulator will provide outcomes like the numbers of opioid overdose deaths, nonfatal overdoses, number of people with oud, and cost-effectiveness of ebp interventions in different communities. it will also allow users to explore economic impact on budgets for specific payers, like medicaid. users will be able to use the simulator to assess the health and economic impact of opioid policy scenarios in other areas of the country impacted by the opioid crisis, allowing translation of hcs results to these communities. the hcs offers an unprecedented opportunity to conduct health economics research on solutions to the opioid crisis. hcs is the largest implementation science, addiction research study ever conducted in the united states. implemented in communities across four states, cth combines a community-level approach with combinations of ebps and communication j o u r n a l p r e -p r o o f campaigns to significantly reduce opioid-related overdose fatalities. the breadth and scale of cth offers a rich environment from which to draw lessons learned for other communities combating the opioid crisis and presents a unique opportunity for health economics analyses. the hcs hes complements the hcs by providing critical estimates of the resources needed to implement cth, its broader impact on societal resources, and its cost-effectiveness for avoiding overdose deaths. simulation modeling will incorporate these results to characterize cth's impact better, allowing for extrapolation of results to non-hcs communities to support planning and policy making around similar interventions. the models developed will evaluate combinations of different interventions and consider synergies across interventions in the continuum of oud prevention, harm reduction, and treatment-an endeavor that no oud simulation model has yet achieved (barbosa et al., ; nosyk, ) . the hcs is a complex and challenging intervention, and the hes has several limitations. first, implementing cth relies on the efforts of numerous and disparate individuals and organizations within and outside cth communities who will be engaged in a variety of different implementation activities. collecting accurate and representative information about how individuals spend their time to implement cth without causing excessive respondent burden is challenging. it requires flexible and tailored data collection instruments and strategies to make use of alternative data sources, including administrative data and literature-based estimates. second, defining measures of resource utilization changes associated with cth is also challenging, especially for the societal perspective. for example, a key limitation of the economic study is that change in health care costs over the study period will be measured using primarily using data on medicaid-enrolled individuals. this population represents a large portion of all individuals with an oud diagnosis or who are receiving oud treatment. (orgera and j o u r n a l p r e -p r o o f ( ) to account for the opioidrelated impacts of the pandemic through simulation modeling. in conclusion, the hcs hes includes economic evaluation and simulation modeling components that will provide valuable insights for both hcs and non-hcs communities. as policy makers and other stakeholders address the devastating effects of the u.s. opioid crisis, these data will show how community and societal resources can be deployed most effectively to reduce opioid overdose deaths. the content is solely the responsibility of the authors and does not necessarily represent the official views of the national institutes of health or its nih heal initiative. iqvia data are not reported in this paper, but iqvia is mentioned as a data source; therefore, the manuscript was reviewed and approved by iqvia. economic evaluation of interventions to address opioid misuse: a systematic review of methods used in simulation modeling studies cost-effectiveness of integrating buprenorphinenaloxone treatment for opioid use disorder into clinical care for persons with hiv/hepatitis c co-infection who inject opioids cost and utilization outcomes of opioid-dependence treatments societal costs of prescription opioid abuse, dependence, and misuse in the united states hospital-based addiction medicine healthcare providers: high demand, short supply missing... presumed at random: costanalysis of incomplete data decision modelling for health economic evaluation cost-effectiveness of emergency departmentinitiated treatment for opioid dependence modelling in economic evaluation: an unavoidable fact of life healthcare costs and utilization associated with high-risk prescription opioid use: a retrospective cohort study prevention of prescription opioid misuse and projected overdose deaths in the united states hep c calculator: an online tool for cost-effectiveness analysis of daas cost-effectiveness of distributing naloxone to heroin users for lay overdose reversal methods for the economic evaluation of health care programmes screening and intervention for suicide prevention: a cost-effectiveness analysis of the ed-safe interventions re-engineering methadone-cost-effectiveness analysis of a patient-centered approach to methadone treatment ispor-smdm modeling good research practices task force, . model transparency and validation: a report of the ispor-smdm modeling good research practices task force- a guide to handling missing data in cost-effectiveness analysis conducted within randomised controlled trials metamodel-based design optimization of structural one-way slabs j o u r n a l p r e -p r o o f based on deep learning neural networks to reduce environmental impact the economic burden of prescription opioid overdose, abuse, and dependence in the united states a cost-effectiveness analysis of opioid substitution therapy upon prison release in reducing mortality among people with a history of opioid dependence economic evaluation in clinical trials costeffectiveness analysis of a large jail-based methadone maintenance treatment continuation program in new mexico the economic burden of opioidrelated poisoning in the united states national and state treatment need and capacity for opioid agonist medication-assisted treatment model and approach to assessing implementation context and fidelity in the healing communities study cost-effectiveness of publicly funded treatment of opioid use disorder in california cost-benefit analysis of communities that care outcomes at eighth grade benefit-cost analysis of a randomized evaluation of communities that care: monetizing intervention effects on the initiation of delinquency and substance use through grade health communication campaigns to drive demand for evidencebased practices and reduce stigma in the healing communities study the economic burden of the opioid epidemic on states: the case of medicaid community engagement to implement evidence-based practices in the healing communities study monetary conversion factors for economic evaluations of substance use disorders they're making it so hard for people to get help:" motivations for non-prescribed buprenorphine use in a time of treatment expansion drug dependence, a chronic medical illness: implications for treatment, insurance, and outcomes evaluation assessment of opioid prescribing practices before and after implementation of a health system intervention to reduce opioid overprescribing is the whole larger than the sum of its parts? impact of missing data imputation in economic evaluation conducted alongside randomized controlled trials medications for alcohol and opioid use disorders and risk of suicidal behavior, accidental overdoses, and crime cost-effectiveness of buprenorphine-naloxone versus extended-release naltrexone to prevent opioid relapse economic evaluations of opioid use disorder interventions cost-effectiveness of extended release naltrexone to prevent relapse among criminal justice-involved a history of opioid use disorder qalys in -advantages and concerns updating cost-effectiveness--the curious resilience of the $ , -per-qaly threshold cost-effectiveness in health and medicine how simulation modeling can support the public health response to the opioid crisis in north america: setting priorities and assessing value long-term effects of the communities that care trial on substance use, antisocial behavior, and violence through age years the opioid epidemic and medicaid's role in facilitating access to treatment modeling health benefits and harms of public policy responses to the us opioid epidemic association of current and past opioid use disorders with health-related quality of life and employment among us adults the prevalence and cost of medicare beneficiaries diagnosed and at risk for opioid abuse, dependence, and poisoning key substance use and mental health indicators in the united states: results from the national survey on drug use and health. center for behavioral health statistics and quality, substance abuse and mental health services administration exploring the economic effects of the opioid epidemic revenue losses to state and federal government from opioid-related employment reductions operationalizing and selecting outcome measures in the healing communities study mortality risk during and after opioid substitution treatment: systematic review and meta-analysis of cohort studies costeffectiveness analysis of alternative naloxone distribution strategies: first responder and j o u r n a l p r e -p r o o f lay distribution in the united states drug and opioid-involved overdose deaths-united states the opioid-overdose reduction continuum of care approach (orcca): evidence-based practices in the healing communities study treatment utilization among persons with opioid use disorder in the united states the substance abuse services cost analysis program (sascap): a new method for estimating drug treatment services costs all authors have contributed materially in developing the overall health economics study design.all authors also contributed to manuscript preparation, reviews, and substantial edits. all authors read and approved the final manuscript. the authors have no conflicts of interest to disclose. all authors have reviewed and approve the final manuscript. we acknowledge members of the hcs health economics workgroup who have provided ongoing feedback and other assistance for the study protocol, data collection approaches and orme, rasika ramanan, jackson luckey, colleen watson, and brent gibbons). we also thank others across the hcs consortium for overall project leadership, supporting the health economics study and thoughtful inputs to help the health economics study coordinate both scientifically and practically with this large and complex study. key: cord- -jck f nx authors: ruffino, paolo; jarre, matteo title: appraisal of cycling and pedestrian projects date: - - journal: nan doi: . /bs.atpp. . . sha: doc_id: cord_uid: jck f nx cycling and walking have gained a prominent role in the mobility policy agenda as awareness has risen over the growing unsustainability of the current transport system and the multiple co-benefits of active mobility. as interest and investments for cycling and walking increase, how active mobility can be appraised becomes a crucial question, which has been tackled over the years through different methods and tools. the aim of this chapter is to provide a structured review of the methods and the practices of appraisal of walking and cycling policies and projects, focusing on both traditional and emerging assessment techniques. at present, much attention has been paid to the application of four main traditional methods: balance sheet calculations, cost-benefit analysis, cost-effectiveness analysis and multi-criteria analysis. we compare and discuss these methods to identify strengths and weaknesses for each of them, as well as their main limitations and knowledge gaps in their application. we conclude that over the last decades much effort has been undertaken to further expand and develop these tools thanks to an increased attention to walking and cycling. however, much research is still needed, particularly in the quantification and valuation of specific effects within cost-benefit analysis and in better integrating different appraisal techniques. finally, the impact of appraisals on decision-making outcomes is still underexplored. over the last two decades, interest in the promotion of active mobility, namely walking and cycling, as alternative form of urban transport has grown consistently (banister, ; buehler and dill, ; buehler et al., ; pucher and buehler, ) . on one hand, concerns have risen about externalities of the current (car-centric) transport system; on the other hand, a better understanding and awareness of the multiple co-benefits of active mobility in terms of health, efficiency and social inclusion is emerging (banister, ; g€ arling et al., ; g€ ossling et al., ; mueller et al., ) . increased planning and financing activity has followed, targeting larger infrastructure projects such as bicycle and pedestrian networks, cycling highways, mass bicycle parking, diffused traffic calming measures and carfree areas, as well as the experimentation of behavioral interventions such as (non) monetary incentives such as bike-to-work or walk-to-school programs (banister, ; bertolini and le clercq, ; braun et al., ; martens, ; pucher and buehler, ; pucher et al., ) . more recently, the covid- pandemic has urged public authorities in many countries to promote walking and cycling even more vigorously as a way to limit the spread of the virus, address physical inactivity and also prevent a mass shift from public transport to private car that would worse preexisting traffic conditions (iea, ; world health organization, ) . for example, in alone the uk government approved a £ billion package for active mobility and green transportation (uk , the italian ministry of transport allocated over € million for urban cycling infrastructure (italian , and many other countries have taken similar initiatives. these developments pose two main challenges: ( ) a planning challenge: as investments and projects' size increase, questions arise on the feasibility, efficiency, and prioritization of measures (aldred et al., ; bloyce and white, ) . ( ) a political challenge: as changes in the functions of public space (e.g., the removal of car parking to add bicycle lanes) have historically been met with suspicion and sometimes public outcry, a need to build greater stakeholder support and acceptance using rational arguments emerges (see for instance: aldred et al., ; bloyce and white, ; oldenziel and albert de la bruhèze, ) . transport appraisal attempts to address these challenges by supporting decision-makers in forming a rational opinion about the strengths and weaknesses of alternative options (priemus and van wee, ) . in many countries in europe, the us and australia, standardized frameworks exist to appraise "traditional" transport infrastructure projects, such as highways or railways lines, using methods such as cost-benefit analysis (cba) mackie et al., ) . however, it is neither a common nor an institutionalized practice to appraise cycling and walking projects, as they usually entail lower costs and risks (van wee and b€ orjesson, ) . this is especially the case for countries with low cycling and/or walking rates, but also in the netherlands and denmark-where active mobility is widespread-thorough evaluations are not regularly performed (ibid.). this can be self-defeating, as either too much or too little resources may be allocated, possibly preventing the realization of greater benefits (b€ orjesson and eliasson, ) . nevertheless, interest on how active mobility projects could be appraised is growing rapidly among governments, practitioners and academics, and multiple approaches are being explored. recently, governments are starting to include walking and cycling in their own evaluation frameworks (see, for example, uk department for transport, ); countries such as denmark, the netherlands, germany and sweden have commissioned guidelines and studies to identify applicable unit costs for cba to cycling and walking schemes within their own territories (see for instance cowi and city of copenhagen, ; decisio, decisio, , . a number of transnational research projects have been funded that among others focus (or have focused) on walking and cycling projects and policies appraisal, such as the eu projects pasta, flow, handshake. international agencies have published tools (such as the heat tool from the world health organization) to support urban planners, professionals, and community leaders in performing economic assessment of the health effects of projects aiming to increase walking and cycling rates (deenihan and caulfield, ) . in this chapter, we examine how appraisal methods commonly used in the transport sector are applied to evaluate walking and cycling projects. the main goal is to observe what strengths, weaknesses, and limitations these appraisal methods entail when applied to walking and cycling. for doing so, we first provide an overview of the main tools used to assess transport policies and projects. secondly, we examine the literature and provide examples of where such techniques have been applied, together with a critical discussion of such application(s). finally, we draw some concluding remarks and implications for further research and policies. . transport appraisal methods and assessment criteria . overview of appraisal methods for transport policies transportation networks provide multiple benefits in terms of accessibility to people, goods and services but they may also be the source of social, environmental and economic impacts. for this reason, decision-makers should, appraise how different policy options trade-off when planning new infrastructure. in a nutshell, project appraisal is the process of evaluating (i.e., attaching a value to) a policy or project outcome with the intent to assess its particular desirability condition (efficiency, effectiveness, etc.) before the implementation, in order to judge the strengths and weaknesses of a particular course of action using a common framework (rossi et al., ) . this should enable decision-makers to rank their preferences and deal with multiple stakeholder interests and perspectives over the same issue. in the field of transport, scholars have proposed several techniques to appraise projects and policies (bakker et al., ; browne and ryan, ; grant-muller et al., ; mackie et al., ) , with the most widely used being: bsc is typically the first step of any assessment upon which other methods are built, while cba, cea and (in most cases) bsc belong to "mono-criterion" assessment methods, as they consider a single and specific objective, mca is "multi-criteria" as it attempts to deal with a plurality of objectives (dean, ) . the main characteristics of each method are summarized in table , and a short description is provided, while we refer to the specific chapters for a more in-depth discussion. the balance sheet calculation (bsc) is the simplest among transport appraisal methods. it consists of the separate observation of a number of selected criteria and effects upon which decision-makers draw their own conclusions (typically the intervention costs, supplied by extra information about specific effects, such as traffic impacts) (bakker et al., ) . balance sheets-particularly the cost-analysis, business cases and technical-financial feasibility studies-represent the basic input to other assessment methodologies discussed in this chapter. in general, this approach has the benefit of being quick and cheap, but the assessment of the broader consequences is often limited to the decision-makers' intuition. cost-benefit analysis (cba, or social cost-benefit analysis for completeness) stands out traditionally as the most common appraisal method for large transport infrastructure projects (see boardman and pearson education, ) . cba is grounded in welfare economic theory and it measures changes in society's welfare (expressed as the aggregation of all individual utilities or willingness-to-pay) resulting from the implementation of a specific project or policy (boadway and bruce, ) . the analysis' object is said to be "desirable" or "socially efficient" if it satisfies the kaldor-hicks criterion, namely if the sum of gains outweighs the sum of losses and therefore losers are theoretically compensated by winners (hicks, ; kaldor, ) . hence, in performing a cba, all the quantifiable effects (direct and indirect) revolving around a policy or a project are listed and monetized (as costs and benefits) during a specific timeframe (usually the project lifespan). these monetized effects, and the associated investment costs, are then discounted to the present value of money (or net present value, npv) and results are typically expressed as a benefit-cost ratio, which is the means to verify if the kaldor-hicks criterion is met (boadway, ) . a cba may be used to compare either an intervention scenario with a do-nothing (or do-minimum) scenario or different courses of action. cba may also be performed ex-post in order to verify the accuracy of the initial predictions and/or to monitor the effects and promote policy learning (i.e., eliasson et al., ) . in the field of transport, the quantification of the effects often relies on transport models, which provide the necessary inputs for a cba (such as changes in travel times, emissions, etc.) (priemus and van wee, ) . cost-effectiveness analysis (cea), also known as cost-utility analysis (cua) in health economics (robinson, ) , is a form of cost-benefit analysis that focuses on a single, non-monetized effect or outcome which is compared to the costs of different courses of action (browne and ryan, ) . in this way, decision-makers are informed about which measure ensures that a goal will be reached at the minimum cost. cba and cea follow similar research techniques and principles; the latter, however, is limited to a narrowly defined goal. in the field of transport and environmental policy, cea is used especially for the so called "optioneering," i.e., the comparison of multiple options with a specific set of outputs in order to rank priorities by cost-effectiveness (bakker et al., ) . a typical example of cea application in transport is the ranking of projects by cost per unit of emission reduced (see kampman et al., , for example). multi-criteria analysis (mca) allows to select alternative projects by considering multiple weighted monetary and non-monetary criteria (bakker et al., ; beinat, ; browne and ryan, ) . the weighing of criteria can be performed in a participatory setting to include expert and stakeholders' opinions in order to balance trade-offs among different goals and needs advocated by different actors (dean, ) . several approaches to mca exist, ranging from formal (continuous and discrete) to simplified methods, the most common being: analytical hierarchy process (ahp), analytical network process (anp) and regime, electre (etc.) (see dean, for a classification of methods). each method presents unique features as well as advantages and disadvantages (ibid.). the process to draft a mca follows in general five main steps: ( ) the project and its alternative(s) are defined; ( ) the judgment criteria, weighing and ranking method are determined; ( ) the impacts of the project and its alternatives are analyzed; ( ) the impacts are categorized in as list of criteria that are weighted; ( ) the judgments may be aggregated into a single criteria depending on the chosen approach (ibid.). in the field of transport appraisal, mca is the most common alternative to cost-benefit analysis as it allows to consider effects that are typically difficult to quantify and monetize (such as social inclusion, aesthetics, image, equity, etc.) (browne and ryan, ) . a significant body of literature has analyzed appraisal methods for active mobility in the past. as the aim of this paper is to focus on strengths and weaknesses of the different methods, we selected relevant papers to illustrate the application of appraisal tools and focused on the methodology and process of construction of the appraisal technique more than on the results of each application to the context of cycling and walking. in this perspective, the results of the evaluations are of less interest than the applicability of the proposed methods. therefore, the selection has discarded papers which did not offer new insights on the choice of appraisal method or discussed its applicability and limitations; the initial selection has built upon previous systematic reviews (mainly brown et al., ; cavill et al., ; mueller et al., ) that have addressed appraisal methods and similar research questions in the past. in addition, such sources were integrated and corroborated by: (d) interviewing active mobility experts from several municipalities within the project civitas handshake in order to have a more comprehensive understanding of how decisions are formed within municipalities and obtain relevant examples. the inclusion criteria that were adopted for selecting papers about each method are summarized in table . despite being the most common way in which public authorities perform appraisals on walking and cycling projects, the academic debate on bsc is surprisingly scarce. in part, this might be due to the fact that bsc usually represents only the first phase of a more thorough appraisal. much of the publicly available knowledge is gray literature in the form of technicalfinancial feasibility studies (see for instance centraal utrecht , ; st. luis (city government), ; opus consultants, ) . study designs also vary depending on the laws and standards applied in each country, as well as the context-specific needs. in general, the content of such studies can be narrowed down to three main components: . a general description of the intervention site. . a preliminary technical design and cost estimate of the proposed solution, and its alternative(s). . occasionally, the previous steps are supplemented by qualitative judgments on strengths and weaknesses, sensitivity analyses, traffic impact studies, environmental assessments, etc. typical examples of "balance sheet calculations" are the financial analyses performed on bicycle sharing projects and large bicycle parking facilities to verify costs and revenues of their operation in order to determine adequate budgeting. for instance, in the municipality of utrecht, dutch railways, and prorail (the owner of the railway lines) performed a business cases and scenario analyses when redesigning utrecht's train station; such redesign included the construction of several large bicycle parking facilities. the bsc was necessary to estimate the financial impacts of different daily/monthly tariffs under several assumptions (daily users, parking duration, quality of service) in order to quantify the costs and revenues and determine a possible management agreement (centraal utrecht . the resulting costs for the different scenarios are summarized in table . this type of analysis considers mainly financial effects. in the reported example, the construction, maintenance, enforcement, exploitation costs, incidental costs etc., were included, while revenues consisted of tariffs, taxes, sales etc. sensitivity analyses were included in order to allow decisionmakers to understand the order of magnitude of the financial implications. bicycle share programs are another example that is typically evaluated using bs studies (st. luis (city government), ) . these studies start by analyzing the potential demand in the area to identify adequate locations of bicycle docking stations; next, the costs of the program for a variable number of years are estimated as well as the revenues of multiple financial plans using scenarios. in most cases, since the demand for such infrastructure projects is complex to determine, significant hypotheses must be introduced and then tested through sensitivity analyses on key parameters (such as duration, trip frequency and modal shift) (ibid.). through bsc the broader social, economic, and environmental effects are not systematically captured, thus leaving the judgment about the merits and flaws of the proposal to intuitive assessments. cba is currently among the assessment techniques that have received the most attention from both practice and academia over the last two decades (van wee and b€ orjesson, ). a "typical" cba study applied to walking and cycling does not substantially differ from its counterpart applied to other modes, as it consists of: . a general description of the intervention site. . an analysis of the reference scenario (which usually accounts for a "donothing" or "do-minimum" policy intervention) to be used as benchmark. . a description of the intervention scenario (including costs and risks), alternatives, and a causal model to quantify the effects. . a monetarization of the expected effects that revolve around a project's lifetime, and a comparison to the costs at the npv. . a (optional) sensitivity analyses to test the effect of some key parameters to the end result. in contrast to bsc, which is limited to an analysis of financial cash flows only, a cba provides a more comprehensive picture of all the welfare effects revolving around a measure which would otherwise be underexposed. an example of cba that included walking and cycling among other modes is the study performed by the city of amsterdam in to appraise different solutions to improve the connection between the city centre and the expanding neighborhood of amsterdam-noord across the river ij (hoefsloot et al., ) . the explored solutions included the improvement of the current ferry system, the construction of a pedestrian and bicycle bridge (including different design variants), the construction of a tunnel under the river ij, the construction of a metro station and pedestrian tunnel, and several "packages" of different measures. in total, (combinations of ) measures were tested in two development scenarios (high and low growth scenario) using amsterdam's transport model. table is an excerpt that illustrates some of the results. in the scientific literature, elvik ( ) was among the first scholars to critically discuss the application of cba on measures designed to improve safety or mobility for pedestrians and cyclists. in doing so, he applied the best available knowledge of the time to a hypothetical case in order to identify a research agenda. what he found is still relevant nowadays and concerns four main aspects: (a) how to determine changes in the amount of walking and cycling; (b) how to value changes in travel time for pedestrians and cyclists; (c) how to measure changes in road user insecurity and feeling of safety; (d) how to determine and value changes in the health state. his analysis indicated that the inclusion of these effects could make a major difference in the results of cba. later, saelensminde ( ) published one of the first "complete" cba study on walking and cycle tracks in three norwegian cities. the study included for the first time (a) the health benefits associated with increased active mobility rates, (b) reduced external costs from motorized traffic and (c) reduced parking costs. as hypothesized by elvik ( ) , the inclusion of these social effects meant that the benefits of investment in active travel networks could be as high as - times the costs. however, the study also acknowledged that improvements in the valuation of some effects as well as more information on the relationship between physical activity and the incidence and costs of different diseases were needed in order to make more accurate estimates. finally, the traffic accidents effects of a modal shift from car and public transport to cycling were deemed unclear (ibid.). multiple studies have since been published that have further explored the application of cba to walking and cycling infrastructure in different contexts and attempted to address various knowledge gaps. three systematic reviews of the literature have been published between and (see brown et al., ; cavill et al., ; mueller et al., ) . in general, most studies have found that investing in cycling and walking usually carries a positive effect on society because of lower road externalities, particularly when the shift occurs from car travels(ibid.). when losses occur, these are usually due to the missed collection of car and fuel taxes or when a policy fails to generate enough demand for a project (g€ ossling et al., ; litman, ) . the effects with the most significant impact are the reduced health-related costs and travel time gains (especially due to decongestion). hence, the results of cbas have been used to harness support among stakeholders by showing that promoting more walking and cycling would create a win-win situation and deconstruct policy frames that marginalize cycling and walking as recreational activities (aldred, ; bloyce and white, ) . however, it is unclear from the literature how these results affect the outcomes of decision-making processes. over the years, a relevant body of research has attempted to fill the knowledge gaps about the estimation and valuation of specific effects of active mobility and other consequences of changes in travel habits. notably, hopkinson and wardman ( ) , wardman et al. ( ) , ramjerdi et al. ( ) and b€ orjesson and eliasson ( ) have focused on estimating the value of travel time reductions and improvements in perceived safety for different types of roadway improvements, finding that cyclists have higher value of times than other mode users due to the physical effort involved. studies in the health and epidemiology domain have found positive effects of walking and cycling in reducing all-cause mortality (kelly et al., ) , lowering absenteeism (de hartog et al., ) , improving fitness and productivity levels (etemadi et al., ; walker et al., ; wattles et al., ) , and reducing the cost of several illnesses (kahlmeier et al., ) . a push to the development and use of cba for walking and cycling projects has indeed come from the health sector. notably, the who made an important contribution by publishing the health economic assessment tool (heat) for walking and cycling (kahlmeier et al., ; world health organization, ) which is grounded in some of the studies cited before (kelly et al., in particular) . this planning-support tool, based on cba principles, aids planners and advocates in estimating the value of reduced mortality and other externalities that results from a shift to regular walking and cycling and compare the monetized effects with the costs of a measure. despite the limitation arising from its "simplified" dashboard-like functioning, the heat tool has contributed to increasing the popularity in both academia and practice of health-economic assessments. for instance, fishman et al. ( ) used heat to quantify the population-level health benefits of cycling in the netherlands, finding that over deaths are prevented each year and dutch people have half-a-year-longer life expectancy thanks to high cycling levels with respect to a non-cycling base. cba has also seen applications to assess non-infrastructure projects such as mandatory helmet laws (sieg, ; taylor and scuffham, ) , programs that encourage active travel habits (beale et al., ) , changes to the built environment (guo and gandavarapu, ) , bicycle share programs (bullock et al., ) as well as integrated active travel policies (chapman et al., ) . moreover, cba has been used to appraise measures at different levels: from site-specific interventions-such as bicycle and pedestrian trails and bridges (hoefsloot et al., ; li and faghri, ) -to changes at the network level (beria and rafaele, ; brey et al., ; gotschi, ) . cba is generally applied in ex-ante, while ex-post cbas of active mobility projects are limited in the literature (one example is chapman et al., ) . moreover, studies that have compared ex-ante with ex-post cba to validate the results of previous appraisals are not present in the literature. cba frameworks have also been used to compare the different societal costs imposed by different transport modes (including walking and cycling) on society in order to advocate in favor of more sustainable transport but also in order to include a wider array of effects in evaluations. for instance, g€ ossling and choi ( ) found that in copenhagen the societal costs borne to society from each km traveled by car is more than six times higher than the same km traveled by bike, if all effects are included (especially health). similarly, g€ ossling et al. ( ) estimated that the total passenger-kilometer driven by car in the european union impose an external cost of more than € billion per year, while cycling and walking kilometers, due to positive health effects, are worth € billion and € billion per year respectively. a major point when it comes to cba is the demand forecasting of future infrastructure projects. in the literature, multiple approaches have been proposed, ranging from simple assumptions to more complex approaches depending on the tackled research question, as well as the level of detail and data available. the approach employed by saelensminde ( ) and gotschi ( ) is most commonly adopted: present volumes of pedestrian and bicycle traffic are estimated using average statistical figures, sometimes supplemented by traffic counting and surveys, whereas future induced volumes are estimated using assumptions accompanied by sensitivity analyses to account for uncertainty (ibid.). more complex approaches involve the use of potential analysis scans to identify short car trips (lovelace et al., ) , system dynamic modeling techniques that capture positive and negative feedback loops (macmillan et al., ) and traditional multi-modal traffic simulation models to better capture changes in consumers' surplus (as in the case of beria and rafaele, ; hoefsloot et al., ) . cea is another common assessment method for appraising walking and cycling measures, more so in the field of health economics then transport economics (abu-omar et al., ) . that is the case because the promotion of safe walking and cycling is seen by many health authorities-such as the who-as a prevention policy to tackle the risks associated to physical inactivity (world health organization, ). in fact, multiple studies over the years have tested the effectiveness of different programs (including the promotion of active mobility) aiming at reducing physical inactivity against their cost (some systematic reviews have been conducted by campbell et al., ; garret et al., ; mueller-riemenschneider et al., ) . for example, wang et al. ( ) performed a cea of bicycle and pedestrian trails to illustrate how cost-effectiveness changed depending on the activity levels of the population. cobiac et al. ( ) performed a cea to measure the health outcomes against the costs of six different physical activity interventions compared to identify the most cost-effective option (the comparison included travel smart programs that rewarded travelers for reducing car trips and choosing to walk and cycle). in the majority of studies produced in the field of health economics, the cost-effectiveness is expressed in terms of a ratio of gained health (usually expressed as quality-adjusted life years or qaly) or averted dalys (disability-adjusted life years) to the costs required to achieve a unit of result. in the field of transport economics, cea considers also other traffic-related effects, namely road crashes costs, pollution, congested hours as goal criterion. for example, hatziandreu et al. ( ) applied cea to three different approaches (law enforcement, community-based and school education) aiming at promoting the use of bicycle helmets among pupils. their study used pre-post data and compared the costs of the program with the effect in terms of bicycle-related head injury and deaths. other studies, such as peters and anderson ( ), wijnen et al. ( ) and jiao et al. ( ) , applied cea to measure the efficacy of traffic calming aiming at reducing accidents costs. others such as gunn et al. ( ) have focused on the effects of sidewalks to increase levels of transport walking and related health effects, while gu et al. ( ) analyzed the cost-effectiveness of bicycle lanes as means to both improve health of the general population and reduce crashes. cea is also often used as an instrument to prioritize program investments. in the field of cycling, a simple example is the study conducted by the city of san donato milanese (italy) (ruffino and jarre, ) in which the investment priority in cycle routes was sorted by means of a cea using an accessibility index as effectiveness criterion. the goal of the administration was in fact to provide a transport option alternative to the car to the largest number of residents, commuters, school pupils etc., at the lowest price. the study therefore followed these steps: . the investment costs for each bicycle route was determined. . an accessibility index of each cycle route was defined that fitted the administration goals. . a ratio between the km-costs and the index was performed in order to determine the cost-effectiveness. . the cycle routes were sorted by least cost in order to determine the intervention with the highest effectiveness at the lower costs. similar studies have been performed in other contexts using more complex methods. for instance, to determine investment priorities in bicycle highways in the haaglanden (conurbation surrounding the hague in the netherlands), a transportation model was used to calculate the costeffectiveness of bicycle highways in terms of reduced short car trips and congested hours as effectiveness criterion (decisio, ) . cea has been applied both in ex-ante and ex-post studies using different methods: ex-post studies have mainly used direct pre-post measurements and/or (interrupted) time series, sometimes complemented by surveys (self-report, etc.); on the other hand, ex-ante studies relied mostly on scenarios, using a variety of statistical techniques (ranging from simple trend analysis to regression analyses and markov models) and applying sensitivity analysis to assess the robustness of the obtained results. for instance, moodie et al. ( ) measured the cost-effectiveness of school programs to increase active mobility among pupils aged - by sharing of a small pilot survey and then extrapolated the results to the entire pupil population of australia. dallat et al. ( ) used a quasi-experimental before-and-after household survey and different scenarios to measure the cost-effectiveness of urban greenways in improving physical activity levels. gu et al. ( ) used regression analysis to calculate the effect of marginal improvements of bicycle lanes in nyc in terms of ridership in order to assess the related health effects. conventional reductionist approaches have been criticized for leading to sub-optimal decisions due to the inherent complexity of sustainability dilemmas, such as transport policies (browne and ryan, ; gasparatos et al., ; omann, ) . in this perspective, mca is increasingly being proposed as a viable alternative also in the field of walking and cycling appraisal (glavic et al., ; gris e and el-geneidy, ) since: • there is a need to include and deal with effects that are typically difficult to quantify and monetize yet relevant for planning walking and cycling infrastructure (such as comfort, aesthetic quality etc.) as well as addressing equity questions. • secondly, there is a need to incorporate opportunities and risks related to the type of infrastructure measure proposed. • finally, stakeholders' views and equity issues can be better represented by assigning weights. in particular, mca integrated with gis (also defined as mcdm-gis) is becoming increasingly popular to appraise walking and cycling projects (larsen et al., ; rybarczyk and wu, ) . for example, larsen et al. ( ) and rybarczyk and wu ( ) were among the first scholars to propose mca and gis to identify and prioritize investments by integrating both supply-and demand-analysis criteria for cycling planning. later, milakis et al. ( ) and milakis and athanasopoulos ( ) expanded on this approach including inputs from cyclists in a participatory setting to plan athens' metropolitan cycle network. guerreiro et al. ( ) applied mca, gis and data mining techniques to plan and compare the investments in a cycling network. canu et al. ( ) proposed spatial mca for the assessment of walkability of intersections and the prioritization of pedestrianoriented policies. kent and karner ( ) explored the application of gis-mca to prioritize low-stress and pleasant bicycle routes. besides traditional infrastructures, spatial mca has been widely applied to bicycle share systems analyses (croci and rossi, ; kabak et al., ; milakis and athanasopoulos, ; moshref javadi et al., ) . however, to the best of our knowledge, no study has compared competing investments in walking or cycling with investments in other modes of transport using mca. in most cases, mca has been applied as an ex-ante appraisal method to either assess planned walking and cycling projects (such as glavic et al., ) or to prioritize investments (guerreiro et al., ; kabak et al., ) . although possible, no study has been performed ex-post, hence no reported experiences of the effects and/or the usefulness of the method at a later stage are available. there is currently no standard framework for mca, which is tailored to address each specific case. criteria included in walking and cycling mca range depending on the planning scale, the method used, the available data and the study design. usually, at a strategic level (such as in gris e and in this paragraph we present and compare the main strengths, weaknesses and limitations of the four methods for appraising walking and cycling projects. table provides a summary. the main advantage of this type of analyses is that it provides a clear summary of the direct financial effects from a specific project and the range of variation across different scenarios and assumptions. this is particularly useful for budgeting and ensuring long-term financial sustainability of a project. however, the social effects are often neglected as they are less relevant for the research objective or too complex to be accounted. even when positive/ negative "social" impacts of the project are considered, these are either qualitative ones (e.g., "bikers will feel safer") or, when quantitative, they are expressed as non-comparable unit of measurements (e.g., "pollution will go down % in the area"). based on investment costs and impacts (if any), relevant actors decide based on their own judgment, i.e., they introduce their own weighs on the importance of impacts for specific stakeholders and value them against the projected costs (bakker et al., ) . many walking and, especially, cycling projects are often evaluated only through a balance sheet calculation; this mainly happens because such method is the quickest and cheapest of all, as it can require, at its minimum, no further analysis besides the financial and technical feasibility studies that are required by the law, and the "appraisal" of the project is done through pure judgment by the decision-maker(s). this allows for ample discretion on his/her side, which of course is an advantage or a disadvantage depending on one's position. even when impacts are considered, the weighing phase introduces a high degree of subjectivity, not only on which impacts are considered relevant but also by which stakeholder(s). in fact, these simple and straightforward tools are also the most limited in scope and objectivity: first, as effects are analyzed separately it is not possible to provide a comprehensive comparability of different options. moreover, the subjectivity of the decisionmaking might accelerate the process only if interests among stakeholders are aligned, which is seldom the case in public policies and even more so for transport projects: when differences of interests emerge, and no clear power structure that can impose a decision exists, the balance sheet calculation method does not contribute in reaching a shared decision, and the process can be slowed down or altogether stopped. hence, in a situation in which budget is limited, the costs are high, the potential number of alternatives increases and/or several stakeholders are involved, the balance sheet calculation approach is usually integrated with other methods such as cost-benefit analysis and/or multi-criteria analysis. cba applied to walking and cycling present similar methodological strengths and weaknesses already discussed by the literature on general transport cba. namely, cba enables the comparison between costs and benefits of policies and programs targeting different travel modes, which can be a straightforward and convincing way to present arguments of economic efficiency as it has an allure of scientific soundness (browne and ryan, ) . currently, most of the cases in literature use cba for this purpose, for example, in g€ ossling and choi ( ) and g€ ossling et al. ( ) . moreover, cba may have potential applications to rank program priorities and projects selection for financing, especially when used at the early stages of the decision-making process (eliasson et al., ; mackie et al., ) . another potential application currently not investigated is the use of cba in policies aiming at internalizing externalities of walking and cycling (i.e., quantify the value of km reimbursement for bike-to-work schemes). finally, cba may prove to be valuable to structure a debate and improve learning, communication and trust among stakeholders when used in a participatory setting (beukers et al., ) , although in this case too the literature lacks concrete examples for walking and cycling. despite the growing literature on the social and economic effects of walking and cycling, there are substantive limitations to the quantifications and valuation of these effects. van wee and b€ orjesson ( ) and decisio ( ) have discussed these in detail. a major weakness of cba is that it is extremely "data hungry"; this is particularly evident when it comes to estimating current and future demand for the infrastructure. demand forecast is a crucial first step also for other assessment methods such as cea. how many cyclists or pedestrians will use the infrastructure once opened? how will the urban traffic change as a result of the pedestrianization of a specific street? what will be the revenues of a bicycle parking at station? these are questions that are impossible to answer without a model. hence, the quality of a cba highly depends on the type of model used as well as the quality of the input data. the integration of walking and cycling into traditional transport simulation models is a "recent and complicated affair" (see for a detailed discussion barnes and krizek, ; buehler and dill, ; hollander, ; porter et al., ; turner et al., ) . in synthesis, current difficulties with cycling and/or walking modeling include the following. • there are many gaps in our understanding of what factors play a role in motivating people to choose to walk and cycling instead of driving and building cycle paths alone does not necessarily explain bicycle use on their own. therefore, a simple correlation between infrastructure quantity and cycling/walking rates is unlikely to be robust. • cycling is much more affected by the interaction with other traffic modes, the environment, seasonality, weather conditions and other factors than car traffic. these factors are typically difficult to include in a model. • bicycle use and behavioral change according to trip purpose, age groups and the level of some benefits depend on the physical activity levels of the targeted population which is often unknown in an origin-destination matrix. • for walking in particular it is difficult to determine what counts as a walking trip and distinguish by motive. • another issue is related to the zoning of the model which needs to be more refined as walking and cycling trips take place usually at short distances meaning that calculations become more cumbersome and data less reliable. • network coding is usually a difficult and lengthy process and the quality of information is not always readily available and requires many more assumptions. • with walking and cycling infrastructure "the devil is in the detail." some slight design choices and infrastructure characteristics (type of pavement, etc.) may have a greater impact on route choice and behavior than on typical road infrastructure. • finally, it is also currently difficult to predict the added value of marginal improvements in cycling infrastructure especially in countries in which these type of infrastructure projects might be common such as in the netherlands and denmark. even though nowadays models have become far better at predicting and estimating the effects of policies and road adjustments to walking and bicycle traffic, in practice these models are not always available, and it is simply impossible to gain a satisfactory level of data coverage. hence, several academics and practitioners recur to other means to predict induced traffic such as potential analysis tools which observe short car trips to enable the testing of modal shift scenarios (one example is the propensity to cycle tool developed in the uk, see lovelace et al., ) . although the uncertainty of these methods is high, the use of models is not necessarily a guarantee of improved accuracy considering that interventions happen in a non-closed system (naess and strand, ) . another prominent issue is related to the quantification and valuation of specific effects. in particular: • limited research is available about specific travel time valuation (vot) of pedestrian and cyclists. related aspects such as the valuation of reliability, waiting time and search time (i.e., when parking a bicycle) have not been investigated either. moreover, vot of different target groups and travel motives (utilitarian vs recreational) could be significantly diverging. moreover, there is limited literature and research on comfort evaluation, travel experience and perception of safety mainly because of the challenge in defining, measuring and attaching a monetary value to this concept (van ginkel, ) . however, it is likely that comfort, along with perception of safety, are important factors in motivating people to travel by bicycle or on foot (handy et al., ) . • although the literature generally suggests that increasing the level of physical activity has positive health consequences (kelly et al., ) . including these effects into cbas presents several uncertainties. the extent to which people actually become "healthier" is strongly related to the individual herself and his/her lifestyles (haskell et al., ; pate, ) : inactive people who start cycling, for example, may have greater health effects than already-active people. moreover, it is important to assess which means of transport is substituted (car, public transport, e-bike, etc.). the extent to which health effects are internalized is also uncertain. b€ orjesson and eliasson ( ) pointed out that most cyclists accounted for health effects when choosing to cycle and argued therefore that there might be a risk of overestimating the size of the external effects. however, it is also unclear to which extent cyclists and walkers are able to quantify the order of magnitude of these effects (ibid.). another issue is related to new mode of travels such as electric bicycles, steps and pedelecs which are becoming increasingly popular among different target groups and require less effort from the user, limiting the magnitude of the health effects. however, such evidences are difficult to collect everywhere, and it is unlikely to get this specific information for a specific intervention site in which a cba may be used as appraisal technique. in addition, it is unclear how to trade-off health effects from potentially increasing pollution intake. finally, there is a lack of understanding on the extent to which increased cycling rates create substitution effects from other sport activities and influences selfselection. since health benefits are usually very high in slow moderelated cbas studies (brown et al., ) . • improving road safety is another important rationale for improving walking and cycling facilities (pucher and buehler, ) . however, including the effects on road safety in cbas on walking and cycling remains tricky as knowledge on road-type specific disaggregated risk factors is often lacking and the use of aggregated statistics may lead to underestimations of the risks effects for short car trips happening in urban areas which are usually the target of cycling policies (stipdonk and reurings, ) . moreover, there is some evidence that increasing cycling levels substantially reduce the risk of accidents due to the so called "safetyin-number," i.e., the fact that vehicle drivers become more accustomed to cycling/walking people and more capable of anticipating their behaviors and, thus avoid accidents ( jacobsen, ; wegman et al., ) . however, it is also true that the relationship is inverse, and that when cycling/walking becomes safer (e.g., thanks to infrastructure improvement), more people start to walk/bike. • there are other intangible effects that are discussed in the literature that can be relevant for the appraisal of cycling and walking project such as increased urban quality and attractiveness (pucher and buehler, ) , increased option value (geurs et al., ; laird et al., ) and reducing transport poverty (martens, ) . the main advantage of cea compared to other methodologies is that it is cheaper and effective as a tool to rank options. this allows decision-makers to easily sort between alternative options that ensure that a goal will be reached at the least possible cost. however, this is also its main limitation as transport policies may not only want to address one objective at a time. typically, there are in fact a number of competing objectives to be balanced such as: improving health, reducing accidents, alleviate congestion and improve environmental quality (litman, ) . hence, cea may not be the most suitable method if the objective is to fully consider a wide range of effects in one decision criterion (browne and ryan, ) . secondly, the results of cea have limited transferability due to heterogeneous study designs and the context-specific nature of its application, as well as the limited number of ex-post assessment which hinders the generalizability of results. thirdly, some long-term benefits of cycling and walking that may not occur immediately and other synergistic effects resulting from an intervention (i.e., installing bike lanes may increase bicycle traffic improving health but also reducing car traffic alleviating congestion and pollution) may be underrepresented due to the static picture that a cea provides. similarly, to cba, there are important limitations related to forecasting which require the analyst to make strong assumption and predictions (ibid.), and uncertainties on how to quantify and value effects (such as s adverted daly or gained qaly). in general, the main advantage of mca is that it can incorporate quantitative and qualitative analysis of economic, environmental and social impacts and, therefore, the results can be more informative than quantitative analysis alone, as is the case in cea or cba (browne and ryan, ) . secondly, mca can account for multiple stakeholders' opinions, leading in principle to more legitimate approaches as it allows for the inclusion of qualitative and process-related aspects which, for example, the cba typically does not (dean, ) . finally, it can be used as a policy learning tool, where the objective is process-oriented rather than result-oriented and can be modified to weight criteria with stakeholder input and explicit opinions or values (ibid.). on the other hand, mca may be subject to ambiguity and subjectivity in applying weights, it holds risks of double counting and it can present lack of consistency (see beria et al., ; dean, ) . in addition, the specificity of the context makes the transferability of the results impossible to generalize and highly subjective. despite this, most studies underlined some important lessons such as the importance of considering the perspective of multiple actors and to choose the appropriate study design. for example, mca has been used by moshref javadi et al. ( ) to identify the most suitable locations for bicycle share stations. they reported that the most import criteria in determining the final location were proximity to bicycle paths, transportation and networks, demand, and use type. milakis and athanasopoulos ( ) included the opinions of cyclists in their study, proposing a four-step methodology for bike-share network planning using multi-criteria and gis methodology. the methodology was considered to be suitable for cities attempting to introduce and prioritize cycling infrastructures, since it focuses on determining where cyclists would prefer to cycle. another positive aspect of mca is the flexibility to tailor the instrument based on the data availability which is typically low for walking and cycling. in addition, the ad hoc definition of criteria may also induce (intended or unintended) manipulations that steer the results to a specific (desired) outcome. furthermore, certain increasingly popular concepts such as "walkability" and "bikeability" and more broader concepts of fairness find hardly a common definition. some scholars (annema et al., ; eliasson et al., ; mouter, a,b) have already investigated the use and view of appraisal methods by politicians. it is argued in this chapter, that appraisals conducted on walking and cycling might be used similarly. in particular, cba may be one of the instruments used by policy entrepreneurs to promote different framings of walking and cycling as transport and instrumentally use cba while harnessing political support as suggested by weber ( ) and aldred ( ) . however, concrete evidence of the views and uses of cba applied to walking and cycling by policymakers and other stakeholders remains underexplored. this constitutes an interesting avenue for further research, considering that most of the use of this appraisal technique on walking and cycling is to promote political debate and enhance a positive public dialog (see for instance ecf, ) . the purpose of the development of methodological guidelines and tools such as heat stems from this very need to provide an instrument to justify investments into active modes from a health-economic standpoint. weber ( ) has pointed out that there may be value in pointing the research into this direction and has proposed the use of the multiple streams framework (msf) and other policy process framework as a possible lens to study the use of cba and other appraisal methods for walking and cycling within decision-making processes. however, multiple other theories and lenses of policy processes could be used as well (see a review by sabatier, ) . from the comparison of multiple lenses, a better understanding of the impact of appraisal methods on decision-making outcome on walking and cycling projects may be identified. filling this knowledge gap may promote a better integration of cba within the decision-making processes, promote communicative rationality in transport planning and support the creation of stronger stakeholder coalitions. the appraisal methods that have been discussed were initially developed for traditional transport projects, by which we mean somewhat large-sized projects, mainly concerning infrastructure for motorized private vehicles and/or public transport systems. such established methods have been adapted to active mobility projects out of reconsiderations about the traditional transport system, which has led to increased interests in active mobility forms and in the methods to evaluate their costs and benefits. at the same time, the last two decades of research are increasingly suggesting that active mobility has a positive impact on society, however, this might be framed, in most contexts. in this sense, appraisal methods have both shaped and are shaped by the increasing interest in active mobility. however, research is unsettled in most, if not all, aspects of evaluation of cycling and walking projects and programs, as the existing methods have been adapted to the new(er) active mobility field with mixed results in terms of analysis capability, applicability, reliability and communicability. the evaluation of projects based exclusively on the costs (and revenues) of the proposal (such as the balance sheet), technical aspects and intuitive assessments of the merits and flaws remains the only practice that is adopted by most decision-makers across the world. on the one hand, its simplicity promises quicker and clearer decisions, as fewer input data are required, and the decision-maker oversees establishing relevant criteria. however, the bs-and in some way even cea-considers only the feasibility of the project, which is hardly a justification for the necessity of implementing it, without touching upon the benefits of active mobility projects and thus reducing the room for discussion about the desirability of a project. although simpler methods promise speed of adoption thanks to the few parameters to be considered and evaluated, this very feature can easily backlash and lead to ill-informed and often inconclusive debate. on the contrary, methods such as cba or mca, though more complex to both develop and explain, explore the full spectrum of possible impacts, thus fostering a more comprehensive and informed discussion about the role of active mobility within society. however, the choice for simpler appraisal methods is mostly driven by considerations of costs, time and increasingly so data unavailability. in particular, cycling and walking demand modeling is probably the largest source of uncertainty and variability to the usability of economic appraisal methods, especially when plans, and not single infrastructures, are concerned. not being able to quickly, cheaply and reliably assess the effects on cycling or walking level of a certain intervention creates a "garbage in, garbage out" type of problem, especially if forecasts are made through the introduction of a significant number of hypotheses. in that sense, the practitioner has fundamentally two opposite possibilities: (a) increase the modeling effort (combined with data collection in most cases) and provide an improved forecast or correlation linking intervention and results; (b) shift the focus from "modeling and forecasting" to "what if" scenarios, which would remain more general but would assess scenario impacts without claims of prediction. however, improved modeling and increased data collection and availability about active mobility would anyway be necessary in most cases in which a certain degree of correlation between the proposed intervention (and relative costs) and expected benefits must be made. this is true for cycling and even more so for walking, for which very little modeling effort has been carried out in the past. one of the most fundamental issues for economic appraisal methods stems from the necessity to monetize the relevant impacts: in particular, many effects have been identified and quantified to a (somewhat) high degree of certainty, such as the value of time, several types of environmental impacts, value of congestion etc. however, even when impacts can be determined and precisely quantified, it is the phase of their monetization that introduces the largest variabilities and uncertainties. these uncertainties stem from three main factors: ( ) the consequences of the impact cannot be determined to the same level of certainty (e.g., the consequences of local air pollution on health); ( ) the consequences of the impact, though determined, cannot be reflected into a direct economic measure (e.g., the loss of biodiversity from a specific eco-system); ( ) the economic measure attached to the impact, though determinable in principle, is highly contextdependent and subjective (e.g., the value of time). for these reasons, the uncertainty that surrounds economic appraisal methods is significant, thus increasing the variability of the results and potentially hindering the model reliability. moreover, the high variability introduces an important drawback for cba in particular: the possibility to quantify and monetize the impacts, although complicated, appears to support the case for a purported objectivity/ neutrality of the method, which seemingly suggests that the decision can be demanded to the results of the cba. on the other hand, room for discretion always exists and lies in the hands of the practitioner performing the analysis, who must ultimately choose, even if within ranges, which values to consider and which ones to discard; this is particularly relevant when considered parameters are highly variable (i.e., the value of co emissions, the value of statistical life, etc.). in this sense, the main difference between the cba and the mca is that for the latter the subjectivity is clearly visible and transparent, in that the weights are openly discussed and assigned according to personal criteria. on the other hand, the subjectivity of cba is somewhat hidden, as the practitioner is forced to choose among possible values or evaluation methods for the considered impacts. the best practice for cba reporting, in fact, is to fully state the introduced hypothesis and communicate median values and variability ranges of the results, as well as providing a sensibility analysis if possible. nevertheless, cba results are often not accepted by the audience because choices about parameters range have been made solely by the practitioner and not by a group of stakeholders. this is particularly relevant in fields of application where choices are often "emotionally loaded," such as the case for cycling infrastructure that diminishes space dedicated to motorized vehicles or policies that disincentive their use. officials from the city of amsterdam and munich within the handshake project, for example, have argued that they prefer to adopt mca rather than cba because of the "perceived subjectivity" of the latter, which induces confusion and hinders the discussion. our own experience suggests that debate about an emotionally loaded project will not be easily solved by the results of a cba, as these would be interpreted differently by different actors in policy controversies (more on this is discussed by rein and sch€ on, ) . quite paradoxically, a higher acceptability could characterize the results from the "highly subjective" mca, because the weighs can be made explicit and part of a transparent participatory process. in that sense, the evaluation of effects within cba could also be part of a participatory process, as it has been underlined by beukers et al. ( ) . one aspect that hinders, or at least slows down, the improvement of economic appraisals is that any attempt to reduce the uncertainty and variability of impact monetization in cba must include a high degree of cross-sectorial expertise, as the process from impact identification to its quantification and then to its monetization requires a very different set of skills. for example, the quantification of air pollution emissions form vehicles is an engineering problem, the diffusion of pollutants depending on the specific context is an environmental scientist-type of problem, whereas the health effects of pollution concentration requires epidemiological studies, whose results must then be assessed in terms of monetized impact through sociological and economic studies about the consequences of increased illness and premature deaths. this type of knowledge-chain is often specific to each identified impact and, in many cases, to each case-study when local context can significantly change conditions. finally, the presented methods have been considered as "alternatives," but they really should not. in fact, these methods can embrace the full spectrum of socio-economically-relevant consequences of increased cycling/ walking conditions only when combined, i.e., only when more than one method is applied to the same case-study. a plurality of methods is seldom applied to a single case-study for obvious resource-scarcity reasons, which force the practitioner or the decision-maker toward the single method that can deliver the best results given the constraints (usually, time and money). bakker et al. ( ) had already suggested to combine, for the evaluation of integrated transport policies, the strengths of mca and cba, specifically for projects where impacts might be harder to monetize. on the other hand, no case-study of such (or other) combinations of methods has been published so far to the best of our knowledge. however, academics and private researchers might aim at constructing new tools that could develop the potential and overcome the weaknesses of the existing methods, which have been historically developed for quite different contexts and might therefore not be the best possible solution. in this regard, the combination of two or more of the presented tools, and even the transposition of a different appraisal method altogether, might benefit the field and increase the possibility for fast, reliable and high-quality appraisals. the costeffectiveness of physical activity interventions: a systematic review of reviews a matter of utility? rationalising cycling, cycling rationalities barriers to investing in cycling: stakeholder views from england cost-benefit analysis (cba), or multi-criteria decision-making (mcdm) or both: politicians' perspective in transport policy appraisal appraisal of integrated transport policies planning for cycling unsustainable transport: city transport in the new century estimating bicycling demand should we invest in environmental interventions to encourage physical activity in england? an economic appraisal multi-criteria analysis for environmental management cost benefit analysis to assess urban mobility plans. in: consumers' surplus calculation and integration with transport models multicriteria versus cost benefit analysis: a comparative perspective in the assessment of sustainable mobility urban development without more mobility by car? lessons from amsterdam, a multimodal urban region using cost benefit analysis as a learning process: identifying interventions for improving communication and trust when transport policy becomes health policy: a documentary analysis of active travel policy in england principles of cost-benefit analysis cost-benefit analysis: concepts and practice. pearson education limited short-term planning and policy interventions to promote cycling in urban centres: findings from a commute mode choice analysis in barcelona is the widespread use of urban land for cycling promotion policies cost effective? a cost-benefit analysis of the case of seville a systematic review of economic analyses of active transport interventions that include physical activity benefits comparative analysis of evaluation techniques for transport policies bikeway networks: a review of effects on cycling reducing car dependence in the heart of europe: lessons from germany, austria, and switzerland the economic contribution of public bike-share to the sustainability and efficient functioning of cities a systematic review and economic evaluation of exercise referral schemes in primary care: a short report a spatial multi-criteria analysis method for the assessment of walkability to guide the design of interventions in urban crossings economic analyses of transport infrastructure and policies including health effects related to cycling and walking: a systematic review a cost benefit analysis of an active travel intervention with health and carbon emission reduction benefits cost-effectiveness of interventions to promote physical activity: a modelling study economic evaluation of cycle projects-methodology and unit prices optimizing the position of bike sharing stations urban greenways have the potential to increase physical activity levels cost-effectively do the health benefits of cycling outweigh the risks? environ. health perspect health impact modelling of different travel patterns on physical activity, air pollution and road injuries for são paulo assessing the applicability of participatory multi-criteria analysis methodologies to the appraisal of mega transport infrastructure. ph.d. dissertation, the bartlett school of planning a practical guide to multi-criteria analysis the social costs and benefits of investing in cycling themaplan snelfietsroutes beter benutten vervolg estimating the health economic benefits of cycling does benefit-cost efficiency influence transport investment decisions? which are the relevant costs and benefits of road safety measures designed for pedestrians and cyclists? a review of the importance of physical fitness to company performance and productivity calculating the economic benefits of cycling in eu- dutch cycling: quantifying the health and related economic benefits handbook of sustainable travel are physical activity interventions in primary care and the community cost-effective? a systematic review of the evidence a critical review of reductionist approaches for assessing the progress towards sustainability option value of public transport: methodology for measurement and case study for regional rail links in the netherlands social impacts of transport: literature review and the state of the practice of transport appraisal in the netherlands and the united kingdom transport transitions in copenhagen: comparing the cost of cars and bicycles contrasts in active transport behaviour across four countries: how do they translate into public health benefits? economic appraisal of european transport projects: the state-of-the-art revisited if we build it, who will benefit? a multi-criteria approach for the prioritization of new bicycle lanes in quebec city the cost-effectiveness of bike lanes data-mining, gis and multicriteria analysis in a comprehensive method for bicycle network planning and design the cost-effectiveness of installing sidewalks to increase levels of transport-walking and health an economic evaluation of health-promotive built environment changes promoting cycling for transport: research needs and challenges physical activity and public health: updated recommendation for adults from the the cost effectiveness of three programs to increase use of bicycle helmets among children the foundations of welfare economics s maatschappelijke kosten en baten sprong over 't ij bijlagen bij nota van uitgangspunten sprong over het ij transport modelling for a complete beginner evaluating the demand for new cycle facilities transport behaviour during the covid- crisis. iea legislative decree safety in numbers: more walkers and bicyclists, safer walking and bicycling cost-effectiveness of neighbourhood slow zones a gis-based mcdm approach for the evaluation of bike-share stations health economic assessment tools (heat) for walking and for cycling: methods and user guide on physical activity, air pollution, injuries and carbon impact assessments. world health organization-european office welfare propositions of economics and interpersonal comparisons of utility cost-effectiveness of co mitigation in transport systematic review and meta-analysis of reduction in all-cause mortality from walking and cycling and shape of dose response relationship prioritizing low-stress and equitable bicycle networks using neighbourhood-based accessibility measures option and non-use values and rail project appraisal build it. but where? the use of geographic information systems in identifying locations for new cycling infrastructure cost-benefit analysis of added cycling facilities integrating public health objectives in transportation planning. victoria transport policy institute quantifying the benefits of nonmotorized transportation. victoria policy institute the propensity to cycle tool: an open source online system for sustainable transport planning transport appraisal revisited the societal costs and benefits of commuter bicycling: simulating the effects of specific policies using system dynamics modeling health cobenefits and transportation-related reductions in greenhouse gas emissions in the san francisco bay area promoting bike-and-ride: the dutch experience role of the bicycle in the limitation of transport poverty in the netherlands what about people in cycle network planning? applying participative multicriteria gis analysis in the case of the athens metropolitan cycle network planning of the athens metropolitan cycle network using participative multicriteria gis analysis assessing cost-effectiveness in obesity: active transport program for primary school children-travel smart schools curriculum program locating of bicycle stations in the city of isfahan using mathematical programming and multi-criteria decision making techniques dutch politicians' attitudes towards cost-benefit analysis dutch politicians' use of cost-benefit analysis health impact assessment of active transportation: a systematic review cost-effectiveness of interventions promoting physical activity what kinds of traffic forecasts are possible? contested spaces: bicycle lanes in urban europe how can multi-criteria decision analysis contribute to environmental policymaking? a case-study on macro-sustainability in germany cadboro bay road bike lane feasibility study physical activity and public health-reply the cost-effectiveness of mandatory mph zones for the prevention of injuries forecasting bicycle and pedestrian travel: state of the practice and research needs international handbook on mega making cycling irresistible: lessons from the netherlands, denmark and germany city cycling infrastructure, programs, and policies to increase bicycling: an international review value of time, safety and environment in passenger transport time frame-critical policy analysis and frame-reflective policy practice cost-effectiveness analysis health economic assessment of a shift to active transport evaluation: a systematic approach valutazione dei costi e dei benefici della mobilità attiva. caso di san donato milanese bicycle facility planning using gis and multi-criteria decision analysis impact of travel mode shift and trip distance on active and non-active transportation in the são paulo metropolitan area in brazil theories of the policy process cost-benefit analyses of walking and cycling track networks taking into account insecurity, health effects and external costs of motorized traffic costs and benefits of bicycle helmet law for germany. institute for transport economics m€ unster the effect on road safety of a modal shift from car to bicycle bike share study new zealand bicycle helmet law-do the costs outweigh the benefits? bicycle and pedestrian travel demand forecasting: literature review, study no. - tag unit a . active mode appraisal billion package to create a new era for cycling and walking the value of time and comfort in bicycle appraisal how to make cba more suitable for evaluating cycling policies the longitudinal relation between self-reported physical activity and presenteeism cost effectiveness of a bicycle/pedestrian trail development in health promotion factors influencing the propensity to cycle to work the relationship between fitness levels and employee's perceived productivity, job satisfaction, and absenteeism the process of crafting bicycle and pedestrian policy: a discussion of costbenefit analysis and the multiple streams framework how to make more cycling good for road safety? update effectiviteit en kosten van verkeersveiligheidsmaatregelen health economic assessment tools (heat) for walking and for cycling-methods and user guide regions and cities of the who european region commit to safe mobility and transport for urban populations the cost-effectiveness of bike share expansion to low-income communities in new york city key: cord- -e koo authors: shank, nancy title: a review of the role of cost–benefit analyses in - - diffusion date: - - journal: american journal of preventive medicine doi: . /j.amepre. . . sha: doc_id: cord_uid: e koo context the - - helpline is a social services innovation that has spread rapidly throughout the u.s. policy diffusion theory suggests that policymakers seek to reduce uncertainty by anticipating the effects of a proposed innovation through tools such as cost–benefit analyses. few policy diffusion studies have examined use of information, such as cost–benefit analyses, in the diffusion process. the purpose of this study is to examine how cost–benefit analyses were used during the rapid diffusion of - - across states. the paper also describes components of - - cost–benefit analyses. evidence acquisition in , cost–benefit analyses of - - and substantive citations of them were identified through scholarly key word searches using academic search premier and web of science, general internet searches using google search terms, and communications with academicians and - - practitioners through personal contact and e-mail discussion groups. to be included in this study, documents had to be related to - - helplines, present information about their costs and benefits, and be formal documents. the documents were catalogued and analyzed for cost–benefit analyses or references to analyses, and stated purpose. evidence synthesis of the documents that met eligibility inclusion criteria, nine were original cost–benefit analyses and ten referenced analyses conducted for other jurisdictions. conclusions the diffusion of - - helplines in the u.s. has been influenced by interjurisdictional exchange of cost–benefit analyses, in both the creation of original analyses and/or the referencing of previous work. i n , the federal communication commission (fcc) set aside the three-digit dialing code - - for community services information and referral. in the decade since, use of - - has spread to each of the states, washington dc, and puerto rico. there are problem/need categories of information and referral provided by - - s (table ). although call proportions vary from state to state, health and mental health calls typically account for % to % of calls. - increasingly, - - helplines are serving as public information points for health concerns including severe acute respiratory syndrome (sars), west nile, and flu shots. these helplines also have participated in community health prevention outreach to vulnerable populations to promote smoking cessation programs, mammography, adult human papillomavirus (hpv) vaccination, and paps. despite the rapid adoption of - - helplines across the u.s., there has been limited scholarly attention paid to this social services innovation. cost-benefıt analyses (cbas), often sponsored by advocates for - - , informed policymakers about likely economic impacts. it is well known in the - - community that cbas conducted for one jurisdiction were sometimes used by other jurisdictions as evidence of a positive economic benefıt. however, no systematic documentation has been conducted on the use of - - cbas from one jurisdiction to take action in another jurisdiction. moreover, there is little guidance in the academic literature generally that documents information flows between jurisdictions in the policy diffusion process. the present study addresses this gap by using - - as a case study to document possible use of cbas done by one jurisdiction to justify policy action about - - in other jurisdictions. this paper also describes use of cba for understanding components of - - and provides one case example of components of a - - cba. the fırst - - started in atlanta ga in . the ability to offer information and referral services through a threedigit dialing code was embraced by organizations across the u.s., who successfully banded together to petition the fcc to reserve - - nationwide for community information and referral. the fcc gave states the responsibility of determining what organization(s) would be responsible for delivering the service. for most states, this determination would be made by a public service or public utility commission. in a limited number of states, the determination was made by a legislative body or delegated to a coalition of information and referral organizations. states have varied in whether they have staged approval for - - geographically or done so in a single action. some states, such as michigan, nebraska, and california, have staged - - successively within their state, adding regions over a number of years. other states, such as hawaii and texas have assigned - - responsibility across their state in a single action. once authorized, - - services typically have been responsible for developing their funding and implementing the service. a limited number of - - services receive substantial state support, but most rely primarily on a combination of public and private support. the - - s have not addressed funding and implementation challenges in isolation. several national not-for-profıt organizations have promoted - - nationally. notably, united way worldwide and the alliance of information and referral systems (airs) have jointly championed - - development. united way worldwide used its nationwide network of united way member agencies to advocate at the national and local levels for adoption of and funding for - - s across the u.s. airs, the association of information and referral organizations, leveraged its network and incorporated - - into its professional standards and certifıcations. these two national organizations, along with local - - s, worked with legislators to propose federal funding through the calling for - - act. the act would have provided substantial federal funds to support - - s throughout the u.s. although the act was not passed, several - - s successfully received targeted congressional appropriations. united way worldwide and airs have served as networks for information sharing about the costs and benefıts of - - , typically by equipping local - - supporters who were in contact with their state and local policymakers. airs and united way worldwide publicized documents and toolkits with information about the service's costs and benefıts through e-mail discussion groups, at national conferences, and through personal communications. in , united way worldwide also commissioned a cba for adoption of - - across the entire u.s. airs' and united way worldwide's dissemination strategies appear to have promoted sharing of cbas in states through local - - supporters. one national - - leader reflected on a typical encounter of a - - supporter with a state policymaker (anonymous, personal communication, ): the rote question [asked by state policymakers] was often "have you done a cost-benefıt analysis?" . . . [the] answer would be, "not yet, although we would be willing to develop one if you could help with the funding of the study. however, cost-benefıt documents have been published by texas and nebraska [for example]. would you like to review those?" types of cost analyses include cbas, cost-effectiveness analyses, cost-utility analyses, and cost-feasibility analyses. cbas judge the worth of a policy based on its economic impact by assigning and comparing the costs of the inputs to the value of the outcomes. cbas have been used in a variety of human service settings, such as education, corrections, poison control centers, services for people with dementia, and health promotion. proponents of cbas claim that the economic valuation of policies provides objective criteria with which to make decisions. they claim that the valuations are transparent methods used to weigh relevant factors. critics charge that it is impossible to monetize all inputs and outcomes of a proposed program and therefore cost analyses ignore important intangible outcomes. despite ongoing debate about cbas, they continue to be a popular tool considered by policymakers. cost-benefıt analyses have been used by advocates of - - to demonstrate its worth. some jurisdictions have developed cbas, and others simply have cited the results of other jurisdictions' analyses. there has been limited scholarly work examining - - cbas. , only two peer-reviewed articles were found that explored costs and/or benefıts of - - s. the fırst focused solely on the benefıts that - - s offer to communities and described the challenges and approaches to quantifying those benefıts. that study was not an analysis of a specifıc - - service; rather, it was an overview of the types of benefıts - - s accrue and a description of how those benefıts might be measured. the second article documented potential costs and benefıts at the individual, organizational, and societal levels and incorporated a temporal dimension to benefıts calculation (i.e., short-, medium-, and long-term). neither of the scholarly articles calculated the costs and benefıts of a specifıc - - service or reviewed the role of cbas in the diffusion of - - across the u.s. policy diffusion has been a topic of keen interest to scholars over the past several decades. researchers seek to explain why some governments adopt innovative public policies whereas other jurisdictions do not. walker's seminal work defıned a policy innovation as any program new to a government, regardless of how many other governments may have adopted it. for him, and the researchers who have followed his lead, it is not the novelty of the idea but the fact that it is new to the jurisdiction that qualifıes it as an innovation. many diffusion researchers have examined the interdependence of governmental bodies when making policy decisions, recognizing that policymakers do not enact policy in isolation. policymakers reduce the risk of making bad decisions by relying on a variety of sources, including those used by policymakers in other jurisdictions. , based in rogers's broader theory of diffusion of innovation, policy diffusion describes how policymakers opportunistically scan their environment for new ideas or solutions to problems. when an idea is identifıed, it is subjected to an information-gathering and testing period. during this period, policymakers may seek to reduce the risk of unexpected consequences by looking to the experiences of others. researchers hypothesize that jurisdictions emulate each other for four reasons: ( ) policymakers face a problem and use solutions already developed by others as a decision-making shortcut; ( ) policymakers want to remain competitive with other states to provide a good economy and quality of life for residents; ( ) policymakers want to conform to the norms of other states or the expectations of the federal government; or ( ) policymakers respond to public pressure from citizens and media. , researchers have studied diffusion in the context of a wide range of policies including lotteries, gaming, school choice, end of life, water fluoridation, living wills, antismoking, and hmos. policymakers may learn of innovations from policy networks, professional associations, epistemic communities, nongovernmental organizations, think tanks, and the media. this transfer of information may take place through face-to-face communications, but it also may occur through documentary and informal evidence. formally documented expert-based information, such as cost analyses, may be particularly desired and persuasive when policymakers are developing policy decisions. , despite the important and powerful role of such information, researchers have noted that few policy diffusion studies have focused on the use of information in the diffusion process. the present study addresses this gap by examining the use of information produced through a popular policy tool (i.e., cbas) in the rapid spread of - - in the u.s. both scholarly and gray literatures were searched. documents identifıed through the searches were included in the current study if they were related to - - , examined - - costs and benefıts, and were formal documents (e.g., not simply website narrative). excluded from the study were analyses not specifıc to - - , information sources that were not formalized, or that presented costs or benefıts only in a very general way. information collected from each document included the author, title, publisher, year of publication, and how costs and benefıts were monetized. the documents were coded independently so no test of inter-rater reliability was required. scholarly literature was identifıed through computerized searches using academic search premier (asp) and web of science (wos). the searches were conducted for the presence of keywords in scholarly, peer-reviewed publication abstracts prior to march . each database was searched using keyword combinations of: and cost; - - and cost; - - and call center; and call center; - - and human services; and human services; human services and telephone; information and referral and human services; and telephone and community-based services. of the articles returned by the searches (duplicates included), were found by asp and through wos. only one of the articles met the eligibility criteria for the present study. gray literature was searched in three ways: ( ) through existing lists of cbas; ( ) via personal communications; and ( ) using internet searches. lists of cbas were obtained from airs, united way worldwide, and the center for excellence in cancer communication research at washington university in st. louis. personal communications in the form of e-mail inquiries were sent on april , , and june , , to - - administrators and staffers through two e-mail discussion groups: the airsnetworker (approximately recipients) and the - - discussion-l list ( recipients). the e-mails requested cost-benefıt or cost-effectiveness analysis, or any other type of analysis that attempted to compare the costs of - - s to positive outcomes. the third information source was web searches. searches were conducted using the terms - - cost benefıt, cost benefıt, feasibility study, and - - business plan. the search identifıed documents for inclusion in the study. , [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] only one document was found through the scholarly database search, whereas the remaining were identifıed through the gray literature search. three documents were from national organizations: two from united way worldwide and one from the national aging information and referral support center. eleven documents covered an entire state or area within a state, with some states represented by multiple documents (number denoted parenthetically): arkansas, california, hawaii, kentucky, maryland, michigan ( ), nebraska, oregon, pennsylvania ( ), texas ( ), and washington ( ). approximately half (nϭ ) of the eligible documents were original cbas. all but one of the cbas were conducted at the state level. of the documents, six were identifıed as cbas, , - two were identifıed as valuation documents, , and one was identifıed as a business plan. table lists the cbas in chronologic order based on publication date. the analyses included those that were developed prior to delivery of - - services (relying on speculative projections) and those developed after - - had been established (incorporating actual delivery experience in projections). two studies, the texas study and the nebraska study, predated the ruling by the fcc. the remaining seven studies were conducted in the years following the fcc designation. five cbas were conducted prior to their states' authorization, and three were conducted subsequent to state authorization. the studies conducted prior to - - authorization within their state were texas ( ); ne- according to the documents, the purpose of these analyses was to estimate the value of the existing service. the national study ( ) was conducted following the fcc decision, but prior to widespread adoption of - - across the u.s. several of the prospective cbas compared varying service-delivery structures. for example, the nebraska study compared three models of situating call centers (i.e., single statewide center, six regional centers open / , and six regional centers operating weekdays and one becoming the overnight and weekend center). in other studies where information and referral services (e.g., texas, ) or actual - - s were already in place (e.g., michigan, ), existing call volumes and organizational structures were analyzed. in all of the cbas, except maryland's, benefıts outweighed costs. (maryland projected a slightly higher per capita cost and monetized a smaller number of benefıts than did other states). some of the cbas were used as models or information sources for work done in subsequent cost analyses. for example, both the maryland and arkansas studies explicitly mentioned that they used aspects of the nebraska study. the texas study and the national study shared several authors and featured similar valuation approaches. the washington study reviewed benefıts that were quantifıed in the maryland and nebraska studies. the hawaii study referenced arkansas, maryland, michigan, and the national analyses. the analyses across the nine cbas varied in depth and scope of costs and benefıts. moreover, there were variations in which costs and benefıts were monetized or simply described in nonmonetary terms. for the purpose of illustration, the costs and benefıts identifıed in texas's cba are listed in table . costs. all the cbas, except the washington study, monetized the costs of implementing - - . none of the analyses presented nonmonetized costs, suggesting that the authors believed all - - costs were able to be valued fınancially. of the monetized costs, most of the studies (maryland, michigan, national, nebraska, texas and ) divided costs into start-up costs and ongoing costs. the arkansas (pre-implementation) and hawaii (post-implementation) studies made no distinction. start-up costs included such items as capital expenses for hardware, software, telephony, databases, and offıce furnishings. the nebraska study also included training as a start-up cost. all studies, except the washington study, calculated operating expenses. operating expenses included personnel costs; materials and supplies; travel; facilities rental; promotion/marketing; training; insurance; technology (e.g., telecommunications, hardware and software); professional services (e.g., legal, information technology, telecommunications, accounting); and equipment. benefits. both monetized and nonmonetized benefıts were reported in most studies. many of the studies categorized monetized benefıts. for example, the texas studies examined benefıts to individuals, government, and society, whereas the nebraska and arkansas studies categorized benefıts to individual citizens, employers, human service providers, planners, and funders. the most commonly monetized benefıts included call avoidance for other three-digit numbers (e.g., , ); personal and professional time-savings in locating services; call avoidance for community human service programs; cost avoidance for needing redundant information and referral lines that could be handled by - - ; improved productivity for providers; tax dollars recovered to the community through referrals to volunteer tax-assistance programs; and cost avoidance for more-expensive interventions that could be addressed through less-expensive programs (e.g., nursing homes, early intervention). most studies also included nonquantifıed benefıts. the most-frequently mentioned nonquantifıed benefıts were reduced frustration for individuals seeking services, reduced cost of services, cost avoidance of more-expensive interventions that could be addressed through lessexpensive programs, improved information about service coverage and needs, and reduced toll-free lines sponsored by governments. some studies did not quantify benefıts that were quantifıed by other studies. for example, the nebraska, national, arkansas, and michigan studies monetized the benefıt of avoiding redundant information and referral calls, whereas maryland mentioned call avoidance as a benefıt but did not attempt to value it. ten documents referred to other jurisdictions' - - cbas and did not present original analyses ( table ). most of the documents (nϭ ) were local or state-focused. two had a national focus. the remaining study was a combination of a national and municipal focus. the experience of other states appears to have been important as most of the reports explicitly stated their desire to learn from other states who had implemented - - . two states (michigan and pennsylvania) did not have state authorization for - - at the time of publication of the documents referring to others' cbas. one state (washington) had authorized a third party to make a decision about what organization would be authorized to operate their - - , but at the time of the statewide plan ( ), had not made an assignment. the one local plan (peninsulas region of washington) had received state authorization to operate a - - but had not yet implemented the service. two states (california and kentucky) had achieved authorization of at least one - - within their boundaries at the time of publication, but only approximately one half of their state's populations had access to - - . most often, the results of other jurisdictions' cbas were reported as a net monetary value. for example, the california study is typical in referring to the national study's fınding that the net present value to the nation over years of complete - - coverage would be $ million. five of the documents reported the results of only one of the cbas (table ) . four studies mentioned two analyses. one study mentioned three analyses. the most-frequently reported analysis was the national analysis, which was reported in eight of the documents. the nebraska analysis was cited in seven reports. one report mentioned the texas analysis. essentially, it appears that the case for - - was established by , as none of the cbas published after were referenced in other documents. many of the documents presented the challenge of identifying fınancial support for - - services. for example, pennsylvania's plan states that building a - - network: requires a signifıcant long-term commitment from state government as well as ongoing support from a broad mixture of private sources-united ways, private foundations and business-and local funding obtained by the regional - - call centers from both public and private sources. an important but largely unexamined question in policy diffusion is how information is used by jurisdictions throughout the policy diffusion process. researchers increasingly are viewing products of policy analyses as contributing to wider processes beyond that for which they originally were authored. , the present study provides evidence that specifıc results produced by cbas were used in the rapid diffusion of - - s across the u.s. the current study begins to fıll a gap in policy diffusion research, because there has been little documentation of the use of information by others. the results illustrate policy information sharing through the example of cbas in the diffusion of - - . as one national leader in - - commented (anonymous, personal communication, ): some leaders in the political sphere and/or major internal government administrators liked to know that they could be a trailblazer in their state with some security that other states had adopted it and that studies were available to back it up. there are several limitations to the present study. first, archival evidence was the sole source of information. thus, study scope limited to documented analyses and their mention in formal documents by other jurisdictions. this approach prevents exploration of the comparative valance of cbas in relation to other information policymakers may have accessed. therefore, this study is not able to describe the persuasiveness of cbas results in the results of the current study point to areas for future research and practice. the rapid diffusion of - - s suggests that the national network of - - s was an effective channel for sharing policy-relevant information. this raises further research questions: what contributed to the effıcacy of this diffusion process? how might the network be used to spread future policy innovations? does diffusion theory help us understand the mechanisms associated with early and late adoption of policies like - - ? why have certain groups not adopted - - ? these questions may benefıt from both qualitative and quantitative (e.g., mixed-methods) approaches. understanding the potential inequities in policy diffusion will be important. diffusion theorists have posited that jurisdictions' internal characteristics (e.g., socioeconomics of the intended population) and propinquity to other innovators are related to likelihood to adopt, suggesting that isolated, economically disadvantaged areas may be less likely to adopt policy innovations, such as - - . , data from united way worldwide indicate that rural communities may be disproportionately represented in areas without - - . rural areas tend to have greater need for, but fewer, social services. this leads to the disquieting possibility that the areas whose residents could most benefıt from - - services may be the very ones lacking access. future research should explore disparities in adoption and access to innovative social policies such as - - , as well as to understanding the mechanisms by which successful adoption is accomplished. as a result, research will benefıt policymakers, policy implementers, and other practice-based professionals who are striving to accelerate the uptake of life-saving - - services. a practical implication of the present study is that - - s are able to work together to share information and mobilize quickly, despite the fact that - - s are not a single national body but rather a collection of independent organizations. there is promising evidence, for example, in the realm of disaster response and recovery, that - - s together form an effective national network that is able to meet unanticipated surges of need. , , , in their role in providing health-related information and referral, and in their emerging role as a proactive source for promoting community and individual health, the - - network comprises an effective vehicle for diffusing new practices. further, - - and health practitioners may fınd fertile ground for furthering mutual goals by working together. united way/airs . nationwide status. us.org/status.htm airs: i&r problem/ needs national categories - - texas. - - texas information and referral network action summary united way of pennsylvania. ways - - works proactive screening for health needs in united way's - - information and referral service the role of information in the policy process: implications for the examination of research utilization in higher education policy - - state by state: a periodic report on the national implementation of three digit-accessed telephone information and referral services indiana partnership, inc. fact sheet - - in the mid south delta: a collective case study collectivecasestudy/ - - % in% the% mid% south% delta % collective% case% study.pdf advancing the common good: united way policy agenda for the th congress cost-effectiveness analysis: methods and applications economic techniques the state of cost-benefıt and costeffectiveness analyses in education does cost-benefıt analysis or self-control predict involvement in bullying behavior by male prisoners? cost-benefıt analysis of a regional poison control center cost-benefıt analysis of assistive technology to support independence for people with dementia-part : development of a methodological approach to the enable cost-benefıt analysis a fınancial cost-benefıt analysis of a health promotion program for individuals with mobility impairments examining the potential benefıts of a - - system: quantitative and other factors - - information services: outcomes assessment, benefıt-cost analysis, and policy issues the diffusion of policy diffusion research. psweb.sbs.ohio-state the diffusion of innovation among the american states diffusion of innovations adoptions as policy innovations: an event history analysis innovations and diffusion models in policy research explaining the diffusion of innovation types amongst high and low innovative localities: a test of the berry and barry model disentangling diffusion: the effects of social learning and economic competition on state policy innovation and expansion policy networks and innovation diffusion: the case of state education reforms promoting a "good death": determinants of painmanagement policies in the u fluoridation: diffusion of an innovation among cities innovation and reinvention in state policymaking: theory and evolution of living will laws the mechanisms of policy diffusion interstate professional associations and the diffusion of policy innovations introduction: epistemic communities and international policy coordination non-governmental policy transfer: the strategies of independent policy institutes international organizations as teachers of norms: the united nations educational, scientifıc, and cultural organization and science policy interstate communication among state legislators regarding energy policy innovation expert-based information and policy subsystems: a review and synthesis information sources in state legislative decision making the value of a comprehensive texas information and referral network final report: survey of existing i&r services and a nebraska system cost/benefıt analysis the value of a comprehensive texas information and referral network maryland - - : benefıts and costs of a - - system for maryland national benefıt/cost analysis of three digit-accessed telephone information and referral services: fınal report win - - : performance evaluation and cost-benefıt analysis of - - i&r systems arkansas system cost/benefıt analysis benefıt/cost analysis of aloha united way's program michigan - - . michigan - - business plan aging network involvement in civil society consulting group llc, venture architects. washington information network business plan. win .org/docs/ win wabusinessplan.pdf united way - - of the peninsulas business plan - - california. - - california partnership business plan michigan association of united ways. michigan - - business plan united way of southeastern pennsylvania. policy brief: how does reduce state spending? . . . /communityimpact/publicpolicy/ does save taxpayers kg pennsylvania - - . pennsylvania - - business plan united way of america. - - fact sheet kentucky - - . kentucky - - strategic business plan after hurricanes, support grows for call service policy analysis for what? the effectiveness of nonpartisan policy research organizations the diffusion of policy innovations-an experimental investigation putting the - - coverage map in context department of agriculture. rural income, poverty, and welfare: summary of conditions and trends after the storms: - - stories from the hurricane season. united way of america trial by fıre: how - - 's regional response to the southern california wildfıres underscored the need for a statewide network publication of this article was supported by funding from the national cancer institute (nci) and the offıce of behavioral and social science research (obssr) of the nih (hhsn p).ns has received consulting fees from the alliance of information and referral systems and united way worldwide, the lead national organizations in the development of - - in north america. key: cord- - zq tw authors: d’acci, luca s. title: urbanicity mental costs valuation: a review and urban-societal planning consideration date: - - journal: mind soc doi: . /s - - - sha: doc_id: cord_uid: zq tw living in cities has numerous comparative advantages than living in the countryside or in small villages and towns, most notably better access to education, services and jobs. however, it is also associated with a roughly twofold increase in some mental disorders rate incidence compared with living in rural areas. economic assessments reported a forecasted loss of more than trillion dollars in global gdp between and and of around trillion for the year alone when measured by the human capital method. if we exclude self-selection processes and make the hypothesis to be able to level down the mental illness rate incidence in urban areas to these of the rural by better urban-societal planning, around € . trillion could be saved yearly worldwide. even a reduction of only % in urban mental illness rate would save around billion dollars yearly. disorders are also the primary cause of disability-adjusted life years worldwide (bloom et al. ) . decades of empirical research shows an association between mental health and urbanicity, especially for the individuals genetically more inclined and those who lived in cities during their early life. links, often proven to be causal by longitudinal and dose-response analysis, between urbanicity and mental illness have been greatly reported such as in these studies: coid et al. ; evans et al. ; vargas et al. ; sampson et al. ; lecic-tosevski ; reed et al. ; evans et al. ; castillejos et al. ; kirkbride et al. kirkbride et al. , cooper et al. ; besteher et al. ; gruebner et al. ; krzywicka and byrka ; vassos et al. ; newbury et al. ; brockmeyer and d'angiulli ; adli et al. ; freeman et al. , wilker et al. peterson et al. ; haddad et al. ; vaessen et al. ; steinheuser et al. ; haluza et al. ; krabbendam et al. ; streit et al. ; calderón-garcidueñas et al. ; heinz et al. ; bedrosian and nelson ; stevens et al. ; tandon et al. ; lederbogen et al. ; fonken et al. ; larson et al. ; galea et al. ; mcclung mcclung , meyer-lindenberg ; park et al. ; bowler et al. ; kelly et al. ; mortensen et al. ; levesque et al. ; gwang-won et al. ; tae-hoon et al. ; peen et al. peen et al. , van os et al. ; kennedy et al. ; bentall et al. ; march et al. ; joens-matre et al. ; fuller et al. ; graziano and cooke ; maas et al. ; mortensen a, b, a, b; weich et al. ; krabbendam and van os ; tsunetsugu and miyazaki ; wang ; sundquist et al. ; van os ; van os et al. ; mcgrath et al. ; harrison et al. ; caspi et al. ; frumkin ; allardyce et al. ; haukka et al. ; torrey et al. ; van os et al. ; eaton et al. ; schelin et al. ; marcelis et al. ; mortensen et al. ; marcelis et al. ; thornicroft et al. ; lewis et al. ; cohen ; eaton ; christmas ; faris and dunham ; white . paykel et al. ( , analysing data from almost ten thousand individuals (household survey of the national morbidity survey of great britain) via a logistic regression, reported "a considerable british urban-rural differences in mental health, which may largely be attributable to more adverse urban social environments". according to vassos et al. ( ) , the rate of incidence of nine types of psychiatric disorders is in average . times higher in the capital city than in the rural areas, with 'schizophrenia and related disorders' even almost double ( . ), while the review of mcgrath et al. ( ) of studies found a schizophrenia incidence rate times higher in urban areas than in mixed rural/urban areas; a rate that rises up to a . times greater risk of schizophrenia when one has lived years of her early life in a capital city rather than a rural area (pedersen and mortensen a) . peen et al. ( ) reported an odds-ratio for mental disorders in very highly urbanized areas of . related to non-urbanized ( . when unadjusted by control variables). an approximatively twofold increase in psychosis risk associated with urbanicity is also confirmed in the following empirical studies: marcelis et al. ( marcelis et al. ( , , mortensen et al. ( ) , schelin et al. ( ) , allardyce et al. ( ) , pedersen and mortensen ( a, b) , van os et al. ( van os et al. ( , , harrison et al. ( ), sundquist et al. ( , pedersen and mortensen ( a, b) , kirkbride et al. ( ) , haukka et al. ( ) and torrey et al. ( ) . an increase as high as fourfold was found in eaton et al. ( ) . a meta-analysis review summarised that urban dwellers have a . times greater risk of mood disorders than non-urban (peen et al. ) . due to the type of the analysis conducted, the causality (rather than a reverse causation) of the nature of this link, emphasising that urbanicity has an etiological effect on mental health, has been underlined, among many, by march et al. ( and lederbogen et al. ( ) . if we shift our attention to people's preferences toward places to live their lives, "many surveys about quality of life in cities invariably suggest that it is in smaller cities that the highest quality of life is achieved" (batty , p. ) . similarly, to european surveys, % of americans voted small towns/rural environments as the best kind of places to live and only roughly one in five ( %) voted cities (knox and pinch ) . another questionnaire (d'acci ) reported that only % of respondents prefers to live in a city rather than (ceteris paribus) in a natural environment ( %), in a town/village ( %), in a suburb ( %), while % of them were indifferent. in line with these stated residential preferences, happiness seems to decrease when urbanicity levels increase (sander ; lawless and lucas ) , and studies about self-declared life satisfaction, psychological well-being in rich countries systematically show lower levels of life satisfaction in urban areas compared to the rural or less urban areas (viganò et al. ; easterlin et al. ; gilbert et al. ; helliwell et al. ; fassio et al. this discrepancy between rural and urban environments' influence on mental health, life satisfaction and happiness suggest that by re-organizing our socioeconomic urban daily life and the physical urban-regional structure itself, there would be a potential margin of reduction in the urban mental illness rates and an increase of life satisfaction and daily mood of urban dwellers. to convince governments, urban and regional planners, stakeholders and the ordinary population about the relevance of the issue, an economic translation of the costs that psychological effects that cities have to us, might help to make the topic more tangible. mental disorder costs go far beyond the direct costs (diagnostic and treatment); their economic costs assessment for the society as a whole should monetarily translate also the following indirect factors: increased chance of leaving school early, lower likelihood of achieving good and full-time employment, reduced quality of life for the individual and her loved ones. the monetary quantification of indirect costs on health usually follows the human capital method which measures the personal direct costs plus the amount of discounted earnings from lost productivity due to several reasons such as those listed above (doran and kinchin ; gustavsson et al. ) . early commencement mental disorders result to be statistically significantly associated with the interruption of secondary education (leach and butterworth ) , which in turn means less likelihood to be employed in higher skilled professions (schofield et al. ) . as expected, psychiatric disorders between the ages of and , after controlling for confounding variables, was statistically significantly (p value < . ) negatively linked with workforce participation, income and economic living standards at age , and, more generally, cumulative episodes of psychiatric disorders negatively affect life outcomes (gibb et al. ). however, a bi-directional causality might appear between mental health and labour force participations as once workforce participation is being affected, a dangerous positive feedback loop could start: you get mentally ill then you work less, and the more excluded from work the more mentally ill you might be (laplagne et al. ). this unemployment rate within the mentally ill population has being quantified to be as high as four times more than the healthy population, and when they work they are more inclined both to presenteeism (work with low productivity) and absenteeism (more leave for illness) (schofield et al. ). this psychological distress cost related to lower productivity has been estimated in to be a$ . billion (equivalent to roughly a$ . billion in ) per annum in australia (hilton et al. ) , and the individuals' loss due to depression has been assessed as a % lower income than their full-time counterparts, while those deciding to retire early because of their mental health issues have % lower incomes, which at a national aggregate level means us$ million in transfer payments, $ million in lost income taxation revenue, and almost $ billion in gdb, just in (schofield et al. ) . reports for canada (smetanin et al. ) assessed that in years ( - ) the cumulative costs related to mental illness could be around us$ trillion (based on us$ )-even if underestimated for the lack of some types of cost and of mental illness-and in - australia spent a$ . billion ( . % of all government health outlay) in mental health services by governments and health insurers. studies also estimate that personal family costs and lost productivity for businesses and other non-government organisation costs, equal, or even surpass, the total government expenditures (degney et al. ; hilton et al. ; jacobs et al. ) . the oecd report estimates as more than % of gdp (around € billion) the costs due to mental illness across europe (oecd ), while gustavsson et al. ( ) estimated it to be around € billion for the , including norway, iceland and switzerland to the european countries. a team of members from the world economic forum and the harvard school of public health (bloom et al. ) used different methods (although non comparable among each other) to estimate mental disorders costs: ( ) direct and indirect costs by human capital approach (the standard cost-of-illness method), ( ) impact on economic growth (macroeconomic simulation), and ( ) value of statistical life (willingness to pay). each method has a different approach: personal versus social, private versus public, yearly costs versus multiple years' cumulative costs. the human capital approach ( ), as we anticipated earlier, considers personal costs such as medical costs, transportation, care, income losses (related also to education loss due to illness), and sometimes it can also add non-personal costs such as public health education campaigns and research. the economic growth method ( ), also called the value of lost output, considers how the investigated diseases diminish labour, capital and any other factors involved in the gdp formation at the country level, focusing on the illness related mortality rates impact on gdp. the value of statistical life method ( ) is based on the people's willingness to pay (a kind of trade off) to lessen the risk of disability or death connected with the analysed illness, therefore by attaching an economic value to health/life itself it goes beyond the practical impact on gdp alone. the quantification is done either by observed trade-offs (e.g. in the labour market the wage premium a worker is willing to receive to take a job with a high injury-death risk, or the extra amount of money an individual spends for healthier food) and hypothetical trade-offs (surveys asking people how much they would pay to elude a risk or how much they would ask to take that risk). by method ( ) the team (bloom et al. ) estimated a world cost for mental illness of us$ . trillion for the year alone, and us$ trillion for the year alone, two-thirds of which for indirect costs. by method ( ) they estimated a world cumulative gdp loss of us$ . trillion (usa dollars ) due to mental health alone over years ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) . by method ( ) they estimated a world output loss of us$ . trillion in , and us$ . trillion in . converted into us$ for the year , they resulted approximatively us$ . trillion of gdp loss during the years between and ; us$ . trillion of human capital loss for the year alone; and us$ . trillion the willingness to pay for the year alone. all these estimates, even if already showing impressively high economic loss translations, are very likely underestimated (whiteford et al. ). it seems clear from decades of a reasonable amount of mutually confirming independent research that urban life has unfavourable (often hidden) effects on our psyche, especially for those genetically susceptible and for those exposed to urban contexts during their juvenile years when the brain is still developing, whose causality has been proven by longitudinal and dose-response studies. most people may not be aware about this psychological damage as it might be that the harm does not reach a sufficient entity to become visible, and that would remain below a certain level implying a manifested invisibility. yet, individuals might still suffer some kind of psychological uncomfortable feeling even without being able to define it, or, if so, to establish the direct link with their urban life. if it is indeed true that it is not 'only' a small percentage of genetically susceptible urban dwellers targeted by statistically significantly higher psychosis risks, but a larger urban population, although with consistent variability in magnitude, we need to include this type of mental costs within any cost-benefit alike analysis. territorial and urban planners cannot ignore the negative consequences that cities and territories have on our psychological well-being and mental health when poorly planned, designed and managed (e.g. endless cementification, lack of daily natural contact, congestion, lack of sky view, crowding, visually and socially boring dormitory areas extended for hectares, …). the same is valid for actions enabling us to change our socio-economic systems toward a more liveable scenario: just to cite an example, teleservices and teleworking (i.e. working remotely from home or wherever), a practice more and more in use and even becoming law (since july , first case in europe and probably in the world) in the netherlands if the worker wishes, would dramatically improve quality of life, free time, work efficiency and productivity, and enormously reduce congestion, daily car use, pollution, car parkstreet space, carbon emissions, and so on. similar effects would be induced by flexible personalized working times (following personal biological circadian rhythmsessential for health and productivity-and private life schedules) and reduction of national daily working hours from, e.g. from to h: equivalent or probably even higher productivity thanks to more efficient use of working time, concentration, positive mood and an overall physical and psychologically healthier population. probably a substantial help will come from medical genetics, pharmaceutics and psychologic-psychiatric progress regarding non-modifiable risk factors such as age, sex and genetic make-up, and from urban-territorial planning and governance, politics and education regarding the modifiable risk factors such as environment (e.g. greener and less crowded-polluted cities) and life style (diet, sport, sleeping, hobbies, sociality, daily natural contact) part of it linked with the environment where one lives. according to the large amount of empirical research evidence we saw, we can quite confidently say that urbanicity determines an approximately twofold increase in psychosis risk. let's speculate that by planning better structural-infrastructural urban environments and forms (d'acci ) and their socio-economic systems/life styles, (eliminating crowd-congestion, pollution, greenless, noise, crime, overwork, stress, over-pace…) of our current cities we are also able to entirely reduce their extra psychosis incidences and then levelling the urban psychosis rate to the rural one. if we refer to the cost-of-illness method (human capital) which directly involves money actually spent in mental illness issues, us$ . trillion would be use for mental urban costs in year alone ( , when, according to un , a . % of urban population is expected). if we prefer to avoid forecasts so far away in time ( ) and refer our thoughts only to real data from the past, in the money actually spent for mental illness was us$ . trillion worldwide ( dollars). an amount also in line with the % of gdp costs for mental illness assessed by the oecd regarding europe: in fact if we use this gdp percentage at the world level, the world gdp in was around (in current dollars) us$ . trillion, whose % means us$ . trillion, namely around us$ . trillion in dollars. by following the previous reasoning about levelling the urban psychosis incidence to the non-urban one thanks to better urban planning and socio-economic life styles, the share of world urban population in was around . %, therefore us$ trillion could have been saved in urban mental illness costs in that year alone. in the world urban population was around . % and the world gdp around us$ . trillion (current dollars), meaning roughly us$ . trillion today, whose % is around us$ . trillion which, following the same approximate reasoning (and assuming a similar percentage of gdp use) means that roughly us$ . trillion could have been not spent in mental illness due to urbanicity issues. if we assume a reduction of "only" % of urban mental illness rate, we would still save around billion dollars yearly. to put these trillions in context, the entire apollo space program ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) , including the walks on the moon, still one of the major humanity achievements, costed only around us$ billion (in dollars); almost times less than what can be saved in only year in mental illness due to urban life. an equivalent program but on mars (sending nine crews), could cost around us$ . trillion, while the mars rover mission costs 'only' between and us$ trillions. another colossal human achievement, the year human genome project costed 'just' us$ . billion ( adjusted into dollars ). cities are a potentially great place to live and achieve our life's goals and progress, both as individual and as a species; however, it has some mental costs for the most susceptible. by planning better cities, territories and socio-economic daily life styles such as teleworking plus flexible working times, weekly working hour national reductions, greening cities and radically transforming the physical structures, forms and functioning of our current urban environments it would have an enormous potential economic impact environmentally, infrastructurally, but also psychologically speaking. from the academic literature we can say that the urban impact on mental health implies an approximately twofold increase in the rate of mental issues compared to the rural environment. reports estimated a us$ . trillion global expense for mental disorders in alone and forecasted to become as high as us$ . trillion in the year alone. if we make a rather imaginative, although not that unrealistic, hypothesis to level the mental illness rate incidence of urban areas to these of the rural by better urban-societal planning, an astonishing amount of more than one trillion us dollars could be saved yearly: we could live happier, and send nine crews to mars every year. neurourbanism-a joint methodological approach between urban planning and neurosciences comparison of the incidence of schizophrenia in rural dumfries and galloway and urban camberwell influence of the modern light environment on mood social predictors of psychotic experiences: specificity and psychological mechanisms dissatisfaction with city life: a new look at some old questions an urban-rural happiness gradient associations between urban upbringing and cortical thickness and gyrification the global economic burden of noncommunicable diseases a systematic review of evidence for the added benefits to health of exposure to natural environments how air pollution alters brain development: the role of neuroinflammation brain immune interactions and air pollution: macrophage inhibitory factor (mif), prion cellular protein (prpc), interleukin- (il- ), interleukin receptor antagonist (il- ra), and interleukin- (il- ) in cerebrospinal fluid and mif in serum differentiate urban children exposed to severe vs. low air pollution influence of life stress on depression: moderation by a polymorphism in the -htt gene a systematic review and meta-analysis of the incidence of psychotic disorders: the distribution of rates and the influence of gender, urbanicity, immigration and socio-economic level psychological stresses of urban living: new direction for mental health services in the inner city calculating people: the spread of numeracy in early america urbanicity and psychosis in a chinese undergraduate population: preliminary findings schizophrenia-risk and urban birth are associated with proteomic changes in neonatal dried blood spots preferring or needing cities? psychology, utility and disutility of urban life counting the cost: the impact of young men's mental health on the australian economy. inspire foundation and ernst and young, canberra doran cm, kinchin i ( ) a review of the economic impact of mental illness the impact of modern economic growth on urban-rural differences in subjective well-being residence, social class, and schizophrenia obstetric factors, urbanization and psychosis urbanicity is associated with behavioral and emotional problems in elementary school-aged children urbanicity, biological stress system functioning and mental health in adolescents health, quality of life and population density: a preliminary study on "contextualized" quality of life air pollution impairs cognition, provokes depressive-like behaviors and alters hippocampal cytokine expression and morphology the stress of the street for patients with persecutory delusions: a test of the symptomatic and psychological effects of going outside into a busy urban area beyond toxicity: human health and the natural environment psychological benefits of greenspace increase with biodiversity the urban environment and mental disorders: epigenetic links burden of psychiatric disorder in young adulthood and life outcomes at age are rural residents happier? a quantitative analysis of subjective wellbeing in scotland parieto-frontal interactions, personal space, and defensive behavior cities and mental health cost of disorders of the brain in europe functional neuroanatomy associated with natural and urban scenic views in the human brain: . t functional mr imaging brain structure correlates of urban upbringing: an environmental risk factor for schizophrenia green perspectives for public health: a narrative review on the physiological effects of experiencing outdoor nature association between psychotic disorder and urban place of birth is not mediated by obstetric complications or childhood socioeconomic position: a cohort study regional variation in the incidence of schizophrenia in finland: a study of birth cohorts born from to urbanicity, social adversity and psychosis how happy are your neighbours? variation in life satisfaction among canadian neighbourhoods and communities using the interaction of mental health symptoms and treatment status to estimate lost employee productivity the cost of mental health and substance abuse services in canada: a report of the mental health commission of canada. institute of health economics schizophrenia and the city: a review of literature and prospective study of psychosis and urbanicity in ireland personal space regulation by the human amygdala heterogeneity in incidence rates of schizophrenia and other psychotic disorders: findings from the -center aesop study city living and psychotic disorders-implications of global heterogeneity for theory development schizophrenia and urbanicity: a major environmental influence-conditional on genetic risk neural effects of the social environment restorative qualities of and preference for natural and urban soundscapes effects of health and education on labour force participation what role can child-care settings play in obesity prevention? a review of the evidence and call for research efforts predictors of regional well-being: a county level analysis the effect of early onset common mental disorders on educational attainment in australia is urban living good for mental health? city living and urban upbringing affect neural social stress processing in humans air pollution and the brain: subchronic diesel exhaust exposure causes neuroinflammation and elevates early markers of neurodegenerative disease schizophrenia and city life green space, urbanity, and health: how strong is the relation? urbanization and psychosis: a study of - birth cohorts in the netherlands urbanization and risk for schizophrenia: does the effect operate before or around the time of illness onset psychosis and place circadian genes, rhythms and the biology of mood disorders circadian rhythms and mood regulation: insights from pre-clinical models a systematic review of the incidence of schizophrenia: the distribution of rates and the influence of sex, urbanicity, migrant status and methodology from maps to mechanisms through neuroimaging of schizophrenia effects of family history and place and season of birth on the risk of schizophrenia psychiatric family history and schizophrenia risk in denmark: which mental disorders are relevant why are children in urban neighborhoods at increased risk for psychotic symptoms? findings from a uk longitudinal cohort study unhappy metropolis (when american city is too big) urbanism and happiness: a test of wirth's theory of urban life the physiological effects of shinrinyoku (taking in the forest atmosphere or forest bathing): evidence from field experiments in forests across japan urban-rural mental health differences in great britain: findings from the national morbidity survey evidence of a dose-response relationship between urbanicity during upbringing and schizophrenia risk family history, place and season of birth as risk factors for schizophrenia in denmark: a replication and reanalysis are the cause(s) responsible for urban-rural differences in schizophrenia risk rooted in families or in individuals? urbanization and traffic related exposures as risk factors for schizophrenia is the prevalence of psychiatric disorders associated with urbanization? the current status of urban-rural differences in psychiatric disorders effects of prenatal exposure to air pollutants (polycyclic aromatic hydrocarbons) on the development of brain white matter, cognition, and behavior in later childhood interaction of childhood urbanicity and variation in dopamine genes alters adult prefrontal function as measured by functional magnetic resonance imaging (fmri) urbanization, urbanicity, and depression location and happiness in the united states regional differences in schizophrenia incidence in denmark how depression and other mental health problems can affect future living standards of those out of the labour force the life and economic impact of major mental illnesses in canada: to . risk analytica, on behalf of the mental health commission of canada rural-urban differences in life satisfaction: evidence from the european union stress and the city: impact of urban upbringing on the (re)activity of the hypothalamus-pituitary-adrenal axis adverse health effects of nighttime lighting: comments on american medical association policy statement a functional variant in the neuropeptide s receptor gene moderates the influence of urban upbringing on stress processing in the amygdala urbanisation and incidence of psychosis and depression human brain activation in response to visual stimulation with rural and urban scenery pictures: a functional magnetic resonance imaging study frequency of parent-supervised outdoor play of us preschool-aged children urban-rural differences in the associations between social deprivation and psychiatric service utilization in schizophrenia and all diagnoses: a caseregister study in northern italy risk factors and confounders in the geographical clustering of schizophrenia measurement of absolute hemoglobin concentrations of prefrontal region by near-infrared time-resolved spectroscopy: examples of experiments and prospects united nations department of economic and social affairs, population division the dopaminergic response to acute stress in health and psychopathology: a systematic review does the urban environment cause psychosis? prevalence of psychotic disorder and community level of psychotic symptoms confirmation of synergy between urbanicity and familial liability in the causation of psychosis the environment and schizophrenia chronic stress, structural exposures and neurobiological mechanisms: a stimulation, discrepancy and deprivation model of psychosis urban-rural differences in incidence rates of psychiatric disorders in denmark urban-rural dwellers' well-being determinants: when the city size matters. the case of italy. city estimating the true global burden of mental illness disease burden and government spending on mental, neurological, and substance use disorders, and self-harm: cross-sectional, ecological study of health system response in the americas rural-urban differences in the prevalence of major depression and associated impairment rural/non-rural differences in rates of common mental disorders in britain: prospective multilevel cohort study the geographical distribution of insanity in the united states challenges to estimating the true global burden of mental disorders long-term exposure to fine particulate matter, residential proximity to major roads and measures of brain structure key: cord- -cbfiqvu authors: as'ad, rami; hariga, moncer; shamayleh, abdulrahim title: sustainable dynamic lot sizing models for cold products under carbon cap policy date: - - journal: comput ind eng doi: . /j.cie. . sha: doc_id: cord_uid: cbfiqvu amid the ever growing interest in operational supply chain models that incorporate environmental aspects as an integral part of the decision making process, this paper addresses the dynamic lot sizing problem of a cold product while accounting for carbon emissions generated during temperature-controlled storage and transportation activities. we present two mixed integer programming models to tackle the two cases where the carbon cap is imposed over the whole planning horizon versus the more stringent version of a cap per period. for the first model, a lagrangian relaxation approach is proposed which provides a mean for comparing the operational cost and carbon footprint performance of the carbon tax and the carbon cap policies. subsequently, a bisection based algorithm is developed to solve the relaxed model and generate the optimal ordering policy. the second model, however, is solved via a dynamic programming based algorithm while respecting two established lower and upper bounds on the periodic carbon cap. the results of the computational experiments for the first model display a stepwise increase (decrease) in the total carbon emissions (operational cost) as the preset cap value is increased. a similar behavior is also observed for the second model with the exception that paradoxical increases in the total emissions are sometimes realized with slightly tighter values of the periodic cap. concerns over global warming and its detrimental consequences continue to rise, making carbon emissions reduction an increasingly critical matter. it is a global consensus that carbon emissions stand out as the prime cause leading to climate change and global warming (ipcc, ) . the recent surge in carbon emissions and its negative impact is mainly attributed to the fast-paced industrial growth and the associated intensive energy consumption. the amount of carbon footprint generated evidently varies across different sectors of the economy, where the potential reduction of such emissions in the industrial, construction, and agricultural sectors was discussed by huisingh et al. ( ) . the authors stated that carbon emissions due to logistics operations in specific constitute a small percentage to possibly more than percent of the global emissions depending on the characteristics of goods as well as the transportation mode adopted. as a matter of fact, in the year , it is reported that % and % of greenhouse gas emissions in the united states were generated from transportation and electricity, respectively . on the notion of goods characteristics, cold supply chain has unique dynamics and peculiarities that sets it apart from other supply chains. cold, or temperature-sensitive, products are typically of perishable nature and encompass a wide range of products such as fresh produce, poultry and dairy products, fishery items, pharmaceuticals, vaccines, among many others. efficiently managing the logistics operations for such products has recently received an ever growing interest from both industrial practitioners and academic researchers alike, which may partially be attributed to two main reasons. first, the expanded market reach of these products and the unprecedented surge in their sales figures due to the advancement in the enabling cooling technologies coupled with the increase in the number of third party logistics service providers specialized in handling this type of products. for instance, it has been estimated that global cold chain market was valued at $ . billion in and it is projected to reach $ . billion by . in the year , the overall global capacity of cold storage reached around million cubic meters . these numbers clearly illustrate the great potential that exists towards optimizing the movement and storage of such products along each step across the supply chain. secondly, since cold products are typically chilled ( ) or frozen ( ), special temperature-controlled ℃ - ℃ transport and storage facilities are needed to handle them making this industry highly energy intensive. in fact, kayfeci et al. ( ) pointed out that energy consumption in cold chains amounts to % of the total world energy consumption. furthermore, this extensive use of energy triggers elevated levels of carbon emissions, where it has been reported by james and james ( ) that the cold chain is held accountable for almost % of the world's greenhouse gas emissions. the intensive energy usage, and accordingly the cost, of cold products logistics operations coupled with their negative environmental impact renders the development of efficient and sustainable replenishment policies of such products an inevitable necessity. a bad selection for the number and capacity of temperature-controlled trucks, the lot size to order, and the number and size of freezer units at the retailing outlets, among other factors, could have a devastating impact on the cost, quality, environment and customer service level. therefore, supply chain practitioners dealing with cold products ought to explicitly account for sustainability aspects as an integral part of their operational decision-making process. muriana, ) . however, the dynamic or stochastic demand patterns are more practical as they better capture real-life dynamics of the demand for many products such as the ones under study here. more specifically, cold or perishable products in general, typically exhibit time varying demand stressing the need for dynamic lot sizing modeling approach to optimize inventory and transportation related decisions on a periodic basis, and accordingly control costs and carbon emissions. for instance, dairy products, fishery items and fresh food exhibit daily or weekly varying demand patterns. in case of daily demand, this demand is much higher during the weekend than during weekdays as most consumers, living in the vicinity of the retail stores, frequently tend to purchase their needs of such items on weekends. the remaining consumers prefer to go to the retail store when the need arises during the weekdays. similarly, the demand varies from week-to-week, where the demand in the first week of the month is especially higher as consumers do more of their purchases during this week after receiving their paychecks. this paper addresses the logistics operations of a cold product exhibiting a discrete time varying demand over a finite planning horizon. as modular temperature-controlled units are used to store and transport such products, the carbon footprint generated due to these activities is taken into account through the carbon cap regulatory policy. the cap is firstly set on the total emissions throughout the planning horizon and then a stricter version is considered with the cap being imposed on each period's emissions. for each of the two cases, a mixed integer linear programming (milp) model is developed along with a solution algorithm that yields the optimal ordering policy. to the authors' best knowledge, this work is the first to tackle this problem in the context of cold products using temperature-controlled transportation trucks and warehousing facilities while accounting for the products limited shelf life aspect. the remainder of this paper is organized as follows. section presents a review of the state-of-the-art literature pertaining to the problem at hand, and highlights the contributions of this work. section provides the mathematical models along with the proposed solution algorithms. the impact of the carbon cap (both total and periodic) as well as other key problem parameters on the lot sizing policy, the operational cost and the carbon footprint generated is assessed in section through sensitivity analysis. concluding remarks in addition to suggestions for future research avenues are presented in section . lastly, the detailed derivation of the economic and environmental objective functions coefficients along with illustrative examples for the total and periodic cap cases are shown in the appendices. illustrated the economic and emissions reduction benefits realized when two buyers cooperate to share transportation paths and handling units. perishability is an important aspect characterizing a wide range of products. due to their widespread applicability, an increasing body of literature has been devoted to modeling the logistical operations of these products while explicitly accounting for their perishable nature. as for perishable products in general, accounting for shelf life aspect is of practical relevance as it directly affects the replenishment strategy and the associated carbon emissions as well as the potential waste generated and ultimately a firm's profitability. despite its paramount importance, it is noted in the review paper of pahl and voß ( ) that little research has been done in the modeling of lifetime restrictions to prevent wastage and disposal of perishable products, especially in a dynamic planning context. as cold products are usually of a perishable nature, this work contributes to such deficiency through addressing the lot-sizing policy of a cold product facing time varying demand while explicitly considering environmental and limited lifetime aspects. stated more formally, we next highlight the three main contributions of this paper. ) it is known that lagrangian relaxation approach provides lower bounds to complex minimization problems. as such, this approach is typically used to generate approximate (nearoptimal) solutions. however, as shown in this paper, the novelty in the proposed hybrid lagrangian relaxation and bisection based solution procedure is that it generates the optimal solution to the optimization problem when a carbon cap is imposed on the total carbon emitted over the planning horizon. in addition, based on the optimal value of the lagrangian multiplier, consider a cold product facing known time-varying demand over a discrete planning horizon of length t periods. due to its perishable nature, the cold product has a pre-determined shelf life and requires refrigerated trucks for its transportation and modular temperature-controlled following are the assumptions underlying the mathematical models developed in this paper: ) orders are delivered only at the beginning of the periods. ) ideal transportation and storage conditions are always maintained which prohibits the deterioration of the cold product during those activities. ) the cold product is perishable in the sense that it loses its value/utility all at once upon exceeding the fixed shelf life (see pahl and voß ( ) for a more detailed discussion). accordingly, a quantity delivered at the beginning of period t can only be held in stock for a limited number of periods equal to its shelf life, sl. demand fulfillment follows a first-in first-out (fifo) stock outflow pattern. in this case, ending inventory left from an order delivered at the beginning of period t, is g periods old at the end of period (t + g - ), where g is less or equal to sl. the lead time associated with shipment delivery from the supplier is negligible. ) demand backordering is not permitted, and holding cost is charged against inventory at the end of each period. the operational status of a freezer unit does not change during the period, and a freezer is switched off only when it is completely emptied. ) there is no restriction on the number of reefers and freezers available to transport and store the cold product, respectively. ) the adopted carbon regulatory policy is carbon cap, which sets a limit on the amount of carbon footprint generated from the transportation and storage activities. ) without loss of generality, the initial (ending) inventory at the beginning (end) of the planning horizon is set equal to zero. the following main notations are adopted in the development of the mathematical models. more notations will be introduced as needed. problem parameters: h: unit storage cost of the cold product for one period, excluding energy consumption cost binary variable, equals to one when an order is placed and received at the beginning of period t and zero otherwise towards curbing carbon emissions, the carbon cap regulation has proven to be an effective regulatory policy that is gaining wider acceptance nowadays. following such policy, a cap is set on the maximum carbon footprint generated by an organization where this cap may span the entire planning horizon (e.g., a year) or could simply be imposed on a shorter-range periodic amounts of carbon emissions (e.g., a month), with the latter being obviously a more strict version of the former. in this paper, we address both cases where the following subsection presents the mathematical formulation and the solution algorithm for the case of a cap being defined for the entire planning horizon whilst the next subsection tackles the shorter interval caps. for a better organization of the paper, the derivation details of the mathematical expressions of the coefficients for the total operational cost function and total carbon footprint constraints are delegated to appendix a. furthermore, illustrative examples on the two optimization problems, with total and periodic caps, are shown in appendices b and c, respectively. using the expressions for the costs associated with the reefers and freezers operations derived in appendix a, the operational cost incurred during period t is where the first two terms are the ordering and holding costs incurred during period t, respectively. the sum of next two terms represents the transportation cost and the last term is the freezers' operational cost. accordingly, the total operational cost over the planning horizon ( ) is given by: the total carbon footprint over the entire planning horizon ( ) is directly attainable from equation (a ) and it is given by: therefore, the constraint stipulating that the allowable carbon cap over the planning horizon is not exceeded would be given by: in the following, an ordering policy is defined by the vector in the = ( , , …, ) sense that once is determined then the remaining decision variables can be found using: hence, the operational cost optimization problem with carbon emission constraint is given by: (cap t ): subject to: ≥ = , , …, , equations ( ) represent the classical inventory balance constraints, while constraints ( ) establish the number of trucks needed to transport the lot size of that period. the number of operational freezers in a particular period is dictated by the beginning inventory as seen in constraint set ( ) . constraints ( ) ensure that the lot size is zero in case no order is placed in period . assuming a fifo consumption pattern, constraints ( ) are the shelf life constraints, which mandate that the inventory at the beginning of a period does not exceed the demand for periods including that own period's demand. the remaining constraints represent the non-negativity, binary and integrality restrictions on the respective decision variables. we next provide some insights leading to the development of an optimal solution algorithm to problem (cap t ). feasibility assumption: the carbon cap should satisfy the condition ≤ ≤ . the lower bound carbon cap, is the minimum carbon footprint over the entire planning horizon obtained by solving the following carbon cap-unconstrained environmental optimization problem, since its objective function is to minimize the total carbon emitted over the planning horizon: cap min,t : min tcf(q) subject to constraints ( ) to ( ) the upper bound, , is the total carbon emissions of the optimal lot sizing policy of the following carbon cap-unconstrained operational cost optimization problem. cap max,t : min toc(q) subject to constraints ( ) to ( ) in order to solve the optimization problem (cap t ), the carbon cap constraint ( ) the objective function of (d) is, therefore, a piecewise linear concave function of . the dual function is shown in figure using the data of the illustrative example presented in appendix b. note that the function z() has a finite number of breakpoints where it is not differentiable, and between each two consecutive breakpoints the functions toc(q) and tcf(q) remain constant. we show in the following lemma that toc(q) and tcf(q) are non-decreasing and nonincreasing functions of , respectively. (i) (ii) let q( ) be the optimal ordering policy for the unconstrained carbon cap problem (cap max,t ). clearly, this unconstrained ordering policy has the least total operational cost, toc(q( )) and the largest total carbon footprint, tcf(q( )). therefore, for a given positive , the total operational costs, toc(q()), of the optimal ordering policy q( to the lagrangian relaxation problem (cap t () cannot be smaller that toc (q( ) ). similarly, it can be seen that tcf(q()) ≤ tcf(q( )). next, consider the optimal solution q(' to the lagrangian relaxation problem (cap t (') , with ' =  +  ≥ . in its expanded form, the objective function of (cap t (') is: ] -} which can be rewritten as: for = and setting the objective function's coefficients for the reefers and freezers' Δ operational costs to and , respectively, the optimal solution to the + , + + relaxed problem is q( with the smallest total operational costs and largest total carbon footprint with these coefficients. consequently, toc(q(+ )) ≥ toc(q( )) and tcf(q( )) ≤ tcf(q( Δ Δ )) for a given Δ > . it can be observed from figure that for a carbon cap of . tons, the range of values for the lagrangian multiplier between the two dotted vertical lines will result in the optimal ordering policy for the primal problem (cap t (   therefore, the smallest  value with the total carbon footprint smaller than the carbon cap can be used to determine the optimal solution. using this observation and assuming that , we develop the following bisection based ≤ ≤ algorithm to generate the optimal ordering policy. obviously, in case , the optimization < problem (cap t ) has no feasible solution. on the other hand, if , then the optimal ordering > policy for the optimization problem (cap max,t ) is also optimal for (cap t ). optimal solution procedure to solve the optimization problem (cap t (). step . set   ,  l =  and solve cap t ( ) step step . . if tcf( < c set  u =  and go to step step . . if tcf( = c set  u =  and go to step step . if  u - l > . , go to step . , otherwise go to step step . . set  = ( u +  l )/ and solve p( using shamayleh et al. ( ) dp algorithm step . . if tcf( > c set  l =  and go to step step . . if tcf( < c set  u =  and go to step step . . if tcf( = c set  u =  and go to step step .  * =  u , q * = q( * ), toc * = toc( * ), and tcf * = tcf( * ) the value of the lagrangian multiplier,    is vital towards assessing the cost and carbon emission performance of the carbon cap regulatory policy when compared to the carbon tax policy. the following lemma shows the relationship between these two commonly used carbon regulatory policies. let be the tax rate charged for each ton of carbon emitted. , then < * ( > * ) a-total carbon emitted under carbon tax policy is larger (smaller) than the one generated by the carbon cap policy. b-the total operational cost of the carbon tax policy is smaller (larger) than the one of the carbon cap policy. the proof is straightforward by referring to figure . in case the cap is imposed on the amount of carbon emitted over each period of the planning horizon, the carbon cap constraints ( ) of the optimization problem (cap t ) are replaced by: where are given in equations (a ) and (a ). , and the optimization problem to be solved is then: subject to: constraints ( ) to ( ) and ( ) feasibility assumption: the carbon cap per period, cp, should satisfy the condition: . the periodic upper carbon cap is the maximum carbon emitted per period over the entire planning horizon when following the optimal lot sizing policy of the unconstrained carbon cap optimization problem, i.e., cp max = max{cf t , t = , , …, t}. on the other hand, the periodic lower carbon cap, cp min , is the optimal objective function value of the following min-max optimization problem: cap min,t : min x subject to in order to solve the optimization problem (cap min,t ), we propose the following dynamic programing based algorithm. we first rewrite the carbon footprint during period t as: ( ) = ⌈ ⌉ + ⌈ - + ⌉ + for = , , …, next, note that the ordering quantity in a period t depends only on the beginning inventory in the same period. for example, when t = t, the ordering quantity is equal to since -- the ending inventory at period t must be zero. for other periods, the beginning inventory, e t- + q t , must satisfy: therefore, we define e t- as the state variable at period t. we also let f t (e t- ) be the minimum of the maximum carbon footprint generated during periods t through t when the beginning inventory at period t is e t- . the periodic lower carbon cap is then f ( ). for period t, we have ( ) on the other hand, for any period t < t, the beginning inventory at period t should satisfy equation ( ) . note the right-hand-side inequality is due to the product shelf-life constraint. under these two inequality constraints, the following recursive equation can be used to solve the t-period problem for cap min,t : and ( ) in order to solve the optimization problem (cap t ), we can use lagrangian relaxation method by relaxing the set of constraints in ( ) . for given lagrangian multipliers associated with constraints ( ) however, the bisection method cannot be used to solve the dual problem to (cap t ( ) as we are dealing with more than one lagrangian multiplier. instead, other solution : = , , …, ) procedures, such as the sub-gradient method, have to be used which may not guarantee the attainment of an optimal lot sizing policy. in the following, we propose a dp based algorithm similar to the one used to solve (cap min,t ) with the same state variable e t- to solve the optimization problem (cap t ). let k t (e t- ) be the minimum total operational costs satisfying the demands and carbon footprint limit during periods t through t when the state variable is e t- at the beginning of period t. the minimum total operational costs over the planning horizon is then k ( ). for a given e t- , the operational costs during period t is: next, for any period t < t, the beginning inventory should satisfy equation ( ) and . ⌈ ⌉ + ⌈ - + ⌉ + ≤ then, the recursive equation to be used is: where y t = if q t > and otherwise. in this section, we analyze the impact of the main problem parameters, such as the imposed carbon cap, inventory related cost (ordering and holding costs), as well as cooling infrastructure related parameters (driver's wage, trucks capacity, fuel price and electricity price) on the resulting lot sizing policy for the two optimization problems (cap t ) and (cap t ). being a key model parameter, we assess in this section the performance of the two mathematical models under various values of the imposed carbon cap (both total and periodic). the purpose is to draw analytic managerial insights pertaining to the impact of the carbon caps on the resulting lot sizing strategy and the associated operational cost and carbon emissions. the importance of conducting such analysis is that it aids the policymakers in setting appropriate values for these caps to effectively reduce the carbon emissions rather than doing so in a complete ad-hoc manner. from the individual corporations' perspective, it greatly helps them capitalize on and maneuver within the caps imposed by the legislative entities through adopting the most-cost effective lot sizing and shipping policy that is in accordance with the allowed caps limits. it shall be noted that the analysis carried hereafter is based on the illustrative example presented in the appendix, wherein the chosen values for the carbon cap respect the established limits. table and figure , respectively. price. as such, as the total cap is relaxed (i.e., assumes a higher value), this resource becomes more abundant and each additional ton of the emissions cap becomes less worthy. that is, it can be seen from table that as the total cap is increased, the optimal value decreases until reaching the * upper limit of the feasibility range, or the unconstrained carbon cap policy, at which point the optimal dual price value . besides the important role of the lagrange multiplier in assessing * = the performance of the carbon tax and the carbon cap policy (as seen in lemma ), it also provides the policymaker with valuable information on the value of each additional ton of emissions cap helping ultimately with the setting of appropriate values for those caps. insight : as the total emission cap becomes less tight, more frequent shipments take place and/or more trucks are utilized in conjunction with fewer number of freezer units turned on (see table ). this pattern is justified as the model takes advantage of the increased permits on the cap and thus strives to minimize the operational cost, through minimizing the holding and freezers operational costs, at the expense of an increase in the ordering and transportation costs, where the savings in the former two cost components outweigh the increase in the latter two. as can be seen from figure , a stepwise decrease (increase) in the operational cost (total emissions) is realized for higher values of the emissions cap. explicitly accounting for the limited trucks and freezers capacities induce such stepwise behavior where the lagrange multiplier, and accordingly the lot sizing policy, remains the same for a range of the emission cap values. insight : through making operational adjustments to the lot sizing strategy, it might be possible to substantially reduce carbon emissions without significantly compromising the operational cost. as can be seen from table table b . we did not present the results for the second case (periodic cap) as we observed almost the same effect for most of the parameters, with few exceptions, where the justification behind such exceptions is provided at the end of the previous section. surprisingly, the increase in the holding cost did not affect the total number of orders placed over the planning horizon. typically, as the holding cost increases, one would expect the lot size to decrease at the expense of an increase in the number of orders. however, it should be noted here that the total cost represents a tradeoff between the holding and ordering cost on one end, and the trucks and freezers cost on the other end. for smaller lot sizes, there will be a need for more trucks due to more orders and, consequently, larger trucks related costs. accordingly, in order to minimize the overall cost function, it is more economical to maintain the same number of orders for larger holding cost rates. additionally, the increase in the holding cost did not have any impact on the total carbon emissions or the number of trucks and freezers used over the planning horizon. this can be explained by the same rational provided above. as can be noted from figure , increasing the holding cost resulted in an increase in the optimal lagrangian multiplier value, which implies that the carbon tax policy will result in smaller (larger) total operational cost (carbon footprint) than the carbon cap policy over a wider range of tax values. such implication is justifiable using the results of lemma . holding cost (h) lagrange multiplier (λ) figure . impact of the holding cost on the lagrange multiplier value ( ) turning to the ordering cost, as anticipated, the obtained results indicate that an increase in this cost would lead to a decrease in the number of orders with an accompanying increase in the number of operational freezers (see figure ). furthermore, increasing the ordering cost triggers a decrease in the optimal value of the lagrangian multiplier, and an increase in the amount of carbon footprint generated (as seen in figure ). as opposed to the holding cost, while the ordering cost is increased, it prevails over the rest of the cost components (trucks and freezers costs) leading the model to prefer a reduction in the number of orders while naturally utilizing more freezers due to the increase in the lot sizes. the heightened levels of carbon footprint is attributed to better utilization of the trucks capacities through ftl shipments, where the co emissions are related to the load as explained in appendix a. lastly, the induced reduction in the lagrangian multiplier results in a smaller range for the tax values for which the carbon tax policy outperforms that of the carbon cap in terms of the operational cost while yielding larger carbon footprint. as for the driver's wage, it turns out that increasing this fixed component of the transportation cost leads to smaller values (see figure ) , while the number of orders, trucks * and freezers utilized as well as the associated carbon footprint remain unchanged. this is caused by the fact that the variable fuel based transportation cost outweighs the fixed transportation cost since the former is a key component that contributes to the carbon footprint, which has to be maintained within the imposed cap. accordingly, all carbon related variables ( , and ) ( ) remain unchanged. for the obtained reduction in the value, the same rationale provided above * for the ordering cost holds true. sizes can be shipped and, therefore, less total number of orders are placed and fewer total number of trucks will be used. as can be noted from table , the medium size truck ( units) = yields the lowest operational cost and co emissions, and shall thus be selected. upon increasing the fuel price (pg), the observed outcome is a decrease in the total carbon footprint as well as the number of trucks and the number of orders whereas the number of freezers increased (see table ). clearly, the reduction in co emissions is attributed to the use of fewer trucks where this reduction outweighs the additional emissions as a result of using more freezers. in order to minimize the fuel consumption related cost associated with larger values of the fuel price, the number of trucks (and orders) have to be decreased, which is what we observed. obviously, decreasing the number of orders implies an increase in the lot size and the initial inventory in each period and consequently the number of freezers needed to accommodate them. finally, it is noted that varying the electricity price (ep) had no apparent impact on the lot sizing policy as it only affects the freezer related cost where this cost is overshadowed by the other cost components. in principle, as the electricity price increases, one may expect the number of freezers to decrease and accordingly smaller initial inventory and lot sizes are realized. however, such a decrease would imply an increase in the number of trucks and the number of orders, and consequently the trucks and ordering related costs. in order for this not to happen, the number of freezers, trucks, and orders remains the same. in adherence to external pressures from legislative entities and conscientious customers with regard to carbon emissions, companies are constantly facing the challenge to adopt environmentally friendly practices that aim at reducing their generated carbon footprint. to that end, this paper tackles the lot sizing problem of a temperature-sensitive product having a limited shelf life while accounting for environmental constraints via the carbon cap regulatory policy. we present mathematical models to address the two scenarios where the carbon cap is imposed on the emissions generated throughout the planning horizon or those generated per period. for each case, solution algorithms that yield the optimal ordering strategy are developed. for the first case, a hybrid lagrangian relaxation and bisection based solution approach is developed to obtain the optimal ordering policy, after establishing the upper and lower bounds on the total emissions. through the use of the lagrangian multiplier, we compare the performance of the carbon tax and the carbon cap policies from both environmental and economic perspectives, which ultimately aids the policymakers in choosing the most effective policy. for the second case, a dynamic programming based solution algorithm is devised which also guarantees the convergence to the optimal lot sizing policy. furthermore, a one-way sensitivity analysis is conducted on the key model parameters to assess their impact on the lot sizing policy, and the resulting operational cost and carbon emissions. the results indicate that operational adjustments to the lot sizing strategy may pose as a viable and a more affordable alternative towards reducing carbon emissions as compared to making substantial investments in costly energy-efficient technology. also, it turned out that making operational adjustments, through modifying the order quantities and accordingly the number of trucks and freezers used, may significantly reduce the carbon footprint generated at the expense of a minor increase in the operational cost. furthermore, it is noted from the computational analysis that the operational cost (carbon emission) is decreasing (increasing) with higher values of the carbon cap, where the later increase could possibly take place on an intermittent basis for the periodic carbon cap case. the work presented in this paper provides the policymakers or legislative entities with a decision making tool to aid them in setting appropriate values for the carbon caps towards effectively reducing the carbon emissions rather than doing so in a complete ad-hoc manner. from the corporate decision making perspective, once those caps have been set, individual corporations may make use of this work to establish the most cost-effective lot sizing and shipping policy that adheres to the imposed cap limits whilst attaining operational efficiency goals. the work presented in this paper may be extended in several directions. one may opt to analyze similar problem settings to the one presented herein but for a more involved supply chain structure encompassing multi stages with one or more firm at each stage. similarly, this work tackled a single cold product situation while an interesting extension would be to explore the multi cold product case which further complicates the analysis and renders a more challenging problem to solve. the consideration of other regulatory policies, such as cap-and-trade, or situations wherein the demand is stochastic, rather than deterministic, pose as other promising future research avenues. finally, our work could be extended to situations of unforeseen disruptions such as the recent covid- outbreak through devising contingency plans that seek to prioritize logistics needs in terms of required storage and transportation capacity against the uncertainties pertaining to the availability and delivery of the items (see ivanov and dolgui, ) . in such times of rising uncertainties, considering the resilience profiles of involved firms as well as the pre-booking of logistics capacity to minimize cost becomes the imperative rather than a choice. to that end, deploying simulation based approaches is of particular significance towards assessing the negative effects of such disruption and developing effective risk mitigation strategies (see aldrighetti the freezers at the retailer's facility are held accountable for high levels of energy consumption and increased amount of co emissions. in particular, the carbon footprint generated per period due to cold storage activities is a function of the number of freezers operational during that period ( ), the energy consumption by one freezer operated for one period, and the carbon footprint of one kwh energy. it is thus given by: * * , or * , where (a ) = * using equations (a ) and (a ), the carbon footprint generated due to transportation and storage activities during period t is: lot sizing with carbon emission constraints the single-item green lot-sizing problem with fixed carbon emissions healthcare supply chain simulation with disruption considerations: a case study from northern italy optimization of vehicle routing with inventory allocation problems in cold supply chain logistics haulage sharing approach to achieve sustainability in material purchasing: new method and numerical applications eoq revisited with sustainability considerations sustainable cold supply chain management under demand uncertainty and carbon tax regulation sustainability in material purchasing: a multi-objective economic order quantity model under carbon trading a sustainable eoq model: theoretical formulation and applications carbon emissions and energy effects on manufacturing-remanufacturing inventory models carbon footprint and the management of supply chains: insights from simple models environmentally responsible inventory models: nonclassical models for a non-classical era including sustainability criteria into inventory models multi-product inventory model for cold items with cost and emission consideration a new inventory model for cold items that considers costs and emissions state-of-the-art literature review on inventory models for perishable products on the consideration of carbon emissions in modelling-based supply chain literature: the state of the art, relevant features and research gaps the carbon constrained eoq multi-period inventory routing problem under carbon emission regulations low carbon supply chain: a state-of-the-art literature review optimal discounting and replenishment policies for perishable products low carbon supply chain: a state-of-the-art literature review emissions targets and the real business cycle: intensity targets versus caps or taxes optimal lot-sizing under strict carbon cap policy considering stochastic demand integrated economic and environmental models for a multi stage cold supply chain under carbon tax regulation operational and environmental decisions for a two-stage supply chain under vendor managed consignment inventory partnership cap-and-trade vs. carbon taxes: a quantitative comparison from a generation expansion planning perspective the economic lotsizing problem with an emission capacity constraint effect of carbon emission regulations on transport mode selection under stochastic demand dynamic economic lot size model with perishable inventory an economic lot size model for perishable products with age-dependent inventory and backorder costs managing carbon footprints in inventory management recent advances in carbon emissions reduction: policies, technologies, monitoring, assessment and modeling climate change : synthesis report viability of intertwined supply networks: extending the supply chain resilience angles towards survivability. a position paper motivated by covid- outbreak the food cold-chain and climate change literature review of deteriorating inventory models by key topics from to a stochastic micro-periodic agebased inventory replenishment policy for perishable goods an approximate periodic model for fixed-life perishable products in a two-echelon inventory-distribution system determination of optimum insulation thickness of external walls with two different methods in cooling applications carbon constrained integrated inventory control and truckload transportation with heterogeneous freight trucks integrated inventory control and transportation decisions under carbon emissions regulations: ltl vs. tl carriers production and transportation outsourcing decisions in the supply chain under single and multiple carbon policies a time-varying lot size method for the economic lot scheduling problem with shelf life considerations importance of the loading factor in transport co emissions, th wctr a literature review on green supply chain modelling for optimizing co emission greening the food supply chain: an optimization model for sustainable design of refrigerated automated warehouses multi-period planning of closedloop supply chain with carbon policies under uncertainty optimal cycle time for production-inventory systems considering shelf life and backordering an eoq model for perishable products with fixed shelf life under stochastic demand conditions the two-level economic lot sizing problem with perishable items the economic lot-sizing problem with perishable items and consumption order preference inventory problems with perishable items: fixed lifetimes and backlogging integrating deterioration and lifetime constraints in production and supply chain planning: a survey order quantities for perishable inventory control with non-stationary demand and a fill rate constraint production planning of perishable food products by mixed-integer programming dynamic lot sizing in biopharmaceutical manufacturing dynamic economic lot size model with perishable inventory and capacity constraints a novel mathematical model for a multi-period, multi-product optimal ordering problem considering expiry dates in a fefo system economic and environmental models for cold products with time varying demand a stochastic reverse logistics production routing model with emissions control policy selection a review on quantitative models for sustainable food logistics management modelling food logistics networks with emission considerations: the case of an international beef supply chain quantifying the environmental and economic benefits of cooperation: a case study in temperature-controlled food logistics research advances in environmentally and socially sustainable operations carbon-regulated eoq models with consumers' low-carbon awareness food transport refrigeration -approaches to reduce energy consumption and environmental impacts of road transport sustainable inventory management with deteriorating and imperfect quality items considering carbon emission joint decisions on inventory replenishment and emission reduction investment under different emission regulations literature review analytics (lra) on sustainable cold-chain for perishable food products: research trends and future directions eoq models for a coordinated two-level international supply chain considering imperfect items and environmental impact optimization of a low-carbon two-echelon heterogeneous-fleet vehicle routing for cold chain logistics under mixed time window supplier-buyer deterministic inventory coordination with trade credit and shelf-life constraint research developments in methods to reduce the carbon footprint of the food system: a review recent advances and opportunities in sustainable food supply chain: a model-oriented review credit author statement rami as'ad formal analysis, software, writing -review & editing conceptualization, methodology, validation, software, writing -review & editing abdulrahim shamayleh conceptualization, data curation, validation; writing -review & editing the authors are indebted to the esteemed anonymous reviewers for their constructive feedback which led to substantial improvement in the content and presentation of the paper. in this appendix, we derive the coefficients of the economic and environmental objective functions. in particular, we show how to compute the transportation and cold storage costs as well as the amount of carbon footprint emitted due to transportation and storage activities. the following notations are adopted in the derivation of these equations: accordingly, in this paper we mark up the available engine fuel consumption figures from the literature by a certain percentage to account for the additional consumption due to refrigeration. in the following, we assume the gallon per mile (gpm) figures include both fuel used for power motive and temperature control.the total fuel consumption is a function of the number of trucks utilized, the distance traveled, the ftl capacity and the actual load in each truck. as the lot size in a specific period ( ) is transported to the retailer via a fleet of identical trucks each having a capacity of units, one may note that trucks will be fully loaded while the load of the last truck is simplyassuming a linear relationship between the gallons of fuel consumed and thetruckload, the total fuel consumption in period for the full ( trucks is , and - )the fuel consumption for the last partially loaded truck is:the total fuel consumption by all trucks during period t is then given by:the fuel consumption cost during period t is obtained by multiplying the last equation by the fuel price per gallon, pg, which after some mathematical simplification can be written asthe total transportation cost is calculated as the sum of the fixed transportation cost given in (a ) and variable transportation cost shown in (a ). it can then be expressed as:similarly, the cold storage cost depends on the electricity consumed by the freezers, where the latter is a function of the number of operational freezers per period ( ) and the periodic energy consumption by each freezer, . as such, the freezers' operational cost during period t is givenusing equations (a ) and (a ), the total operational costs due to transportation and storage activities of the cold product during period t is given bygiven that and , where is the smallest integer larger than or equal to x, it isclear that the above function given in (a ) is a stepwise function of the ordering quantity with discontinuity that are multiples of the truck capacity and freezer capacity (see figure a ). in order to demonstrate the solution procedure for the optimization problem (cap t ), we solve the following example in which some of the parameters values are adapted from the cold supply chain literature and are cited when used. for the sake of clarity, we provide detailed calculations of the objective function and carbon cap constraint coefficients.a cold product, with a shelf life of sl = , is facing a discrete time-varying demand over a planning horizon of weeks, as shown in table b . all administrative activities associated with the placement of an order is estimated to be $ per order. the distance between the supplier and the retailer is miles ( km). it is assumed here that the fuel type used is diesel, which costs however, since we are dealing with refrigerated trucks, there will be an increase in energy consumption to preserve the truckload at the desired temperature. as mentioned in appendix a, tassou et al. ( ) pointed that the average fuel consumption of the refrigeration systems varies between % and % of the engine fuel consumption. this percentage is much higher for countries experiencing high ambient temperatures and humidity during most months of the year, such as middle east countries. accordingly, we adopt herein a % fuel consumption markup, which results in gallon per mile values of . and . for full and empty trucks, respectively.the refrigerated truck used is ' long with internal dimensions of ' " * ' " * ' ". given the truck and crate dimensions, this translates into a loading capacity of units.the amount of carbon emission generated per gallon of fuel consumed, cegf, is . ton . the freezer's capacity is m which is enough to store a maximum of u = crates with the same dimensions as above. it is assumed that a single stage recuperating compressors and evaporative condensers is used for the refrigeration system with an energy consumption of . optimal ordering policy resulted in a reduction of . % in the total carbon emissions when compared to the optimal solution of the unconstrained carbon cap policy. furthermore, in an attempt to reduce the carbon emissions due to transportation activity, it is noted that less frequent orders are made for a cap value of . tons along with better utilization of the trucks' capacity when compared to the unconstrained policy. on the other hand, a limit of . tons on the total carbon emitted over the planning horizon yields a reduction in carbon footprint by . %. as it can be noticed from table b , the reduction in co emissions is mainly due to the drop in the total number of trucks used over the entire planning horizon. this reduction in the carbon footprint due to the use of fewer trucks clearly outweighs the additional emissions induced by the use of more freezer units.  a cold product exhibiting a discrete time varying demand is considered. two models are presented one for the total carbon cap and the other for periodic cap. a lagrangean relaxation and a bi-section method based algorithms are developed for model . model is solved via a dynamic programming based algorithm. sensitivity analysis is performed on the effect of the preset total and periodic caps. key: cord- - goaqir authors: maudgil, d.d. title: cost effectiveness and the role of the national institute of health and care excellence (nice) in interventional radiology date: - - journal: clin radiol doi: . /j.crad. . . sha: doc_id: cord_uid: goaqir healthcare expenditure is continually increasing and projected to accelerate in the future, with an increasing proportion being spent on interventional radiology. the role of cost effectiveness studies in ensuring the best allocation of resources is discussed, and the role of national institute of health and care excellence (nice) in determining this. issues with demonstrating cost effectiveness have been discussed, and it has been found that there is significant scope for improving cost effectiveness, with suggestions made for how this can be achieved. in this way, more patients can benefit from better treatment given limited healthcare budgets. pre-covid- , approximately % of uk gross domestic product (gdp) was spent on healthcare, with planned spending for department of health and social care for / being £ billion: contrast this with the . % of gdp spent for the first full year of the nhs in (£ million; approximately £ billion in today's money). the impact of covid- to the national economy has been estimated to add at least £ billion to government borrowing in alone. clearly, there is currently extreme pressure on spending on public finance, so it is important to spend on healthcare as effectively as possible, and demonstrate that this has been done. nice (the national institute of health and care excellence) was set up to help with this (box ). cost-effectiveness studies (ces) can be helpful in guiding how we should spend public resources, as : ces provide an objective system to compare the complete range of relevant alternatives, from invasive treatments to conservative management. costs for the same procedure can vary widely, e.g., the "getting it right first time" (girft) vascular surgery report noted that reported cost for elective endovascular aortic repair (evar) varied between £ , and £ , for no apparent reason and with no indication that lower cost procedures were less effective , ; ces encompass a wider societal perspective than just the clinician's or patient's point of view alone, helping demonstrate equitable resource allocation in a publicly funded service ; ces allow evaluation of short-and long-term costs and benefits, which are often under-or overestimated; and ces provide an explicit and accountable framework for decision making, which can be re-examined as data accumulate, particularly important with evolving techniques and experience as in interventional radiology (ir). healthcare costs continue to grow faster than the economy as measured by gdp: since public spending on health in the uk has increased by . % per annum on average while gdp has grown by . %, with similar trends in other european countries , , due to: demographic change, chronic medical conditions and rising cost of medical infrastructure and medical technology. as mean population age has risen due to increasing life expectancy and falling fertility rates, so has healthcare expenditure. the precise reasons for this are complex, but broadly divide into the "sisyphus effect" (more elderly expect to be fit and independent into older age, requiring more medical resources, creating more elderly), and the "multimorbidity hypothesis" (decreasing mortality rates create a larger pool of less fit multimorbid elderly ). most costs arise in the last months of life regardless of age, , where ir may be a highly acceptable substitute for surgery; for example, embolisation for gastrointestinal bleeding, or an useful option for symptom palliation, such as placement of ascitic or pleural drains. , chronic medical conditions accelerating rates of obesity and diabetes worldwide are significantly increasing the incidence of vascular disease and cancer. the number of adults with diabetes is projected to increase worldwide by % by . the risk of peripheral arterial disease in patients with diabetes is increased by a factor of . , more than smoking ( . ), with a prevalence of % in those aged > yearshttps:// paperpile.com/c/e smlw/ hzys, and more diffuse and infrapopliteal disease, which is challenging to treat. the increasing demand for rapid diagnosis, treatment, and palliation means interventional oncology is now regarded as a vital and highly cost-effective component of cancer care. rising cost of medical infrastructure and medical technology about half of the increased overall healthcare expenditure in high-income countries is due to increasing costs of medical technology, including equipment and drugs. , e when analysed in more detail, costs have actually fallen in some conditions but increased disproportionately in conditions with high-technology interventions, including ir. these factors indicate increased demand for, and cost of, ir in the future. ces can consider not only costs to the healthcare system, but also costs incurred by the patient, such as loss of ability to look after family, and wider societal costs. the benefits calculated have developed from unadjusted life years to quality adjusted life years (qalys) gained, where quality of life is gauged using questionnaires such as eq- d or sf- d, , and multiplied by years of survival, so that qaly¼ year of perfect health, reducing to . for poor health and zero for death. qalys provide a consistent and transparent means of comparing the outcomes of different surgical, ir, or medical procedures. ces often show states of health using decision trees (fig ) . these represent the points of treatment decision or choice (with probability of choice) as decision nodes, leading to various outcomes represented by branches, which each lead to a chance node with probabilities of different clinical outcomes, repeating until the patient has no further decisions or changes in risk, represented by a terminal node. costs and health outcomes are ascribed to each branch. the tree thus generated can be "rolled back" to calculate the overall costs and outcomes for each treatment option. in the healthcare scenario, the patient may have a relapsing or recurrent condition and can transition back and forth from different health states, e.g., remission and active disease, so decision trees that only allow one-way progression become unmanageable; these are better demonstrated by markov models (fig ) , which allow two-way progression and map the health states and the probability of transitioning between these states after an intervention during a given time cycle. the time spent in each health state is associated with a cost and outcome, and these can be aggregated to calculate overall costs and qalys for each treatment option. the incremental cost effectiveness ratio (icer) is calculated as the difference in cost divided by difference in qalys between different strategies, compared to an alternative. for example, the evar and dream trials showed an icer of £ , per qaly for evar in patients unfit for surgical repair. in the national health service (nhs), a icer threshold of £ , to £ , per qaly (and up to £ , for end-of-life treatments) has been considered reasonable to decide whether a treatment is cost effective versus baseline, and although it has been emphasised by nice that thresholds are not fixed, they have attracted attention for seeming to put an arbitrary price on health, or being associated with rationing of resources. there is no explicit rationale for the cost per qaly, although some commentators relate it to the share of gdp per person in an economy, or average household income. the threshold can cause issues with new ir technologies where initial prices may be higher reflecting development costs or low volume production runs. notably the affordable care act in the usa forbad the use of cost per qaly as a threshold, to counter accusations of enacting "big-government healthcare" or setting up "death panels". fig illustrates the icer cost-effectiveness plane, which plots cost in pounds sterling against effect in qalys. the icer threshold represents the acceptable threshold to decide whether the increased effect of a procedure justifies its cost. compared to the baseline procedure o at the origin, treatments a and b are more effective but more costly, c is more costly and less effective, d is less costly and less effective, but above the threshold, and e is less costly and more effective, so the clear winner. the usual situation is a or b: in this case, it can be seen that b would be preferable as it is below the icer threshold. sensitivity analyses (sa) vary key inputs (e.g., probability of treatment efficacy or price of equipment) where there is uncertainty of real or estimated parameters and can determine which factors in the model are the main drivers in cost effectiveness. if the icer varies little when inputs are varied, the findings can be considered more robust. sa can yield unexpected insights into cost effectiveness and are thus a powerful tool in designing pathways to maximise benefits gained for a given budget and improve a procedure's icer. for example, contrast-enhanced follow-up post-evar with abdominal radiography and ultrasound has been shown to be more cost-effective than computed tomography (ct). although widely used, qalys are inherently subjective as they are based on valuing the patient's quality of life (qol), raising ethical, methodological, and disease-specific issues. , ethical issues include valuing another's qol when one can have no experience of having the specific physical disorder. methodological issues involve relating this subjective judgement to one or more of many different health utility scores, which introduces further intra-and interobserver variability (e.g., varying among different age groups, nationalities). disease-specific issues arise in less common conditions where estimates of the impairment caused vary widely. uncertainties also arise in costing items, such as time off work after surgery or an interventional procedure is highly variable between patients and difficult to quantify. the rapid growth in health ces has created the need for a checklist to ensure that different ces are complete and contain enough information to be compared and combined with each other for meta-analysis. the cheers (consolidated health economic evaluation reporting standards) checklist is recommended when planning ces to ensure quality and completeness. many recent radiological studies assessed by this checklist are incomplete. the quality of the conclusions drawn from ces are only as robust as the data used, and if these are drawn from a wide range of disparate studies, or are biased or not generalisable, this will limit the strength of the study. choice of control groups or comparison strategies will influence the icer and sensitivity results. the time horizon will also have a large effect, especially because of compound discounting. there are already some conditions where ir has been shown to provide a more cost-effective solution (table ) . areas where ir can improve and better demonstrate its cost effectiveness are discussed below. robust trials with clear outcomes are the basis of ces, and many trials comparing ir are underpowered or incomplete for the following reasons: study power, recruiting team factors, and non-robust outcome measures. recruitment into trials is often expensive and difficult, and many fail to achieve planned recruitment. nihr found that less than one third of clinical trials achieved their recruitment target. from to , % of cardiovascular trials registered on clinicaltrials.gov terminated early, mainly due to poor recruitment or high patient dropout. the latter often arises as some patients prefer a specific option, often what they perceive as more "active" treatments such as angioplasty or stenting over medical therapy, e.g. in the exact (exercise versus angioplasty) only % of screened patients agreed to be randomised and the trial was terminated early. other reasons are unwillingness to undergo treatment regarded as "experimental", and non-compliance with follow-up. e complex trial protocols increase non-completion rate. recruitment may also prove prohibitively expensive: a study on peripheral arterial disease demonstrated a wide range of recruitment cost from $ (at a community event) to more than $ , (radio advertising) per randomised participant: the same study spent more than $ , on recruiting participants, a mean of approximately $ , . a business model approach can help to improve recruitment, retention, and trial completion (table ) . specialist research nurses can make a significant contribution in recruiting, retaining, and following up patients. research teams often have to balance research and clinical commitments. clinician participation may be table examples of interventional radiology procedures proven more cost effective than surgery. suboptimal where the clinician feels options are not in equipoise or is unfamiliar with, unconvinced by, or actively resistant to a novel technique. pseudorandomisation may occur whereby sicker patients with more comorbidities are referred for ir procedures and fitter ones for surgery, thereby introducing bias into the long-term outcomes and decreasing generalisability of the results to the general population. a whole team approach should be used, with wide involvement in auditing, presenting, and publishing the results of the work. clinical endpoints need to be objective, robust, and verifiable, or bias may be introduced. there needs to be clear guidance of who assesses, when and how. nonblinded observers have been shown to exaggerate treatment effect by up to % versus blinded. lack of definition of "best supportive care" in oncology studies has been shown to distort interpretation of outcomes. lack of long-term follow-up is a common issue in interventional procedures where patients revert back to the original referring consultant for ongoing review. some evar studies found only % of patients having complete surveillance as per the protocol. , ir involvement in follow-up would not only improve the evidence base for long-term efficacy, but would also afford ir the opportunity to advise or intervene if late complications arise. learning curve and procedure volume effects medical outcomes improve as teams complete a "learning curve" and start performing larger volumes of procedures regularly. , interventional radiologists perform a wide variety of emergency and elective procedures, which are being continually developed and elaborated. it is challenging to master and maintain competence in these different procedures, which may be infrequently performed. these factors can reduce outcomes and cause delayed recognition of complications, which may require expensive retreatment and further surveillance, e thus reducing cost effectiveness. two strategies can help here. first, increasingly sophisticated simulation training (st) has been developed, which provide detailed real time audiovisual and haptic feedback. simulation training has already been embraced by cardiologists (for arterial puncture and coronary artery catheterisation), neurosurgeons (for neurovascular procedures such as aneurysm coiling and stroke thrombectomy) and vascular surgeons (for renal, carotid and peripheral vascular procedures), e and st has been explicitly incorporated into their curricula. st allows the trainee to experience a standardised set of scenarios designed to cover a range of teaching points and embed useful skills in the most efficient way, without relying on random caseload in a particular centre, at a convenient time and setting conducive to learning, with objective and supportive feedback. it has been shown to reduce procedure time and radiation dose, improve outcomes and patient safety, and operator confidence. the operator can learn at their own pace, and pay attention to areas where they need more training, or where they feel less confident. st can be particularly helpful to teach the basic skills rapidly, such as arterial puncture and selective catheterisation, allowing the trainee to concentrate on the more advanced aspects of the procedure. it reduces the burden on trainers, , and increases patient safety. st can also be valuable for experienced operators to maintain and update their skills, with objective feedback. st is still expensive, but resources can be shared at a regional or college level, as most trainees only require e days of practice initially with shorter further sessions as required. secondly, hub and spoke models with concentration of more advanced services in a hub with higher volumes have been noted by the girft report on vascular surgery to achieve better outcomes with better use of healthcare resources and opportunities for cost saving due to improved procurement. , there is already a tendency for hub and spoke working to provide -h ir availability, and increasing use of this model will improve outcomes at relatively minor increased cost, thus improving costeffectiveness. interventional procedures lend themselves to shorter length of stay, due to smaller incisions, decreased use of general anaesthetic, and faster recovery times, versus surgical alternatives. this leads to shorter operating time, length of stay in hospital, time away from home and back to work and therefore significant cost savings, especially if one considers an overnight stay costs £ on a ward, and £ , in a critical care unit (nhs reference costs ). thus moving from evar with open surgical femoral exposure to percutaneous radiological femoral access was shown to table improving the evidence base: a marketing approach. building brand value and defining purpose of the trial gain legitimacy and prestige (coordinated by an academic centre, funding by a noncommercial body, signal worthiness (that benefits to trial participants will outweigh costs) marketing and product planning adopt an explicit marketing plan with stakeholder engagement, local and regional champions, and strategies for overcoming resistance (address concerns), providing a complete administrative process with easy data collection and transfer "making the sale" deliver a targeted multi-level approach to multiple different audiences (with appropriate language) and achieving "buy-in" (confirmed commitment), through websites and communications back to trial participants maintaining engagement especially important when follow-up or supplementary studies are envisaged. key points: deal with feedback constructively, continue to provide reinforcement, and communicate findings and positive learning points reduce operating theatre time by %, length of stay by % and overall cost by %. the audit commission noted in that switching to day-case surgery nationwide would allow up to , patients more to be treated annually without extra expenditure, and similar calculations would show significant cost savings from increased use of ir. furthermore, in the current covid era, many ir units have been able to continue doing day-case and outpatient treatments without cancellations due to ward or critical care shortages, also helping to reduce waiting lists, which have recently risen sharply. a substantial proportion of ir procedure cost is the kit used, which may be substantially more than the surgical equivalent (e.g., average price of evar graft and wires £ , versus surgical aortic graft and consumables £ ), due to small production volumes and requirement to recoup development costs. the girft vascular surgery report noted a nearly -fold variation nationwide in cost of evar grafts with no apparent difference in effectiveness, implying a massive opportunity to reduce costs with judicious procurement. equipment prices tend to fall with increasing production and commercial competition (as with coronary angioplasty balloons and stents) so it is important to continually review the market to see if more costeffective kit or cheaper alternatives become available. with increasing experience, less kit is generally used, which underlines the value of st as above. costs vary significantly between countries, e.g., coronary stents cost six-times more in the usa versus uk and germany, so local costs should be obtained and compared. complications can massively increase the cost of a routine procedure, and therefore relatively small additions to a procedure may prove highly cost effective if they reduce complication incidence. this has been an important factor in evar, especially in those patients with relatively long lifespan post-procedure, and may easily tip the balance of cost-effectiveness between ir and surgery. some radiological interventions provide short-term benefit but not longterm durability: for example, in uterine artery embolisation (uae), the rest trial showed the initial cost benefit of uae over surgery at months was eroded by a higher rate of treatment failure rate in the embolisation group ( %) versus surgery ( %), which reduced the cost effectiveness to equipoise by years. similarly, the hopeful study showed improved quality of life initially but with erosion of benefits over time. for aortic aneurysm, evar was noted to have short-term benefits, but lack of benefit in the long term (as well as increasing costs from re-intervention) leading nice to approve it only for acute rupture , and not elective repair of uncomplicated aortic aneurysms. , the reasons for complications and lack of durability need to be acknowledged and addressed. with increasing experience and training, complications can be recognised and mitigated earlier and more cost effectively. developments in technique frequently occur: for example, in uae, the role of anastomotic vessels causing regrowth of fibroids has been recognised and can be treated primarily or with reintervention. long-term studies can also help define which subgroup of patients are likely to benefit most, so patients can be better informed of the options most likely to suit them and therefore prove cost-effective. sometimes a different treatment paradigm may prove more suitable. for example, magnetic resonance imaging (mri)-focussed ultrasound may prove more effective for certain groups of patients with fibroids, and this is currently being trialled. similarly, the evar strategy of internal graft fixation by radial force and non-abolition of the aneurysm sac may be superseded. so far, different approaches such as endovascular aneurysm sealing (for example with nellix) have not proved successful, although eventually a hybrid technique with evar graft plus sac filling may prove the best solution. imaging follow-up adds significant costs, especially if cross sectional, such as ct, which costs three times the cost of ultrasound. follow-up consultations can often be done more conveniently for patients and ir teams by telephone rather than face to face, and by different team members such as specialist nurses, , and large societal cost savings can be obtained due to reduced hospital usage, fewer missed appointments, and decreased transport costs. the duration of follow-up for all procedures should be continually under review as better and longer-term data becomes available, and tailored to where re-intervention can make a significant difference. ir is a rapidly evolving field with many potential clinical advantages for patients and cost-effectiveness advantages for the whole healthcare system. ir can demonstrate and promote its value with high-quality robust long-term data, using criteria developed by nice, which is recognised worldwide as a leader in objective evaluation of healthcare effectiveness, working in the best interests of patients. the methodology of cost-effectiveness analysis and areas where improvements could be made have been discussed. it is hoped that ir teams will use these and work together to expand this evidence base, which can potentially be of great benefit to all. data briefing. why budgets have always been a bugbear the economic response to coronavirus will substantially increase government borrowing. insitute for fiscal studies methods for the economic evaluation of health care programmes vascular surgery girft programme national specialty report. girft (getting it right first time) fiscal sustainability and public spending on health. obr office for budget responsibility machines that go "ping": medical technology and health expenditures in oecd countries is there a sisyphus syndrome in health care? multimorbidity, health care utilization and costs in an elderly community-dwelling population: a claims data based observational study population ageing and health care expenditure: new evidence on the "red herring a longitudinal study of the effects of age and time to death on hospital costs endovascular therapy for palliative care of cancer patients palliative care for interventional radiology: an oncologist's perspective international diabetes federation. idf diabetes atlas. th edn. international diabetes federation comparison of global estimates of prevalence and risk factors for peripheral artery disease in and : a systematic review and analysis peripheral arterial disease in patients with diabetes peripheral arterial disease in diabetic and nondiabetic patients: a comparison of severity and outcome developing a roadmap for interventional oncology the ageing report: economic and budgetary projections for the eu member states the drivers of public expenditure on health and long-term care: an integrated approach medical care costs: how much welfare loss? medical technology as a key driver of rising health expenditure: disentangling the relationship changes in the costs of treatment of selected illnesses eq- d for the assessment of health-related quality of life and resource allocation in children: a systematic methodological review validation and comparison of euroqol- dimension (eq- d) and short form- dimension (sf- d) among stable angina patients. health qual life outcome abdominal aortic aneurysm endovascular stentgrafts final appraisal determination national institute for clinical excellence and its value judgments the limits of cost-effectiveness analysis. circ cardiovasc qual outcome the cost of cures: what does zolgensma teach us about the use of cost-effectiveness assessments? legislating against use of cost-effectiveness information use of colour duplex ultrasound as a first line surveillance tool following evar is associated with a reduction in cost without compromising accuracy the limitations of qaly: a literature review retaining, and enhancing, the qaly. value health consolidated health economic evaluation reporting standards (cheers)dexplanation and elaboration: a report of the ispor health economic evaluation publication guidelines good reporting practices task force cost-effectiveness analysis in radiology: a systematic review the clinical effectiveness and costeffectiveness of ablative therapies in the management of liver metastases: systematic review and economic evaluation metastatic colorectal carcinoma: cost-effectiveness of percutaneous radiofrequency ablation versus that of hepatic resection cost-effectiveness of hepatic arterial chemoembolization for colorectal liver metastases refractory to systemic chemotherapy cost-effectiveness of percutaneous radiofrequency ablation for malignant hepatic neoplasms cost-effectiveness of hepatic resection versus percutaneous radiofrequency ablation for early hepatocellular carcinoma cost-effectiveness of radiofrequency ablation and surgical therapy for small hepatocellular carcinoma of cm or less in diameter radiofrequency ablation versus nephron-sparing surgery for small unilateral renal cell carcinoma: cost-effectiveness analysis mortality, morbidity and healthcare expenditures after local tumour ablation or partial nephrectomy for t a kidney cancer trials update: ufe versus myomectomy versus hysterectomy comparative clinical effectiveness and cost effectiveness of endovascular strategy v open repair for ruptured abdominal aortic aneurysm: three year results of the improve randomised trial the importance and challenge of recruitment for peripheral artery disease randomized clinical trials recruitment to randomised trials: strategies for trial enrollment and participation study. the steps study prevalence, characteristics, and predictors of early termination of cardiovascular clinical trials due to low recruitment: insights from the clinicaltrials.gov registry the exercise versus angioplasty in claudication trial (exact): reasons for recruitment failure and the implications for research into and treatment of intermittent claudication barriers to clinical trials: part ii: knowledge and attitudes of potential participants barriers to clinical trials: part iii: knowledge and attitudes of health care providers barriers to clinical trials. part i: reimbursement problems recruiting african americans with peripheral artery disease for a behavioral intervention trial using a business model approach and marketing techniques for recruitment to clinical trials selection bias in clinical trials observer bias in randomized clinical trials with measurement scale outcomes: a systematic review of trials with both blinded and nonblinded assessors defining best supportive care compliance with long-term surveillance recommendations following endovascular aneurysm repair or type b aortic dissection endovascular aneurysm repair: is imaging surveillance robust, and does it influence long-term mortality? volume-outcome relationships for percutaneous coronary interventions in the stent era the relationship between coronary angioplasty procedure volume and major complications identifying the learning curve for uterine artery embolisation in an interventional radiological training unit learning curve of radiology residents during training in fluoroscopy-guided facet joint injections operator learning curve for transradial liver cancer embolization: implications for the initiation of a transradial access program training in interventional radiology: a simulationbased approach high-fidelity endovascular simulation simulation reduces navigational errors in cerebral angiography training review of simulation training in interventional radiology using fully-immersive simulation training with structured debrief to improve nontechnical skills in emergency endovascular surgery development of simulator system using endovascular evaluator for catheter intervention training endovascular resuscitation and trauma management: education and simulation does simulation-based training improve procedural skills of beginners in interventional cardiology?da stratified randomized study training on a vascular interventional simulator: an observational study does simulation-based medical education with deliberate practice yield better results than traditional clinical education? a meta-analytic comparative review of the evidence national tariff payment system outcomes and cost comparison of percutaneous endovascular aortic repair versus endovascular aortic repair with open femoral exposure audit commission. a short cut to better services: day surgery in england and wales. london: h.m. stationery office response of uk interventional radiologists to the covid- pandemic d survey findings. cvir endovasc covid- : nhs leaders braced for longer waiting times as service deals with fallout the uk endovascular aneurysm repair (evar) randomised controlled trials: long-term follow-up and cost-effectiveness analysis prices for cardiac implant devices may be up to six times higher in the us than in some european countries randomised comparison of uterine artery embolisation (uae) with surgical treatment in patients with symptomatic uterine fibroids (rest trial): -year results uterine artery embolisation or hysterectomy for the treatment of symptomatic uterine fibroids: a costeutility analysis of the hopeful study cost-effectiveness of open versus endovascular repair of abdominal aortic aneurysm in the over trial endovascular treatment of abdominal aortic aneurysm: a nice u-turn abdominal aortic aneurysm: diagnosis and management nice guideline. nice guideline uterine artery embolization: state of the art the positive effect of targeted marketing on an existing uterine fibroid embolization practice review of magnetic resonance-guided focused ultrasound in the treatment of uterine fibroids the firstt: comparing mrgfus (mr-guided focused ultrasound) versus uae (uterine artery embolization) for uterine fibroids is this the end for evar? new technology failures: who to blame or time to be cautious? the costs, resource use and cost-effectiveness of clinical nurse specialist-led interventions for patients with palliative care needs: a systematic review of international evidence the potential economic impact of virtual outpatient appointments in the west midlands d a scoping study the authors declare no conflict of interest. key: cord- -x i xb authors: faugere, louis; klibi, walid; white, chelsea; montreuil, benoit title: dynamic pooled capacity deployment for urban parcel logistics date: - - journal: nan doi: nan sha: doc_id: cord_uid: x i xb last-mile logistics is regarded as an essential yet highly expensive component of parcel logistics. in dense urban environments, this is partially caused by inherent inefficiencies due to traffic congestion and the disparity and accessibility of customer locations. in parcel logistics, access hubs are facilities supporting relay-based last-mile activities by offering temporary storage locations enabling the decoupling of last-mile activities from the rest of the urban distribution chain. this paper focuses on a novel tactical problem: the geographically dynamic deployment of pooled relocatable storage capacity modules in an urban parcel network operating under space-time uncertainty. in particular, it proposes a two-stage stochastic optimization model for the access hub dynamic pooled capacity deployment problem with synchronization of underlying operations through travel time estimates, and a solution approach based on a rolling horizon algorithm with lookahead and a benders decomposition able to solve large scale instances of a real-sized megacity. numerical results, inspired by the case of a large parcel express carrier, are provided to evaluate the computational performance of the proposed approach and suggest up to % last-mile cost savings and % capacity savings compared to a static capacity deployment strategy. cost-efficient way. in , % of the world's population lived in urban areas (up to % in north america) . the united nations ( ) predict that global urbanization will reach % by , with an increasing number of megacities (cities of +m inhabitants). increasing population density is a challenge for city logistics in terms of traffic congestion, vehicle type restrictions, limited parking spaces, expensive and rare logistic facility locations, and is further complex in megacities due to their extremely high density (fransoo, blanco, and argueta ) . for urban parcel logistics systems, the growth of e-commerce is currently one of the main challenges to tackle with an annual growth over % on the - period, projected to be over % until % until (emarketer . online-retailing with goods being transported to consumers' homes increase the number of freight movements within cities while reducing the size of each shipment (savelsbergh and van woensel ) which makes first and last mile logistic activities harder to plan. moreover, consumers' desire for speed (i.e. same-day delivery and faster) has yet to be met by online retailers (emarketer ) . with promises as fast as -hour delivery (e.g. amazon prime in select u.s cities), the cost of last-mile logistics becomes an ever more critical part of urban parcel logistics. these trends have been accelerated due to attempts to mitigate the impacts of the covid- pandemic (e.g., sequestering in place), requiring companies to increase their last-mile delivery capabilities and to deal with the dramatic shift to online channels (wade and bjerkan ) . to tackle these challenges, a number of innovations have emerged from academia and industry. savelsbergh and van woensel ( ) provide an overall view of recent innovations and modeling of solutions such as multi-echelon networks, dynamic delivery systems, pickup and delivery point networks, omni-channel logistics, crowd-sourced transportation and the integration of public and freight transportation networks. many of these innovations are considered in the physical internet initiative, introduced in montreuil ( ), which seeks global logistics efficiency and sustainability by transforming the way physical objects are handled, moved, and stored by applying concepts from internet data transfer to real-world shipping processes. a conceptual framework on the application of physical internet concepts to city logistics was recently proposed in crainic and montreuil ( ) , in particular the concepts of pooling and hyperconnectivity in urban multi-echelon networks. as underlined by savelsbergh and van woensel ( ) , city logistics problems integrating real-life features such as highly dynamic and volatile decision making environments, sharing principles or multi-echelon networks, offer a fertile soil for groundbreaking research. inspired by the case of a large parcel logistics company operating in megacities, this paper examines a novel tactical optimization problem in urban parcel logistics. it consists in the dynamic deployment and relocation of pooled storage capacity in an urban parcel network operating under space-time uncertainty. it builds on the recent proposal of a hyperconnected urban logistics network structure (montreuil et al. ) in line with the new challenges of the parcel logistics industry. the proposed network structure is based on the pixelization of urban agglomerations in unit zones (clusters of customer locations), local cells (cluster of unit zones) and urban areas (cluster of local cells). it is composed of three tiers of interconnected logistics hubs: gateway hubs (gh), local hubs (lh) and access hubs (ah) respectively designed to efficiently handle inter urban areas, inter local cells, and inter unit zones parcel flows. beyond the realm of an urban agglomeration, the network of gateway hubs connects to a network of regional hubs (rh) covering entire blocks of the world (e.g. north america), and these regional hubs connect to a worldwide network of global hubs. this paper focuses on access hubs which are small logistics hubs located at the neighborhood level within minutes of customers, enabling parcel transfer between different vehicle types temporarily holding parcels close to pickup and delivery points. access hubs are to be used by logistics carriers, and not by consumers as smart lockers are. access hubs can materialize in many forms including a parked trailer, a smart locker bank, or a storage shed as illustrated in figure . trailer based solutions like figure parcel logistics networks have undergone significant changes in the last years, notably in urban contexts as seen in janjevic and winkenbach ( ) , and have received an increasing attention in the academic literature. strategic and tactical network design problems such as the ones examined by smilowitz and daganzo ( ) , winkenbach, kleindorfer, and spinler ( ) approximate operations costs when designing and planning for multi-echelon networks. while network design problems are complex due to intricate interdependencies between strategic, tactical and operational decisions, continuum approximations (see ansari et al. ( ) ) are useful to capture operations complexity and take informed decisions. however, such approximations are typically used to estimate travel distance and cost, but not travel time and operations synchronization. this paper considers access hubs to be modular in storage capacity similar to designs proposed in faugère and montreuil ( ) , such that capacity modules can be removed/added to adapt access hub's storage capacity. at the tactical level, capacity modules are to be deployed over a network of access hub locations; at the operational level, capacity modules are to be allocated to serve their access hub's need or neighboring locations via capacity pooling. in a dynamic setting, the associated problem can be related to a multi-period location-allocation problem which belongs to the np-hard complexity class (manzini and gebennini ) . once the capacity of the network of access hubs is adjusted, each access hub plays the role of a transshipment location between couriers performing pickup and delivery services within minutes of the access hub and riders transporting parcels between local hubs and a set of access hubs. such transshipments require tight synchronization of the two tiers so as to provide efficient and timely pickup and delivery operations. this operational context mimics, on a hourly basis, a two-echelon pickup and delivery problem with synchronisation, which is a complex routing problem (see for instance cuda, guastaroba, and speranza ( ) ). thus, the integration of operations in the tactical decision model leads to better capacity deployment decisions (klibi, martel, and guitouni ) , yet induces solvability challenges due to its combinatorial and stochastic-dynamic structure. this paper studies a novel tactical optimization problem: the dynamic deployment of pooled storage capacity in an urban parcel network operating under space-time uncertainty. its contribution is threefold: ( ) the characterization of a new tactical problem for capacity deployment, motivated by dynamic aspects of urban parcel logistics needs, ( ) the modeling of the access hub dynamic pooled capacity deployment problem as a two-stage stochastic program with synchronization of underlying operations through travel time estimates, and ( ) the design of a solution approach based on a rolling horizon algorithm with lookahead and a benders decomposition able to solve large scale instances of a real-sized megacity. numerical results, inspired by the case of a large parcel express carrier, are provided to evaluate the computational performance of the proposed approach and suggest up to % last-mile cost savings and % capacity savings compared to a static capacity deployment strategy. section summarizes the literature relevant to this type of problem, section describes the problem and proposes a mathematical modeling, section presents the proposed solution approach, section provides an experiment setup and discusses results, and section highlights key takeaways and managerial insights, and identifies promising research avenues. multi-echelon network for urban distribution have received a lot of attention in the academic literature (e.g. benjelloun and crainic ( ) , mancini ( ), janjevic, winkenbach, and merchán ( ) ), commonly using urban consolidation centers (ucc) to bundle goods outside the boundaries of urban areas. as reported in janjevic and ndiaye ( ) , several micro-consolidation initiatives have been proposed to downscale the consolidation effort by bundling goods at the neighborhood level using capillary networks of hubs located much closer to pickup and delivery points, defined as access hubs in the conceptual framework proposed by montreuil et al. ( ) . examples of such initiatives are satellite platforms (e.g. benjelloun and crainic ( ) ), micro-consolidation centers (e.g. leonardi, browne, and allen ( ) ), mobile depots (e.g. marujo et al. ( ) ), and micro-depots stodick and deckert ( ) . most of the focus has been on location and vehicle routing aspects (e.g. anderluh, hemmelmayr, and nolz ( ) and enthoven et al. ( ) ) and cost and negative externalities assessment (e.g. verlinde et al. ( ) , arvidsson and pazirandeh ( ), marujo et al. ( ) ) in solutions using depots and cargo-bikes. to the best of the authors' knowledge, the dynamic management of access hub capacity for urban parcel logistics has not yet been studied in the academic literature. the problem studied in this paper involves modular capacity relocation and a capacity pooling recourse mechanism impacting the operations of a two-echelon synchronization problem. in this section, a literature review on dynamic capacitated facility location problems and integrated urban network design problems is presented. dynamic facility location problems where systems are subject to varying environments (e.g. non-stationary demand) allow the relocation of facilities over time. arabani and farahani ( ) provide a literature review on facility location dynamics, including problems with and without hub relocation. innovations in the manufacturing industry have motivated the study of modular and mobile production and storage. marcotte and have presented various threads of innovations such as distributed production, on-demand production, additive production, and mobile production, that would motivate and benefit from hyperconnected mobile production systems. marcotte, montreuil, and coelho ( ) and malladi, erera, and white iii ( ) proposed mathematical modeling for production and inventory capacity relocation and allocation to manage multi-facility network facing stochastic demand. however, they examine small to medium networks far from the scale of urban parcel logistics networks and do not study operations synchronization. aghezzaf ( ) studied storage capacity expansion planning coupled to dynamic inventory relocation in the context of warehouse location allocation problems, but did not consider capacity reduction or relocation. ghiani, guerriero, and musmanno ( ) , melo, nickel, and da gama ( ) , and jena, cordeau, and gendron ( ) modeled dynamic facility location problem where not only sites could be permanently or temporarily opened or closed, but also resized by adding or removing modular capacity. melo, nickel, and da gama ( ) proposed models capturing modular capacity shifts from existing to new facilities. however in these problems, capacity relocation is generally not managed jointly with capacity allocation or its impact on underlying operations. dynamic facility location literature partially covers the tactical capacity relocation problem studied in this paper, but does not integrate underlying operations dynamics at the urban logistics scale. integrated network design problems typically deal with a combination of strategic decisions such as facility location, tactical decisions such as resource allocation and scheduling, and operational decision such as vehicle routing. the integration of these different levels of decisions can be found in two main problem classes: service network design problems and location routing problems. service network design problems deal with the selection and scheduling of services such as hub operations, shipping lines and routing of freight (e.g. , hewitt et al. ( ) ) while location routing problems combine facility location-allocation decisions with associated freight routing decisions. drexl and schneider ( ) provide a recent survey of variants and extensions of the location routing problem. the dynamic location routing problem (francis, smilowitz, and tzur ( ) ) considering the assignment of demand to locations over multiple periods, is similar to the problem studied in this paper: it aims at minimizing network and routing costs over a multiperiod location and routing decision vector. however, multi-echelon location routing problems (e.g. crainic, ricciardi, and storchi ( ) , perboli, tadei, and vigo ( ) ) have only recently gathered attention in the literature. although multi-echelon networks are relevant to postal and parcel delivery distribution systems (gonzalez-feliu ( )) where fine time constraints and synchronization have become an essential consideration, most papers studying multi-echelon networks are concerned with the two-echelon case and ignore temporal aspects (drexl and schneider ( ) ). when allowing inter-location capacity pooling, underlying operations described in section are impacted. couriers perform pickup and delivery tours starting and ending in their reference access hub, while riders visit access hubs starting and ending their routes in their reference local hub. the impact of capacity pooling can be measured by modeling its impact on the route of parcels, couriers and riders. however, when taking decisions at the tactical level, explicitly modeling routes is not necessary. tsp and vrp continuous approximations have been introduced by daganzo ( daganzo ( , to embed operations in strategic and tactical logistics problems (e.g. erera ( ) , franceschetti et al. ( ) ). a recent literature on variants of this approach can be found in ansari et al. ( ) . smilowitz and daganzo ( ) , winkenbach, kleindorfer, and spinler ( ) , bergmann, wagner, and winkenbach ( ) adapted these continuous approximations to the context of parcel express logistics to approximate distance traveled and cost. however, the aspect of synchronization using travel time continuous approximations has not yet been studied. to the best of the authors' knowledge, this paper is the first to study a capacity relocation problem with the synchronization of two-echelon routing operations through travel time estimates. a parcel logistics company provides pickup and delivery services to customers in a region covered by a network of access hubs. the network of access hubs may be dedicated to the parcel logistics provider, or shared between several companies as suggested by the concept of open networks in the physical internet. figure provides a conceptual illustration of the network of access hubs and the relocation of capacity modules over two deployment periods. once the network capacity is set, pickups from customers are dropped off by couriers in access hubs and will occupy a certain storage volume for some time until a rider picks them up to perform outbound activities. to-be-delivered parcels are dropped off by riders in access hubs and will occupy a certain storage volume for some time until a courier picks them up to perform the delivery to customers. to provide good service, the company must ensure that parcels flow rapidly and seamlessly between couriers and riders, which requires the sound management of storage capacity deployed in access hubs. storage volume requirements vary depending on the fluctuation of demand for pickup and delivery services over time and are observed over a discrete set of operational periods (e.g. hourly). access hubs are composed of modular storage units that can be assembled and disassembled relatively easily, enabling rapid relocation of storage capacity in the network. during each deployment period (e.g. week or day), storage capacity can be relocated within the network of access hubs, or to/from a depot where additional capacity modules are stored when not in use. figure illustrates demand variability and the relocation of capacity modules within the network of access hubs over two deployment periods. for instance, unit zones with increasing demand (and therefore increasing capacity requirements) from period t to t + receive capacity module(s) from the depot of from locations that have decreasing capacity requirements (e.g. lower left unit zone in figure ). the relocation of capacity modules over the network adjusts the storage capacity available in each ac-cess hub for the following period. in this study, we assume capacity module relocation is performed by a separate business unit whose routing decisions are out of the scope of the research reported in this paper. the objective is to minimize the cost incurred by operating such a network of access hubs without disrupting underlying operations. the decision scope is tactical (capacity deployment) and requires the integration of operational decisions. however, since the main interest is a set of tactical decisions, there is no need to explicitly model operations, but only to approximate the impact of deployment decisions on routing cost and time synchronization. let l be a set of access hub locations and w a set of external depots composing a network g = (n = l ∪ w, a) where a is the complete set of directed arcs between locations in n. a capacity deployment of i capacity modules in time t over the network is represented by a vector s(t) = (s l (t), ∀l ∈ n). the relocation of capacity modules can be represented as vectors r(t) = (r a (t), ∀a ∈ a). accordingly, there are ) possible arrangements of i modules over |l| locations. in the case where i ≥ |l| and that each location gets at least one module, there are ( i − |l|− ) possible arrangements. in this realistic context, access hub networks are expected to be composed of a high number of locations (i.e., hundreds). thus, state and action spaces would be significantly large-sized, which results in curse of dimensionality issues (powell ( ) ). moreover, a set of realization scenarios ω ∈ Ω with probability φ ω is considered. the number of pickups and deliveries as well as the storage volume requirements are observed hourly and respectively represented as a vectors ρ p (τ, ω) = (ρ p l (τ, ω), ∀l ∈ l), ρ d (τ, ω) = (ρ d l (τ, ω), ∀l ∈ l) and d(τ, ω) = (d l (τ, ω), ∀l ∈ l), for every operations hour τ ∈ t t , where t ∈ t is an operations horizon between two deployment periods (e.g. a week). if a courier or rider observes a lack of storage capacity when visiting an access hub, the courier or rider can perform the following recourse actions: pool capacity by making a detour towards a neighboring access hub with extra capacity or consign its load to a nearby third-party business (e.g. local shop) for a certain price agreed upon (uncapacitated recourse). once volume requirements are observed, recourse actions are taken for each operational period τ: capacity pools as a vector p(τ, ω) = (p a (τ, ω), ∀a ∈ a pool ) where a pool is the set of arcs on which capacity can be pooled, and consignments as a vector z(τ, ω) = (z l (τ, ω), ∀l ∈ l). at any time τ in scenario ω, the system can thus be represented as a state s t = s(t) s.t. τ ∈ t t and an action based on the optimisation framework proposed in powell ( ), our stochastic optimization challenge for the access hub dynamic pooled capacity deployment problem can be formulated as follows: where x(τ) is the set of feasible actions at time τ, s is the initial state of the system, and c τ (·) is the cost function at time τ. figure illustrates the dynamics of the problem with the tactical decision timeline: before each period t, a network deployment strategy s(t) is decided through relocation decisions r(t) and implemented right before the beginning of period t. then, demand realized and recourse actions are taken in each period τ ∈ t t . at the end of periods t t , a network deployment strategy s(t + ) is decided through relocation decisions r(t + ) and implemented right before the beginning of period t + and the process repeats. once decisions on capacity deployment are set for a given period t, they strongly impact the quality of operations performed by couriers and riders. more specifically, capacity at each location impacts the number and costs of detour and perturb the synchronisation of the operations between couriers and riders at each location. accordingly, the surrounding objective of integrating routing operations is to evaluate the performance of the capacity deployment in minimizing the detours due to an underestimation of the capacity needs and in guaranteeing the synchronisation of the operations between couriers and riders at each location. to do so, this subsection proposes to develop routes with detours cost approximations, and travel time approximations. it builds on a refined granularity of routing operations periods (hourly) and uncertain storage volume requirements. it is clear that capturing the dynamics of underlying operations when taking capacity deployment decisions leads to better solutions. however, the pickup and delivery problem with transshipment is np-hard (rais, alvelos, and carvalho ( ) ) and including it explicitly in the tactical model would make it intractable. since the goal is to foster best capacity deployment decisions, it is sufficient to anticipate the operations costs and time synchronisation constraints using scenario-based continuous approximations. accordingly, hereafter is proposed a tractable approximation of each period τ pickup and delivery problem with transshipment by developing deterministic continuous approximations of vehicle routing problems. the starting point of the proposed approximations is the estimation of the vehicle routing problem length when the depot (from which vehicles start their routes) is not necessarily located in the area where customers are located as proposed in daganzo ( ) : where r is the average distance between the depot and the customer locations, m is the number of routes required to serve all customers, n is the number of points to be visited, k is a constant parameter that can be estimated through simulation (daganzo ( ) ), and δ is the density of points in the area. an a priori lower bound on the number of routes required to serve all customers, m, is n/q where q is the capacity of one vehicle in terms of customer locations. the first term of approximation ( ) represents the line-haul (back and forth) performed by vehicle to travel from the depot to the area where customers are located, and the second term represents the tour performed by traveling between each successive stops. based on these seminal works, the next subsection proposes an adaptation of these equations to the operational context of riders and couriers, and develops an explicit time-based estimation of their operations. riders work in local cells, which are clusters of access hubs served by the same upper level local hub(s) as illustrated in figure . riders visit a set of n lc access hubs within their local cell of area a lc (and density δ lc = n lc a lc ) to pickup and deliver parcels as part of a defined route (e.g. planned beforehand based on averaging network's load). at the time of deployment, underlying riders' routes are not known with certainty, but need to be estimated in order to anticipate operations performance. when a rider makes his tour in period τ under scenario ω two cases are possible: (i) the tour is operated as planned because sufficient capacity is deployed at all visited access hubs in the route or because the detours are assigned to access hubs that are already in the remaining itinerary of the rider (bold lines in figure ); (ii) the rider tour is perturbed due to a lack of capacity at an access hub, and thus has to perform an immediate detour to a neighboring access hub before pursuing the rest of the regular tour (dash lines in figure ). ( ), if the number of detours performed by riders in local cell lc in period τ in scenario ω is n r lc (τ, ω), the route length estimation with detours of riders' operations is: where n lc is the total number of access hubs in local cell lc, r lc is the average distance between lc's local hub(s) and its access hubs, and m r τ (ω) is the number of riders' operating. the cumulative time (in time-rider) necessary to perform tours approximated in ( ) is: where the first term is the time spent to setup tours (t r s per tour) and perform the line-haul at a speed of s r , the second term represent the travel time between stops at a speed of s r and the stopping time t r a per access hub, and the third term represents the service time (handling) t r u per pickup and delivery. thus, the cost associated with riders' operations in local cell lc in period τ in scenario ω is: where the first term represents the fixed, c r f , and variable, c r v in line-haul and c r v in tour, costs associated with vehicles, and the second term represents the variable labor cost c r w of m r τ (ω) riders. since the nominal routing cost (with no detours) is a sunk cost incurred regardless of the capacity deployment, the marginal cost is sufficient to inform the tactical decision of the impact of recourse actions. the marginal cost of the detours induced by the tactical decisions, or difference between the rider routing cost with detours and the nominal rider routing cost, is: where the time associated with performing detours is the time needed to perform detours: couriers operate in unit zones, which are clusters of pickup and delivery points served by access hub(s). couriers leave their reference access hubs to visit customers and perform pickups/deliveries before returning to their access hub. when a courier arrives at the courier's access hub with picked parcels, if the courier observes a lack of capacity, the courier can be immediately directed to available capacity in some neighboring access hub. then, the courier will perform a detour (out and back) to the assigned neighbour access hub before starting their next tour from their reference access hub. figure illustrates a courier's tour and a detour as described. since access hubs are located in the same area as pickup/delivery locations, the line-haul distance at this echelon is negligible, which eliminates the first term of approximation ( ). if the number of detours performed by couriers on arc a ∈ a pool (l) = {a = (l, j), ∀j : (l, j) ∈ a pool } of length d a in period τ under scenario ω is n c a (τ, ω), the route length estimation with detours of couriers' operations is: where the first term represents the total length of tours performed by couriers to visit pickup/delivery locations, and the second term represents the detours (out and back) performed between access hub l and its neighboring access hubs. the cumulative time (in time-courier) necessary to perform courier tours is based on the approximation in ( ) as follows: where the first term represents the travel time between pickup/delivery locations at a speed of s c and the stopping time t c a per stop, the second term represents the travel time during detours to neighboring access hubs at a speed of s c plus a stopping time t c a , and the third term represents the service time (handling) t c u per pickup and delivery. thus, the cost associated with couriers' operations at access hub l in period τ under scenario ω is: where the first term represents the variable travel costs, respectively c c v between pickup/delivery locations and c c v between access hubs, and the second term represents the variable labor cost c c w of m c τ (ω) couriers. again, since the nominal routing cost (with no detours) is a sunk cost incurred regardless of the capacity deployment, the marginal cost is sufficient to inform the tactical decision of the impact of recourse actions.the marginal cost of the detours induced by the tactical decisions, or the difference between the courier routing cost with detours and the nominal courier routing cost is: where the time associated with performing detours is: recall that a key objective of integrating routing operations with the capacity deployment problem is to guarantee the synchronisation of the operations between couriers and riders at each location. to do so, this subsection proposes to develop time-based synchronisation constraints based on the travel time approximations ( ) and ( ), developed above. parcels transshipped from riders to couriers and couriers to riders through access hubs must be transshipped during the period of time the parcels are within the network. that is, the length of a courier's (respectively rider's) original tour, plus the added detour(s) must not exceed the maximum length feasible within one operational period. for riders' operations, at the local cell level, this tour length can be expressed, based on the number of riders (m r τ (ω) in period τ under scenario ω, as follows: where ∆ τ is the length of period τ. similarly, for couriers' operation, at the access hub level, synchronization can be expressed, based on the number of couriers (m c τ (ω) in period τ under scenario ω, as follows: in this section, a stochastic programming formulation is proposed to tackle the optimization problem ( ) presented in section . . we remark that the stochastic optimisation problem ( ) can be modeled as a multistage stochastic program based on a scenarios tree. however, this program would be intractable for realistic size instances, due to its combinatorial structure and non-anticipatory constraints (schültz ) . under a rolling horizon framework, the model is built here on the relaxation approach (shapiro, dentcheva, and ruszczynski ) that is applied to transform the multi-stage stochastic program to a two-stage stochastic program with multiple tactical periods. more specifically, it consists in transferring all the capacity deployment decisions of the t periods to the first-stage in order to be set at the beginning of the horizon. in this case, only first-stage design decisions (t = ) are made here and now, but subsequent capacity deployment decisions (t > ) are deferrable in time according to their deployment period. hereafter are introduced the additional sets, input parameters, random variables and decision variables that formulate the overall model. s l maximum number of capacity modules that can be placed in location l v volume provided by a capacity module v r volume that a rider can carry on a tour v c volume that a courier can carry on a tour d l (τ, ω) volume requirements in location l in scenario ω in period (τ) s l (t) number of capacity modules available in location l for period t r a (t) number of capacity modules relocated through arc a at the beginning of period t p a (τ, ω) volume shared from location i to j, a = (i, j) ∈ n − ω,τ (i) in period τ under scenario ω z l (τ, ω) lack of capacity in volume at location l in period τ under scenario ω n r a (τ, ω) number of detours performed by riders on arc a in period τ under scenario ω n r lc (τ, ω) number of detours performed by riders in local cell lc in period τ under scenario ω n c a (τ, ω) number of detours performed by couriers on arc a in period τ under scenario ω n c l (τ, ω) number of detours performed by couriers from location l in period τ under scenario ω model min ∑ t∈t ∑ l∈l∪w h l s l (t) + ∑ a∈a r a r a (t) s.t.: inventory balance of capacity modules at all locations: total capacity module inventory constraint: spatial constraint at all locations: volume requirements satisfaction constraints: synchronization constraint for riders' operations: ( ) synchronization constraint for couriers' operations: ( ) rider's detours count: courier's detours count: integrality and non-negativity constraints: minimizing expression ( ) corresponds to minimizing the last-mile cost, defined in this paper as the cost of deploying capacity modules in each access hub locations (holding costs) and the relocation costs for each capacity module movement for each reconfiguration period, and the marginal cost incurred by recourse actions (capacity pool from neighboring location and consignment). constraints ( ) and ( ) ( ) and ( ) are the synchronization constraints for the underlying riders and couriers problems as developed in section ( . ). constraints ( ) and ( ) in this section, our rolling horizon solution approach is presented, which builds on solving sequentially the two-stage model presented above using scenario sampling, benders decomposition and acceleration methods. it approximates optimization problem ( ) by planning for one capacity deployment period, t, at the time and deferring subsequent capacity deployment decisions to the following iterations of the algorithm. in order to enhance the quality of the solutions produced at each iteration, a θ tactical lookahead is considered to plan for + θ tactical periods, where only the first period is implementable and the subsequent ones are used as an evaluation mechanism. the proposed rolling horizon solution approach is described in algorithm . here, the length of the sub-horizon is controllable; it can represent one tactical period (i.e. myopic, θ = ) or several of them (i.e. lookahead, θ ≥ ). of course, when dealing with large-scale networks, the selection of the lookahead length is part of the trade-offs necessary to make in order to keep the model tractable. in order to enhance the solvability of the optimization model ( - ), for each subhorizon [t, t + θ], a tailored benders decomposition approach is developed, that fits with the two-stage and multi-period setting of our formulation. it is applied under a large sample of multi-period scenarios. the following subsections address the decomposition approach as well as the associated acceleration methods developed. benders decomposition is a row generation solution method for solving large scale optimization problems by partitioning the decision variables in first stage and second stage variables (benders ( ) ). the model is first projected onto the subspace defined by the first stage variables, replacing the second stage variables result: s l (t), r a (t) suppose the capacity deployment and relocation decisions (first stage decision variables) s l (t), s l (t + ), ... ,s l (t + θ) and r a (t), r a (t + ), ..., r a (t + θ) are given with values s l (t), s l (t + ), ... , s l (t + θ) and r a (t), r a (t + ), ..., r a (t + θ). then, the subproblem can be defined as taking recourse action decisions (i.e. second stage decisions; capacity pooling) to minimize the approximate overall operations costs. the subproblem can be decomposed per scenario ω, operational period τ and local cell lc into a set of independent subproblems as follows: s.t.: volume requirements satisfaction constraints: synchronization constraint for riders' operations: ( ) synchronization constraint for couriers' operations: ( ) detour linking constraints: ( ), ( ), ( ), ( ) it is important to notice that the defined subproblems are feasible regardless of the value of the tactical decisions (first stage variables); this is possible thanks to the variables z l (τ, ω) that compensate for any lack of capacity in the network by incurring a large cost. solving each subproblem using a dualization strategy, one can identify the following optimality cuts for each local cell, operational period τ and scenario ω: where j ∈ j, the set of extreme points of the dualized subproblem; π j l (τ, ω), µ j lc (τ, ω) and λ j l (τ, ω) are the dual values respectively associated with constraints ( ), ( ) and ( ). finally, the restricted master problem, whose objective minimizes the cost of deploying capacity modules in each access hub and the relocation costs for each capacity module for each period subject to the optimality cuts, can be formulated as follows: s.t.: inventory balance of capacity modules at all locations: ( ) total capacity module inventory constraint: ( ) spatial constraint at all locations: ( ) optimality cuts: ( ), ∀j ∈ j ⊂ j ( ) solving the restriced master problem with added optimality cuts provides new values s l (t) and r a (t), and a new incumbent solution. this process can be executed iteratively until the incumbent solution equals the subproblem value, indicating optimality. the following subsection describes acceleration methods developed to improve the performance of the proposed solution approach on large instances. the acceleration techniques retained are those that improve significantly the convergence speed of the benders decomposition algorithm for the proposed model. the proposed implementation of the benders decomposition can be improved using pareto-optimal cuts, which requires to solve two linear programs: the original subproblem ( ), and the pareto subproblem. the result is the identification of the strongest cut when the original subproblem solution has multiple solutions. a pareto-optimal solution produces the maximum value at a core point, which is required to be in the relative interior of the convex hull of the subregion defined by the first stage variables. the pareto subproblem can be decomposed per scenario ω, operational period τ and local cell lc in a set of independent pareto subproblems as follows: modified synchronization constraint for riders' operations: modified synchronization constraint for couriers' operations: detour linking constraints: ( ), ( ), ( ), ( ) where v sp is the value of the corresponding original subproblem and s l (t) a core point of the current solution to the restricted master problem. solving each pareto subproblem using a dualization strategy, one can identify strengthened optimality cuts ( ) by assigning π j l (τ, ω), µ j lc (τ, ω) and λ j l (τ, ω) the dual values respectively associated with constraints ( ), ( ) and ( ). the proposed implementation also updates the core point, which can be seen as an intensification procedure: locations that are rarely given capacity modules decay toward low values while locations with consistent capacity module presence in every solution are assigned a high coefficient in pareto solutions. the update rule was introduced in papadakos ( ), and consists of updating the core point at iteration k, s (k) by combining it with the solution of the master problem at this iteration, s (k) , using a factor λ. maheo, kilby, and van hentenryck ( ) suggest that a factor λ = / yields the best results. the update rule is defined as follows: where k is the current iteration of the benders algorithm. when dealing with large-scale instances, the -optimal method as described in rahmaniani et al. ( ) has proven to speed up the proposed benders decomposition algorithms by avoiding to solve the restricted master problem to optimality at each iteration, while guaranteeing an optimal gap within . it is not necessary to solve the restricted master problem to optimality at each iteration to generate good quality cuts, and there is no incentive to do so at the beginning of the algorithm because the relaxation is weak. instead, the restricted master problem can be solved with a relaxed optimality gap by adding a constraint forcing the objective value to be improved by at least percent compared to the previous solution. then, when no feasible solution is found, is decreased. the same mechanism is applied until is reached; the algorithm terminates when no feasible solution is found to the restricted master problem, guaranteeing that the current solution is within of the optimal. in this section, the results of numerical experiments are presented in order to validate the developed modeling and solution approaches, and to analyze the performance of the proposed capacity deployment strategy for urban parcel logistics. after describing the test instances which are inspired from the real data of a large parcel express carrier, experimental results about the computational performance of the solution approach are presented. then, the performance of the dynamic pooled capacity deployment strategy is exposed and compared to its static counterpart. finally sensitivity analyses are conducted on the capacity pooling distance and the holding costs to derive further insights. table summarizes the characteristics of the considered instances: number of access hub locations, number of local cells, and area and population covered by the network. non-stationary demand scenarios are generated randomly from given distributions at the hourly level with monthly, weekly, daily and hourly seasonality factors. figure illustrates demand dynamics by displaying access hub volume requirements box plots and snapshots of demand levels in two consecutive tactical periods as seen in figure for a sample local cell from instance e. the number of scenarios is chosen to ensure tactical decision stability with a reasonable in-sample statistical gap ( . %) and coefficient of variation ( . %) as detailed in a. the considered planning horizon spread over two months, with weekly tactical periods and hourly operational period. each week is composed of seven days of ten operating hours each. the -method is implemented with a guaranteed optimality gap of . %. as benchmark solutions, static capacity deployments are considered for each instance. such static ca-pacity deployment represents the minimum capacity module deployment required over the network of access hub locations to satisfy storage requirements for all operational periods within the planning horizon t without being able to update capacity over time or use capacity pooling recourse actions. benchmark solutions are found by solving min{∑ t∈t ∑ l∈l∪w h l s l (t)} such that s l (t) ∈ {vs l (t) ≥ d l (τ, ω), ∀l ∈ l, τ ∈ t t , t ∈ t, ω ∈ Ω} over the entire planning horizon with no relocation or recourse by relaxing spatial constraints to ensure feasibility. an instance has more or less savings potential depending on its demand dynamics and network configuration. although assessing the potential of capacity pooling a priori is non trivial, the potential of capacity relocation can be assessed by a lower bound to the dynamic capacity deployment problem with no capacity pooling. defines l (t) as the maximum number of capacity modules required at location l in any operational period associated with tactical period t in all considered scenarios; that iss l (t) = max( d l (τ, ω)/v , ∀τ ∈ t t , ω ∈ Ω). then, an instance's capacity relocation cost savings potential can be computed by factoring in holding costs while ignoring relocation costs, producing a lower bound for the dynamic capacity deployment problem with no capacity pooling, with objective value ∑ t∈t h lsl (t). benchmark solutions and relocation potential for the considered instances are summarized in table . the initial capacity deployment is defined by running the proposed solution approach for the tactical period immediately preceding the studied planning horizon by relaxing constraint ( ). the default values of input parameters are estimated relying on company experts and presented in a. each instance is assigned one depot in one of its local hub locations to store unused capacity modules at no cost. the number of modules available i and the penalty cost p l are set to large values (respectively modules and $ , per modules in order to prevent full recourse actions by lack of capacity and focus on feasible capacity deployments with capacity pooling. as suggested by winkenbach, kleindorfer, and spinler ( ) (through simulation) when studying a french parcel express company, this paper considers the value of the k constants to be . for riders and . for couriers. all experiments were implemented in python . using gurobi . as the solver and were computed using logical processors on an amd epyc processor @ ghz. the experiments presented in this section study the computational performance of the proposed solution approach when tackling instances of different sizes. the first experiment aims at validating the efficiency of the proposed acceleration methods in section ( . ) for the benders algorithm. it examines the impact of combinations of the acceleration methods on the runtime of the benders algorithm for solving the optimization model ( - ) for one relocation period with no lookahead. figure display the runtimes for instances c with a capacity pooling distance of km and a time cutoff of hours; b represents the original benders algorithm developed in section ( . ); bp represents benders with pareto-optimal cuts; be represents benders with the -optimal method; and bpe represents bp with the -optimal method. figure suggests that pareto-optimal cuts have the strongest impact on computational performance as it allows the bp algorithm to converge in seconds when the b algorithm did not converge within the time limit. the -optimal method suggests a significant improvement compared to the original benders algorithm, and has an advantage over bp when close to optimality (while guaranteeing a solution within . % of optimality). similar behaviors can be observed for larger instances, with bpe outperforming b, bp and be. presented in the paper. the first observation is that the proposed solution approach is efficient in solving large-scale instances considered in this paper ( access hubs), with a maximum runtime around hours (with weeks lookahead); this result suggests tractability for most urban area sizes, including megacities. the second observation is that adding tactical lookahead reasonably increase runtime: week and weeks lookahead runtimes are respectively at most . times and . times as long as no lookahead runtimes wihtin the range of network sizes considered. the results presented in this section highlight the benefits of relocating capacity dynamically over time and allowing capacity pooling compared to a static capacity deployment with no capacity pooling. results are summarized in table for different lookahead values and capacity pooling distance (in km). table presents total costs of the network, deployed capacity (maximum number of modules), relocation share (average number of relocations per period as a share of capacity), and cost and capacity savings with respect to the static counterpart. first, cost and capacity savings are observed in all the instances. maximum cost savings of . % and capacity savings of . % are reached for instance a with a capacity pooling distance of km and a weeks tactical lookahead. most of these savings are a result of the capacity pooling recourse as savings with capacity pooling of km indicate a much lower savings (maximum of . % cost savings). note that for each instance, savings with no capacity pooling are less than potential savings presented in table (where relocation costs are not accounted for). the average number of relocations per period represent up to . % of the capacity, and is decreasing as more tactical lookahead is added; capacity deployments are gradually reconfiguring networks. capacity savings indicate that the total number of modules required (both deployed and stored at a depot) is inferior to the number of modules required in static counterparts. capacity savings also increase as capacity pooling is available, making the total capital invested in capacity modules inferior than in static counterparts. furthermore, the results show that adding tactical lookahead is beneficial for all instances with and without capacity pooling by improving cost savings and decreasing the number of relocations. the role of tactical lookahead is to anticipate future needs and avoid relocations that will be reverted to in the future. lookahead can be seen as the flexibility hedging of the solution approach to avoid relocations under uncertainty. however, the difference between one week and two weeks of tactical lookahead is more subtle with smaller cost improvements. these results suggest that solution's quality increase with lookahead (θ), offering extra cost savings. tactical lookahead anticipates for future relocations therefore decreasing relocation share at the cost of slightly higher capacity deployments. however, there does not seem to be significant improvements from extending the lookahead from one week to two weeks, especially when considering the additional computational runtime. lastly, capacity pooling brings significant value to instance a, b, and c, but less cost savings improvements for instance d and e. this is probably due to the the fact that instances d and e have lower hub density, increasing the distance between access hubs (see table ). section . examine the impact of capacity pooling distance in more details by focusing on instance c. this experiment examines the effect of capacity pooling as a way to further decrease costs. table summarizes the effect of different capacity pooling distances (in km) on instance c's solutions. it presents average additional rider and courier travel (induced by detours), and cost and capacity savings for instance c. the increase in capacity pooling distance allows to produce superior solutions but only until a maximum of . % is reached with a pooling distance of km. this trend can clearly be seen in figure . indeed, no matter how large capacity pooling pooling neighborhoods are, constraints ( ) and ( ) relocation share), requiring more capacity deployed and therefore limiting cost savings ( . %). similar behavior can be observed on the other instances. overall, this experiment indicates that denser urban environment (high holding costs) tend to be better candidates for dynamic capacity management of access hub networks. moreover, low relocation costs (i.e. easy installation and good mobility of capacity modules) can make any urban environment a worthy candidate for such capacity management strategy. finally, the combination of lesser dense urban environment and high relocation costs significantly limits opportunity for cost savings. this paper defines and formulates the dynamic pooled capacity deployment problem in the context of urban parcel logistics. this problem involves a tactical decision on the relocation of capacity modules over a network of discrete locations associated with stochastic demand requirements. to improve the quality of the capacity deployment decisions, the proposed model integrates an estimate of the difference of operations cost, which includes capacity assignment decisions with the possibility of capacity pooling between neighboring locations. it also integrates synchronization requirements of the -echelon routing subproblems, using an analytical derivate from the route length estimation function. the dynamic problem is modeled and approximated with a two-stage stochastic program with recourse, where all capacity deployment decisions on a finite planning horizon are moved to the first stage. due to the uncertainty of capacity requirements and the challenges of solving the mip formulation for realistic networks of several hundreds locations, a roll-out approach with lookahead based on a benders decomposition of the finite planning horizon problem coupled with acceleration methods is proposed. five instances of networks of different sizes are presented to perform computational experiments to test the performance of the proposed approach and assess the potential of the defined capacity deployment strategy. results show that the proposed approach produces solutions in a reasonable time even for large scale instances of up to hubs. they suggest that a dynamic capacity deployment strategy with capacity pooling has a significant advantage over a static capacity deployment strategy for access hub networks, with up to % cost savings and % capacity savings. results also show that one-week lookahead helps producing superior solutions by anticipating future relocations, but adding a two-weeks lookahead does not make a significant improvement. increasing the capacity pooling distance, while increasing computing time, tend to increase opportunities for cost savings by allowing more locations to pool capacity until an operational feasibility threshold is reached. dynamically adjusting workforce assignment in the network was not explored but could potentially overcome this limitation. denser urban environments (i.e. with higher real estate costs) are natural candidates for dynamic capacity deployments as relocation costs are more easily overcome by holding costs. however, relocation costs are the most limiting when it comes to cost savings. technology solutions featuring cheaper installation costs and high degree of mobility make it more interesting to consider periodic network reconfigurations. the implementation of such innovation also has management challenges not studied in this paper. for instance, implementation may require a more agile workforce, specialized training and targeted hiring enabling a data-driven approach to managing network capacity. management challenges also need to be considered by decision makers along with the potential reduction of fixed-assets offered by capacity savings when evaluating the solution for implementation. finally, there are numerous research avenues around reconfigurable networks, dynamic capacity management and access hubs in urban parcel logistics. where technology allows for very frequent network reconfiguration, solutions featuring not only modular but mobile capacity (e.g. on wheels) and near real-time capacity relocation can become relevant as a complement to the proposed dynamic capacity deployment strategy. moreover, the possibility of updating operations planning as needed (e.g. dynamic routing, dynamic staffing) can unlock the potential of capacity pooling not only as a recourse but as an integral part of network design and operations planning. in-sample variability was tested with no lookahead for instance a with capacity pooling limited to km for samples. results are presented in table . coefficient of variation represent the ratio between the standard deviation and the average of solutions' total cost. statistical gap represent the ratio (ub − lb)/lb where ub and lb are respectively the highest and lowest total cost in the sample. the capacity module relocation costs r a include an operational cost of $ . per kilometer, and a fixed cost of two operators for two hours at a rate of $ per hour to uninstall/install modules once at the desired locations: r a = . d a + , ∀a = (i, j) ∈ a where d a is the distance between location i and j such that a = (i, j). the holding costs are computed from an amortized acquisition cost of $ over years ( weeks long years), and from a rent cost of $ per square meter times a location specific factor ( + f l ) randomly generated to represent the real estate difference between locations. h l = * + ( + f l ), ∀l ∈ l where f l is randomly generated from a uniform distribution over [ %, %] . it is also assumed that modules do not depreciate when stored at depots (h l = , ∀l ∈ d). capacity planning and warehouse location in supply chains with uncertain demands synchronizing vans and cargo bikes in a city distribution network. central european advancements in continuous approximation models for logistics and transportation systems: - facility location dynamics: an overview of classifications and applications an ex ante evaluation of mobile depots in cities: a sustainability perspective partitioning procedures for solving mixed-variables programming problems trends, challenges, and perspectives in city logistics. transportation and land use interaction integrating first-mile pickup and last-mile delivery on shared vehicle routes for efficient urban e-commerce distribution citylog project bentobox service network design with resource constraints advanced freight transportation systems for congested urban areas a survey on two-echelon routing problems the length of tours in zones of different shapes logistics systems analysis a survey of variants and extensions of the location-routing problem items that internet users in the united states want to have vs. have received via same-day delivery as of annual retail e-commerce sales growth worldwide from the two-echelon vehicle routing problem with covering options: city logistics with cargo bikes and parcel lockers design of large-scale logistics systems for uncertain environments smart locker bank design optimization for urban omnichannel logistics: assessing monolithic vs. modular configurations strategic fleet planning for city logistics the period vehicle routing problem and its extensions. the vehicle routing problem: latest advances and new challenges reaching million nanostores: retail distribution in emerging megacities the capacitated plant location problem with multiple facilities in the same site the n-echelon location routing problem: concepts and methods for tactical and operational planning scheduled service network design with resource acquisition and management under uncertainty development and application of a transferability framework for micro-consolidation schemes in urban freight transport characterizing urban last-mile distribution strategies in mature and emerging e-commerce markets integrating collection-and-delivery points in the strategic design of urban lastmile e-commerce distribution networks dynamic facility location with generalized modular capacities the impact of operations anticipations on the quality of stochastic location-allocation models before-after assessment of a logistics trial with clean urban freight vehicles: a case study in london benders decomposition for the design of a hub and shuttle public transit system a dynamic mobile production capacity and inventory control problem multi-echelon distribution systems in city logistics optimization models for the dynamic facility location and allocation problem introducing the concept of hyperconnected mobile production modeling of physical internet enabled interconnected modular production assessing the sustainability of mobile depots: the case of urban freight distribution in rio de janeiro dynamic multi-commodity capacitated facility location: a mathematical modeling framework for strategic supply chain planning toward a physical internet: meeting the global logistics sustainability grand challenge parcel logistics hub design: the impact of modularity and hyperconnectivity integrated airline scheduling the two-echelon capacitated vehicle routing problem: models and math-based heuristics approximate dynamic programming: solving the curses of dimensionality a unified framework for stochastic optimization the benders decomposition algorithm: a literature review new mixed integer-programming model for the pickup-and-delivery problem with transshipment th anniversary invited article-city logistics: challenges and opportunities stochastic programming with integer variables lectures on stochastic programming: modeling and theory continuum approximation techniques for the design of integrated package distribution systems sustainable parcel delivery in urban areas with micro de-pots dutch hub company presents micro hub solutions for city logistics tnt express is using a mobile depot in brussels to enable more efficient delivery revision of world urbanization prospects letzte meile: studie zu micro-hubs does a mobile depot make urban deliveries faster, more sustainable and more economically viable: results of a pilot test in brussels three proactive response strategies to covid business challenges. mit sloan management review enabling urban logistics services at la poste through multi-echelon locationrouting key: cord- -hs wfffs authors: lambert, stephen b; allen, kelly m; carter, robert c; nolan, terence m title: the cost of community-managed viral respiratory illnesses in a cohort of healthy preschool-aged children date: - - journal: respir res doi: . / - - - sha: doc_id: cord_uid: hs wfffs background: acute respiratory illnesses (aris) during childhood are often caused by respiratory viruses, result in significant morbidity, and have associated costs for families and society. despite their ubiquity, there is a lack of interdisciplinary epidemiologic and economic research that has collected primary impact data, particularly associated with indirect costs, from families during aris in children. methods: we conducted a -month cohort study in preschool children with impact diary recording and pcr testing of nose-throat swabs for viruses during an ari. we used applied values to estimate a virus-specific mean cost of aris. results: impact diaries were available for % ( / ) of community-managed illnesses between january and january . the mean cost of aris was au$ ( % confidence interval $ to $ ). influenza illnesses had a mean cost of $ , compared with rsv, $ , the next most expensive single-virus illness, although confidence intervals overlapped. mean carer time away from usual activity per day was two hours for influenza aris and between and minutes for all other ari categories. conclusion: from a societal perspective, community-managed aris are a significant cost burden on families and society. the point estimate of the mean cost of community-managed influenza illnesses in healthy preschool aged children is three times greater than those illnesses caused by rsv and other respiratory viruses. indirect costs, particularly carer time away from usual activity, are the key cost drivers for aris in children. the use of parent-collected specimens may enhance ari surveillance and reduce any potential hawthorne effect caused by compliance with study procedures. these findings reinforce the need for further integrated epidemiologic and economic research of aris in children to allow for comprehensive cost-effectiveness assessments of preventive and therapeutic options. respiratory virus infections are a major cause of morbidity and healthcare usage in children, resulting in substantial costs for families and society [ ] [ ] [ ] [ ] [ ] . given their ubiquity, there has been surprisingly little research examining the costs associated with childhood respiratory infections that has involved collecting primary data from families. even for influenza, the most studied of all respiratory viruses, cost-of-illness and vaccine cost-effectiveness evaluations in children have tended to rely on assumptions or use retrospectively collected estimates, often from surveys, for resource utilisation, such as carer time away from work in seeking healthcare or caring for an ill child [ ] [ ] [ ] [ ] . there are three pieces of evidence required by those developing health policy in assessing whether to recommend or implement a publicly-funded vaccination program, or any intervention, against respiratory viruses: epidemiology of the targeted illness, the efficacy of the intervention, and the cost-effectiveness of the intervention [ ] . all interventions to prevent or treat infections will be associated with a cost of implementation, but cost-effectiveness is determined not only by the cost of the intervention, but also by costs arising from the illness. getting these data for respiratory viruses, particularly information on indirect costs incurred by families, requires a conjunction of epidemiologic and economic research [ ] . the prospect of new and improved influenza vaccines [ ] , the hope of new vaccines against other respiratory viruses [ ] , development of novel therapeutic possibilities [ ] , and the possible use of nonpharmaceutical interventions to contain virus transmission [ ] [ ] [ ] [ ] all underline the need to more critically weigh the costs and benefits of prevention and treatment for common respiratory tract viruses. we present here findings from a community-based cohort study using parent-collected specimens for etiologic assignment and diary recording of impact data. these data have been used to calculate virus-specific costs of illness from a societal perspective, including often neglected indirect costs. the study cohort and acute respiratory illness surveillance details of recruitment, composition, and maintenance of the dynamic study cohort have been published elsewhere [ ] . ethics approval for the study was given by the royal children's hospital ethics in human research committee, melbourne, and written informed consent was obtained from parents before participation. this dynamic cohort consisted of one healthy child less than five years of age at time of recruitment from each study family. children involved in this study were recruited from a number of sources. in victoria, australia, maternal and child health nurses (mchns) provide support to families during the early childhood years, particularly on issues to do with general health and vaccination. based on a model used by our group for community vaccine studies [ ] , mchns from local councils assisted with recruitment by providing study information to parents of eligible children. advertising material for the study was placed in child care and playgroup centers and, because of proximity, we also used bulletin boards and staff e-mail lists at the royal children's and the royal women's hospitals in melbourne. details about the study child and household demographics were collected at an enrolment home visit, including annual gross household income collected in / australian dollar values (aud$). income was separated into four bands, roughly dividing the study households into quartiles: band , less than $ , ( % of study households); band , $ , to $ , ( %); band , $ , to $ , ( %); and band , $ , or greater ( %). the approximate proportions for australian households during the same period were: band , %, band , %, band , %, and band , % [ ] . parents undertook daily respiratory symptom surveillance of the study child using a diary card and collected a combined nose-throat swab (nts) and completed a summary impact diary when the child had an acute respiratory illness (ari). for this study we used a sensitive ari definition that had previously been used in an influenza vaccine efficacy study [ ] and our pilot study [ , ] . symptoms were classified as category a (fever, wheezing, shortness of breath, pulmonary congestion or moist cough, pneumonia, or ear infection) and category b (runny nose or nasal congestion, sore throat, cough, muscle aches, chills, headache, irritability, decreased activity or lethargy or weakness, or vomiting). an ari of interest required one category a or two category b symptoms on a single day [ ] . other than pneumonia, which we asked parents to record only if supported by a health care professional's diagnosis, no illness or symptom details, including a report of otitis media, were validated by study staff or health care professionals. a new ari could not commence unless there were three symptom free days since the end of the previous ari. this meant an ari could contain no more than two consecutive symptom-free days. study families were asked to continue normal healthcare seeking behaviour and treatments, and were not alerted about the start of the influenza season or asked to alter surveillance during the winter season. pre-stamped envelopes were provided and families were asked to return all completed study documents (daily symptom diary, impact diaries) at the end of each month. ari duration was calculated using symptom diary data and aris were classified by study staff as being simple (no fever or otitis media recorded), or occurring with fever, otitis media, or with both fever and otitis media [ ] [ ] [ ] . the nts was couriered to the victorian infectious diseases reference laboratory (vidrl) where it was tested for a number of common respiratory viruses using a polymerase chain reaction (pcr) method for adenoviruses and reverse transcription (rt) pcr for rna viruses: influenza a virus, influenza b virus, respiratory syncytial virus (rsv), parainfluenza viruses i, ii, and iii (pivs), and picornaviruses [ ] . a letter outlining these test results was sent to parents when these details became available. at completion of the study all specimens were transported to the queensland paediatric infectious diseases (qpid) laboratory where they were tested for human metapneumovirus (hmpv) and human coronavirus nl (hcov-nl ) using rt-pcr [ ] . a summary impact diary was used to collect details of resources used during the study child's ari and was based on an impact diary used in a pilot study [ , ] , with some simplification. the units of resource use requested were: ▪ health care visits: number and timing of primary care (general practice) visits, hospital presentations and admissions, and visits to other providers (such as naturopaths, homeopaths); ▪ use of prescribed antibiotics; ▪ laboratory tests performed to investigate the illness; ▪ carer time consumed during the illness seeking health care; and ▪ excess carer time during the illness spent caring for the ill child. we did not collect information about some items that were shown not to be major cost drivers in the pilot study: non-antibiotic prescription medication, over-the-counter and other medication, paid childcare for other children whilst normal carers were spending time caring for the ill study subject, and travel costs seeking health care. the average total cost for these items in the pilot study [ ] was aud$ per ari. time values were captured in hours and minutes. parents were not given instructions about when or how frequently they should capture time data during an ari. for both carer time spent seeking healthcare and excess time spent caring for an ill child, time was recorded as a total value for the ari in two categories: time away from work and time away from usual, non-work activities. all costs were incurred over a day period between january and january . costs are reported in this manuscript using australian dollar values, with taken as the reference year for reporting unit prices. the mean exchange rates for major currencies during the study were: united kingdom (uk) pound £ = aud$ . , euro € = aud$ . , and united states (us) $ = aud$ . [ ] . discounting costs for time preference is not routinely considered for periods of time less than months, and as this study period barely exceeds this time frame, no costs have been discounted. details of the source and value for all costs are provided ( table ) . applied costs were retrieved, where possible, from published sources, and where no standard published cost was available we used costs derived from the pilot study. resource costs were allocated as being borne by either the 'patient and family' sector, the 'healthcare' sector, or the 'employer' of absent staff. the proportions of time away from work seeking healthcare or time away from work caring for an ill child that were incurred by either the patient and family sector or met by an employer were not collected, and these values have been derived from the same proportions in the pilot study, based on illnesses (table ) [ ] . we applied a sex-weighted hourly wage rate derived from the australian bureau of statistics average weekly full-time adult total earnings for all reported times [ ] . we calculated mean costs (total and by categories) with % confidence intervals ( % ci) and median costs with interquartile ranges for aris in study children. data were analysed using stata . for windows (statacorp, texas, usa). there were children, one from each study family, progressively enrolled in the study and we identified aris in , child-days of follow-up [ ] . of these, aris ( %) had at least one specimen and an impact diary available, ( %) had an impact diary returned but no specimen, ( %) had at least one specimen returned but no impact diary available, and ( %) had neither a specimen nor impact diary returned. children aged between one and two-years of age contributed the most person-time to the study ( % of all child-days) and had the highest acute respiratory illness (ari) incidence rate ( . aris per child-month). contribution by males and females was equivalent, and children who attended some form of out-of-home care were responsible for % of all person-time [ ] . the illnesses with a diary returned that did not involve a hospital admission had a total cost of $ , (table ) , and mean cost of $ ( % ci $ to $ ). as our particular interest is in the cost of communitymanaged aris, that is, those illnesses that do not require the mean and median costs by virus identification, including co-identification and specimen availability, are provided ( table ). the differences between the mean values and the median values demonstrate the right-skewed nature of these data, similar to other health-related costs [ ] . whilst confidence intervals overlap, the point estimate of the mean cost of an influenza a ari, $ , is three times higher than the next most expensive single virus ari: rsv $ . of the aris where more than one virus was identified, influenza a virus was present in four: two illnesses with co-identification with a picornavirus alone, one illness with hcov-nl alone, and one illness with a picornavirus and piv. these four illnesses had a mean cost of $ . there were no illnesses where influenza b virus was identified. three children had received influenza vaccine in the year prior to the study and none had an influenza-positive ari. as the difference in mean cost between the most expensive (rsv: $ ) and least expensive (hmpv: $ ) noninfluenza single virus ari falls within a comparatively narrow band ($ ) we collapsed these data into a single category for further comparisons ( there was little difference in the mean duration of influenza a illnesses and other single virus illnesses, but coidentifications were . and . days longer than each of these respectively ( table ). the mean delay between illness onset and a result letter being sent was shortest in influenza illnesses at . days (table ). in this study we present the costs associated with community-managed respiratory viral infections in healthy preschool aged children. these costs are based on the direct recording of impact information captured by parents when the study child was unwell. the study has a unique combination of features including a sensitive definition for ari, parent-collected specimens, laboratory testing for respiratory viruses using sensitive molecular methods, and, based on findings from our pilot study, comprehen- sive collection of costs, including the previously neglected indirect cost, time away from a usual, non-work activity. we found, from a societal perspective, the point estimate for the mean cost of all aris ($ ; % ci $ to $ ) was not dissimilar to the mean value we calculated from the pilot study ($ ; % ci $ to $ ) [ ] using the same ari definition and a slightly modified impact diary. the use of pcr testing for diagnosis on collected specimens allowed us to assign impact and costs to specific viral agents. for all but influenza a illnesses, the cost of community-managed aris in healthy preschool-aged children fell within a relatively narrow $ range. despite overlapping confidence intervals, the finding of most note in this study was the dramatically higher point estimate of the mean cost of influenza a aris, being three times higher than illnesses caused by rsv and the other common respiratory viral infections of childhood. the presence of fever and/or otitis media generally increased the mean cost of illness; but despite having a high prevalence of fever, a longer mean duration, and higher primary care usage [ ] , adenoviral infections, for example, did not have the cost burden of influenza infections. this highlights the pivotal contribution of excess carer time away from usual non-work activity to total costs, making it the key cost driver for all aris in children and differentially amplifying the total costs of influenza illnesses. whilst the confidence intervals for mean cost of influenza a aris and other single virus aris overlap, due to the relatively small number of influenza illnesses available for costing, we believe it is unlikely that chance could account for such an extreme difference. the availability of preventive vaccines and specific therapeutic options makes influenza the most studied of respiratory viruses in all age groups; no other virus is more predictably disruptive year-on-year than annual interpandemic influenza [ ] [ ] [ ] [ ] . studies conducted in the second half of last century [ ] [ ] [ ] [ ] and recent observation [ , ] and intervention [ ] [ ] [ ] ideally further studies in other countries should be conducted to allow for an examination of how impact and cost data vary with the nature of the healthcare system, local virus epidemiology, and other societal factors, including household structure. despite lower mean costs than influenza illnesses and the lack of population-based prevention options, the importance of working towards the prevention of other respira-tory viral infections is obvious. picornavirus aris, though typically milder and more difficult to be certain of a causal association with illness [ , ] , were associated with the highest overall costs of any viral group totalling over $ , or one-third of all costs, for the -month study period. in the absence of specific vaccines and therapies for other viruses, the application of nonpharmaceutical interventions at a population level, such as improved hand and respiratory hygiene, may have an important place in reducing illness due to respiratory viruses [ ] . our findings reinforce the importance of virus testing in such studies to accurately estimate epidemiology and costs [ ] . these data add to accumulating evidence that laboratory confirmation of influenza, in particular, is required, rather than less specific influenza-like illness (ili) definitions or hospital coding. other recent studies have found laboratory-confirmed influenza hospitalizations were two to four times more costly [ ] [ ] [ ] than shown in previous studies using coding-based estimates [ , [ ] [ ] [ ] . when ili definitions or coding are used, rather than laboratory confirmation, a lack of specificity results in influenza illnesses being mixed with other agents, thereby considerably diluting cost differences [ , ] . a direct comparison of parent-collected nts specimens with collection of a more invasive specimen, such as a nasopharyngeal aspirate, by a healthcare worker at the time of an ari was beyond the scope of this study. any reduction in sensitivity caused by the type of specimen used is likely to minor: our finding that % of all specimens collected from children in this study were positive for at least one virus is within the range of values from recent home visit studies which also used pcr for diagnosis and nasopharyngeal aspirates ( %) [ ] or nasal lavages ( %) [ ] . there are clearly some issues about the cost of illnesses caused by respiratory viruses in children unresolved by our study, and some issues that need to be considered before interpretation. despite being a relatively large cohort the number of illnesses on which to make costing estimates for some virus types is quite small. further community-based estimates are required to not only confirm our findings but to improve precision around point estimates. compared with the australian population, households with lower incomes were under-represented in our study sample, and, despite overlapping confidence intervals around income band point estimates of mean costs, this may have lead to an overestimation of total costs. however, this may be balanced somewhat by the over-representation of households from the top income band which had a relatively lower mean ari cost ($ ). this pattern of household income distribution was similar to that found in the pilot study [ ] . for this study we sought to make our study sample more representative of the general community by focusing our recruitment efforts in local council areas with a higher proportion of lower income households. we have no empiric data available that would allow us to quantify the effect of any potential bias resulting from this skewed sample. other recent burden studies do not report similar household level income data to allow for comparison [ ] [ ] [ ] . it may be the case that lower income households are under-represented as they do not have the spare capacity required, in time or other resources, to allow for study involvement. we received impact diaries for just over % of all aris identified by daily symptom surveillance. aris without a diary were more likely to be shorter and without fever or otitis media; any information bias resulting from this would likely be in the direction of inflating mean illness costs. our study only captured information from a single season with higher than normal influenza activity with h n influenza a (drifted strain subtype a/fujian/ / -like) being the predominant circulating type [ ] . variations in incidence and severity year-by-year for all respiratory viruses make it difficult to directly translate our findings to other years. we believe documenting all time spent on caring for an ill child is important, even when taken away from a usual activity. we appreciate that applying standard wage rates to leisure time is not a straightforward issue in economics. this approach values carer leisure time and non-paid working time in a similar way to a worker's time consistent with neoclassical theories of labour economics [ ] . in attaching value to leisure time and using sex-weighted wage rates, we have made our assumptions explicit, and provided sufficient detail (table ) so that others can adjust unit prices using different approaches. previous burden data [ ] have been used to assess the cost-effectiveness (c/e) of using influenza vaccine in children [ ] . if our cost values, incorporating these indirect costs, were used in the numerator of c/e calculations, there is a distinct possibility of double counting [ ] . double counting is likely where the denominator is a utility measure that incorporates a quality assessment (such as the quality adjusted life year or qaly), and most economists would see leisure time as a logical component of the qaly. there is also debate [ ] about the inclusion, measurement, and valuation of lost working time in economic evaluations, with the debate centring on whether in practice qaly instruments capture income effects related to absenteeism. for all illnesses where a specimen was tested, parents received a result letter by mail. the delay between illness onset and posting the letter was shortest for influenza illnesses, but for most illnesses parents would have been aware of the result before illness end. pandemic influenza was not being widely discussed in australia during , but interpandemic influenza does receive media coverage annually encouraging vaccine uptake, and this may have caused parents to overestimate key parameters associated with their child's influenza-positive illness. however, if such a bias was in operation it might also be expected that time values for illnesses where no virus was identified may be relatively understated when compared to aris with one or more viruses present. we did not find such a phenomenon; aris with no virus identified had a higher mean cost than those with a single virus present, and for the key cost driver of excess carer time away from a usual activity, no cause illnesses had higher values than both single and multiple virus aris. despite the impact of respiratory viral infections in children there are relatively few burden comparisons available that collect primary data from ill children. an italian study examining the impact of hmpv, rsv, and influenza in children less than -years of age presenting to an emergency department found hmpv illnesses to be significantly more burdensome than rsv, having a similar impact to influenza [ ] . in our study hmpv was the least expensive single-virus illness. this finding may be due to the different nature of illnesses that result in hospital presentation or hospital admission, compared with community managed illness. of the aris in this study only . % (n = ) prompted hospital presentation, with less than % (n = ) requiring admission. an excellent community-based finnish study describing the burden of influenza in children -years of age or younger over two seasons, with child-seasons of data, also contrasts this imbalance between community-managed and hospitalized cases of influenza, with only three emergency department referrals and one hospital admission in children less than three years of age with influenza [ ] . this study differed from ours in that whilst it used laboratory confirmation, it did not employ more sensitive molecular diagnostics [ ] , families were required to visit the study clinic when the study child had fever or signs of respiratory infection, indirect costs did not include nonwork time away from a usual activity, and the study did not provide a comparison with other viral acute respiratory illnesses [ ] . the findings from the finnish study reinforce the need to follow children for aris over more than one season, with different rates of influenza infection from year-to-year in each age group. these differences extended to changes in likelihood of infection between age groups: for example, the rate of laboratory-confirmed influenza increased by one-third from season one ( ) ( ) to season two ( ) ( ) for children less than three years of age, but the rates for three to six year olds and seven to year olds fell % and %, respectively. a german study, recruiting children less than three years of age with lower respiratory tract infection (lrti) through office and hospital-based paediatricians, collected cost of illness from a societal perspective, including loss of work days by caregivers [ ] . this study showed that non-hospitalized cases of influenza lrti had twice the cost of piv lrti and were one-third more costly than rsv lrti, with this difference made up entirely by indirect costs [ ] . whilst methods vary, previous cost effectiveness studies of influenza vaccine in children are characterised by two findings: first, that cost-effectiveness is unsurprisingly enhanced by taking a societal perspective through the inclusion of indirect costs [ , , , , , ] . our findings reinforce the importance of indirect costs [ ] , and highlight a previously inadequately measured layer of burden -carer time away from a usual, non-work activity. second, the potential cost-effectiveness of implementing a vaccination program is improved by flexible or non-individual based delivery programs [ , ] . vaccine delivered through pharmacies for a small service fee -improving access and negating the time and costs associated with a primary care visit -or large school-based programs, are likely to be acceptable to parents and providers. it is likely that the cost benefits of preventing influenza in children would extend beyond the targeted age-group [ ] , similar to the indirect effects in older age groups seen following the introduction of childhood conjugate pneumococcal vaccination in the us [ ] . our study reinforces the costly impact of all respiratory viruses, but particularly interpandemic influenza, on children, their families, and society. efforts to further explore the costs associated with community-managed illness over a number of seasons for all respiratory infections are needed. similar to recent hospital-based findings, using laboratory-confirmation to specifically identify influenza appears to increase the cost of illness many fold; a finding that may make population-based vaccination programs a more cost-effective proposition. we believe the use of parent-collected specimens may have important effects in reducing bias in both the epidemiologic and impact data collected. not requiring parents to either present with their ill child to a health clinic or host a home visit by study staff may result in enhanced ari surveillance, but more importantly, allows for the reporting of impact data uncontaminated by compliance with study procedures, thereby reducing any impact a hawthorne effect may have. further studies that collect primary, integrated epidemiologic and economic data, particularly indirect costs, directly from families about community-managed aris in children, are required. such data would allow for a more informed exploration of the cost-effectiveness of vaccine programs and other interventions designed to reduce the morbidity associated with aris in children. the economic burden of non-influenza-related viral respiratory tract infection in the united states socioeconomic impact of influenza on healthy children and their families burden of influenza in healthy children and their households clinical and socio-economic impact of influenza and respiratory syncytial virus infection on healthy children and their households clinical and economic impact of influenza vaccination on healthy children aged - years economic impact of influenza vaccination in preschool children population-wide benefits of routine vaccination of children against influenza potential cost-effectiveness of annual influenza immunization for infants and toddlers: experience from canada costeffectiveness of adjuvanted influenza vaccination of healthy children to months of age influenza vaccination: policy versus evidence: no gap between policy and evidence prill mm: interdisciplinary epidemiologic and economic research needed to support a universal childhood influenza vaccination policy a one-size-fits-all flu vaccine? science respiratory syncytial virus and parainfluenza virus inhibition of respiratory viruses by nasally administered sirna world health organization writing group: nonpharmaceutical interventions for pandemic influenza, international measures world health organization writing group: nonpharmaceutical interventions for pandemic influenza, national and community measures respiratory infections during sars outbreak nonpharmaceutical interventions implemented by us cities during the - influenza pandemic community epidemiology of human metapneumovirus, human coronavirus nl , and other respiratory viruses in healthy pre-school aged children using parent-collected specimens melbourne. victorian infectious diseases bulletin australian bureau of statistics: household income and income distribution the efficacy of live attenuated, cold-adapted, trivalent, intranasal influenzavirus vaccine in children the cost of seasonal respiratory illnesses in australian children: the dominance of patient and family costs and implications for vaccine use respiratory illness during winter: a cohort study of urban children from temperate australia the distribution of health care costs and their statistical analysis for economic evaluation a study of illness in a group of cleveland families. xvi. the epidemiology of influenza, - modification of an outbreak of influenza in tecumseh, michigan by vaccination of schoolchildren infection with influenza a/victoria virus in houston families influenzavirus infections in seattle families, - . i. study design, methods and the occurrence of infections by time and age influenza and the rates of hospitalization for respiratory disease among infants and young children the effect of influenza on hospitalizations, outpatient visits, and courses of antibiotics in children effectiveness of influenza vaccination of day care children in reducing influenza-related morbidity among household contacts effectiveness of influenza vaccination of children with recurrent respiratory tract infections in reducing respiratory-related morbidity within the households the effect of mass influenza immunization in children on the morbidity of the unvaccinated elderly human picornavirus and coronavirus rna in nasopharynx of children without concurrent respiratory symptoms predominance of rhinovirus in the nose of symptomatic and asymptomatic infants cost of influenza hospitalization at a tertiary care children's hospital and its impact on the cost-benefit analysis of the recommendation for universal influenza immunization in children age to months epidemiology, complications, and cost of hospitalization in children with laboratory-confirmed influenza infection direct medical cost of influenza-related hospitalizations in children the economic impact of pandemic influenza in the united states: priorities for intervention cost-effectiveness analysis of an intranasal influenza vaccine for the prevention of influenza in healthy children an economic analysis of annual influenza vaccination of children influence of clinical case definitions with differing levels of sensitivity and specificity on estimates of the relative and absolute health benefits of influenza vaccination among healthy working adults and implications for economic analyses icd- codes for identifying influenza hospitalizations in children role of respiratory viruses in acute upper and lower respiratory tract illness in the first year of life: a birth cohort study frequency of detection of picornaviruses and seven other respiratory pathogens in infants burden of influenza in children in the community impact of human metapneumovirus in childhood: comparison with respiratory syncytial virus and influenza viruses economic impact of community-acquired and nosocomial lower respiratory tract infections in young children in germany higher than normal seasonal influenza activity in victoria the role and estimation of productivity costs in economic evaluation costeffectiveness of influenza vaccination of healthy children avoiding double-counting in pharmacoeconomic studies diagnosis of influenza in the community: relationship of clinical diagnosis to confirmed virological, serologic, or molecular detection of influenza the efficacy, effectiveness and cost-effectiveness of inactivated influenza virus vaccines herd immunity in adults against influenza-related illnesses with use of the trivalent-live attenuated influenza vaccine (caiv-t) in children control and prevention: direct and indirect effects of routine vaccination of children with -valent pneumococcal conjugate vaccine on incidence of invasive pneumococcal disease -united states medicare statistics: average patient contribution per service medicare benefits schedule book operating from november medicare benefits schedule book operating from november content/health-pbsgeneral-pubs-manual-appendix .htm] . national hospital cost data collection national hospital cost data collection the work was supported by project grants from the victorian department of human services, the murdoch children's research institute, and the university of melbourne. stephen lambert was the recipient of a national health and medical research council public health postgraduate scholarship. we thank all children and families who volunteered to participate in the study. this study would not have been possible without the generous terence nolan and stephen lambert have, in the past five years, received research grants for epidemiological and vaccine related research from csl limited, medimmune, gsk biologicals, wyeth, and merck. kelly allen and robert carter have no competing interests to declare. all authors were involved in the study design and approach and sbl and tmn developed the original protocol. kma and sbl were responsible for the day-to-day conduct of the study. sbl performed the analysis and drafted the article. all authors contributed to and approved the final manuscript.publish with bio med central and every scientist can read your work free of charge http://respiratory-research.com/content/ / / key: cord- - yivxfi authors: robertson, christopher t; schaefer, k aleks; scheitrum, daniel; puig, sergio; joiner, keith title: indemnifying precaution: economic insights for regulation of a highly infectious disease date: - - journal: j law biosci doi: . /jlb/lsaa sha: doc_id: cord_uid: yivxfi economic insights are powerful for understanding the challenge of managing a highly infectious disease, such as covid- , through behavioral precautions including social distancing. one problem is a form of moral hazard, which arises when some individuals face less personal risk of harm or bear greater personal costs of taking precautions. without legal intervention, some individuals will see socially risky behaviors as personally less costly than socially beneficial behaviors, a balance that makes those beneficial behaviors unsustainable. for insights, we review health insurance moral hazard, agricultural infectious disease policy, and deterrence theory, but find that classic enforcement strategies of punishing noncompliant people are stymied. one mechanism is for policymakers to indemnify individuals for losses associated with taking those socially desirable behaviors to reduce the spread. we develop a coherent approach for doing so, based on conditional cash payments and precommitments by citizens, which may also be reinforced by social norms. in the medium-term time horizon until a vaccine is developed, and perhaps thereafter, management of the covid- pandemic will largely depend on broad behavioral changes at the population level. under a strategy of social distancing, lockdown, or quarantine, individuals are directed or suggested to exercise precautions including staying home, closing businesses, wearing masks, and avoiding physical proximity to other persons. presymptomatic patients are a reservoir for spread, but such precautions depend on them believing that spread is both important and preventable. individuals have inherent incentives to undertake such measures, because they are self-protective, reducing the chance of the actor herself becoming infected and suffering. accordingly, policymakers can expect a substantial degree of voluntary compliance, as long as the public receives, and believes, accurate information about the risk. that is, regrettably, a nontrivial assumption, but that problem is not our focus here. notwithstanding the self-protective effects of these measures, microeconomic analysis suggests a likely market failure, due to heterogeneity in the population. part of the heterogeneity is biological-some individuals (eg younger persons and women) appear less likely to suffer harm from infection, whereas other individuals (eg older persons, men, and those with other medical conditions) face greater risk of harm if infected. the costs of precaution are also heterogeneous. for individuals who would otherwise be actively building careers or businesses, searching for romantic partners, or providing for dependents including children, the costs of staying at home are much more substantial, compared with others who are already retired from work, settled with spouses and partners, and no longer have dependents. certainly, the consequences of loss of life are more dire for the younger as they have more years of life to lose. however, younger people tend to discount the future more heavily than older individuals and may not place much value on the additional years of life at risk. perhaps most important, the greatest risk to individuals in this younger group, particularly teenagers, is the loss of the social interactions through school, sports, and other activities that are essential to combat depression, among other risk factors for health. suicide and suicide attempts in younger individuals dwarf the magnitude of risk from covid- , which makes the pandemic precautions particularly burdensome for this population, even if yielding spillover benefits for everybody else. perilously, young people are starting to venture into 'covid- parties' organized, so that people can mingle with infected people to get the illness 'out of the way' and carry on with life. in rough terms, for the age gradient in particular, these observations suggest a stark misalignment of incentives-younger persons personally receive the fewest benefits from precautions, but suffer the greatest costs of taking them. on the margin, some individuals will see the personal costs of taking a precaution to be greater than the personal benefits. the heterogeneity in the population suggests a classic market failure, known in the law and economics literature as an 'externality' . those who do not comply with precautions disproportionally impose the costs of noncompliance on others, who are more likely to suffer from infection. similarly, for seasonal flu where vaccination is a primary precaution, the rate of vaccination in the - age bracket has not in recent years exceeded %, which is just over half the rate of older adults. in a recent survey, young millennials were the least well informed about influenza, and the most likely to believe anti-vaccine rhetoric. accordingly for covid- , we can expect that for those who can get infected without bearing healthcare costs, suffering, or unemployment, precautions may seem unnecessary. this is especially true given that persons have private information about their preventive measures, but may not possess information about whether they are infected. how can governments effectively regulate this situation in the complex environment of an infectious disease, like covid- ? we make the nontrivial assumption that governments can promulgate policies about what activities may be optimal at a given point in time (ie whether to return to work), but we focus on mechanisms of compliance, recognizing that many micro-level decisions are difficult for the government to monitor, much less to enforce. from a normative perspective, the goals are clear: to minimize the net harm to aggregate social welfare caused by the pandemic, accounting for both the public health losses (mortality and morbidity, and the costs of treating or suffering with the same) and the economic losses associated with taking precautions (lost wages and forgone consumption of welfare-enhancing goods and services). we generally assume, for the sake of argument, that certain levels of specific precautions are worthwhile from this social utility perspective. yet, for the reasons just stated, the costs and benefits of taking those precautions are poorly distributed. we do not offer a normative theory of optimal distribution, but rather a behavioral one: how to align the costs and benefits of precautions so that the precautions will be sustainably undertaken. one mechanism to address that problem is for policymakers to indemnify individuals for losses associated with taking those socially desirable behaviors to reduce the spread. we discuss a coherent mechanism for encouraging the taking of costly precautions, which may also be reinforced by social norms. the concept of 'moral hazard' is that individuals are more likely to take risky or costly behaviors when those costs are borne by others. in health policy, most of the focus on moral hazard has been in the design of health insurance policy, using deductibles and copays, to insure that once ill or injured, a patient does not wastefully consume healthcare. that form of moral hazard is sometimes distinguished as 'secondary', or 'ex post', after the illness arises. notwithstanding an overwhelming emphasis in health policy, the literature suggests that 'ex post' moral hazard is a relatively small driver of healthcare consumption for several reasons, including patients' lack of agency in making many of their own healthcare choices in practice. accordingly in the covid- infectious disease context, where underconsumption is a greater risk than overconsumption, congress provided that tests be covered without cost exposures and insurance companies have largely covered associated treatments. in contrast, 'primary', or 'ex ante', moral hazard applies where individuals who have healthcare costs largely externalized to the insurance pool may undertake risky behaviors (eg smoking or skydiving) increasing chances of having an injury or illness in the first place. for health insurance design, primary moral hazard may have a relatively small effect on risk-taking behavior, because individuals personally suffer many of the other risks associated with illness or injury (including pain, suffering, lost work, chance of death). these other costs are likely more salient to a person selecting a risky behavior than is the fact that some of the healthcare costs will be insured. for the policy of managing infectious disease, primary moral hazard is likely to be a more important driver of behavior where the chance of suffering any disutility whatsoever is heterogeneous. in this case, although the risk does not approach zero, it is as if relative youth provides partial indemnity insurance against not only the healthcare costs, but also the pain, suffering, lost work, and chance of death that are associated with covid- infection. traditional health insurance policy has conceived healthcare as a cost when a risk materializes, but healthcare is often itself a precaution against a greater future risk. baicker and colleagues have coined the term 'behavioral hazard' to refer to the phenomenon of people declining to consume worthwhile healthcare. this body of research in the health insurance domain demonstrates the more general phenomenon of how policy may strike the wrong balance, for example, if actors are biased away from optimal precaution-taking decisions, because the immediate costs of care (eg copayments for insulin) seem more salient than the longer-term costs (eg treating neuropathy for uncontrolled diabetes). of course, that problem is exacerbated if the costs of precaution are so high that individuals simply cannot afford to take them, even if they would prefer to do so. accordingly, the affordable care act now requires coverage of certain preventive services without cost exposure at all-indemnifying the cost of precaution-an approach that may improve welfare and reduce spending overall. in the recent past, few global diseases in human health have been comparable to covid- with respect to the simultaneous level of transmissibility and pathogenicity. yet policymakers are not 'flying blind' . in other infectious disease contexts, moral hazard and risk-seeking behavior have been confirmed empirically. for example, in the uk, policymakers have struggled to manage an outbreak of bovine tuberculosis. the primary policy mechanism is a surveillance program, where agricultural agents visit farms and test animals for the disease. if it is discovered, the animal is killed, and the farm is put on a disease-restricted status (like a quarantine or lockdown) for days. the farmers are able to take various costly precautions (eg maintaining fences, disinfecting trailers) to reduce the risk of infection. one might suppose that the risks of having beef purchasers reject infected meat, of having an infected animal discovered and slaughtered, and having the farm put on lockdown would be sufficient to induce optimal rates of precaution-taking. however, the policy also includes a provision to indemnify the farmer for the loss of the infected animal that must be destroyed by law. forthcoming empirical work shows that higher levels of indemnity may lead to higher rates of infection likely because farmers take fewer precautions when facing larger indemnities. in this case, the government has in part exacerbated the moral hazard problem, by paying an indemnity that lowers the farmer's risk exposure, reducing the inherent incentive he would otherwise have to keep his cattle healthy. this indemnity is similar to the young, female, otherwise healthy human in the age of covid- , who may be less likely to suffer adverse effects of infection, and thus has reduced inherent incentives to take precautions. as in human health, a farm experiencing a disease breakdown generates a negative externality for its neighbors through spatial disease transmission. by taking steps to minimize the likelihood of a breakdown on her own farm, the farmer also reduces the likelihood of infection for her neighbors. the indemnity payment relieves the farm of a portion of the costs of contracting the disease and, by doing so, reduces the incentive to invest in on-farm biosecurity. consequently, not only do high indemnity rates lead to higher rates of disease, directly, by disincentivizing precautionary steps, but also the spatial feedback effect runs the risk of generating further infections in the surrounding area. again, the example of animal disease management echoes the human experience. in the agricultural settings, the unfortunate policy of indemnifying farmers against the costs of their failing to take precautions may reflect a constitutional, legal, or political limit to governmental enforcement. the payment of an indemnity may reflect a political bargain, allowing a relatively intrusive regulatory mechanism, such as mandatory testing, to be agreeable to the agricultural lobby, if accompanied by a payment from the treasury to offset some of the costs thereof. similarly, under the us constitution's due process clause, the government must generally compensate individuals for 'takings' of their property. moreover, under international commercial agreements, businesses may enjoy a remedy against excessive regulatory interventions that cause economic injury. more generally, the us constitution limits the powers of the state and federal governments to restrict certain liberties, but the supreme court has generally upheld reasonable public health interventions, especially in times of emergency. in the case of jacobson v. massachusetts, for example, the supreme court upheld a local massachusetts law requiring smallpox vaccination. the courts have, however, struck down public health enforcement actions that were motivated by racial animus, or which were not appropriately tailored to the risks at hand. nonetheless, the government has broad powers in this domain. accordingly, to enforce orders for social distancing, or staying at home in particular, some governmental actors made strict orders backed by the coercive power of the state. on april , for example, police officers throughout the state of california began issuing citations for those who refused to close businesses or maintain social distance on beaches. if these sorts of citations, or more severe sanctions such as imprisonment, have a deterrent effect, they may solve the collective action problem. however, deterrence theory suggests that to be effective, enforcement has to find the optimal mix between detecting noncompliance and then sanctioning it once found. for a rational actor, the risk of suffering the sanction is the product of these two variables. although it may be relatively easy for the police to detect and enforce against business operators and beach goers, it will be much more difficult to detect individuals who meet to have sex, for example. similarly the war on illicit drugs has been stymied by such difficulties of detection. whether such black markets can be kept small enough to keep infection rates at manageable levels, is a key question. in this case, deterrence is also stymied by the realities of an infectious disease, such as covid- . law enforcement typically employs escalating consequences for violations, starting with the threat of fines and culminating with imprisonment as the ultimate consequence short of death. yet, fines may affect communities that have already been hit hard by the pandemic and are generally hard to enforce, when individuals have little or no funds to pay (a problem known as 'judgment proofing'). indeed, the power to fine is ultimately backed by the power to imprison those who refuse to pay. moreover, prisoners are particularly vulnerable to covid- due to overcrowding, poor ventilation, unsanitary facilities, and poor access to healthcare services. prisons easily become epicenters of disease, threatening the health and safety of the inmates as well as the surrounding community. inmates from prisons to home confinement. some prisoners have reportedly even tried to get themselves infected, as a basis for then getting early release. thus, governments are hamstrung in how they are able to enforce compliance with social distancing and quarantine requirements. fines are regressive and imprisoning an individual violating quarantine guidelines would be counterproductive as they could introduce the disease into a particularly vulnerable population where the chance of spread is high. governors and law enforcement agencies have dramatically reduced arrests. by may , some local governments were already refusing to enforce 'stayat-home' orders for political, ideological, or other reasons. hence, governments will have to look beyond the threat of fines and imprisonment to compel compliance with precautionary measures. these difficulties may explain why, in march and april , the usa saw private actors merely encouraging people to take precautions, and even some governmental actors, such as state governors, who had the power to exercise the coercive role of the state, declined to do so, in favor of mere exhortation and encouragement. the literature on 'private ordering' and the broader literature on social norms, together suggest that individuals may provide socially desirable behaviors, even when it is not compelled to do so under strict rationality. economics literature suggests the possibility of repeated interactions with individuals and the threat of social retaliation may be sufficient to establish and uphold social norms (ie 'don't burn bridges'). for instance, adherence to government recommendations to social distance may be achieved simply through the universal fear of loss of status or reputations if one is discovered to disregard recommended behavior. in hong kong and elsewhere, internet users shamed people they believed had flouted lockdown orders, traveled, or socialized recklessly during the pandemic. vendors may also voluntarily embrace precautions to burnish their own reputations for safety. for covid- , government directives were initially crude-just a binary decision to allow businesses to remain open (if essential) versus mandate closure (if not). yet, the market responded innovatively: in some 'essential' stores, there were no limits on the number of shoppers in the store, no systematic disinfection of carts or baskets or checkout counters, no masks or other protective equipment of store personnel, and the like. conversely, some stores voluntarily created waiting lines outside the store, with tape demarcating foot intervals, allowing entry into the store only with exit of another shopper, thoughtful distancing of carts from other shoppers on checkout, a designated position for the customer while items are scanned, and more. stores may have enacted voluntary precautions in hopes of boosting business by being seen as a safe place to shop, reducing liability exposure in the event of infected employees, or simply to minimize the chance of disease spread for safety reasons. eventually, governments may catch up to prescribing the sort of granular policies for managing covid- as they have, for example, for handling food-borne illness in foods. yet, reputational sanctions tend to work best when used against members within identifiable groups, such as a close-knit religious or ethnic community, but can cut against compliance of formal rules. in communities in which government authority conflicts with strong social norms or fear of loss of status or reputations, governmental orders can also be harder to enforce. it has been reported that an important factor contributing to the outbreak in south korea included mistrust by a close-knit community. economic and behavioral insights are powerful for understanding the problems inherent in managing a highly infectious disease, but what do they recommend as solutions? the overarching principle is to make it easier (ie less costly) for individuals to take reasonable precautions than to not take reasonable precautions. in this essay, we are primarily focused on regulatory mechanisms prior to a vaccine becoming available. however, the american experience with vaccine mandates is illustrative of our economic and behavioral analysis so far. consider a parent's choice about whether to vaccinate her child prior to covid- . her state may nominally require vaccination unless an exemption applies, but an indifferent or harried parent may find it easier to scribble her name on an exemption form rather than go to the hassle of finding a pediatrician, making an appointment, and securing the vaccination. indeed, scholars have found that vaccination rates vary substantially depending on the proce- dure required to invoke such an exemption. for example, in washington state implemented a requirement that parents have a counseling session with a physician prior to securing such an exemption, and exemption rates went down by %. in addition to any information delivered, these requirements function as behavioral speed bumps for the parent, with the inconvenience (ie non-monetary cost) serving to ration the scarce resource of non-vaccinated parents. in this way, the precaution is made less costly than not taking the precaution. for covid- , when a vaccine is available, it may begin as a scarce resource to be rationed across many people desiring it, but once it achieves a certain saturation, there will likely remain a minority of persons who resist vaccination, unless properly incentivized. the usa has generally focused only on childhood vaccinations, but strategies for incentivizing adults will then be required as well. for example, participating in some valuable but risky activities, such as airline travel, could be conditioned on vaccine compliance. returning to our focus on the prevaccine era, in order to diminish the negative impact on the economy and citizens' finances, some governments are considering issuing 'immunity certificates' to individuals who have recovered from the disease and are presumed to be immune from reinfection as well as unable to spread disease. these individuals with immunity certificates would then be able to re-enter society and resume their employment as well as patronize businesses and continue participating in the economy. assuming that fundamental questions regarding immunity can be answered affirmatively by medical science, and that a technological solution can be employed to ensure validity over counterfeiting concerns, such a policy could then receive support in judicial review, as discriminating against those with and without immunity, not unlike the way extant and potential policies discriminate against those who are unvaccinated. still, this sort of certificate policy could backfire, especially as it creates an incentive to contract the disease-a 'get-out-of-jail-free' card, creating a risk of community spread, prior to the hypothesized benefit of immunity being secured. intentionally seeking out disease like in the form of 'covid- parties' has precedents in the form of 'chicken pox parties' in order to obtain immunity as a young child as well as the concept of 'gift giving', the intentional transmission of hiv. unless the infection is secured in controlled conditions (eg exposure followed by weeks of quarantine), the net effects may be negative overall. this analysis suggests that the immunity certificates policy could exacerbate the same moral hazard problem, which already discourages some people from taking reasonable precautions. this problem may be overstated, to the extent that individuals are already actively flouting stay-at-home orders, immunity certificates may do little to alter their private risk calculus toward less precaution. thus, the primary concern is the population on the margin, who otherwise would have abided by shelter-in-place and social distancing, but due to immunity certificates, will be tipped toward reckless behavior. those who are extremely financially insecure, those who perceive themselves to not be at risk of serious adverse consequences of the disease, or those who believe contracting covid- is inevitable have an incentive to eschew precautionary measures and seek out the disease in order to shorten the length of time they are out of work. this problem reinforces the more fundamental need, discussed below, to further indemnify individuals who are taking precautions, including through substantial increase in unemployment compensation, as already beginning in the cares act (discussed below). speaking more broadly, while the immunity card solution allows identification of individuals who are no longer disease susceptible, it fails to solve the primary imperfect information problem-identifying who is contagious. expanded testing capacity and faster tests would allow for more direct targeting of the moral hazard problem. improved testing would dramatically reduce the need for population-wide 'shelter in place' strategies. under this scenario, quarantine restrictions could be limited only to individuals that test positive and their contacts. this strategy imposes the economic costs of contagion on precisely the right individuals-those that are contagious. and unlike the immunity card strategy, it does not create a perverse incentive to become infected, but rather imposes the costs of quarantine on those who become infected. note that such an analysis need not suggest that all individuals who become infected are in some way at 'fault' for having become infected (though some may have been, if they failed to take reasonable precautions to avoid infection). once infected due to any cause, in economics jargon, the infected individuals generate a negative externality, and, thus, they are the appropriate population to be isolated. regardless of the source of infection, it may then be unreasonable for such a person to engage in activities that risk harm to others. we also emphasize below that the individuals under targeted quarantine could be compensated for the private costs of quarantine through private or public insurance programs. a test-and-quarantine strategy would allow the economy-susceptible and immune individuals alike-to resume normal life while actively and effectively mitigating the spread of the disease. of course, this targeted quarantine strategy still requires compliance with guidelines to stay at home on the part of the covid- -positive individuals. compliance may be achieved more easily in this scenario as subjects of quarantine will know the restrictions are temporary. additionally, because those quarantined have tested positive or come in contact with a positive case, they will face more severe social pressure to avoid contact with others. however, unless widespread testing could be implemented and made mandatory (at least conditional on certain risky behaviors), other imperfections in the healthcare system confound this strategy. if tests were not mandatory, symptomatic individuals could simply avoid testing so as not to be detected as a positive case and removed from employment. for instance, those who are financially insecure and without any sickleave program that would provide compensation during the quarantine period would have incentive to avoid the test to prevent loss of income. for covid- across all these particular policy tactics, the fundamental mechanism must be to reduce the costs associated with individuals taking precautions, so that the net individual cost of precaution is lower than not taking the precaution. the most costly form of precaution is to stay home from work. one obvious solution is to effectively indemnify persons who lose income due to lost work. prior to the covid- crisis, this was a weakness of us policy, compared with other industrialized countries, which tend to have robust provisions for paid sick leave and unemployment insurance. for people who actually lose their jobs, unemployment insurance in the usa pays for up to weeks and is even then capped to cover only a portion of prior income. in one study, those receiving unemployment insurance only recouped half of their lost wages on average. the $ trillion package passed by congress in mid-march provides a one-time payment at about $ plus $ per child for most families (and less or none for others). it is worthwhile that these payments went to both workers and nonworkers, because even nonworkers face substantial disutilities in complying with social distancing protocols. however, for workers at least, the payments are far from sufficient to indemnify the costs of precaution: the median american family would need about $ , to compensate for a single lost month of work. additionally, the usa has increased weekly unemployment benefits by $ . even with the increased weekly payment for the unemployed, the benefit may end in the near future and it does not scale with income to the detriment of those living in high cost-of-living areas. this suggests that noncompliance will remain a very substantial problem (not to mention the substantial disutility suffered by those under the current income shock and the larger economic effects due to contraction). of course, with broader public spending on income replacement and broader compensation for taking precautions, someone must pay those costs. as younger persons today may be disproportionately taking precautions for the sake of protecting older persons (as we suggest in the introduction), it may not be sensible to use deficit financing, where younger persons would ultimately bear those costs. however, we set aside the point of intergenerational justice. behaviorally, it may be more sensible to distribute the burdens across current taxpayers more rationally. current us tax policy is relatively flat, unlike the wartime era of the s, when marginal tax rates went as high as % for the highest-income earners. we cannot here make a comprehensive case for the optimal tax policy, recognizing complications like tax avoidance. however, it remains true that wealthier individuals suffer less disutility for each dollar expropriated through taxation, compared with poorer individuals who suffer greater deprivations on the margin. this 'diminishing marginal utility' of money (as it is known in the economics literature) suggests a substantial opportunity to reallocate the costs of precaution to minimize the disutility. most importantly, this special pandemic situation inverts some of the classic arguments against progressive taxation as undermining incentives to work. here, the rush to get back to work may impose more social costs (in terms of infections) than social benefits (in terms of economic productivity). in this way, government payments to those who are complying with social distance or quarantine protocols are designed to indemnify them against the costs of taking those prosocial precautions. of course there is a risk of mismatch-people receiving the payments who are nonetheless not complying with distancing/quarantine guidelines. to make the bargain more explicit, the payments could be made conditional on agreeing to comply with stay-at-home or social distancing orders. given the extensive literature on the deep evolutionary basis of reciprocity for human behavior, we hypothesize (subject to testing) that such an explicit promise to comply as a reciprocal condition of accepting funds (a 'carrot' strategy) may have behavioral advantages over other ('stick') forms of enforcement, such as fines or imprisonment. our point is not about the cost of noncompliance (assume equal probability of detection and an equal-size penalty), and we are not suggesting a differential framing according to prospect theory (in both cases it may be framed as a loss conditional on noncompliance, rather than a gain). rather, asking individuals to explicitly commit to compliance as a condition of accepting payment reinforces a social norm of compliance. moreover, the breaking of the promise creates a cognitive dissonance, which itself may be experienced as a disutility. the explicit promise and size of these payments are likely to make them more effective than the sorts of small fines that may have perverse effects. in addition to self-policing by people feeling motivated to keep their own promises, compliance may be buttressed by other members of the public shaming those who do not comply or even reporting them to authorities, because they represent not only a hypothetical risk for infection but also expropriation of public funds. in this way, the noncompliance can also be connected to standard law enforcement mechanisms, where an individual convicted of violating a legitimate order suffers revocation of the payment as a penalty, with full due process protections of course. even with such a standard (rational) deterrence theory, the explicit promise to forfeit the funds upon noncompliance may make that risk seem more salient to the citizen and thus increase compliance. overall, these economic and behavioral principles suggest that to manage a pandemic prior to dissemination of a vaccine, policymakers should use conditional payments to encourage compliance with social distancing, stay-at-home, and quarantine directives. the key is to make compliance easier and cheaper than noncompliance, especially for those that would otherwise feel the greatest compulsion to not comply. adolescents are paying a high price for covid- prevention (suicide is the second leading cause of death for individuals - years of age new source of coronavirus spread: 'covid- parties', nytimes fatal tradeoffs: public and private responsibilities for risk flu vaccination coverage insurance coverage of emergency care for young adults under health reform families first coronavirus response act health insurance providers respond to coronavirus (covid- ) anticipatory ex ante moral hazard and the effect of medicare on prevention, health econ health insurance increases preventive care, not risky behaviors diabetes outpatient care and acute complications before and after high-deductible insurance enrollment: a natural experiment for translation in diabetes (next-d) study. the theory of demand for health insurance effects of a cost-sharing exemption on use of preventive services at one large employer, health aff bovine tb strategy review: summary and conclusions, department for environment food & rural affairs actions once tb is suspected or confirmed, tb hub cruel to be kind: moral hazard in british animal disease management memorandum on prioritization of home confinement as appropriate in response to covid- pandemic inmates tried to infect themselves with the coronavirus to get early release, los angeles county sheriff says portland police will not take misdemeanor arrests to jail during the covid- outbreak, to avoid overcrowding, willamette week northam encourages police to avoid arrests, imprisonment in wake of covid- virus coronavirus: officers make fewer arrests during covid- emergency two arizona sheriffs say they will not enforce governor's stay-at-home order, the hill to-know-about-face-masks-and-staying-home-as-virus-spreads#transcri pt (national institute for allergies and infectious diseases director anthony fauci interviewed: 'q. but when lives are at stake, why aren't command measures requiring people to do this appropriate now? a. that generally is not the way things operate in the relationship between the federal government and the states order without law: how neighbors settle disputes credible commitments: using hostages to support exchange social norms and community enforcement medical-exemptions-questions-and-answers/ (a california law allows the health department to report physicians to the medical board if they issue too many such exemptions can rationing through inconvenience be ethical? vaccines and airline travel: a federal role to protect the public health. covid- see robertson supra note chickenpox transmission, ctrs. for disease control & prevention the intentional transmission of hiv by hiv-positive men to their hiv-negative sex partners, automobile accidents, tort law, externalities, and insurance: an economist's critique introduction to unemployment insurance senate passes $ trillion coronavirus relief package, npr budget office, average household income, by income source and income group extra $ unemployment benefits will start flowing as early as this week for a lucky few unemployment benefits will be reduced after federal individual tax rates history, nominal dollars joel slemrod & shlomo yitzhaki, tax avoidance, evasion, and administration, in handbook of public economics social welfare and the rate structure: a new look at progressive taxation, cal. l. rev. ( ) (arguing that the diminishing marginal utility of money makes it efficient to take more from higher-paid individuals); see also leo p. martinez see generally, philippe choné & guy laroque, optimal incentives for labor force participation the case for a progressive tax: from basic research to policy recommendations, perhaps the most remarkable aspect of evolution is its ability to generate cooperation in a competitive world thus, we might add 'natural cooperation' as a third fundamental principle of evolution beside mutation and natural selection erc: a theory of equity, reciprocity, and competition prospect theory: an analysis of decision under risk applying behavioral economics to public health policy: illustrative examples and promising directions, legal promise and psychological contract the neural circuitry of a broken promise numerous psychological and economic experiments have shown that the exchange of promises greatly enhances cooperative behavior in experimental games fine enough or do not fine at all, a fine is a price. the authors thank andrea sharp and jacqueline salwa for research assistance, and bert skye for administrative support. the manuscript benefited from two anonymous peer reviewers. key: cord- - su m authors: alam, aatif; jiang, linda; kittleson, gregory a.; steadman, kenneth d.; nandi, somen; fuqua, joshua l.; palmer, kenneth e.; tusé, daniel; mcdonald, karen a. title: technoeconomic modeling of plant-based griffithsin manufacturing date: - - journal: front bioeng biotechnol doi: . /fbioe. . sha: doc_id: cord_uid: su m griffithsin is a marine algal lectin that exhibits broad-spectrum antiviral activity by binding oligomannose glycans on viral envelope glycoproteins, including those found in hiv- , hsv- , sars, hcv and other enveloped viruses. an efficient, scalable and cost-effective manufacturing process for griffithsin is essential for the adoption of this drug in human antiviral prophylaxis and therapy, particularly in cost-sensitive indications such as topical microbicides for hiv- prevention. the production of certain classes of recombinant biologics in plants can offer scalability, cost and environmental impact advantages over traditional biomanufacturing platforms. previously, we showed the technical viability of producing recombinant griffithsin in plants. in this study, we conducted a technoeconomic analysis (tea) of plant-produced griffithsin manufactured at commercial launch volumes for use in hiv microbicides. data derived from multiple non-sequential manufacturing batches conducted at pilot scale and existing facility designs were used to build a technoeconomic model using superpro designer(®) modeling software. with an assumed commercial launch volume of kg griffithsin/year for . million doses of griffithsin microbicide at mg/dose, a transient vector expression yield of . g griffithsin/kg leaf biomass, recovery efficiency of %, and purity of > %, we calculated a manufacturing cost for the drug substance of $ . /dose and estimated a bulk product cost of $ . /dose assuming a % net fee for a contract manufacturing organization (cmo). this is the first report modeling the manufacturing economics of griffithsin. the process analyzed is readily scalable and subject to efficiency improvements and could provide the needed market volumes of the lectin within an acceptable range of costs, even for cost-constrained products such as microbicides. the manufacturing process was also assessed for environmental, health and safety impact and found to have a highly favorable environmental output index with negligible risks to health and safety. the results of this study help validate the plant-based manufacturing platform and should assist in selecting preferred indications for griffithsin as a novel drug. griffithsin is a high-mannose binding lectin found natively in the marine red alga griffithsia (mori et al., ) . the protein is composed of amino acids and its monomer has a mass of approximately kda. griffithsin forms a homodimer with six binding pockets with high affinity for mannose, a common sugar found at the terminal end of oligosaccharides on the surface of many enveloped viruses. the protein is thought to inhibit the entry of enveloped viruses into host cells as well as viral maturation and transmission events by binding to oligosaccharides on the viral envelope surface. native griffithsin and its analogs are the most potent hiv- entry inhibitors yet described, with ec values in the picomolar range (mori et al., ; o'keefe et al., ) . griffithsin also effectively inhibits transmission of hsv- (nixon et al., ) , hcv (meuleman et al., ) , sars-cov (o'keefe et al., ) , ebola (barton et al., ) , and possibly other viruses yet to be studied. importantly, griffithsin appears devoid of cellular toxicity that is associated with other lectins. o'keefe et al. conducted studies with explants of macaque and rabbit vaginal tissues ex vivo and showed that griffithsin did not induce changes in the levels of cytokines or chemokines, nor did it alter lymphocyte levels in human cervical tissue nor elicit inflammatory responses in rabbit tissue (o'keefe et al., ) . the combination of extremely wide viral target range and demonstrated preclinical safety makes griffithsin potentially useful as a prophylactic and/or therapeutic in multiple and diverse antiviral indications. the potential indications for griffithsin as a human prophylactic or therapeutic include its use as an active pharmaceutical ingredient (api) in vaginal and rectal microbicides. in spite of the value shown by pre-exposure prophylaxis (prep) drugs to prevent hiv transmission, issues of cost, side effects, the potential for development of viral resistance through chronic use of antiretrovirals (arv) as prevention modalities, and access to prep drugs by underresourced populations remain. these unmet needs could be met by the availability of affordable, safe and effective "on demand" antivirals, especially with griffithsin as the api and its potential to control co-transmitted viruses such as hiv- , hsv- and hcv during intercourse. adoption of griffithsin as a new biologic drug, especially in cost-constrained products such as microbicides, is predicated on the feasibility of a scalable manufacturing process that can supply market-relevant volumes of the api at an acceptable cost of goods sold (cogs). previously, we showed that recombinant griffithsin can be expressed and isolated with high efficiency using transient gene expression in green plants (fuqua et al., a,b) . although the process described can be further optimized, the achieved pilot-scale expression yields of > . g griffithsin per kg of fresh (hydrated) green biomass ("fresh weight"; fw), recovery efficiencies of - % overall, and griffithsin purity of > % of total soluble protein (tsp) are already impressive. in this study, we developed a technoeconomic model for griffithsin manufacturing using a plant-based system with the goal of estimating api manufacturing cost and determined the factors that have the greatest impact on cogs. the output of our study should serve as a basis for additional process improvements, selection of a commercial-scale manufacturer, and should assist in the identification of future product targets for cost-sensitive markets such as prophylactic microbicides as well as those for less cost-constrained therapeutic indications. technoeconomic modeling was performed with the widely used superpro designer modeling software (intelligen, inc., scotch plains, nj, usa). the main analysis in this study was conducted using data available from pilot-scale manufacturing of griffithsin in nicotiana benthamiana plants using tobacco mosaic virus (tmv)-induced transient gene expression, and assuming that manufacturing would take place in an existing and fully equipped state-of-the-art plant-based biomanufacturing facility. modeling costs based on existing resources of a contract manufacturing organization (cmo) instead of a "greenfield" build of a new facility was seen as the most likely scenario for launch of a new product. our reasoning was that dedicated infrastructure could be built subsequently depending on market demand for the drug. as a result, we did not estimate capital equipment or total capital investment costs, and neglected depreciation, insurance, local taxes and factory expenses in the manufacturing operating cost analysis as these investments would have been made by the cmo. our analysis assumed a % net profit margin/fee (sood et al., ) assessed by the cmo and this figure was added to the production cost of the product to arrive at the final total product cost. in addition to the technoeconomic analysis, an environmental health and safety assessment (ehsa) of the designed process was conducted using the method described by biwer and heinzle ( ) to evaluate the environmental, health and safety impact of griffithsin manufacturing using the plant-based system, with the goal of assessing the sustainability of the process. the technoeconomic modeling for this study was performed using superpro designer ("superpro"), version . (intelligen, inc., scotch plains, nj; http://www.intelligen.com/), a software tool for process simulation and flowsheet development that performs mass and energy balances, equipment sizing, batch scheduling/debottlenecking, capital investment and operating cost analysis, and profitability analysis. this software has been used to estimate cost of goods in a variety of process industries including pharmaceuticals produced by fermentation (ernst et al., ) and plant-made pharmaceuticals (evangelista et al., ; zapalac and mcdonald, ; tusé et al., ; nandi et al., ) . it is particularly useful at the early, conceptual plant design stage where detailed engineering designs are not available or warranted. superpro was chosen because it has built-in process models and an equipment cost database for typical unit operations used in the biotechnology industry, such as bioreactors, tangential flow ultrafiltration and diafiltration, chromatography, grinding or homogenization, and centrifugation. there are some specific unit operations and processes used in this study that are currently not included in superpro, such as indoor plant cultivation, transplantation, plant harvesting and screw press/disintegrator. such unit operations were addressed through the "generic box" feature of the application. unless otherwise noted, the maintenance costs of major equipment, unit operation-specific labor requirements and costs (e.g., operators, supervisors), pure components, stock mixtures, heat transfer agents, power and consumables (e.g., filter membranes, chromatography resins) used in the analysis were determined using the superpro built-in equipment cost model and default databanks. additional case study specific design parameters were selected based on experimental data from journal articles, patent literature, the authors' laboratories, interviews with scientists and technologists conducting the work cited, technical specification sheets or correlations, heuristics, or assumptions commonly used in the biotechnology and/or agricultural industry. process flow and unit operations were derived from published methods and unpublished results obtained by the authors and collaborators who have participated in the development and scale-up of the process described and in the development of griffithsin products. on the basis of this information, the superpro software was used to select and size equipment for each of the unit operations to achieve the desired production target ( kg of purified griffithsin/year), simulate the operations by performing material and energy balances, and specify and schedule all operations taking place within each piece of equipment to calculate material inputs and outputs and process times. costs for raw materials, utilities, consumables, labor, laboratory qa/qc, waste disposal and equipment maintenance were then used to determine annual operating costs, and per-unit mass or per-dose costs ($/kg or $/dose). the main case study model was based on an existing plantbased manufacturing facility, operating in batch mode, and excluded new capital investments and other facility dependent costs, except for equipment maintenance costs, which were included. for the downstream portion of the griffithsin manufacturing process, an annual available operating time of , h ( days, -h operation, or % available operating time per year) for the facility was used with indoor-grown nicotiana benthamiana plants. operating time was based on holtz et al. ( ) for a similar facility, which was designed with overlapping utility capacity and in which the largest single utility unit can be down for maintenance and/or repairs and the utility loads can be maintained with redundant (spare) equipment. likewise, per nandi et al. ( ) it was assumed that the plants would be grown continuously throughout the year ( , h, or days, -h operation, or % available operating time per year). land costs, upfront r&d, upfront royalties, and regulatory/certification costs were neglected in the model as these costs can vary widely. griffithsin protein for this modeling study was produced in nicotiana benthamiana host plants. this host is preferred for indoor protein manufacturing due to its metabolic versatility, permissiveness to the propagation of various viral replicons, and high expression yields achievable with a wide range of targets, as reviewed by pogue et al. ( ), de muynck et al. ( , thomas et al. ( ) , gleba et al. ( ) , and others. griffithsin protein can be produced in plants in a number of ways. these include (a) stable expression in recombinant plants; (b) inducible expression in transgenic plants; (c) transient expression induced directly by tobacco mosaic virus (tmv) replicons; or (d) via agrobacterial vectors introduced into the plants via vacuumassisted, or surfactant-assisted, infiltration (gleba et al., ) . relative to stable transgenic plants, the advantages of speed of prototyping, manufacturing flexibility, and ease of indoor scale-up are clearly differentiating features of transient systems and explain why this approach has been widely adopted in the manufacture of many plant-made pharmaceuticals (gleba et al., ) . in our base-case analysis, we modeled expression of griffithsin using tmv induction described in fuqua et al. ( b) and results from pilot-scale manufacturing runs because these batches provided the most extensive and complete data set; however, this process has been corroborated in additional manufacturing runs at pilot-scale or larger. nicotiana benthamiana host plants are generated from seed and propagated indoors under controlled environmental conditions until sufficient biomass is obtained for inoculation with the tmv vector carrying the griffithsin gene. the process is summarized as follows. an n. benthamiana master seed bank is generated from seeds obtained from the u.s. department of agriculture (usda) repository. for biomanufacturing, seeds from the tw- line are obtained in bulk and stored securely. the master seed bank is qualified for germination rate (> %), freedom from disease, and genetic uniformity, and stored in sealed containers under temperature-controlled conditions ( ± • c). if the seed batch passes release tests, it becomes the production seed batch and is used in the designated production run ("working seed lot"). seedlings are allowed to grow for days under controlled environmental conditions ( ± • c and % rh per holtz et al., ) . at this stage, the seedlings are transplanted to accommodate their larger size and moved to another growth room to await inoculation, as described in the following sections. the expression vector is constructed as described in o' keefe et al. ( ) . briefly, a synthetic cdna (genbank no. fj ) encoding the -amino acid griffithsin amino acid sequence is cloned into a tmv-based expression vector. in posttranslational processing in planta, griffithsin's amino-terminal methionine is cleaved and the n-terminal serine is acetylated. the construct containing the tmv vector backbone and griffithsin gene insert are built into a plasmid that is propagated in the e. coli host strain dh b (fuqua et al., b) and constitutes a master plasmid bank. the master plasmid bank is maintained in stocks at − • c and is checked periodically for stability and insert fidelity. excision via t polymerase produces free tmv transcript, which constitutes a working transcript batch used to inoculate n. benthamiana plants days after sowing and generate a tmv virion inoculum batch days post infection, which is checked for conformance to quality control criteria (e.g., infectivity, message fidelity, bioburden, stability) per fuqua et al. ( a) . the tmv inoculum is then applied to host plants to initiate expression. the tmv inoculant is applied to the -day-old plant host production batch by high-pressure spray with an abrasive (diatomaceous earth), to introduce the virus into plant tissue. once the tmv vector gains access to plant tissue, the virion decapsidates and the genomic rna encodes for a polymerase/replicase to multiply the message. as described in shivprasad et al. ( ) and pogue et al. ( ) , subgenomic promoters also drive expression of a movement protein (mp) to translocate the transcript throughout the plant, and a coat protein (cp) that encapsidates the rna and reconstructs the virion that then self-propagates throughout the plant. simultaneously, a subgenomic promoter (tmv u ) also drives expression of the griffithsin gene, which is translated into griffithsin protein. plants at this stage are therefore induced to synthesize the api. using this expression method, griffithsin concentration in planta reaches a maximum without further increase typically days post inoculation (optimized internally based on the amount of inoculum used). at this stage, the plants are ready for api extraction. the api extraction procedure modeled is per holtz et al. ( ) except that a : ratio of biomass:buffer is used. briefly, the aerial parts of the plants (i.e., leaves, stems) containing accumulated griffithsin are mechanically inverted and cut with a mechanical cutter. the harvested biomass is collected in baskets for transport to the extraction suite, to initiate downstream processing. the harvested biomass fresh weight (fw) is determined to calculate the volume of extraction buffer to be added, typically at a rate of kg biomass fw: l buffer mix ( mm sodium acetate, mm sodium chloride, mm ascorbic acid, mm sodium metabisulfite). the ph is adjusted to . and the mixture is heated to • c for min to help precipitate major host plant proteins. the heated mixture is passively cooled and filtered ( . µm cellulose filter) to yield a crude extract. the crude extract is stirred overnight at • c in the presence of bentonite and mgcl . this procedure helps remove tmv coat protein (cp), which at this step represents the largest protein impurity in the extract. the suspension is filtered ( . µm cellulose filter) to remove aggregated tmv cp, yielding a clarified and partially purified api-containing solution and is then sterilefiltered ( . µm polyethersulfone filter). in-process controls are applied throughout downstream processing unit operations to determine reagent volumes and assess yield and quality at key steps. the partially purified extract is subjected to capto r mmc multi-modal chromatography (ge healthcare) per fuqua et al. ( b) using a -step pbs gradient ( % and % phosphate buffered saline [ mm nacl, . mm kcl, mm nah po , mm kh po ]) at ph . . the purified product consists of the drug solution in pbs, which is considered the drug substance (ds). the ds is release-tested per specification and is typically > % pure (fuqua et al., b) . the ds solution is typically bulk-packaged in inert bottles with screw cap closures per usp class vi guidelines. because container options vary, the final packaging step was not included in the model. results used for modeling purposes were averages from non-sequential manufacturing runs at pilot scale conducted at kentucky bioprocessing llc ("kbp, " owensboro, ky, usa), the first of which was described in fuqua et al. ( b) . these results have been corroborated by additional production runs since. under the conditions described, . g griffithsin is expressed per kg plant biomass fw. overall recovery efficiency by the method described is typically ≥ %, or ≥ . g griffithsin/kg fw biomass. to adequately meet the projected initial annual market demand for a rectal microbicidal formulation in the united states, approximately . million doses of griffithsin api at mg/dose would be needed. this translates into a production rate of kg of purified griffithsin api per year. the manufacturing facility to produce the required kg of api per year was assumed to segregate production operations into two broad categories; namely, upstream production and downstream recovery and purification. to accommodate a large number of plants, the facility uses a vertical (layered) cultivation design with integrated irrigation and runoff collection system. each rack is compatible with an integrated transportation infrastructure to move each tray to the next phase of the growth cycle. the upstream portion of the facility houses unit operations for n. benthamiana propagation, inoculation with tmv vector, and griffithsin protein expression and accumulation. these processes begin with seeding and end when the biomass is taken to harvest. the downstream portion of the facility begins at harvest and continues through purification of the griffithsin ds. upstream processing is assumed compliant with good agricultural practices (gap), whereas downstream processing is subject to fda current good manufacturing practice (cgmp). the general layout of the upstream growth rooms was adapted from holtz et al. ( ) , and includes one germination chamber for seeds, one pre-inoculation room for biomass growth, and an isolated post-inoculation chamber where n. benthamiana inoculated with tmv expresses and accumulates griffithsin. all plant growth was modeled to occur indoors using a vertical rack system with hydroponic irrigation. plants are arrayed in equally sized trays under light-emitting diode (led) light systems tuned to the optimized photosynthetic absorbance spectrum of n. benthamiana (composite blue/red spectrum: % ± nm wavelength/ % ± nm wavelength; holtz et al., ) and are continuously illuminated. the plants are rooted in rock wool cubes held in the trays by polystyrene foam floats and perfused with a nutrient solution (the components of which are listed in supplementary table in supplementary material). hydroponic irrigation is on a -h cycle and is accomplished via nutrient film technique (holtz et al., ) . we modeled a hydroponic system because the nutrient solution is recycled; hence, water is conserved, and fertilizer runoff is reduced although not eliminated. the mass of nutrient solution taken up by the plants, the cost of the nutrient solution per liter, and the mass of residual nutrient solution that goes to the wastewater treatment system are shown in supplementary table in supplementary material. to ensure consistency of the nutrient solution, all water was assumed to be treated by reverse osmosis (ro) with solution-monitoring for proper ph and dissolved solids content. a semi-quantitative environmental health and safety assessment was conducted by determining the hazardousness and mass of input materials used in the described upstream and downstream manufacturing operations as well as the hazardousness and mass of waste products generated. the method is referred to as "semi-quantitative" because the amounts of input and output components are quantified from superpro, but the "hazardousness" of each component is determined from the properties of the component (e.g., thermophysical properties, material safety data sheets, national fire protection association (nfpa) ratings, etc.) per three qualitative classifications followed by assignment of a numerical value based on the classification (see biwer and heinzle, ) . the three phases of plant growth (i.e., seed germination, seedling growth and development pre-inoculation, and post-inoculation maturation) require a total batch time of days in the upstream portion of the facility. due to the protracted and continuous nature of plant cultivation, the upstream portion of the facility contains multiple concurrent batches staggered at different stages of growth. when one batch graduates to the next step of production (every . days), the trays containing the batch's biomass are cycled out and the corresponding rack space is immediately filled with a new rotation of trays. we divided the -day growth period into concurrent batch periods, with one batch ready to enter downstream purification every . days. table is a summary of the number of plants, trays and batches that comprise the upstream facility at any given moment. for model building, batch schedules were calculated under the initial assumption of / operation for days per year. plant uptake of nutrients and growth were assumed to be linear reaching g fw per plant at viral inoculation and then increasing in mass to reach g fw per plant at harvest. a % failure rate of tmv inoculation was assumed (pogue et al., ) . the griffithsin expression rate was fixed at . g/kg fw the results generated by the software for the upstream operations are shown in figure , with scheduling shown in the equipment occupancy chart in figure . the following descriptions elaborate on the schema presented in each figure. griffithsin recovery and purification was modeled as a batch process in a facility with an available operating time of days a year for h a day and days a week. in each year, there are batches total to produce kg of purified griffithsin api. since the recovery and purification process only takes . days, the downstream facility has a significant down time of . days between batches. overall, each batch requires . days from seed planting to formulating the final product, with days upstream and . days downstream. in figure , the upstream processes are dictated by concurrent batches (represented by generic boxes) with each batch being . days apart from each other. a batch basis of . days was chosen to decrease equipment idle time and thereby increase downstream equipment utilization efficiency. despite the . -day batch period and a -day operating year, in the model the batch time upstream was reduced to approximately days and the operating year was increased to days to reach the desired batches per year. this was done because superpro reproduces uniform results for each year. the goal of the upstream process operations is to produce sufficient biomass to enable isolation of kg griffithsin per annum. the modeling results show that each batch would produce kg of biomass containing g of griffithsin, assuming an expression yield . g api/kg fw biomass (fuqua et al., b) . because induction was modeled using infection with recombinant tmv vector, the three main phases in upstream are germination, pre-inoculation, and postinoculation. the duration of the phases in the model are days, days, and days, respectively. each batch of n. benthamiana plants goes through a germination phase of days and the germination room is designed with a capacity to grow the , plants necessary to reach the production goal. this step of the process uses germination trays, each holding about plants, distributed among batches in the germination room. after days post germination, the n. benthamiana plants are transplanted to a lower density to enable further growth. thus, seedlings from one germination tray are transplanted into three grow trays (with plants per grow tray), meaning that there are three times the number of trays in pre-and post-inoculation, individually, than in germination. the plant density is plants per m in the germination trays and plants per m after transplantation. in practice, during transplantation each plant will spend only a few minutes away from its growth environment to minimize transplant shock and undue stress. in the model, the overall time was overestimated to be h to accommodate other necessary procedures, such as moving the plants back to the tray stacks. the transplanted trays are relocated to pre-inoculation rooms that are designed to accommodate the increased area from transplanting for ∼ days. the pre-inoculation room contains batch, each containing trays with plants per tray. recombinant tmv for inoculation is produced in and isolated from n. benthamiana. the plant growth model is the same as the rest of n. benthamiana plants. by using infected plants and the purification model defined by leberman ( ) , mg of pure tmv per gram of infected plant material can be recovered (leberman, ; bruckman et al., ) . each batch is equivalent to , plants distributed on trays. less than microgram of tmv virion is needed to inoculate each plant (pogue et al., ) . thus, approximately . mg of tmv is needed per batch and the necessary amount of tmv to inoculate a batch can be produced from a single n. benthamiana plant. multiple batches of tmv solution can be made simultaneously and stored at − • c (fuqua et al., b) . tmv production can be done at lab scale and equipment, labor and material costs are negligible (∼$ , ) compared to the overall cost of plant maintenance. the isolated tmv is incorporated in diatomaceous earth buffer solution at a concentration of micrograms per . ml of diatomaceous earth buffer solution, which contains % by volume diatomaceous earth and % by volume of sodium/potassium-based buffer (pogue et al., ) . the selected inoculation volume of . ml is a safe middle value from the range suggested in the literature (e.g., - ml, pogue et al., ). in the model, the estimated mixing and transfer time for the solution is h, which starts at the beginning of post-inoculation, so the plants and solution enter the same stage together. a forklift is used to transport the plants into the inoculation room. the plants are inoculated with the diatomaceous earth buffer solution described above with a high velocity spray. inoculation machines are often custom made and consist of a conveyor traveling through an enclosed cylinder equipped with high pressure spray nozzles aimed at the plants' aerial structures. once the inoculation is complete, the trays are conveyed to the post-inoculation growth room, which is similar in design to the pre-inoculation growth room; the main difference being its size. the post-inoculation room contains batches at any given time for a total of trays with plants per tray. the scheduling of batches is summarized in the equipment occupancy chart in figure . as shown, seeding, germination, transplant, pre-inoculation, inoculation, and post-inoculation occur sequentially, and the batches are staggered by . days. the downstream unit operations developed in superpro are shown in figure , with scheduling summarized in the equipment occupancy chart shown in figure . the following descriptions elaborate on the schema presented in each figure. at the end of each . day growing rotation cycle upstream, one batch of n. benthamiana plants is ready to be transferred to downstream processing. this is done by placing each tray of plants onto a conveyor system which leads them to the first phase of downstream operations. the matured plants are first harvested for the green biomass from which the majority ( %) of griffithsin can be recovered with a single extraction. additional griffithsin could be recovered from fibrous material by reprocessing (o'keefe et al., ) and from roots (which are not harvested); however, reprocessing was not included in this model. the automated harvester processes the kilograms of biomass at a rate of kilograms of biomass per hour. with an operational buffer time of h, this process is thus expected to take h. as the biomass is processed by the harvester, it is directly fed into a shredder which further comminutes the biomass to improve griffithsin recovery. the shredder operates at a capacity of kg of harvested biomass per hour for . h. the shredded biomass is then mixed with an extraction buffer in a buffer addition tank. for every kilogram of plant material, l of extraction buffer is added. thus, for kg of n. benthamiana in a batch, approximately l of extraction buffer are added. the resultant solid-liquid mixture has a total volume of about , l and is sent through a screw press, which is represented as a generic box in the model. the screw press separates the solidliquid slurry leaving a main process fluid stream of plant extract and a waste stream of biomass. the extract solution contains griffithsin as well as the host and viral protein impurities. a loss about % of the original starting griffithsin was modeled assuming it to be non-liberated from the homogenized biomass. the removal of the biomass leaves a main process stream that contains about l ( kg/batch). to facilitate the aggregation of proteinaceous impurities, the extract solution is transferred into a mixing tank and heated to • c for min. the mixture is passively cooled and simultaneously transferred out of the tank and fed into the first . µm plate-and-frame filter. the extract solution is filter-pressed at - psig to remove the aggregated protein impurities. filtering has a process time of h and requires a filter area of m to handle the kg/batch of the process stream. at this stage, the process loses a further % of the griffithsin but removes all the rubisco (ribulose- , bisphosphate carboxylase/oxygenase) and % of the tmv coat protein impurities. the filtrate from this step is transferred to a second mixing and storage tank, mixed with bentonite clay and magnesium chloride, and stored at • c for a -h period. this stage is the bottleneck operation for the downstream process. after the -h incubation, the solution is filtered through a second . µm filter press and a . µm inline sterilizing filter. these operations remove the remaining protein impurities leaving a griffithsin extract with greater than % purity but at the cost of losing % of the griffithsin. the second plate-and-frame filter has a filter area of about m and will process all of the extract in h. there is approximately g of griffithsin per batch at the end of the filtration phase. following the filtrations steps, the griffithsin extract solution is collected in a storage tank and further purified using an axichrom column with capto mmc resin to remove residual color and potential non-proteinaceous impurities. to accommodate the g of griffithsin in solution, . l of mmc bed resin is needed at a mg/ml binding capacity (per product specification sheet). the order of the operations for this chromatography step are: equilibrate, load, wash, elute, and regenerate. in total, chromatography requires h with the load step taking the longest, at h, because approximately l of solution are processed. chromatography is necessary to decolorize the extract at the expense of losing % of the griffithsin, giving a remaining griffithsin mass of g per batch. the l of eluant process fluid is sent through a viral clearance filter and transferred into a pool/storage tank. subsequently, the extract is sent through an ultrafiltration/diafiltration cycle to remove salts introduced in the chromatography column. after ultrafiltration, the product is transferred into a storage tank to be mixed with the final formulation components. the concentrated griffithsin is diluted to give a concentration of g/l griffithsin in mm na hpo , . mm kh po , . mm kcl and mm nacl at ph . . the final volume of the ds is l per batch. as shown by figure , each batch in the downstream requires h of process time which includes all sip and cip operations. as batches move from the upstream portion of the facility every . days, the remaining time left over in the downstream is set as slack time in the model that may be dedicated toward repair, maintenance, etc. the assumptions and results developed in superpro were used to calculate the economics of the process described. table shows the total operating costs segregated individually for upstream and downstream components. figure displays process category cost contributions graphically, including percentages of total costs. in upstream operations, the largest cost components are utilities ($ , ) and labor ($ , ), representing % and % of total upstream costs, respectively. in downstream operations, labor-dependent costs ($ , ) are the highest contributors at % of total downstream costs, followed by consumables ($ , ) at % of total downstream costs. overall, the upstream component represents nearly % of the total griffithsin production cost, which is calculated as just over $ /g protein. for a microbicide dose of mg, the per-dose manufacturing cost is $ . , excluding any cmo fee. an environmental health and safety assessment was also conducted for this case study following the method of biwer and heinzle ( ) and the results are found in supplementary tables - in supplementary materials. overall, the process uses chemicals that are not harmful to people or the environment, as can be seen by the low magnitude of input and output environmental factor values (typically less than . on a - scale) in supplementary table . the biggest causes for concern (based on the environmental indices) are tmv in the residual biomass, and sodium hydroxide and phosphoric acid used in clean-in-place operations, if released to the environment; however we included costs for a thermal or chemical deactivation step for the tmv-contaminated biomass and ph neutralization for the acid and base cleaning agents which would eliminate the environmental impact of these components. it should also be noted that the upstream nutrient compounds can be more efficiently recycled to increase nutrient utilization by the plants and reduce water/soil impact. waste compounds in the downstream process are disposed of through wastewater and biowaste treatment. an aggregate disposal cost of $ . per liter of non-tmv-contaminated aqueous streams and $ . per kg of biowaste is assigned in superpro for expenses related to wastewater disposal and thermal/chemical deactivation of biowaste streams. compounds introduced during or after the post-inoculation step in the upstream facility are considered as biowaste since they may contain tmv. this includes spent nutrient solution in the post-inoculation step and retentate streams from plate-andframe and dead-end sterilizing filtration skids. disposal of tmv-contaminated materials poses low environmental risk. there is extensive industrial experience in disposing of tmvcontaminated materials, which can be rendered non-infective by treatment with bleach, heat or detergents, diluted and disposed of as municipal waste (pogue et al., ) . the facility modeled can annually produce kg of the potent antiviral griffithsin for use in microbicide products. the host used in our modeling was nicotiana benthamiana. this species was selected because of its aforementioned productivity, but also because our previous report on technoeconomic modeling of nicotiana-produced therapeutic and industrial products nandi et al., ) prefaces the work reported herein. in addition, the use of nicotiana for production of clinical trial materials is also familiar to fda and other regulatory agencies, thus facilitating nicotiana's acceptance in regulation-compliant manufacturing (streatfield and howard, ; mccormick et al., ; bendandi et al., ; tusé, ; gleba et al., ) . the api is manufactured in the host nicotiana benthamiana using tobacco mosaic virus (tmv) as the expression vector. the upstream plant growth and griffithsin production operations are adapted from the facility layout detailed by figure | upstream and downstream cost contributions by process category (units in $ ). holtz et al. ( ) . over , plants are housed in vertically stacked hydroponic grow racks, fitted with high-efficiency led lights. the environment is controlled and monitored for compliance with good agricultural practices (gap). each batch of , plants grows over the course of days and yields a total of kilograms (fresh weight) of biomass. ninety-five batches are seeded and grown annually, with one batch reaching harvest every . days. the downstream griffithsin extraction and purification process is scaled up from the pilot industrial scale process presented by fuqua et al. ( b) . an expression rate of . grams of griffithsin per kilogram of biomass (fresh weight) and a downstream recovery of % were used in the base case and give a combined yield of . grams of griffithsin per kilogram of harvested biomass. sterile filtration and cip/sip systems facilitate compliance with cgmp guidelines. downstream processing commences upon the completion of an upstream batch and takes . h. the stable final formulation is > % griffithsin as the api with negligible endotoxin levels. in the model, the upstream costs account for nearly % of the total cost of griffithsin production. containing both upstream and downstream losses of the protein could significantly reduce cogs. approximately % of the protein api is non-liberated from the homogenized biomass (reprocessing was not modeled) and % is lost during downstream polishing steps. based on the data and assumptions employed in the current analysis, the unit production cost of griffithsin is estimated to be $ . per dose ( milligram). the model was based on published designs for a commercialscale facility and pilot-scale data on griffithsin production adapted to the facility described. this type of modeling is useful for determining ranges of api selling price, production capacity and expression level requirements for commercial supply and profitability. in this study we modeled the manufacturing of griffithsin through a contract manufacturing organization instead of a greenfield build of a new facility because we assumed that that would be the most prudent approach to launching a new product. if the product manufactured using the process modeled is used directly as a vaginal rinse or rectal enema, the additional costs post manufacturing would include transportation, storage, insurance, distribution, marketing, etc., none of which were modeled in this manufacturer-level analysis. if the drug substance produced via the process analyzed is further formulated (e.g., as the api in gels, suppositories, or condom additives), or used as a component of another device (e.g., vaginal ring), those costs and other product-specific costs would be additive and were also excluded from our manufacturer-level analysis. the cost of goods calculated by the current model reflects the manufacturer's cost of production. we are less certain about the wholesale price of the drug because there is no standard "offthe-shelf " profit margin that can be added to toll manufacturing cost to arrive at a standardized answer. often scale up to commercial launch volumes of a product requires additional process development and optimization, validation batches, etc., which lead to negotiated transfer prices depending on volume, duration of engagement, license fees, export duties, and other factors, all of which would impact the cost of bulk griffithsin. nevertheless, for this discussion we assumed a manufacturer's fee of % of cogs for a total production cost of bulk griffithsin drug substance of $ . /dose. additive formulation, storage, distribution, insurance, marketing, sales margins and other costs could lead to a consumer-level use cost of $ - /dose (i.e., ∼ to -times the production cost and < to times the price of a male condom, which varies widely depending on material, features and quantities purchased). this technoeconomic analysis emphasized griffithsin's use in microbicides because such products arguably represent the most price-constrained applications of this new drug. we cannot define the target retail price of a griffithsin microbicide; there is no market reference price for microbicides since no commercial microbicides yet exist. for perspective, the user cost of a griffithsin microbicide can be benchmarked against pre-exposure prophylaxis (prep) with traditional male condoms and prep with microbicides containing antiretroviral (arv) drugs as a newer alternative. analyses have been conducted on the cost of prevention modalities and the cost savings to the healthcare system enabled by preventing hiv transmission, with prevention being far more cost effective than treatment in most scenarios (e.g., pretorius et al., ) . walensky et al. ( ) conducted an analysis of the cost-effectiveness of a tenofovirbased prep microbicide in south african women. in their cost modeling of a vaginal gel, they multiplied the product cost of $ . /dose times (product must be applied twice, pre-and postintercourse) and by . (average sex acts per woman-month) to arrive at a product use cost of approximately $ /womanmonth. however, the price of the microbicide gel used in the study was assumed and region-adjusted and hence pricing in other countries may be different. terris-prestholt et al. ( ) estimated tenofovir gel prices of $ . - . per dose, provided that the gel was used in combination with a condom ($ . - . each; planned parenthood, ), from which an additive cost of use (single-use condom plus double-dose microbicide gel) of $ -$ /person-month can be derived. assuming the same average use rate ( . applications/month) of a griffithsincontaining microbicide applied singly without a condom and priced at $ . -$ . per dose, the cost of use would be $ -<$ /person-month. whether a higher cost of use discourages adoption of griffithsin-based microbicides by men and women remains to be shown. a market study by darroch and frost ( ) of the alan guttmacher institute consisted of detailed interviews of a cross-section of , sexually active women aged - in the continental united states. their statistically rigorous survey identified levels and predictors of women's concerns about stds (including hiv transmission) and interest in microbicides, as well as their preferences regarding method characteristics and likelihood of usage versus price of product, with survey sample results extrapolated to the national level. the results showed that of the estimated . million women aged - interested in microbicides and concerned about stds, including hiv, . million ( %) would still be interested in the method even if it were not % effective, and . million ( %) would remain interested even if the microbicide did not protect against stds other than hiv. the same study found that women's predicted use of a microbicide was affected by price, but interest was still high at $ per application, or roughly up to -times the average price of a male condom. the survey concluded that more than seven million sexually active women in the usa would be interested in a vaginal microbicide even if the product only protected against hiv, was only - % effective and cost them $ per application (darroch and frost, ) . that conclusion was arrived at in ; the $ per application cost back then would be $ . in . one can conclude from these results that there is interest in effective yet inexpensive, self-administered hiv and std prevention modalities even if such products might cost more than conventional prevention methods. the darroch and frost analysis was conducted nearly years ago, and the interviews were limited to women practicing vaginal intercourse. to our knowledge, a more recent study linking likelihood of product use and price sensitivity has not been conducted, or at least not reported, to include other populations of potential microbicide users such as heterosexual couples practicing anal sex or gay men practicing unprotected rectal intercourse. nevertheless, the study established an initial price point and price sensitivity for potential users of microbicides in the usa. griffithsin has a broader spectrum of antiviral activity than hiv-specific prep agents, including activity against hsv- and hcv, which are co-transmitted with hiv- (meuleman et al., ; nixon et al., ) . hence, griffithsin might command a higher price due to its broader antiviral activity and its potential to obviate prevention and treatment costs for co-transmitted viruses. in the usa, the cost of the oral prep drug truvada (emtricitabine and tenofovir disoproxil fumarate) ranges from $ , to over $ , per month (https://www.goodrx.com/ truvada) for the uninsured, but treatment is typically covered by insurance with user co-payments of $ -$ per month. so even if a griffithsin-containing microbicide sold for $ per application (e.g., $ per -use pack), a user of packs per month would pay $ for the microbicide, which is in the range of prep, with the potential added benefit of controlling co-transmitted viruses. consumers in wealthier economies might be receptive to microbicides costing $ - or even more per dose; however, consumers in lesser-developed economies might find $ - /dose to be prohibitive. hence, absent subsidies, there exists a continuing need to lower cogs for apis such as griffithsin. we can conclude that a cogs of <$ . /dose of griffithsin ds as determined in this study, and an estimated user cost of $ - /dose, might enable at least some simpler formulations of the drug (e.g., rinses or enemas) to be economically marketed. for more complex formulations and delivery systems, or for higher doses of the drug, lower cogs for bulk griffithsin would be desirable. the environmental assessment of the plant-based production of griffithsin indicates low impact, particularly if the plant nutrient solutions are recycled in a hydroponic system and if waste streams containing tmv are treated in a biowaste heat or chemical treatment process. the assessment method used, although semi-quantitative, utilizes mass input and output stream data generated by superpro, along with independent assessment of compound toxicity and/or environmental impact (for example using material safety data sheet information), and allows comparison between alternative production strategies, process configurations or chemical components used in the manufacturing process. our low environmental impact assessment for plant-based manufacturing should compare favorably with fermentationbased approaches to producing griffithsin (giomarelli et al., ) . in the latter, the complexities of purification suggest less efficient utilization of materials and higher disposal volumes, although a side-by-side environmental analysis between the two platforms was not conducted in this study. upstream, griffithsin expression rates were based on empirical findings using tmv whole virion as the expression vector, which can achieve typically . - . g griffithsin/kg plant biomass (fuqua et al., a ). an average pilot-scale expression rate of . g/kg was used in our model (fuqua et al., b) . although this expression level is quite good for tmv, higher griffithsin expression levels can be achieved with different technology. for example, nomad bioscience gmbh (halle, germany) has achieved griffithsin expression in n. benthamiana exceeding . g griffithsin/kg fw biomass using nomadic tm agrobacterial vectors applied to plants either through vacuum infiltration or agrospray (hahn et al., ) , albeit these results were obtained in small-scale studies. the utilization of such an induction process instead of tmv virions could further improve process economics. for example, even with the same recovery efficiency of % assumed in the current model, the output of griffithsin at the higher expression level would be . g api/kg plant material, instead of the current . g/kg; this represents more than . times the modeled output of protein per kg biomass. under such conditions, the costliest parts of the current process, namely biomass production and upstream procedures, would be lowered by the reduced biomass needs to produce the required kg/year of api. although a full analysis of the cost of agrobacterial inoculation for griffithsin production needs to be conducted, we know from similar analyses (e.g., nandi et al., ) that economics can be favorably impacted by higher expression efficiencies. we can therefore envision that by using a more efficient induction process the per-dose production cost could be less than the current $ . . still other gene expression methods can be considered, including using transgenic plants expressing griffithsin either in constitutive or inducible systems (werner et al., ; gleba et al., ) , which could also lead to higher api accumulation in host plant biomass and potentially lower cogs . increasing expression yield upstream might shift costs to downstream operations to handle process streams with higher concentrations of api. definition of the comparative cost benefits of these improvements relative to the current process modeled awaits a subsequent evaluation. from a process standpoint, improvements in the efficiency of lighting technologies and/or incorporating solar panels would reduce upstream utilities costs, one of the major contributors to the upstream operating costs. improving hydroponic nutrient utilization through recycling and minimizing runoff in the simulation model will reduce raw material costs as well as aqueous waste disposal costs, thereby reducing the cogs. in the downstream portion of the process consumables play a major role, particularly dead-end filters and plate-andframe filters; if these could be replaced with tangential flow filtration systems that utilize reusable, cleanable ceramic filters, downstream operating costs could be further reduced. at the time of this writing, such systems were being considered and their impact on griffithsin cogs will be the subject of a future analysis. aa, lj, gk, ks, sn, and km contributed to the conception and design of the study. aa, lj, gk, and ks conducted initial modeling calculations, provided information for supplementary tables - , organized preliminary results and prepared an initial report on the findings. dt, jf, and km provided additional data inputs and further refined the scope of the model. km developed the final superpro model and tea results. dt wrote initial and final drafts of the manuscript, including the cost-of-use analysis and consumer price sensitivity discussion, with primary editorial input from jf, kp and km. kp provided critical reading of the manuscript. all authors contributed to manuscript revision, read and approved the submitted version. the participation of aa, lj, gk, ks, sn, and km in this project was supported by the university of california, davis; however, the model developed, results presented, and outcomes of this study are the personal views of independent authors. the participation of dt, jf, and kp was supported in part by us niaid u program project grant no. u ai and funds from the helmsley charitable trust. activity of and effect of subcutaneous treatment with the broad-spectrum antiviral lectin griffithsin in two laboratory rodent models rapid, high-yield production in plants of individualized idiotype vaccines for non-hodgkin's lymphoma environmental assessment in early process development biodistribution, pharmacokinetics, and blood compatibility of native and pegylated tobacco mosaic virus nano-rods and -spheres in mice women's interest in vaginal microbicides production of antibodies in plants: status after twenty years process simulation for recombinant protein production: cost estimation and sensitivity analysis for heparinase i expressed in escherichia coli process and economic evaluation of the extraction and purification of recombinant beta-glucuronidase from transgenic corn bulk production of the antiviral lectin griffithsin improving the large scale purification of the hiv microbicide, griffithsin recombinant production of anti-hiv protein, griffithsin, by auto-induction in a fermentor culture plant viral vectors for delivery by agrobacterium a novel and fully scalable agrobacterium spray-based process for manufacturing cellulases and other cost-sensitive proteins in plants commercial-scale biotherapeutics manufacturing facility for plant-made pharmaceuticals the isolation of plant viruses by means of "simple" coacervates. virology plant-produced idiotype vaccines for the treatment of non-hodgkin's lymphoma: safety and immunogenicity in a phase i clinical study griffithsin has antiviral activity against hepatitis c virus isolation and characterization of griffithsin, a novel hiv-inactivating protein, from the red alga griffithsia sp techno-economic analysis of a transient plantbased platform for monoclonal antibody production griffithsin protects mice from genital herpes by preventing cell-to-cell spread broad-spectrum in vitro activity and in vivo efficacy of the antiviral protein griffithsin against emerging viruses of the family coronaviridae scaleable manufacture of hiv- entry inhibitor griffithsin and validation of its safety and efficacy as a topical microbicide component how do i get condoms? how much do condoms cost making an ally from an enemy: plant virology and the new agriculture production of pharmaceutical-grade recombinant aprotinin and a monoclonal antibody product using plant-based transient expression systems evaluating the cost-effectiveness of pre-exposure prophylaxis (prep) and its impact on hiv- transmission in south africa heterologous sequences greatly affect foreign gene expression in tobacco mosaic virus-based vectors the flow of money through the pharmaceutical distribution system plant-based vaccines cost-effectiveness of tenofovir gel in urban south africa: model projections of hiv impact and threshold product prices evolution of plant-made pharmaceuticals safety of plant-made pharmaceuticals: product development and regulatory considerations based on case studies of two autologous human cancer vaccines manufacturing economics of plantmade biologics: case studies in therapeutic and industrial enzymes the cost-effectiveness of pre-exposure prophylaxis for hiv infection in south african women high-level recombinant protein expression in transgenic plants by using a double-inducible viral vector economic and environmental assessment of the purification of alpha- -antitrypsin from transgenic plant cell suspension cultures the authors thank trena tusé of intrucept biomedicine llc for critical reading of the manuscript and for document formatting in compliance with editorial guidelines.superpro designer is a trademark of intelligen inc. bpg, capto, and sepharose are trademarks of ge healthcare limited. sartopore is a trademark of sartorius-stedim biotech gmbh. truvada is a trademark of gilead sciences, inc. the supplementary material for this article can be found online at: https://www.frontiersin.org/articles/ . /fbioe. . /full#supplementary-material key: cord- - ryhc authors: kettani, maryème; bandelier, philippe title: techno-economic assessment of solar energy coupling with large-scale desalination plant: the case of morocco date: - - journal: desalination doi: . /j.desal. . sha: doc_id: cord_uid: ryhc this paper examines the cost competitiveness of an extra-large-scale ( , m( )/d) solar-powered desalination, taking as a case study the chtouka ait baha plant in morocco. it assesses the conditions at which solar photovoltaics (pv) and concentrated solar power (csp) would be competitive with a grid (mainly fossil) driven desalination plant for the reference year and by . the paper considers also a scenario where battery storage complements pv power generation. to conduct the analysis, a simple model of water cost calculation is built. second, the cost related to energy consumption is calculated for different power supply options to evaluate the impact of energy provision cost on the final cost of water. the first main result of this paper is that desalinated water can be obtained at an acceptable cost of around $/m( ). the second one is that pv without storage remains the cheapest power supply option today and by . storage based solution appears less competitive today but can be more attractive in a framework of increasing electricity grid prices and higher flexibility requirements in the future. the paper gives recommendations regarding the implication of different technology choices in the framework of the future moroccan energy system. having access to safe drinking water is one of the human basic rights that are still unequally distributed on earth. still out of persons lack access to drinking water services in the world [ ] . these "left behind" people are a prioritized target of the agenda for sustainable development which aims at achieving global access to water and sanitation for all human, especially in the highly water stressed countries [ ] . in particular, the middle east and north africa (mena) region figures on the top of the very highly stressed zones in the world, due to water supply and demand imbalance [ ] . driven by the joint effect of increased urbanization, population growth, industrialization and climate change, this gap may further increase in the coming decades. it is expected that unmet water as percentage of total demand in the region will increase from % in the period - to % in - [ ] . bridging the water gap in the mena region would have an average annual cost estimated to reach more than us$ billion during - [ ] . this cost largely varies across the countries and is influenced by the climate change projection scenario that is considered (wet, dry). the distribution of adaptation costs would appear to be very skewed as > % of the total cost needs would be covered by three countries, namely iraq ( %), saudi arabia ( %) and morocco ( %) [ ] . to overcome present and future water shortage challenges in the mena region, many alternatives can be experienced such as water productivity upgrade (especially in agriculture), water demand management and water supply diversification. besides traditional water supply sources such as lakes, rivers or groundwater reservoirs, seawater and brackish water desalination can be a valuable option in many countries to diversify the water mix. desalination is a water treatment technique that consists in removing salts from water. two main techniques are commonly used to desalt water. historically, distillation was the leading process in the desalination market but in the recent decade membrane based processes largely overcame distillation covering now > % of current installed capacity [ ] . up to , > million m /d of commissioned desalination capacity in the world (supplied by > , plants) covered a part of water needs for around million people [ ] . between and , global installed capacity of https://doi.org/ . /j.desal. . received april ; received in revised form june ; accepted june desalination units grew at almost % per year [ ] . though its long history and its upward trajectory in terms of installed capacity, desalination supplies only % of drinking water in the world [ ] . indeed, the high energy requirements to power a desalination unit increases the final cost of desalinated water, which may negatively affect the economic attractiveness of desalination projects. the type (electrical or thermal) and amount of specific energy consumption vary according to the considered desalination technique. for example sea water membrane separation process requires between and kwh/m whereas for distillation processes, both thermal ( - kwh/m ) and electrical energy ( . - kwh/m ) are required [ ] . aside from the amount of energy consumed, the price of the purchased energy to power the desalination unit has also a significant impact on the final cost of desalinated water. the calculation of the cost of water depends on several factors and assumptions which makes the fair comparison of results a hard task. however, a comprehensive survey gives interesting ranges of the estimated cost of water reported in the literature [ ] . for large sea water ro plants with a capacity ranging between , m /d and , m /d, the reported cost of water varies between . $/m and . $/m , whereas for medium plants with a capacity between , m /d and , m /d, the cost of water ranges between . $/m and . $/m . for small plants of m /d to m /d, the cost of water ranges between . $/m and . $/m . for electrodialysis plants, the cost of water ranges between . $/m and . $/m [ ] . regarding distillation techniques, the cost of water for multistage flash distillation plants with a production capacity between , m /d and , m /d varies between . $/m and . $/m . for multi-effect distillation plants with a capacity higher than , m /d, the cost of water ranges between . $/m and . $/m . however, for lower capacities ( , m /d to , m /d), the reported cost is higher ranging between . $/m and . $/m . contrarily to previous desalination techniques that rely on large capacities, vapor compressor usually have small capacities and higher costs. for example, the reported cost of water ranges between $/m and . $/m for a capacity around m /d [ ] . available information reveals that % of total desalinated water around the world is powered by fossil energies which are harmful to the environment [ ] . the use of renewable energy sources (res) such as solar, wind, wave, geothermal can therefore bring new opportunities for a sustainable growth of large-scale cost effective renewable powered desalination systems. a variety of coupling options exists between renewable energy sources and desalination techniques [ ] [ ] [ ] [ ] . these coupling options are shown to be technologically and economically feasible but have today various levels of technological maturity from basic research to the application stage [ ] . among all renewable driven desalination plants, solar energy is the dominant power source. between and , % of the res desalination plants installed in the world were powered by solar and only % by wind. almost one third of res powered desalination plants are reverse osmosis (ro) units powered by photovoltaics (pv), making this configuration (pv/ro) the most dominating option currently, especially for small scale installations [ ] . a large part of the literature has investigated the potential of small scale ro systems driven by pv in different locations such as greece, italy, spain, saudi arabia or egypt either as part of theoretical studies or pilot experiments [ ] . for example, based on a techno-economic feasibility study of a m /d ro plant in abu dhabi, it was found that a fully solar driven unit without batteries with a cost of water of . $/m would be competitive compared with a fully diesel powered unit ( . $/m ) [ ] . in general, when solar resources are abundant, pv-ro is found to be more competitive than diesel powered units [ , ] . however one limit of pv-ro design is the intermittency of solar pv generation that can be smoothed using storage device to allow a h steady state operation mode of the pv-ro desalination unit [ ] . introducing lead-acid batteries to a pv system is found to be the most cost competitive solution to power a ro unit compared to a diesel generator or a diesel/wind power supply [ ] . similar conclusions are drawn in [ ] with a % cost reduction when batteries are integrated to remote pv-ro systems. however, batteries have limited lifetime which could be even reduced further in hot climate, requiring regular replacement and thus increasing the overall desalination operation cost. concentrated solar power (csp) is being more and more considered for powering desalination units. the literature remains however much less abundant as csp technology for large-scale solar developments has only started to gain interest in the last decade. a roadmap for short medium and long term large scale csp powered desalination in the mena region [ ] revealed that a potential of billion m per year by could be expected for the whole mena region among which million m per year only for morocco. egypt, saudi arabia, libya, syria and yemen are expected to dominate the csp desalination market in the coming decades [ ] . however, despite this huge potential, electricity from csp is still high which limits the large development of this solar technology. conducting a technical and economic feasibility of distillation process integration to a mw elec csp plant in namibia, [ ] concluded that water and electricity cogeneration could be an interesting option in the country, but is still less attractive than a grid powered ro design. in the case of a csp+ distillation plant, the cost of desalinated water ranges between . $/m and . $/m depending on the brine temperature (respectively °c and °c). the high capital cost of central csp receiver plants, and the cost of inland seawater pumping would be the main limiting factors affecting the economics of such a design. as csp electricity generation is still high, the introduction of high feed-in tariffs for csp generation makes large solar fields with large thermal storage systems becoming economically interesting compared to conventional designs. to breakeven with conventional distillation cogeneration and for a , m /d capacity plant, feed-in tariffs for csp generation would have to reach . $/kwh and . $/kwh respectively for jordan and israel, whereas for ro-csp systems, they would be at a minimum of . $/kwh and . $/kwh respectively [ ] . the use of csp powered desalination plants with heat storage would provide an opportunity to the large-scale deployment of csp technologies while supplying sustainable fresh water in arid areas (chile, namibia, australia, and mena region). compared to a pv+ battery system which generally needs high investment because of expensive storage requirements, cogeneration of desalinated water and electricity using csp allows the use of a relative cheap thermal energy [ ] . even if critical factors such as the csp plant location (desert or coast) and design may prevent this development, future cost reduction in csp technologies and the wish to encourage sustainability of industrial processes could boost large-scale csp powered desalination plants. several [ ] is expected to contribute to achieve the moroccan desalination objectives of , m /d [ ] and even . mm /d according to another source [ ] . based on the chtouka ait baha specifications as a case study, this paper examines the cost competitiveness of an extra-large scale ( , m /d) solar powered desalination. it provides an analysis of the conditions for which solar photovoltaics (pv) and concentrated solar power (csp) would be competitive with a grid (mainly fossil) driven seawater reverse osmosis (swro) desalination plant. the paper considers also a scenario where battery storage complements pv power generation. the objectives of this paper are threefold. first it analyses the economics of large-scale solar ro plants under present cost assumptions and by . second, it provides estimates of solar electricity production cost in the country and assesses the economics of lead-acid battery storage in the moroccan context. third, it discusses the implication of solar technology choice for desalination purpose in the framework of the moroccan energy policy. the paper is organized as follows: in section , the main chtouka ait baha plant characteristics are described and an overview of the methodology adopted in this paper is given. in section , results are presented and discussed. section provides a prospective analysis on the evolution of the cost of water by . main conclusions of this study are summarized in section . desalination in morocco constitutes a market segment with high potential for solar electricity deployment. in order to evaluate this potential rigorously, representative data of solar irradiation, desalinated water capacity and associated electricity needs are required. in this work, localization and specifications of the chtouka ait baha desalination plant (technology and capacity) are chosen to conduct the study. in order to fully assess the competitiveness of solar driven desalination, two steps are followed. first, a simple model that assesses the final cost of desalinated water is computed. second, the cost related to energy consumption is calculated for different power supply options to assess the impact of energy provision on the final cost of water. this section briefly describes the chtouka ait baha project, introduces the cost of water model (detailed in the appendix a), gives an overview of the methodology used to calculate the cost of solar electricity and the cost of storage, and finally displays the four main scenarios of power supply retained in this paper. the chtouka ait baha seawater desalination plant is based on ro technology and is designed for a production capacity of , m m /d from which , m /d are dedicated to fresh water needs and , m /d to irrigation (area of , ha) [ ] . aman el baraka, which is the moroccan subsidiary of the project concession company abengoa, will build, operate and maintain the desalination plant over a period of years. first construction works started in july . the plant will be located at m from agadir coast at an altitude of m. the project gathers together different local and foreign partners: ministry of economy and finance, ministry of agriculture, maritime fisheries, rural development and water and forests, national energy and water utility (onee), spanish companies (abengoa sa company, abengoa water international company, seda, sa) and the moroccan bmce bank, who all signed a convention agreement in . the project cost is about $m of which $m are dedicated to the freshwater production. onee is expected to devote additional investments of $m for the construction of km of pipelines, a , m drinking water tank, the installation of high voltage power lines over km from the tiznit source station which is already connected to the noor ouarzazate solar complex and the construction of pumping stations and loading tanks [ ] . the specifications of the chtouka ait baha project that have been used as input data for this work are the following (table ) : a simplified model is used to determine the daily production and the water cost of a reverse osmosis (ro) desalination plant powered by any energy source. due to the relatively quick dynamic of such a system, simulation over a given day is calculated as a succession of steady states, hour per hour. for the purpose of this study, ro is supposed to be operated h a day. when solar resource is not available or insufficient, energy storage (electricity or heat for csp) or the grid takes over. developed with microsoft excel™, it includes the following calculation modules: validation of the ro model has been carried out using experimental results from literature and a benchmark with different professional tools (deep, iaea, wave) available shows that performance and cost are predicted with a deviation better than %. in most reviewed paper on desalination literature, the cost of desalinated water is computed using the "amortization factor" or "annualized life cycle cost method [ ] [ ] [ ] . this calculation method is generally known as simplified cost of water (scow) (eq. ( )) and is based on initial capital costs annualization using an amortization factor [ ] (eq. ( )): investment costs (capex) include four major types of expenditure: specific capital cost (pre-treatment, desalination system, post-treatment and other auxiliary equipment), water distribution and storage equipment and finally other capital costs: permitting, land, construction, design and engineering [ , ] . operating costs (opex) include spare parts such as membrane replacement, maintenance cost, chemicals, insurance and labor, and energy cost. the cost of desalinated water is very sensitive to assumptions, input data and methodologies, which challenges the suitable perception of water cost across studies and experiments [ ] , calling for a clear understanding of desalination cost determinants. such a clarification, in the case of morocco, will be one of the major aims of this paper. therefore, a significant part of the paper is devoted to a thorough description and assessment of both the costs of energy and capital (including technological aspects). it should be reminded that the calculation of operation costs (opex) in the model excludes the cost of energy since it is the principal parameter of the study. indeed, as the paper mainly focused on the impact of different power supply options on the competitivity of a large scale desalination plant, the specific consumption of the swro (kwh/m ) is first calculated, then the power required (kw) and finally the corresponding energy (kwh) over a given period (a year) are computed. since the plant operates at full capacity h a day, days a year ( % of availability factor), these values are fixed. depending on the power supply option considered, the only cost that changes is the one related to energy consumption. the energy consumption includes the energy consumption of the reverse osmosis unit (high-pressure pumps and booster downstream of the dres) but also that of the seawater supply pumps, taking into account the pressure losses in the pipes and in the seawater pre-treatment. the back pressures of the brine and freshwater discharges produced (thus including post-treatment) are also taken into account. as mentioned previously, capex and opex (excluding energy) are derived from the cost of water model and are fixed whatever the power supply considered. what varies across all power options is related to the cost of electricity consumption which is obtained by multiplying the quantity of electricity consumed (kwh) by the cost of electricity. even if criticized by some authors, the levelized cost of electricity (lcoe) is a commonly used metric to assess the economic feasibility of an electricity production project [ ] . it corresponds to the present value of the project costs (in $) per unit of electricity generated (in kwh) over the plant lifecycle [ , ] (eq. ( )). for pv systems, capex includes the cost of modules with the balance of systems (mounting structure, electrical connections, power electronics, tracking systems, etc.) as well as the land cost. opex include insurance costs and maintenance, regular cleaning and performance monitoring of the device's costs. for csp plants, capex includes the solar field (mirrors and receivers), the thermal storage and the power block. opex include thermal storage and solar field maintenance (including mirror's periodic cleaning). today, pv is the leading solar technology in terms of installed capacity and is the most cost competitive one. driven by continuous technological improvements and large-scale deployment, pv lcoe has dropped in average from . $/kwh to . $/kwh between and , following the irena database statistics [ , ] . in the same period, average csp lcoe has decreased at a slower rate from . $/kwh to . $/kwh. long term forecasts and roadmaps suggest a further decrease in the cost of solar technologies [ , ] . by , expected lcoe for solar technologies would reach . $/kwh for pv and . $/kwh for csp systems [ ] . these values correspond to the "aggressive cost targets" of the u.s. department of energy (doe). when pv systems include storage devices, the cost of stored electricity can be calculated using the levelized cost of storage (lcos) metric (eq. ( )) which is derived from the lcoe method. the lcos is defined as the total lifetime cost of the investment in electricity storage technology divided by its cumulative delivered electricity [ ] . capex is added to the annual cost of storage a t discounted over the lifetime of the project. this sum is then divided by the discounted sum of annual electricity output w outt . a t is composed of operation expenditures opex, reinvestments in storage components capex ret ,and the cost of electricity supply determined by the price of electricity c el times the amount of electricity delivered to the battery w in (eq. ( )). a t : annual cost of storage in $ in year t, capex ret : reinvestment in storage system components in year t, c el : price of electricity supply, w in : annual electricity input. several studies assessed the lcos of different energy storage technologies [ , ] . for example, [ ] found a lcos for lead-acid batteries of . $/kwh whereas the lcos for li-ion batteries is found to vary between . $/kwh and . $/kwh by lazard [ ] . these costs are expected to decrease strongly in the next decade due to technological developments and decreasing capex. by , the lcos of li-ion batteries at cycles per year would vary between . $/kwh and . $/kwh whereas the lcos for lead acid batteries is expected to range between . $/kwh . $/kwh [ ] . four case studies of power supply have been considered in this work. case supposes that the desalination plant is powered by the national grid. case assumes that the desalination energy requirements are satisfied jointly by the grid and by a nearly located pv power plant. case considers that the desalination plant is powered by the grid, the nearly located pv power plant and lead-acid batteries. case assumes that the desalination plant is powered by the grid and an inland csp plant located in the desert. more details of each case are given below: ▪ case is based on a fully "grid "power supply. the plant is powered by the national grid and purchases its electricity at present utility tariffs which are hourly dependent and taken from the onee website [ ] . ▪ in case , part of the grid power supply is replaced by solar electricity coming from a pv plant located on the same site as the desalination plant. to design the pv plant, the chosen area of solar panels (m ) is selected so as the maximum solar electricity production equals the nominal maximum power of the desalination plant. the pv plant design is based on the day of the year where solar irradiation is the maximum over the year. under this design, there is no excess pv production in the year and the grid is supposed to complement solar electricity supply whenever needed. ▪ in case , lead-acid batteries are introduced for peak shaving. as the moroccan load curve shows a high peak load during evening hours ( pm to pm in the winter and pm to pm in the summer), using batteries can increase the desalination plant selfconsumption and contribute to maintaining grid stability. moreover, electricity tariffs are the highest during this period of the day. batteries are designed only to power the plant during these peak hours. to determine the additional area of solar panels required to feed the batteries, the marginal area needed to supply the five peak hours consumption is added to the pv panel area calculated in case . cost data of batteries are taken from the relevant literature. ▪ in case , a future unit of the noor csp complex (including thermal storage) is supposed to be entirely dedicated to the power the desalination unit. it is assumed that the csp unit powers the ro plant from am to pm ( h of operation) in average, whereas the grid powers the desalination plant for the remaining hours of the day. in this case, the cost of csp power generation is based on present csp lcoe in morocco. all cases are compared through the sum of all variable charges related to the different power supply sources used to power the desalination plant. for each case, an average cost (variable) of electricity is deduced and used as an input to calculate the final cost of desalinated water depending of the power configuration adopted. (table ). the modelling approach developed aims at analysing the interrelations between the cost of desalinated water and the cost of energy in a large-scale desalination with % availability. in order to evaluate rigorously the impact of the technology on the desalinated water cost, it is mandatory to have an accurate description of the technical choices made for building the desalination plant. as these details are not public, technical characteristics of a virtual desalination plant that could fulfil the global production specification of the chtouka ait baha plant has been determined and sized (table ) using the model described in detail in appendix a. these technical components and their dimensioning have been used to determine the capex for building the desalination plant and its opex, excluding energy consumption cost, that is studied in the next section. calculated data are gathered in table . to calculate the contribution of capex ($/m ) to the lcow, the amortization rate α is calculated and found to be . using eq. ( ), and assuming a discounting rate of %, an amortization time of years and % plant availability (a). using eq. ( ) below, the capex contribution in ($/m ) is calculated to reach . $/m . excluding the cost of energy, the contribution of opex to the lcow is estimated to be . $/m . total (daily) capex: total capex expressed in a daily base in $/(m /d), α: amortization rate, a: plant availability in %. thus, the capex+opex cost (without energy) of the optimized plant appears to be . $/m . the fixed costs (capital and most part of the operation costs) represent about half of the water production costs. this implies that it is very important to use the desalination plant continuously over the year, if it is kept in mind that there is always a strong demand for water. this result is the base of all the following computations where the desalination plant is supposed to run all the yearlong. besides, knowing that energy needs are about kwh per m , and that the public cost of electricity is close to . $/kwh (in order of magnitude), it can be concluded first of all that the cost of energy is a very important parameter (representing roughly half of the cost) and second of all, that the production cost of water will be about $/m , which is a very competitive/interesting result, in line with the range found in the literature. such a good result is obtained because of two mains causes: the high performance of the reference technology (which is fully up to date) and the high rate of average sun irradiation in morocco. as the final cost of desalinated water depends heavily on the cost of electricity, the following part of the paper will focus on this item. several solar based power supply configurations considered in this paper and respective costs are calculated and discussed in the following section. this section mainly focuses on the cost of electricity generation from the four different power supply options retained in this paper. in morocco, electricity tariffs are administrated and regulated by the government to protect the economic competitiveness of local industries and the purchasing power of low-income households. under the current pricing scheme, subsidies and associated electricity tariffs depend on the consumer category (urban/rural household, energy-intensive industries…) and the daily time slot of electricity consumption [ ] . for the medium voltage (case of the desalination plant), subsidies and associated electricity tariffs depend on the season (summer, winter) and the hour considered: i.e. peak, full, and off-peak. as displayed in table , the off-peak tariff is . times the full hour tariff whereas the peak tariff is . times higher than "the full hours" tariff. in addition to variable electricity tariffs, an annual power charge of $/kva/y is applied (tariffs include a % added value tax applied for electricity products). based on the above pricing and since the electricity consumed by the desalination plant is supposed to be "flat" (i.e. identical at each hour at a level of mwh), the average electricity cost is assumed to be around . $/kwh (corresponding to the onee "full hours" tariff). the average electricity cost is simply obtained by dividing the sum of hourly grid charges in m$ (hourly consumption * hourly tariff) with annual desalination consumption (kwh). considering that the city of agadir (near chtouka ait baha) benefits from excellent solar irradiation and that the cost of solar pv modules decreases continuously, it is worth investigating the economic feasibility of a pv + grid configuration to power the desalination plant. to estimate the average yearly energy production of a pv system connected to the electricity grid in agadir, the pvgis online tool was used for input solar irradiation hourly data. the calculations consider solar irradiation, and type of pv module. the modules are supposed to be mounted on a free-standing rack, without any tracking system. as it is proved to be the most cost-effective solution for solar power generation in morocco [ ] , polycrystalline silicon technology is considered. in pvgis, the optimum slope and orientation that maximizes the yearly energy production based on typical meteorological year data ( ) are automatically generated. table displays input data for the pv plant design. as the module efficiency depends on irradiation and temperature, an average efficiency module value is selected. the performance ratio takes into account all system loss. as polycrystalline silicon is a very robust technology, it is assumed that availability is also included in the performance ratio. the typical meteorological year (tmy) database from pvgis displays annual representative solar data for the chosen location based on selected data over a long period ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) . this annual data set is used to calculate pv electricity production based on irradiation data. according to the download tmy for agadir location, the yearly solar irradiation reaches kwh/m /y which corresponds to kwh/m /y of pv electricity output ( × . × . ). to size the pv system a first reference case is considered. it supposes that all the pv output is self-consumed by the desalination plant (no excess pv production during the year). the area of pv modules is determined such as the maximum pv power ( . kw/m ) during the solar peak hour ( april at noon) equals the nominal power of the desalination plant ( mw) . the required area is about , m ( / . ) which implies a surface ratio of . m /kwp. this represent less than a square km, with the current densities of panels in solar farms in the maghreb. even if this surface is not negligible, it is no difficulty to find it near agadir. under this design, all the desalination load is satisfied by pv on the solar peak hour, whereas for the remaining hours of the year, depending on the pv output, the desalination electricity requirements are partially or fully met by the grid (e.g. at night). fig. displays an example of the respective contributions of pv and grid power supply to satisfy the desalination plant power consumption on representative days of high and low irradiance days. with this first hypothesis, the ratio between annual electricity consumed from pv and total desalination electricity needs is around %. this ratio is quite low, which shows that this case appears not to be far from the case , with a "marginal" addition of solar energy. the ratio of pv consumption/pv production is exactly % as all the pv production is used. to assess the cost of pv production, pv lcoe is computed based on eqs. ( ), ( ) [ ] , ( ) and input data displayed in table . indeed, the annual electricity output m elec is calculated by integrating the instantaneous pv production of eq. ( ) over one year, corrected with eq. ( ) to take into account ageing. m elec is then introduced into eq. ( ) to obtain the lcoe over the entire lifetime of the pv farm. doing so all contributions to electricity cost such as maintenance or amortization are included. pv outputt : electricity output in time t, ghi t : global horizontal irradiance in t in kwh/m , pv eff : efficiency of polycrystalline silicon modules in %, pr: performance ratio in %, a: total solar panels area (m ). ageing can be modelled very simply by eq.( ) as following: the calculated pv lcoe is found to reach . $/kwh. in order to capture the change in lcoe caused by relative percentage change in input parameters, a sensitivity diagram was constructed (fig. ) . a variation of ± % on the reference value of capex, opex, pv yield, output degradation and lifetime (as displayed in table ) has been conducted. among the technical parameters, the pv yield is the most important factor that influences the lcoe. the lower the pv yield, the higher the lcoe. capex is also a determining factor in the lcoe calculation. the higher the capex, the higher the lcoe. this result reveals the importance of capex cost reductions and technological improvements to achieve lower lcoe. output degradation, lifetime and opex variation seem to have a very limited impact on the lcoe within a variation range of − % + % compared to the reference value. the chosen value of the discount rate is also a determining factor that influences the value of the calculated lcoe. the reference discount rate used in this study is assumed to be %, which is a relevant level for a strategic investment largely driven by the national administration. for a minimum discount rate of % (close to a fully public rate), the lcoe would decrease by % whereas for a discount rate of % (more private rate), the lcoe would increase by almost %. the calculated pv lcoe ( . $/kwh) represents % of the average electricity price of the first case of this study, namely the fully grid power supply option. in case , using pv by a share of % would drop the average cost of electricity from . $/kwh in the case of a % grid option to . $/kwh (− %) in the present case of hybrid pv + grid power supply. this comes from a cost of solar energy which is half of the average cost of the national grid in average. it is worth noting that the cost of pv electricity ( . $/kwh) is low compared to the values classically found in the literature [ ] . such a value is possible because of the high solar irradiation in morocco, with advanced hardware. in this study, local pv production competes with electricity of the national grid, the price of which, even if subsidized, includes significant transport costs. the first conclusion that can be drawn is that there is an economical interest in linking the desalination plant to a solar farm. in addition, this case , as described above, offers the best yield from each solar panel, because all the production is used. the drawback is that the share of solar electricity remains low ( %). in fact, this case does not represent the minimum cost without storage. if part of the solar production is curtailed, the share of solar electricity available to power the desalination plant could significantly increase. however, this configuration would lead to an additional charge to be "paid" for this larger share of solar pv production, as part of the production has to be curtailed and wasted. fig. displays the impact of doubling and tripling the reference pv surface on the average electricity cost (which is the weighted cost due to pv and grid electricity). the least cost option appears to be close to the one where the pv surface is doubled compared to the reference case. in addition, the minimum seems to be rather flat and it can be considered to be close to the optimum. consequently, in the following and for economical comparisons, the retained pv surface for case is the one resulting in the installation of , m of pv panels. as the pv area is doubled, more solar electricity is produced. therefore, the pv electricity contribution to power the desalination plant illustrated by the ratio pv electricity consumption/desalination total electricity demand also increases. doubling the pv surface results in an increase in this ratio from % to %. however, the ratio of pv consumption/pv production decreases from % to % since the surplus pv production is curtailed and wasted (no storage considered, nor selling to the national grid). when tripling the initial pv surface, the contribution of pv electricity to the desalination electricity demand increases to %, and the ratio of pv consumption/pv production decreases further to % (same hypothesis). table summarizes the results of this sensitivity analysis. one of the main results is that the use of pv (without storage) allows to decrease the total cost of electricity by around % ( . $/kwh instead of . $/kwh). a remaining question, with the important share of curtailed electricity quoted above, would be to implement a module of power electronics and sell power to the national grid. however, selling the peak production to the grid, in a future where morocco plans to rely more and more on solar energy (with a significant share of pv) does not look a sustainable option: the future value of pv in peak production hours could decrease rapidly over time. another relevant option for avoiding wasting pv electricity is to consider the storage of surplus electricity. this option is studied in case . this case supposes the introduction of lead-acid batteries (lower capex than li-ion; heavier, but without consequences for the purpose of this paper) to meet the desalination load during the evening peak hours when electricity tariffs are the highest in morocco and when there is no sun in winter. starting from the case (where no excess pv is produced with the installation of mwp), an additional area of pv panels is introduced to power the five evening hours demand of the desalination unit. batteries are designed for behind-the-meter peak shaving especially when peak demand occurs in the evening as for morocco. excess pv electricity produced by these additional pv panels during sunny hours is stored in the batteries and then reused to power the desalination plant using an efficiency ratio of around % (see hereafter). during off-peak and full load hours, the grid completes the pv output. based on available data and in order to give global estimates on the effect of introducing batteries on the desalination electricity cost, the design of batteries is based on a simple approach that considers a typical spring day ( april ) for the calculation even if the resulting size of batteries is not expected to be the exact optimal one (as for pv sizing). in winter days, solar irradiance is the lowest and the energy produced by pv panels is low, so batteries are not (fully) charged in this period of the year. in summer days, energy produced by the pv plant is the highest resulting in excess energy production that need to be (partially) stored in batteries. during the summer, days are longer and solar irradiation is higher so there is paradoxically not a strong need to charge batteries. as an intermediate solution, choosing a day of spring (mid-season) appears as a reasonable option to dimension the batteries in order to assess the profitability of introducing batteries to power the desalination. table summarizes basic parameters considered in this case . as the hourly consumption of the desalination plant is identical throughout the year, and based on the reference spring day, the daily electricity that needs to flow out from the batteries e(out) is mwh ( × ). considering a roundtrip efficiency of %, excess pv production that flows in the batteries needs to be . times higher ( × . ), reaching mwh. considering a depth of discharge of % for limiting battery degradation, the size of batteries needs then to be doubled, reaching mwh ( × ). using the bisection method and considering the daily pv surplus (including efficiency losses), it is found that , m additional pv panels are required to satisfy the desalination plant consumption during the additional peak hours (fig. ) . pv direct corresponds to the pv energy produced. e(in) is the surplus energy that flows in the battery and e(out) is the energy that flows out of the battery. it is worth noting that the additional pv surface obtained (based on mid-season) is found to be close to the one calculated previously as the pv surface that lowers the average annual cost of electricity without any storage. results of direct and surplus pv generation are displayed in table based on simulation on the mid-season. the cost of direct and excess pv production is based on the pv lcoe, whereas the cost of electricity stored in batteries and used to power the desalination plant during peak hours is based on the lcos which is computed on the base of eq. ( ) and input data displayed in table . the calculation of the lcos is based on different days (representing various seasons) for battery sizing. the lowest lcos is obtained when designing batteries on the mid-season ( th april). the calculated lcos is found to be equal to . $/kwh (< . $/kwh found in [ ] ). the average electricity cost is found to reach . $/kwh. this cost is calculated as the sum of (pv consumption×lcoe pv), (grid consumption × hourly tariff), and (consumption from batteries ×lcos), divided by the total desalination demand. this result, even if it is not fig. . impact of pv installed capacity on the average electricity cost ($/kwh). case is based on a % grid power supply. in case , the pv surface is varied from mwp (reference case) to mwp (tripling of the reference pv surface). the exact minimum, shows clearly that for a quasi-similar surface of pv, introducing batteries rises the average cost of electricity by almost %. this result suggests that for large scale ro, batteries may not be a relevant option for lowering the cost of power supply. however, this statement might change when taking into account the remuneration for the service batteries provide to the system, but this analysis is out of the scope of this study. the results advocate for future research, based on detailed data, for assessing if there will be room (and when) for batteries, especially for allowing to reduce the net evening demand for electricity ("net" meaning that the apparent load of the desalination plant could be decreased when using the batteries; the question is certainly difficult, because the evening peak demand occurs in winter, when the solar production is the lowest). this question is in fact a broader one, because it addresses the value of storage for the moroccan grid. since its new energy strategy in , morocco has gained experience in csp project development. the first csp plant noor i ( mw, h of molten salt storage) was operational in and produced gwh in [ ] . noor ii ( mw, h of storage) and noor iii ( mw, . h of storage) are expected to be operational soon. fig. shows the expected reduction of the purchasing power agreement (ppa) of these three power plants in morocco [ ] , which is considered as an indicator of the lcoe of csp plants. csp ppas follow a downward trajectory dropping from almost . $/kwh for noor i to . $/kwh for noor ii (both based on parabolic trough which is a commercially proven and the dominant technology in the market with a % market share in ). the lcoe of noor iii (tower technology) is estimated to reach . $/kwh which is slightly higher than noor ii because tower technology is less commercially developed than parabolic trough. however, as tower technology allows operation at higher temperatures and pressures (increasing operational efficiency), significant scale economies are expected from wider tower development in the coming years. as a general trend, the lcoe of both csp technologies are expected to fall down in the coming years. along with the development of csp power plants, this case considers that instead of using batteries for peak shaving, an inland csp plant located in ouarzazate will be dedicated to meet the desalination plant for h of operation. the power plant, with its heat storage capacity, is supposed to operate from am to pm, during "full hours" and "peak hours", where electricity tariffs are the highest and to take advantage of csp storage during evening peak hours. it is thus considered that for h, the desalination plant electricity requirements are met by solar production coming from the inland desert. in this case , the reasonable assumption that the cost of this solar electricity would be equal to noor iii lcoe ( . $/kwh) is made. for this cost of electricity generation, a rough estimation of the cost of electricity transmission from ouarzazate to agadir ( km) is conducted based on technical data for the french flamanville line project [ , ] and data on transmission line cost published by etsap [ ] . table summarizes input data used in the estimation of the cost of transmission. a discount rate of % and a lifetime of years are supposed for the calculation. supposing that the transmission line capacity is mw for a tension of kv (this capacity represents more than the double of the desalination capacity, because for long distance few projects would imply too high costs). it is assumed that the remaining capacity is used -at cost -for linking the ouarzazate production complex to the coastal grid near agadir, and that substations are required. the breakdown of the cost of the transmission line is given as follows: the total cost of the transmission line is therefore roughly estimated at m$ which corresponds to , k $/mw-km. assuming an annual production factor of around %, the transported energy is estimated to reach gwh ( × × . ). transmission losses are fixed at %. the obtained cost is around . $/kwh. the total cost of producing and transporting electricity from ouarzazate to agadir is therefore fixed to . $/kwh ( . + . ). based on the estimated cost of csp generation (including electricity transport) and the onee tariffs, the average cost of electricity in the case of a csp + storage + grid power supply option reaches . $/kwh. this cost is calculated as the sum of csp consumption × csp lcoe (including transport) and grid consumption × hourly tariff divided by the desalination total consumption. based on all the above cases, it is now possible to compare the different possibilities of supplying the desalination plant with solar electricity, in order to assess the relevance of the different power supply options, in the moroccan context. table summarizes the average electricity cost in each of the cases covered in this study. the lower cost option is achieved when pv is combined to the grid without using batteries (around mwp of installed pv), whereas "pv + storage + grid" is found to be the less cost competitive option compared to all the power supply options considered. when comparing solutions with storage, csp + storage + grid is slightly more competitive than pv + storage + grid. but introducing storage whatever the solar technology adopted, is less competitive than a grid-based solution or only a pv+ grid solution. as all the studied cases rely on the grid to fully or partially satisfy the desalination load. a variation of the average grid electricity tariffs would impact the average cost of electricity to the desalination plant. as it is well known that electricity tariffs in morocco are below the cost of electricity generation, it may be expected that if electricity prices were cost reflective, the competitiveness between all cases would change. the estimation of the recovery cost of electricity in morocco will be the subject of a forthcoming paper but the result is important in the framework of this work. in particular, the decision of building a desalination plant such as chtouka ait baha relies to the moroccan administration. thus, this body will have to examine the global economic balance of the different electric supplies also in term of impact on the national budget by considering the "real" cost of electricity production and distribution, which is higher than the present tariffs. all the preceding results are based on a cost of electricity derived from the existing tariffs. all the costs are then based on market prices and the spirit of this assessment is to choose the lower cost electricity mix for a private firm investing in the desalination plant. this can be discussed since energy decision largely refers to the choice of the government. indeed, the moroccan government is the major decision maker, the main ruler and main investor (even if investments are done by private firms, but the capital of which is largely state owned) is the state itself. indeed, the onee tariffs do not take fully account of the real cost of electricity production, as they can be viewed from the state (because of the presence of subsidies, and because of external costs). for instance, if these "hidden" costs are high, a decision based on a myopic analysis could favour an electric mix which implies important losses for the government (subsidies) or the collectivity (external costs). because of a limited public data access, the estimation of the recovery cost of electricity in morocco is a difficult task. though, the world bank recent report on mena utilities performance revealed that the ratio of total energy sales out of total costs for onee is around % in [ ] . this ratio is considered as a financial indicator of utilities performance. the ratio is constructed as revenues related to energy consumption and service ($) / (opex + depreciation of fixed assets+ other depreciation and provisions -net interest). in the absence of detailed data, a reasonable assumption is to consider that electricity subsidies vary between % and % of the generation cost, which supposes that the average electricity tariff would range between . $/kwh and . $/kwh without considering any additional cost for co emissions. if a carbon tax of $/tco is applied to the mtco generated in (data applied to because of data availability) by the power sector to generate the twh of electricity generation in morocco ( data) [ ] , an additional cost of . $/kwh would lead to higher tariffs, varying between . $/kwh and . $/kwh. fig. displays the variation of the electricity price based for three different subsidy rates: %, %, % and % and taking into account the co tax. as shown, electricity prices could increase from . $/kwh up to . $/kwh depending on the taken hypothesis on the rate of electricity subsidies and on the value of the co tax. with electricity subsidies varying between % and % and a $/tco tax, the cost of electricity would range between . $/kwh and . $/kwh resulting in a cost of water between . $/kwh and . $/kwh in the case of a % grid power supply. these costs could be interpreted as the cost of desalinated water from the "public" point of view as compared to the previous costs calculated from a private firm standpoint (not a very significant different option). because of a high incertitude on the real increase rate for which onee tariffs would be cost reflective (real subsidies, future carbon tax), fig. represents the expected increase on the average cost of electricity for different power supply to the desalination plant with an increase in onee tariffs between of % and %. the relative sensitivity to the grid price increase differs across the cases. for a % increase in the grid average electricity price, the average cost of electricity of the four power supply options ranges between + % to + %. obviously the most affected option is the one that relies the most on the grid which the % grid power supply option. an increase of % of grid electricity prices would rise the average cost of electricity for the pv + grid option by %. however, both cases with storage are less sensitive to the variation in electricity prices resulting in an increase of % and % in case of pv + storage + grid and csp + storage + grid respectively. it is worth nothing that if grid electricity prices increase by %, the cost of electricity from a % grid power supply, pv + storage + grid and csp + storage + grid would be almost equal (around . $/kwh) which reveals that the use of solar technologies for desalination is all the more relevant in a context of rising electricity prices which is a very realistic scenario in the moroccan context. even if the pv + grid alternative is more sensitive to the two storage-based solutions, it remains the least cost option even with an increase of % in grid electricity prices. compared to the three other power supply alternatives, the pv + grid option results in an average cost of electricity of . $/kwh which is % lower than the average cost of all other alternatives ( . $/kwh). as stated above, the remuneration from demand curtailment due to storage is not taken into account but if considered, it will probably increase the attractiveness of storage-based solution compared to battery-less power supply alternative especially in the context of increasing electricity tariffs. as stated above, even if electricity remains subsidized in morocco, the level of cost to be taken by the public authorities has to take account both the subsidy costs and the external costs. based on the average electricity tariffs for each of the power supply options calculated previously, fig. displays the total cost of water, distinguishing between the energy cost related part and the non-energy cost part. results reveal that the cost of desalinated water would range between . $/m (pv + grid) and . $/m (pv + storage + grid) with the share of electricity cost in the total cost of water varying between % and % respectively. to summarize, it has been shown that for all the power supply options considered in this paper, the cost of electricity ranges between . $/kwh and . $/kwh which results in a cost of water ranging between . $/m and . $/m . it is interesting to compare the cost of water from desalination with the current prices of water in the region. the example of the impact of desalination for farmers can be a good illustration especially that the souss region is known for its high agricultural output. for farmers, the selling tariff of desalinated water fixed by oman el baraka fell from . $/m to . $/m before it was finally fixed at . $/m /y plus an fig. . variation of electricity cost under different subsidies with a $/tco tax on fossil generation. (here it is not considered that the supply to national grid could be more decarbonized in the future, because it is assumed that this trend will occur much slower than the start of the desalination plant.) additional subscription fee of $/ha/y. each farmer has to commit to a m /ha/y ( m /ha/y during six irrigation campaigns) . for example, a farmer who owns ha will commit to , m /y ( , × ) and will pay . $/m /y (( × )/ , ) of fixed subscription fee plus a variable fee of . $/m . per year, the farmer will thus pay . $/m ( . + . ). if this cost of water is supposed to already include the value added tax of % [ ] applied to water and energy products, then the cost of desalinated water would range between . $/m and . $/m depending on the power supply considered. the selling price of desalinated water as proposed by oman el baraka would represent only % to % of the estimated cost range of desalinated water according to the power supply considered in this paper. this result suggests that the difference between the cost of desalinated water and the price at which it is intended to be sold would probably be compensated by public subsidies varying between % and % of the estimated cost of desalinated water. in addition, it is interesting to compare the estimated cost range and the expected selling price of desalinated water to the present water tariffs as applied by the local distributor of water. to be coherent with the previous paragraphs, the same farmer is considered. water tariffs according to ramsa (for industrial clients because there is no specific tariff for agricultural clients) include a fix tariff of . $/month (including the % value added tax rate) and a variable tariff of . $/m (including the % value added tax rate). for a farmer who needs the same water consumption as previously mentioned ( , m /y), the total cost of water consumed will be around . $/m (( . × )/ , ) plus a variable tariff of . $/m . thus, for a farmer, the expected selling price of desalinated water is almost % higher than the cost of water based on present ramsa tariffs, whereas the estimated cost range of desalinated water based on the different power supply options covered in this paper are % higher to % higher than the cost of water based on present ramsa tariffs. fig. displays graphically the cost and selling price of desalinated water compared to the present selling prices of water. as noticed earlier the cost of water from desalination is higher than present water tariffs which may be unfavourable to the promotion of desalination as a future technique to cope with the increasing water pressure in morocco. however, as water is increasingly becoming a scarce resource in the country, one may expect that water tariffs will increase in the coming years (market signal price for scarcity and incentive to water conservation). water supply diversification using desalination (even it appears costly today) can therefore contribute to avoid future water shortage that could hamper the socio-economic development of a whole country. detailed results from the previous section, describing a large set of solar electricity production technologies, reveal that the cost of electricity can significantly impact the cost of desalinated water: ▪ under present conditions, solar use for large scale desalination projects is already competitive for solar pv without storage, compared to grid-only powered alternative. in this case, only part of the supply comes from solar (about one third). larger use of pv supply would increase the total cost of water under the assumptions retained in this paper. ▪ moreover, the introduction of batteries (with pv and the grid) is found to lead to higher costs of water (+ % compared to battery- fig. . impact of the increase in grid tariffs (+ %, + %, + % and + % from current onee tariffs) on the average cost of electricity in all cases considered in this study. less pv option). the extreme design which would be based on pv only + very large batteries set has not been studied. but, as its cost would lie much above the results obtained in the paper, it appears somewhat out of the scope. ▪ the use of csp (with storage and the grid) appears also not yet competitive compared to pv only option and to the grid alternative. even if the pv + storage + grid option results in a slightly higher cost of water, this difference with the csp based alternative(s) is not significant. this indicates that the choice between both these technologies with their respective storage solution for future desalination applications is not much evident based on present costs and assumptions. from the "private" point of view, it is of the first interest to examine how cost competitiveness between all the power supply options would change by . to conduct this prospective analysis, three scenarios are considered. business as usual ("bau") scenario is related to the present costs used in this study. the scenario "high" refers to the aggressive target costs of pv and csp by published recently by the nrel, and to the target cost in reference [ ] and reference [ ] for lead-acid storage (average value). the "mid" scenario considers that only one third cost reduction by is achieved. for csp costs in , a value of . $/kwh (as calculated previously) is added to the target cost of electricity generation to account for electricity transport as previously done in this paper. table summaries cost assumptions used in this prospective analysis. regarding the electricity cost of the grid, bau scenario supposes no relative change in prices by whereas "mid" and "high" assume an increase of . %/y and %/y respectively. as this paper focuses mainly on the impact of different electricity costs on the relative costs of desalinated water, it is assumed that by , the capex and opex (without electricity) would remain at the same level. the only component of the cost of water that is assumed to vary is therefore the cost of electricity consumption, based on different assumptions on the increase in electricity prices and on the pace of solar and storage technologies cost reduction by . in the case of storage, lead acid batteries are taken as the reference storage technology but the reduction of the cost of storage could be reach by other storage technologies (li-ion…). many combinations of cost reduction assumptions can be computed for the different technologies compared in this paper. some cases are however not plausible, as for example an increase in electricity tariffs with a strong cost reduction of solar technologies. future electricity prices depend on the future electricity mix. with a strong cost reduction of solar technologies and their increase proportion in the mix, the price of electricity is not expected to increase by a large amount. for this reason, a reasonable assumption has to be retained for the two prospective scenarios: ▪ scenario : supposes moderate increase of the grid electricity prices (mid) and high achievements of cost reduction targets by for solar technologies and batteries (high) ▪ scenario : supposes high increase of the grid electricity tariffs (high) but no further cost reduction for solar technologies and batteries (bau). as displayed in fig. , the main trend in the scenario is the contrary effects of technological progress and an increase in the electricity tariffs, leading to a globally stable desalinated water production cost. but the rank of the technology's changes. under this scenario, the most competitive solution remains the pv + grid alternative, the cost evolution of which is rather flat from the base year up to . however, beside this "flat" evolution, there are two different trends for the other technologies. the first is the logical increase in the cost of water for a % grid based solution that gives by the highest cost of water ( % increase compared to the lowest cost provided by pv + grid in ). the second trend is the decreasing evolution of the cost of water in the case of csp + storage + grid and pv + storage + grid alternatives. the two storage-based solutions, owing to the assumed decrease of the costs of solar technologies and storage, achieve grid parity by and respectively. in addition, these two solutions lead to the same cost of water (parity) a little before . the scenario in some way is closed to the analysis from the "public" point of view here above, because the main drivers of the evolution in electricity tariffs take into account the real production cost of the electricity and the co externalities. the scenario supposes in fact a high increase in electricity tariffs but no further cost reduction of the other technologies. as displayed in fig. , the cost of water increases by for all solutions compared to the reference level as all alternatives are grid-dependent. the least cost option is always pv + grid whereas the power alternative with the highest cost of water is pv + storage + grid. under the assumptions of this scenario, csp + storage + grid achieves full grid parity by . table summarizes the changes in the cost competitiveness of the different power supply alternatives studied in this paper from the reference year up to . it shows that pv+ grid is still the cheapest power supply alternative for which the lower cost of water is achieved on the full period of time. it is worth nothing that the cost competitiveness between pv + storage + grid and csp + storage + grid appears to be not too discriminant which suggests for a more comprehensive multi-criteria analysis to clearly discriminate between those power supply options. practically, in terms of decisions by the moroccan state, these results are clearly in favour of a policy aiming to develop a first programme of ro desalination with an increasing share of electricity from a local pv source (the bulk of supply coming from the national grid). this share is easy to adapt, whereas it does not exceed one third to half of the total, taking account the technology and regulation evolutions. the parallel development of centralized csp appears also a reasonable source of diversification and could supply, in a second programme, competitive electricity to future desalination plants, taking advantage of their increasing size and scale effects (csp plants and electric lines) that are more important than the expected improvement of the local pv fields. when referring to the two programmes, it is not proposed to rely change in the ranking of the power supply alternatives by , ( is the best cost option, is the most expensive option). only on pv+ grid in the first one (namely in the ies) and to shift (in the ies) to csp. clearly, after a delay of one decade the costs of the two technologies will be in the same range. the system costs of pv will begin to increase significantly with its share in total electricity production. thus, a new generation of systems, based on csp, will have the great advantage to allow to diversify the behaviour of both electric plants, when adding a (very) significant capacity of storage. in order to prepare this situation, it appears important that morocco continues its efforts to develop this technology, in order to be fully ready, at a large scale, in about years. but, once more, the results of this paper suggest strongly to begin the development of "low carbon" desalination with pv. however, it is more difficult today to precisely assess how the trends exposed in this paper could practically be developed because of the sars-cov- pandemic the world is experiencing today. it could cause many delays before such large scale plants become operational. in addition, under the current situation, it is plausible that energy subsidies in morocco would still be maintained at least for the next two years or more to support the revival of national industries that stopped their activities during the lockdown period. however, in a context of higher uncertainties, a diversification of electric mix as long as an increase of domestic (non-fossil) electric production could offer an interesting way of robustness. aside from this specific economic context that may cause many disruption in the desalination market (in the short and medium term), large scale desalination plants powered by solar energy suffer from inherent challenging issues that lead by the past to the give up of solar powered desalination units. for example, a first obstacle can be related to the reliability of contractors and project partners. for example, the al khafji plant (saudi arabia) which is the first large scale solar swro plant of a capacity of , m /d and powered by a pv plant of mw, experienced a delay because initially only one contractor (abengoa which almost went bankrupt in ) had to do everything (desalination unit + pv plant). this is not the case for the chtouka ait baha plant where several partners are engaged in the project under a public private partnership. in addition, the desalination unit and the power supply unit are not built by the same operator. a second limiting factor of large-scale solar powered desalination plants is related to logistics considerations, among which the spatial distribution of solar energy and saline water (generally inland versus along the coast) and grid interconnection policy and infrastructure constitute the main factors. in addition, as large scale desalination plants have high energy demands, the required area can be a significant issue (land availability and conflict with other land usage) especially if both solar and desalination plants are located along the coast. another limitation of large scale solar powered desalination unit is related to their optimal operation that requires accurate solar production forecasting models. in this paper, the economic competitiveness of large scale solar powered reverse osmosis desalination plant has been investigated taking as a case study the chtouka ait baha desalination ( , m / d) plant in morocco. the bulk of the paper consists in the assessment of desalinated water cost for different power supply alternatives (using pv, csp, batteries technologies and the grid). this analysis allows to compare the cost of desalinated water for these four power supply alternatives for the reference year ( ) and by . based on a global water cost model, this paper demonstrates at first that desalination, with the last up to date technologies, is affordable at an acceptable cost of around $/m (range of . $/m and . $/m depending on the power supply option). in addition, the results show that the selling price of desalinated water as proposed by oman el baraka ( . $/m , vat included) would represent only % to % of the estimated cost range of desalinated water according to the power supply considered in this paper. this result suggests that the difference between the cost of desalinated water and the price at which desalinated water is expected to be sold would probably be compensated by public subsidies varying between % and % of the estimated cost of desalinated water. when assessing the impact of the cost of electricity supply on the cost of desalinate water, this paper demonstrates that solar energy is a competitive way to feed large scale reverse osmosis plant today. pv without storage appears today the most competitive solution. electricity storage is not yet fully competitive (in this paper, storage services to the national grid are not considered). csp appears today to be more costly, partly because of the need in the moroccan context, to build a high voltage electric line to transport electricity from the desert. the results of the prospective analysis towards shows potential benefits of csp + storage + grid and pv + storage + grid alternatives even if the most competitive solution remains pv + grid. this study not only illustrates the difficulty to determine the best option to supply a large-scale desalination plant somewhere in africa, but also points out some promising technical combinations. indeed, one of the main results of this paper is that both pv and csp technologies may be part of the game of ro desalination in a quite short future. in fact, pv + grid appears regularly the best solution. from a "public point of view", csp appears not to be much more expensive. moreover, the capacity of csp to store heat at low cost offers an important source of flexibility to this technique, which allows too deeply decrease the share of electricity from the grid in the total consumption, if needed. in addition, with a large deployment of desalination plants, it becomes possible to build high power electric lines with lower costs, to transport electricity from the inland desert to coastal territories where the demand for water is peaking. beyond the optimal technological choice, the results of this paper advocate for the necessary upstream integration of energy requirements from the water sector in the future energy planning schemes in morocco. to conclude, and based on the results of this study, a first programme of desalination plants based on pv + grid should be launched, while a second programme based on csp should be prepared. however, this study faces some limitations. one limit is related to the only decision criterion selected to assess the competitiveness of the retained power supply options that is the minimum cost of water. beyond this metric, a discriminatory criterion could be the reliance on the grid or the global co emissions resulting from each alternative. a multi-criteria analysis with a cost/advantage approach taking into account technical, economic (including job creation or local industry potential), and environmental aspects would provide a more complete picture of the attractiveness of each power supply option for the whole system including. indeed, environmental externalities include for example brine rejection that harms small marine organisms, especially in regions that highly rely on fishing activities. it is also worth noting that this analysis is based on a simplified approach with approximations especially for solar pv sizing. optimizing pv design could allow getting more precise storage requirement specifications and evaluate more accurately the weight of storage in the average electricity cost. in addition, future works could also consider different feed-in tariffs schemes for solar electricity surplus sold to the national grid. finally, morocco will have to cope with the consequences of the sars-cov- epidemic. the impact on the economy will certainly be hard, in the range of − % in gdp, or more. this situation may require several years to be fully stabilized and overcome. the assumptions on costs retained in this study will probably remain robust, but one could expect important changes or delay in economic growth of agriculture and demand for electricity, inducing major inflexions in the moroccan water and energy policies. the authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Énergies alternatives (cea/des/itese) and funded by the "laboratoire d'innovation pour les technologies des Énergies nouvelles et les nanomatériaux" (cea/drt/liten). a simplified model is used to determine the daily production and the water cost of a reverse osmosis (ro) desalination plant powered by any energy source. due to the relatively quick dynamic of such a system, simulation over a given day is calculated as a succession of steady states, hour per hour. for the purpose of this study, ro is supposed to be operated h a day. when solar resource is not available or insufficient, energy storage (electricity or heat for csp) or the grid takes over. fig. a shows ro desalination system. one pass ro unit is considered, without brine recirculation. it comprises a number of trains in parallel. each train is composed of a number of pressure vessels in parallel containing membrane elements in series. seawater pumps feed the trains after pretreatment (filtration and addition of chemicals). their number is to choose (one or several trains per pump). high pressure pumps feed the vessels. each train includes a high-pressure pump and a set of energy recovery devices to recover high pressure from the brine. a sea water pumping system and pre-treatment facility are supposed to be common for the whole desalination plant. before simulating the operation of ro plant over a day, a preliminary step is required to size the ro system. to this effect, two options are implemented in the model: sizing or operating the ro plant. developed with microsoft excel™, it includes the following calculation modules: • ro module: this is the most important and complex module because it controls or uses all the others. it comprises two user modes: sizing of the ro unit for a required desalting capacity and determination of the operating point of the previously sized unit when available power varies. to size the ro system, a membrane element is selected from the supplier's catalogue. their characteristics are entered: surface, permeability, salt rejection, etc. a default reference is proposed and related characteristics are the default values in the calculations. seawater characteristics are also required (temperature, salinity, flow factor). calculations are based on the method described in filmtec™ reverse osmosis membranes technical manual [ ] . the basic equation to calculate the separation of water and salt is: references [ , ] give more details about the calculation method and the determination of different factors. the module is used to size the ro system. for a given desalination capacity, it returns the number of standard membrane elements per pressure vessel, the number of pressure vessels per train and the number of ro trains. the number of required energy recovery devices (erd) is determined according to a selected unit capacity in a proposed pressure exchanger's catalogue. operating conditions -feed pressure and flow rate, conversion ratio -are chosen to minimize either specific energy consumption or water cost. the calculation of opex in the model excludes the cost of energy since it is the principal parameter of the study. first, the specific consumption of the swro (kwh/m ) is calculated, then the power required (kw) and finally the corresponding energy (kwh) over a given period (a year) are computed. since the plant operates at full capacity h a day, days a year ( % of availability factor), these values are fixed. the different electricity cost scenarios combined with this required energy give the contribution of the energy cost to the water price. the energy consumption includes the energy consumption of the reverse osmosis unit (high-pressure pumps and booster downstream of the erds) but also that of the seawater supply pumps, taking into account the pressure losses in the pipes and in the seawater pre-treatment. the back pressures of the brine and freshwater discharges produced (thus including post-treatment) are also taken into account. ▪ cost module: usual cost functions for ro are used for capex and opex calculations. they are taken from different sources [ , [ ] [ ] [ ] [ ] [ ] [ ] . coupled with the performance module, it provides a detailed calculation of capex, opex and finally the levelized cost of water (lcow). the effective daily operation time is taken into account as it has a strong effect on the contribution of capex to the water cost via the capital amortization. this module gives water cost breakdown. validation of the ro model has been carried out using experimental results from literature [ ] [ ] [ ] [ ] [ ] . a benchmark with different professional tools available shows that performance and cost are predicted with a deviation better than %. tools are deep v . code updated by iaea in , wave . edited by dow chemical in (update of rosa . merged with others water processing simulation tools) and gwi online desalination cost estimator. transforming our world, the agenda for sustainable development climate change in mena: challenges and opportunities for the world's most water stressed region recent developments in solar thermal desalination technologies: a review past present and future roadmap for the development of desalination powered by renewable energy energy consumption and water production cost of conventional and renewable-energy-powered desalination processes water desalination using renewable energy recent progress in the use of renewable energy sources to power water desalination plants a critical overview of renewable energy technologies for desalination, , renewable energy technologies for water desalination energy minimization strategies and renewable energy utilization for desalination: a review renewable energy-driven desalination technologies: a comprehensive review on challenges and potential applications of integrated systems solar-driven desalination with reverse osmosis: the state of the art economic feasibility of alternative designs of a pv-ro desalination unit for remote areas in the united arab emirates photovoltaic reverse osmosis -feasibility and a pathway to develop technology photovoltaic powered water purificationchallenges and opportunities a direct coupled photovoltaic seawater reverse osmosis desalination system toward battery based systems -a technical and economical experimental comparative study optimal design of stand-alone reverse osmosis desalination driven by a photovoltaic and diesel generator hybrid system design optimization of batteryless photovoltaic-powered reverse osmosis water desalination in remote areas aqua-csp concentrating solar power for seawater desalination concentrating solar power for seawater desalination in the middle east and north africa integrating desalination with concentrating solar thermal power: a namibian case study techno-economic analysis of combined concentrating solar power and desalination plant configurations in israel and jordan pv and csp solar technologies & desalination: economic analysis desalination plants in australia, review and facts industries et de services d''agadir, dossier de presse, chtouka ait baha la nappe phréatique enfin protégée des abus ministère délégué auprès du ministre de l''energie, des mines et de l''environnement, ème communication nationale du maroc à la convention cadre des nations unies pour le changement climatique adaptation du système eau-énergie au changement climatique: etude nationale -maroc conventions pour la réalisation de la plus grande unité mutualisée de dessalement de l'eau de mer pour l'irrigation et l'alimentation en eau potable de la région de souss massa filmtec™ reverse osmosis membranes, technical manual, form no. - - dow filmtec™ membranes -flow factor for ro system design, form no. - - membrane types and factors affecting membrane performance water reuse association, desalination committee, seawater desalination cost, white paper full cost in desalination. a case study of the segura river basin investment and production costs of desalination plants by semi-empirical method cost analysis of seawater desalination with reverse osmosis in turkey abdel-fatah, comparison between reverse osmosis desalination cost estimation trends the economic feasibility of small solar med seawater desalination plants for remote arid areas socio-economic & technical assessment of photovoltaic powered membrane desalination processes for india economic analysis of desalination technologies in the context of carbon pricing, and opportunities for membrane distillation assessment of methodologies and data used to calculate desalination costs desalination projects economic feasibility: a standardization of cost determinants electricity cost from renewbale erngy technologies in egypt levelized cost of electricty, renewable energy technologies, fraunhofer institute for solar energy systems ise is concentrated solar power (csp) a feasible option for sub-saharan africa?: investigating the techno-economic feasibility of csp in tanzania renewable power generation costs in the potential role of concentrating solar power within the context of doe's solar cost targets comparison of electricity storage options using levelized cost of storage (lcos) method energy storage systems-characteristics and comparisons electrical energy storage systems: a comparative life cycle cost analysis dynamische berechnung der stromgestehungskosten von energiespeichern für die energiesystemmodellierung und -einsatzplanung lazard's levelized cost of storage analysis -version energetic, economic and environmental ( e) analyses and lcoe estimation of three technologies of pv grid-connected systems under different climates photovoltaic software solar photovoltaic system cost benchmark: q local cost of seawater ro desalination based on solar pv and wind energy: a global estimate document d'évaluation du projet concernant une proposition de prêt d'un montant de millions de dollars et une proposition de prêt du fonds pour les technologies propres d'un montant de , millions de dollars à l'office national de l'électricité et de l'eau potable (onee) avec la garantie du royaume du maroc a l'appui d'un projet « énergie propre et efficacité énergétique levelized cost of electricity for solar photovoltaic and electrical energy storage announces financial close of concentrated solar power (csp) projects noor ouarzazate ii un kilomètre de ligne à haute tension: , millions € electricity transmission and distribution shedding light on electricity utilities in the middle east and north africa: insights from a performance diagnostic energy policies beyond iea countries low energy consumption swro energy recovery devices for high pressure hydraulic retrofits to improve desalination plants operating experience of the dhekelia seawater desalination plant using an innovative energy recovery system energy recovery devices in seawater reverse osmosis desalination plants with emphasis on efficiency and economical analysis of isobaric versus centrifugal devices, graduate theses and dissertations the authors would like to thank dr. jean-guy devezeaux de lavergne (cea), dr. florence lefebvre-joud (cea) and dr. elisabeth le net (cea) for academic advice and fruitful discussion. this research is part of a phd thesis conducted at "institut de technico-Économie des systèmes Énergétiques du commissariat à l'Énergie atomique et aux key: cord- -pnjhi cu authors: foreman, stephen; kilsdonk, joseph; boggs, kelly; mouradian, wendy e.; boulter, suzanne; casamassimo, paul; powell, valerie j. h.; piraino, beth; shoemaker, wells; kovarik, jessica; waxman, evan(jake); cheriyan, biju; hood, henry; farman, allan g.; holder, matthew; torres-urquidy, miguel humberto; walji, muhammad f.; acharya, amit; mahnke, andrea; chyou, po-huang; din, franklin m.; schrodi, steven j. title: broader considerations of medical and dental data integration date: - - journal: integration of medical and dental care and patient data doi: . / - - - - _ sha: doc_id: cord_uid: pnjhi cu dental health insurance coverage in the united states is either nonexistent (medicare and the uninsured), spotty (medicaid) and limited (most employer-based private benefit plans). perhaps as a result, dental health in the united states is not good. what public policy makers may not appreciate is that this may well be impacting medical care costs in a way that improved dental benefits would produce a substantial return to investment in expanded dental insurance coverage. have been rising at double digit rates. most employers have been dropping health care coverage rather than expanding it ( kaiser family foundation ) . medicare trust funds are bankrupt (social security and medicare boards of trustees ). adding coverage would exacerbate an already alarming problem. medicaid funding is a major source of state government defi cits. many states are slashing medicaid coverage during this time of crisis (wolf ). improving medicaid dental coverage during times of budget crisis would meet substantial political resistance. strikingly, strong and increasing evidence suggests relationships between oral health and a range of chronic illnesses. for example, recent fi ndings show relationships between periodontal infl ammatory conditions and diabetes, myocardial infarction, coronary artery disease, stroke, preeclampsia and rheumatoid arthritis. this suggests that improved oral health may well have the potential to reduce the incidence of chronic diseases as well as their complications. if chronic disease incidence is reduced it may be possible to avoid medical care costs related to treating them. it would be important to know more about the extent to which improved oral health could reduce health care costs and improve lives. there are few, if any, studies of the costs of providing medicare dental benefi ts, the costs of improving the medicaid dental benefi t or the cost of providing dental insurance to the uninsured. there are a few studies that indicate that periodontitis increases medical care costs, perhaps by as much as % (ide et al. ; albert et al. ) . ideally there should be a controlled study to assess the benefi t of providing dental coverage through a government payer system. for a preliminary inquiry we can consider work already done and using some cost and benefi t estimates, determine whether it is possible that benefi ts of extending dental coverage may outweigh costs. the failure of medicare to cover dental care has engendered some (albeit not much) public debate. in , congress enacted the medicare prescription drug, improvement, and modernization act (medicare part d). by medicare provided $ . billion in benefi t payments for outpatient prescription drugs and medicaid paid . billion for outpatient prescription drugs (center for medicare and medicaid services ) . benefi ciaries provided billions more in the form of monthly part d premiums. the expense of the medicare prescription drug program and the controversy surrounding its enactment may well have eroded public support for increased medicare coverage. so while there has been no shortage of effort paid to improving medicare, the one common theme in all of the recent initiatives is that dental care has been conspicuously a new study by hedlund, jeffcoat, genco and tanna funded by cigna of patients with type ii diabetes and periodontal disease found that medical costs of patients who received maintenance therapy were $ . per year lower than patients who did not. cigna, research from cigna supports potential association between treated gum disease and reduced medical costs for people with diabetes, http://newsroom.cigna.com/newsreleases/research-from-cigna-supports-potential-association-between-treated-gum-disease-and-reduced-medical-costs-for-people-with-diabetes. omitted. as a result, million medicare recipients in (us census bureau ) continue to have no dental insurance coverage through medicare. medicaid dental coverage is an optional benefi t that states may or may not elect to provide. in medicaid, both the state and the federal government provide funds to cover healthcare services to eligible patients. the bulk of the money comes from the federal government. because the medicaid dollars are limited and coverage for systemic diseases has precedence, medicaid coverage of dental care has been spotty. even where it has been provided, payments to dental providers have been so low as to make it diffi cult or impossible for medicaid benefi ciaries to obtain adequate dental care (broadwater ) . the recession increased the number of medicaid eligible individuals nationwide. further, the federal budget defi cits of the past few years have reduced the federal contribution to state medicaid programs. the combination of increases in the number of benefi ciaries and diminished revenues has caused a number of states to eliminate or curtail medicaid dental coverage (ehow ; mullins et al. ) . the result, million medicaid benefi ciaries in the us (us census bureau ) in either had no dental insurance coverage or inadequate coverage. approximately million people in the united states do not have health insurance (kaiser family foundation ) . presumably, they have no dental insurance either. further, not every employer provides dental insurance. a cdc survey found that . % of adults do not have dental insurance coverage (centers for disease control ) . a montana survey found that % of employers who offer health insurance do not offer dental insurance coverage (montana business journal ) . in there were approximately million people enrolled in health insurance plans (us census bureau ) . if half (a rough combination of the cdc and montana percentages) of them do not have dental insurance it is likely that an additional million (nonelderly, non-poor) people in the us do not have dental insurance coverage. finally, the term "dental insurance" is actually a misnomer. dental policies cover routine treatments, offer discounts for more complex treatment and impose a low yearly on total payments. in fact, it has been called "part insurance, part prepayment and part large volume discount" (manski ) . effectively, many (if not most) people who have dental insurance fi nd it coverage to be quite restrictive. for example, many impose a small yearly cap ($ , is common) or large coinsurance amounts ( % for orthodontia, for example) (rubenstein ) . even with discounts it is easy for many people to exceed the annual limit. given the lack of dental insurance coverage it is not surprising that the status of oral health in the us is not particularly good. in approximately . % of adults between the ages of and had untreated caries, % had decayed, missing and fi lled tooth surfaces and more than one-half of adults had gingival bleeding (dental, oral and craniofacial data resource center of the national institute of dental and craniofacial research ) . three fourths of adults in the us have gingivitis and % have periodontitis (mealey and rose ) . if these levels of untreated disease were applied to most systemic diseases, there would be public outcry. over the past decade evidence has been building that there is a relationship between dental disease, particularly periodontal disease, and chronic illnesses. mealey and rose note that there is strong evidence that "diabetes is a risk factor for gingivitis and periodontitis and that the level of glycemic control appears to be an important determinant in this relationship" (mealey and rose ) . moreover, diabetics have a six times greater risk for worsening of glycemic control over time compared to those without periodontitis and, periodontitis is associated with an increased risk for diabetic complications. for example, in one study more than % of diabetics with periodontitis experienced one or more major cardiovascular, cerebrovascular or peripheral vascular events compared to % of the diabetic subjects without periodontitis (thorstensson et al. ) . also, a longitudinal study of type diabetics found that the death rate from ischemic heart disease was . times higher in subjects with severe periodontitis and the death rate from diabetic nephropathy was . times higher (saremi et al. ) . clinical trials have demonstrated that treatment of periodontal disease improved glycemic control in diabetics (miller et al. ) . moreover, investigations have found an association between periodontal disease and the development of glucose intolerance in non-diabetics (saito et al. ) . while it is diffi cult to establish causality and it is possible that other factors infl uence periodontal disease and medical complications, these studies suggest that treatment of periodontitis substantially improves health and greatly reduces medical complications related to diabetes. similarly, periodontitis is associated with cardiovascular disease and its complications including ischemia, atherosclerosis, myocardial infarction and stroke. a study by slade and colleagues found both a relationship between periodontitis and elevated serum c-reactive protein levels (systemic marker of infl ammation and documented risk factor for cardiovascular disease) as well as a relationship among body mass index, periodontitis and crp concentrations (slade et al. ) . hung and colleagues evaluated the association between baseline number of teeth and incident tooth loss and peripheral arterial disease. they determined that incident tooth loss was signifi cantly associated with pad, particularly among men with periodontal disease potentially implying an oral infection-infl ammation pathway (hund et al. ) . the same group of researchers used the population enrolled in the health professionals' follow-up study ( , men free of cardiovascular disease and diabetes at baseline) to assess the relationship between tooth loss and periodontal disease and ischemic stroke. controlling for a wide range of factors including smoking, obesity, and dietary factors, the researchers found a "modest" association between baseline periodontal disease history and ischemic stroke . as early as destefano and colleagues found that among subjects, those with periodontitis had a % increased risk of coronary heart disease relative to those without. the association was particularly high among young men. the authors questioned whether the association was causal or not, suggesting that it might be a more general indicator of personal hygiene and possibly health care practices (destefano et al. ) . in wu and colleagues used data from the first national health and nutrition examination survey and its epidemiologic follow-up study to examine the association between periodontal disease and cerebrovascular accidents. the study found that periodontitis was a signifi cant risk factor for total cva, in particular, for non-hemorrhagic stroke (wu et al. ) . in addition to diabetes and coronary artery disease, associations have been found between periodontal disease and rheumatoid arthritis and respiratory disease. this is not surprising given the role of periodontal disease in the production of infl ammation related proteins. dissick and colleagues conducted a pilot study of the associate ion between periodontitis and rheumatoid arthritis using multivariate regression and chi square tests. they found that periodontitis was more prevalent in patients with rheumatoid arthritis than in the control group and that patients who were seropositive for rheumatoid factor were more likely to have moderate to severe periodontitis than patients who were rf negative and also that patients who were positive for anti-cyclic citrullinated peptide antibodies were more likely to have moderate to severe periodontitis (redman et al. ) . paju and scannapeico investigated the association among oral biofi lms, periodontitis and pulmonary infections. they noted that periodontitis seems to infl uence the incidence of pulmonary infections, particularly nosocomial pneumonia in high-risk subjects and that improved oral hygiene has been shown to reduce the occurrence of nosocomial pneumonia. they found that oral colonization by potential respiratory pathogens, for possibly fostered by periodontitis and possibly by bacteria specifi c to the oral cavity contribute to pulmonary infections (paju and scannapeico ) . the implications for these fi ndings are profound. professionally, they suggest that managing patients with chronic illness and periodontal disease will require teamwork and a deeper knowledge base for dentists and for physicians (mealey and rose ) . dentists will need to be alert for early signs of chronic illness among their patients and physicians will need to be alert for signs of dental disease. both will need to consider wider treatment options than their specialty indicates. dentistry and medicine have operated as professional silos in the past. the relationship between dental disease and chronic medical conditions suggests that continued separation is detrimental to patient centered care. beyond treatment implications, there are extremely important health policy concerns. if treatment of periodontitis and other dental problems leads to reduced incidence of chronic illness, fewer complications from chronic diseases and reduced morbidity among chronically ill patients, increased access to dental services could signifi cantly reduce health care costs. the diseases associated with periodontitis are among the most common illnesses, the fastest growing and the most expensive diseases that we treat. a recent robert wood johnson report notes that approximately million americans have one or more chronic conditions, that the number of people with chronic conditions is expected to increase by % per year for the foreseeable future and that the most common chronic conditions include hypertension, disorders of lipid metabolism, upper respiratory disease, joint disorders, heart disease, diabetes, cardiovascular disorders, asthma and chronic respiratory infections (anderson ) (see fig . . ). one in four americans has multiple chronic conditions. ninety-one percent of adults aged and older have at least one chronic condition and % have two or more of them (anderson ) . people with chronic conditions account for % of all healthcare spending. seventy eight percent of private health insurance spending is attributable to the % of privately insured persons with chronic conditions. seventy three percent of healthcare spending for the uninsured is for care received by the one third of uninsured people who have chronic conditions. seventy nine percent of medicaid spending goes to care for the % of non-institutionalized benefi ciaries who have chronic conditions (anderson ) (see fig. . ). further, health care spending increases with the number of chronic conditions (anderson ) (see fig. . ). more than three fi fths of healthcare spending (two thirds of medicare spending) goes to care for people with multiple chronic conditions. those with multiple chronic conditions are more likely to be hospitalized, fi ll more prescriptions, and have more physician visits (anderson ) . in the american diabetes association estimated direct medical expenditures for diabetes at $ . billion: $ . billion for diabetes care, $ . billion for chronic complications and $ . billion for excess prevalence of general medical conditions. approximately % of direct medical expenditures were incurred by people over . indirect expenditures included lost workdays, restricted productivity mortality and permanent disability -a total of $ . billion. all told, diabetes was found to be responsible for $ billion of $ billion in total expenditures. per capita medical expenditures totaled $ , annually for people with diabetes and $ for people without diabetes (hogan et al. ) . more recently, dall and colleagues estimated that the us national economic burden of prediabetes and diabetes had reached $ billion in , $ million in higher medical costs and $ billion in reduced productivity. annual cost per case was estimated at $ , for undiagnosed diabetes and , for type diabetes (dall et al. ) . the costs of caring for people with diabetes have risen both because the numbers of diabetics has been increasing and because the per capita costs of care have increased. the number of diabetics increased from . million on to . million in (ashkenazy and abrahamson ) . a recent report by the unitedhealth group center for health reform & modernization provides a dire estimation -that more than % of adult americans could have diabetes ( %) or prediabetes ( %) by at a cost of $ . trillion over the decade. this compares with current estimates of % of the population with diabetes and % with prediabetes, or %. these estimates conclude that diabetes and prediabetes will account for % of total healthcare spending in at an annual cost of $ billion, up from an estimated $ billion in (unitedhealth center for health reform and modernization ) . average annual spending over the next decade by payer type is $ billion for private health insurance, $ billion for medicare, $ billion for medicaid and $ . billion for the uninsured. what about cardiovascular disease and rheumatoid arthritis? among the top ten health conditions requiring treatment for medicare benefi ciaries in approximately % of benefi ciaries suffered from hypertension, % from heart conditions, % had hyperlipidemia % had copd, % had osteoarthritis and % had diabetes (thorpe et al. ) . the american heart association estimates the cost of cardiovascular disease and stroke to be $ billion in direct expenditures and $ . billion for productivity losses due to morbidity and $ . billion in lost productivity due to mortality (present value of lost wages at %) (lloyd- ) . the centers for disease control estimates that during - million americans had selfreported doctor diagnosed arthritis, million of them with activity limitations (cheng et al. ) . cisternas and colleagues estimated that total expenditures by us adults with arthritis increased from $ billion in to $ billion in . most of the increase was attributable to people who had co-occurring chronic conditions (cisternas et al. ) . the cisternas study appears to aggregate all medical care expenditures by people with arthritis (which would include expenditures to treat diabetes and cardiovascular disease). an earlier cdc study focused on the direct and indirect costs in attributable to arthritis that estimated $ . billion in direct costs (medical expenditures) and $ billion in indirect costs (lost earnings) (yelin et al. ) . in short, current cost estimates for direct health care expenditures (excluding productivity losses) related to diabetes are approximately $ billion, for cardiovascular treatment, $ billion, and for rheumatoid arthritis, approximately $ billion (estimating that the $ . billion in costs have grown approximately % per year), a total of $ billion of the $ . trillion that will be spent in the us in . moreover, given current growth in the prevalence of diabetes, the unitedhealth estimate of $ million in spending for diabetes alone is not unreasonable. if health care costs attributable to diabetes, cardiovascular disease and rheumatoid arthritis only increase by % over the next decade (even given added demand produced by the aging baby boomer population), annual costs of these chronic diseases will exceed $ . trillion in . if we use the unitedhealth estimates for the proportions of diabetes costs paid by private insurance ( %), medicare ( %), medicaid ( %) and the uninsured ( %) and estimate total costs based on the studies projecting a % increase in years and a % increase in years we can obtain an estimate of future costs for treating diabetes, cardiovascular disease and arthritis. table . set forth below, summarizes these cost estimates. by medicare costs for these chronic illnesses would be approximately $ billion. the estimated costs to medicaid will be approximately $ billion. the costs for the uninsured will be approximately $ billion. any intervention that has the potential to substantially reduce these costs will produce meaningful results. unfortunately, even though there had been a substantial numbers of studies that show relationships between dental disease and chronic illness that are have been very few studies that actually test whether improved dental treatment reduces the incidence of chronic illness and complications due to chronic illness. the potential for large health care cost savings through an active and aggressive program of dental care is so large that such studies are clearly indicated. suppose, for example, that % of all medical care costs required to treat diabetes, cardiovascular disease and arthritis could be avoided through an active aggressive program of dental care. what this would mean is that in private health insurers could see a $ billion reduction in healthcare costs, medicare would see a $ . billion reduction and medicaid pay $ . billion reduction. recent health reform has provided for the issuance of health insurance to the uninsured by state exchanges. aggressive dental care that saved % of costs attributable to diabetes, cardiovascular disease and arthritis could save the exchanges $ billion per year. and, if greater proportions of costs can be saved or if the estimates of costs are low, potential benefi ts will be even larger. once again, it would be important to know whether aggressive dental care could produce such savings and how much. ide and colleagues found that people who were treated for periodontitis incurred % higher health care costs than those who were free of periodontal disease (ide et al. ) . similarly, albert, et al., found medical costs associated with diabetes, cardiovascular disease and cerebrovascular disease were signifi cantly higher for enrollees who were treated for periodontitis than for other dental conditions (albert et al. ) . additional studies of this nature would be important to support a measured approach to expanding dental coverage. so what do we mean by an aggressive dental treatment plan? suppose we were to provide dental insurance to all medicare benefi ciaries at the level of current private dental insurance coverage and strongly encourage benefi ciaries to receive dental treatment. suppose we were to provide for medicaid payment for all benefi ciaries at the level of current private dental insurance coverage. suppose health care insurers provided dental coverage in order to reduce their costs and that such coverage was consistent with current private dental insurance coverage. suppose health insurance companies, understanding the benefi ts from dental care, were to require their private employer customers to cover the costs of dental care. how much would all of this cost? how would it compare to the benefi ts that may be available? in order to estimate the potential costs of providing enhanced coverage for dental care we start use the cms estimates of national health care spending for dental services and statistical abstract of the us estimates for medicare enrollment, medicaid enrollment, private health insurance enrollment and uninsured persons. based on the estimate that half of private employers with health insurance provided dental insurance coverage we estimate that of the private health insurance enrollment one half would have dental insurance coverage and one half would not. table . sets forth the national health care expenditures for dental services in millions and enrollment in private dental plans, medicare, medicaid, the uninsured without health insurance and dental insurance, the uninsured with health insurance and dual eligibles. from this we derive a cost per enrollee for private dental insurance, medicare dental benefi ts and medicaid dental benefi ts. in order to estimate the annual cost of providing full dental coverage to medicare benefi ciaries we subtracted dual eligibles (who receive some dental insurance) from total medicare enrollees to determine the number of persons who would need coverage. in our example there were million medicare benefi ciaries including million dual eligibles. accordingly, the estimates would cover the million medicare benefi ciaries that are not dual eligible at a cost equal to the per capita cost of private dental insurance ($ . ) less amounts that medicare is already paying for dental services ($ . per person). the result provides an estimate of the cost of covering all medicare benefi ciaries for dental services at a level equivalent to private health insurance. using the example the cost of providing full dental insurance coverage to medicare benefi ciaries would have been $ . billion. in addition, we used the cms national health expenditure fi gures to determine administrative costs for private health insurance, medicare and medicaid as a percentage of program expenditures for medical care. we found that the administrative costs of the medicare program were . % on average for - . in order to fully estimate the cost of medicare dental coverage we added . % to the cost health insurers will be in the same position as medicare and medicaid regarding dental coverage. if quality dental coverage saves health care costs attributable to diabetes, cardiovascular disease and rheumatoid arthritis then the exchanges will have an incentive to provide quality dental coverage to reduce costs. accordingly, we estimated the cost of providing dental coverage equivalent to private dental insurance coverage through the exchanges. again we assume that the costs of such coverage will be equivalent to the number of uninsured persons multiplied by the annual per capita cost of coverage. for the example, this would refl ect coverage for million people at $ . per person, a total of $ . billion. with administrative costs, the cost of providing dental insurance coverage to the uninsured at a level equivalent to private dental coverage would be $ . billion. finally, given the evidence that improved dental care has the potential to reduce health care costs private health insurers may wish to expand health insurance to cover dental care. here, we estimate the cost of providing dental insurance to the % of the workforce whose employers currently do not provide dental insurance benefi ts. once again, we multiply the number of covered lives by the estimated annual per capita cost. for the example we estimate million adults will receive dental coverage at $ per person: $ billion for dental services and $ . billion for administrative costs or a total of $ . billion. of course, as noted a number of times above, these estimates are based on providing full "universal" dental insurance coverage at levels equivalent to current benefi t levels for private dental insurance. it may be that an appropriate package of dental services that deals specifi cally with periodontitis can be provided for less than the full cost of private dental insurance. once again, further research should provide better information. the health reform law does not attempt to provide coverage to all million people without health insurance. estimates are that only million people will be covered by the bill. even though this is the case we prepare our estimates using all million uninsured americans. indeed, the failure of % of employers to cover dental services may well constitute a classic externality in the market for health insurance. internalizing this externality may well provide better effi ciency. it is also possible that dental care for persons with greater incidence of chronic illness as is the case with medicare benefi ciaries may require even higher levels of spending per benefi ciary. again, it would be good to know scientifi cally if this is the case. as noted in sect. above, costs for diabetes, cardiovascular disease and arthritis will be $ billion for private health insurance, $ billion for medicare, $ billion for medicaid and $ billion for the uninsured. costs of providing "full" dental coverage will be $ . billion for medicare, $ . billion for medicaid, $ . billion for the uninsured and $ . billion for private health insurance. given this, if . % or more of the medicare costs can be "saved" through improved dental care, medicaid dental insurance will pay for itself and will provide a positive return on investment. see table . . similarly, private health insurers could justify providing dental insurance coverage to employees who do not have it so long as they spend . % or more of their chronic care costs for diabetes, cardiovascular disease and arthritis. on the other hand, it would appear that medicaid expansion would require cost savings of approximately % and that health care insurance coverage of the uninsured would require savings of approximately % in order to justify coverage. while it is possible, it may not be likely that full dental coverage would be justifi ed for these programs. of course, these estimates do not consider indirect costs in the form of lost wages or premature death. these costs are externalities to the health insurance programs. to the extent that they represent a social benefi t that a national dental insurance program might internalize, it would be appropriate to consider their impact in the cost-benefi t analysis. in any event, better understanding of the potential for deriving savings in health insurance costs related to chronic diseases like diabetes, cardiovascular disease and arthritis would be crucial to any determination whether to expand insurance coverage for dental care. heretofore the case for expanding medicare coverage to include dental care has taken the form of "benefi t" to patients rather than benefi t to health insurance programs and society and has been cast in emotional and political terms. for example, oral health america grades "america's commitment to providing oral health access to the elderly" (oral health america ) . in truth, there is no american commitment to providing oral health access to any age group, much less the elderly. rubenstein notes that "at least one commentator has suggested that the dental profession should join with senior citizen groups when the time is right to ask congress to expand medicare to cover oral health" (rubenstein ) . rubenstein emphasizes that "calls for action" are "mere words" unless they are accompanied by political actions that health policy professionals and the dental profession must help promote (rubenstein ) . another commentator has suggested that "as soon as the debate over medicare prescription drug coverage and, the debate to provide dental care coverage for the elderly may soon begin" (manski ) . rubenstein, again suggests that "the dental community must convince americans, and particularly aging boomers, that oral health is integral to all health, and for that reason, retiree dental benefi ts are an important issue". in truth, a decade of defi cit spending and public distaste for out of control program costs in the medicare and medicaid programs as well as the unpopularity of the process that was used to provide medicare prescription drug coverage (with perceived abuses by the health insurance and drug lobbies) and national health reform makes it unlikely that the public would be willing to approve expansions in insurance coverage for dental care "for its own sake" or "as the right thing" or to "benefi t seniors." what this political climate has produced is an arena in which a good idea that could provide appropriate return on investment for society might well be rejected out of hand based on political history of health insurance coverage. as a result, it is incumbent on policymakers, medical and dental research scientists and health economists to investigate and confi rm the potential savings that expansion of dental insurance coverage has the potential to produce and to develop hard evidence regarding potential costs of the expansion prior to, not as a part of, political efforts aimed at dental coverage expansion. a responsible, well informed effort to expand dental coverage may well go far to restore public confi dence in the health policy process. joseph kilsdonk and kelly boggs the adage of "putting your money where your mouth is" is often referenced when being challenged about public statements or claims. in this instance, we use it literally. in health care costs in us were $ . trillion. there have been numerous reports on health disparities, the burden of chronic diseases, increasing healthcare costs and the need for change. long-term economic benefi ts associated with the cost of care are dependent upon integrating oral health with medicine. this is particularly true as it relates to the management of those conditions which impact the economics of healthcare the most. as examples, % of medicare costs and % of medicaid costs are in managing chronic health conditions (partnership for solutions national program offi ce ) . more than % of the u.s. population has one or more chronic condition (cartwright-smith ) and in , % of medicare spending was on patients with fi ve or more chronic diseases (swartz ) . effective management of health care resources and information are critical to the economic well-being of our healthcare system. we can no longer afford to manage care in isolation. integration of care between medicine and dentistry holds much promise in terms of reducing the cost of care and an integrated medical-dental electronic healthcare record (iehr) is the vehicle that will lead to downstream cost savings. in the united states the center for medicare & medicaid services (cms) has conducted demonstration projects around chronic disease management. section of the benefi ts improvement and protection act of mandated cms to conduct a disease management demonstration project. april , , as an effort to reduce the cost of care and improve quality associated with chronic diseases, cms partnered with ten premier health systems to effectively manage chronic diseases in a medicare physician group practice demonstration (pgp). it was the fi rst pay-for-performance initiative for physicians under the medicare program (center for medicare and medicaid services ) . it involved giving additional payments to providers based on practice effi ciency and improved management of chronically ill patients. participants included ten multispecialty group practices nationwide, with a total of more than , physicians, who care for more than , medicare benefi ciaries (frieden ) . the chronic diseases that were targeted were based on occurrence in the population and included diabetes, heart failure, coronary artery disease, and hypertension (frieden under the pgp, physician groups continued to be paid under regular medicare fee schedules and had the opportunity to share in savings from enhancements in patient care management. physician groups could earn performance payments which were divided between cost effi ciency for generating savings and performance on quality measures phased in during the demonstration as follows: year , measures, year , measures and years and having quality measures. for each of the years only the university of michigan faculty group practice and marshfi eld clinic, earned performance payments for improving the quality and cost effi ciency of care. a large part of the success of this project was attributed to being able to extract, evaluate, and monitor key clinical data associated with the specifi c disease and to manage that data through an electronic health record (table . ). during the third year of the demonstration project marshfi eld clinic, using a robust electronic health record succeeded in saving cms $ million dollars; that's one clinic system in year. as a result of such demonstration projects and as of this writing, cms is looking to establish accountable care organization's as the medical front runners to new care delivery methods for quality and cost control. accountable care organization (aco) is a term used to describe partnerships between healthcare providers to establish accountability and improved outcomes for the patients. in a cms workshop on october , , don berwick, the administrator of cms, stated "an aco will put the patient and family at the center of all its activities…" an emerging model of an aco is the patient-centered medical home (pcmh). pcmh is at the center of many demonstration projects. acos were derived from studies piloted by cms. since funds provided by cms, do not cover routine dental care as part of the patient management or quality and cost objectives cms aco studies are limited if they become models for the pcmh, due to the exclusion of dental. more recently, organizations representing the major primary care specialtiesthe american academy of family practice, the american academy of pediatrics, the american osteopathic association, and the american college of physicianshave worked together to develop and endorse the concept of the "patient-centered medical home," a practice model that would more effectively support the core functions of primary care and the management of chronic disease (fisher ) . in geisinger health system, kaiser permanente, mayo clinic, intermountain healthcare and group health cooperative announced they will be creating a project called the care connectivity consortium. this project is intended to exchange patient information. although progressive in their approach their project does not include dental. these benefi ts however, are yet to be adapted in the arena of oral health. as of this writing, dentistry remains largely separate from medical reimbursement mechanisms such as shared billing, integrated consults, diagnosis, shared problem lists, and government coverage. for example, cms does not cover routine dental care. dentistry is also working to establish its own "dental home" with patients. however to reap the economic benefi ts of integrated care, a primary care "medical-dental" home is what needs to be created. according to an institute of oral health report ( ) it is widely accepted across the dental profession that oral health has a direct impact on systemic health, and increasingly, medical and dental care providers are building to bridge relationships that create treatment solutions. the case for medical and dental professionals' comanaging patients has been suggested for almost the past century, in william gies reported that "the frequency of periodic examination gives dentists exceptional opportunity to note early signs of many types of illnesses outside the domain of dentistry" (gies ) . as described by dr. richard nagelberg, dds "the convergence of dental and medical care is underway. our patients will be the benefi ciaries of this trend. for too long, we have provided dental care in a bubble, practicing -to a large degree -apart from other health-care providers. even when we consulted with our medical colleagues, it was to fi nd out if premedication was necessary, get clearance for treatment of a medically compromised patient, or fi nd out the hba c level of a diabetic individual, rather than providing true patient co-management. we have made diagnoses and provided treatments without the benefi t of tests, reports, metrics, and other information that predict the likelihood of disease development and progression, as well as favorable treatment outcomes. we have practiced in this manner not due to negligence, but because of the limitations of tools that were available to us" (nagelberg ) . integrated medical/dental records need to be a tool in a providers' toolbox. in the case of marshfi eld clinic, dental was not included in their past cms demonstration project as dental is not a cms covered benefi t, and thus not part of the demonstration. however, as a leader in healthcare, the marshfi eld clinic recognizes the importance of data integration for both increased quality and cost savings. "marshfi eld clinic believes the best health care comes from an integrated dental/medical approach," said michael murphy, director, business development for cattails software. integration enhances communication between providers and can ultimately lead to better management of complex diseases with oral-systemic connection, avoidance of medical errors, and improved public health. while the cms pgp and other demonstration projects along with independent studies have shown to improve quality and reduce costs through integration, greater results may be afforded if studies are not done in isolation from dental data. in fact, if healthcare does not fi nd a way to manage the systemic nature of the pathogens known to the oral cavity the economic impact and cost savings around chronic disease management will hit a ceiling. the economic opportunity of having clinical data for integrated decision making is readily identifi ed by the insurance industry. the effective management of clinical data around chronic and systemic oral and medical disease as part of an iehr is the greatest healthcare cost savings opportunity associated with such a tool. the insurance industry sustains itself through risk management [obtaining best outcomes] using actuarial analysis [data] and controlling costs [reduction of costs] in order to ensure coverage [profi tability]. as such they have pursued the economic and outcome benefi ts of integrated medical -dental clinical decision making. as an example, in there was a study conducted by the university of michigan, commissioned by the blue cross blue shield of michigan foundation ( ) , the study included , blue cross blue shield of michigan members diagnosed with diabetes who had access to dental care, and had continuous coverage for at least year. with regular periodontal care, it was observed diabetes related medical costs were reduced by %. when compounding chronic health complications were also examined, the study showed a % reduction in cost related to the treatment of cardiovascular disease in patients with diabetes and heart disease. a % reduction in cost related to treatment of kidney disease for patients with diabetes and kidney disease. and a % reduction in costs related to treating congestive heart failure for patients with diabetes and congestive heart failure. according to a joint statement by lead researchers, and blue cross blue shield of michigan executives, "our results are consistent with an emerging body of evidence that periodontal disease…it addresses quality of care and health care costs for all michigan residents." also, at the institute for oral health conference in november joseph errante, d.d.s., vice president, blue cross blue shield of ma reported that blue cross blue shield of massachusetts claims data showed medical costs for diabetics who accessed dental care for prevention and periodontal services averaged $ /month, while medical costs for diabetics who didn't get dental care were about $ /month (errante ) . similarly insured individuals with cardiovascular diseases who accessed dental care had lower medical costs, $ /month lower than people who did not seek dental treatment (errante ) . the cost is $ less per visit for those diabetics who accessed prevention and periodontal services. those savings could be translated into access to care or additional benefi ts for more individuals. in the case of neonatal health there is similar research. over % of all births in the u.s. are delivered preterm, with many infants at risk of birth defects ( martin et al. ) . according to a january statement issued by cigna, announcing their cigna oral health maternity program, "the program was launched in response to mounting research indicating an increased probability of preterm birth for those with gum disease. these research-based, value-added programs are designed to help improve outcomes and reduce expense" (cigna ) . the program was initially designed to offer extended dental benefi ts free of charge to members who were expecting mothers, citing "research supporting the negative and costly impact periodontal disease has on both mother and baby." according to research cited by cigna, expecting mothers with chronic periodontal disease during the second trimester are seven times more likely to deliver preterm (before th week), and the costs associated with treating premature newborns is an average of times more during their fi rst year, and premature newborns have dramatically more healthcare challenges throughout their life. cigna also cited the correlation between periodontal disease and low birth weight, pre-eclampsia, gestational diabetes as additional rationale to support extended dental benefi ts to expecting mothers. six months later cigna initiated well aware for better health, an extended benefi ts free of charge program for diabetic and cardiovascular disease patients aimed at "turning evidence into action by enhancing dental benefi ts for participants in disease management" programs. it is interesting to note, not only does cigna offer extended dental benefi t to targeted groups, they also reimburse members for any out-of-pocket expenses associated to their dental care (co-pays, etc.) in , columbia university researchers conducted a -year retrospective study of , aetna ppo members with continuous medical and dental insurance, exhibiting one of three chronic conditions (diabetes mellitus, coronary artery disease, and cerebrovascular disease) (aetna ) . researchers found members who received periodontal treatments incurred higher initial per member per month medical costs, but ultimately achieved signifi cantly lower health screening (episode risk group/erg) risk scores than peers receiving little or no dental care. convinced by the data and understanding lower risk scores ultimately leads to healthier people and cost savings, aetna initiated the dental/medical integration (dmi) program in . aetna's dmi program offers enhanced benefi ts in the form of free-of-charge extended benefi t dental care to aetna's . million indemnity, ppo and managed choice medical plan members, specifi cally targeting members deemed at-risk, including those who are pregnant, diabetic, and/or have cardiovascular disease and have not been to a dentist in year as a result of various outreach methods during the pilot, % of at-risk members who had not been to a dentist in the previous months, sought dental care (aetna ) . "the fi ndings from this latest study we conducted continue to show that members with certain conditions who are engaged in seeking preventive care, such as regular dental visits, can improve their overall health and quality of life," said alan hirschberg, head of aetna dental (aetna ) . delta dental of wisconsin understands the connection between oral and systemic health and has created a program that is designed to offer members with certain chronic health conditions the opportunity to gain additional benefi ts. more than , groups now offer delta dental of wisconsin's evidence-based integrated care plan (ebicp) option (delta dental of wisconsin ) . ebicp provides expanded benefi ts for persons with diseases and medical conditions that have oral health implications. these benefi ts include increased frequency of cleanings and/or applications of topical fl uoride. they address the unique oral health challenges faced by persons with these conditions, and can also play an important role in the management of an individual's medical condition. ebicp offers additional cleanings and topical fl uoride application for persons who are undergoing cancer treatment involving radiation and/or chemotherapy, persons with prior surgical or nonsurgical treatment of periodontal disease and persons with suppressed immune systems. the ebic offers additional cleanings for persons with diabetes and those with risk factors for ie, persons with kidney failure or who are on dialysis and for women who are pregnant. the iehr provides the insurance industry in partnership with the healthcare industry an integrated tool to facilitate these health and subsequently economic outcomes across medicine and dentistry. in addition to the anticipated savings through better outcomes using integrated clinical data, an example of a positive economic outcome associated with an integrated record as related to increased effi ciency and patient safety is found in the united states veterans administration (va) hospitals and clinics. the va is one of the few institutions that have implemented the shared electronic medical-dental record successfully. the va has the ability to be the "one stop shop" for their patients. an april press release published on the department of veterans affairs website highlighted the success of va's health information technology in terms of cost reductions and "improvements in quality, safety, and patient satisfaction" (department of veterans affairs ) . the press release spotlighted a recent study conducted by the public health journal, health affairs, which focused on va's health it investment from to . the study confi rmed that while va has spent $ billion on their technology initiative, a conservative estimate of cost savings was more than $ billion. after subtracting the expense of the it investment, there was a net savings of $ billion for the va during the years covered by the study (mcbride ) . furthermore, the study estimated that "more than percent of the savings were due to eliminating duplicated tests and reducing medical errors. the rest of the savings came from lower operating expenses and reduced workload." independent studies show that the va system does better on many measures, especially preventive services and chronic care, than the private sector and medicare. va offi cials say "its [integrated] technology has helped cut down hospitalizations and helped patients live longer" (zhang ) . recently, the journal of obstetrics and gynecology reported on a tragic loss of life due to the systemic nature of oral health. a study found oral bacteria called fusobacterium nucleatum was the likely culprit in infecting a -year-old woman's fetus through her bloodstream (carroll ) . the doctors determined that the same strain of oral bacteria found in the woman's mouth was in the deceased baby's stomach and lungs. integrated records would provide critical data to the obstetrician including oral health issues and when the patient had her last dental exam. how does one measure the economic impact of a life not lived and another derailed by such tragedy? in a randomized controlled study, lopez et al. ( ) determined that periodontal therapy provided during pregnancy to women with periodontitis or gingivitis reduced the incidence of preterm and of low birth weight. the institute of medicine and national academies estimate that preterm births cost society at least $ billion annually . data integration of the iehr enables the effective management between the dentist and obstetrician to ensure proper periodontal therapy has been provided during pregnancy. such management based on the lopez et al. study, will have direct impact in reducing the prevalence per preterm births leading to reduced health care costs. there have also been studies indicating a correlation between poorer oral hygiene or defi cient denture hygiene and pneumonia or respiratory tract infection among elderly people in nursing homes or hospitals (rosenblum ; ghezzi and ship ; scannapieco ) . one such study of elderly persons in two nursing homes in japan (adachi et al. ) concluded that "the number of bacteria silently aspirating into the lower respiratory tract was lower in the group who received professional oral care, which resulted in less fatal aspiration pneumonia in that group." over the month period of the study, of the patients receiving professional oral care, % died of pneumonia versus . % of the patients that died of the same cause who maintained their own oral hygiene. lack of access is certainly a key factor to consider. however, lack of available data respective to the interrelationship between oral health and systemic health also contributed to the apathy in these cases. as identifi ed above, complications are correlated to cost. as conditions compound, costs go up. marshfi eld clinic, as part of their iehr is creating a shared problem list that identifi es both oral and medical conditions and history to recent visits and medication lists for monitoring at point of care [be it a medical or dental visit], such cross access to clinical data and care management milestones serves as a tool to prevent conditions from compounding and escalating costs such as those described above. several other areas of economic impact will be seen as iehr's become broadly deployed. some of these are listed as follows: medication management. a great deal of provider and allied support time is • spent obtaining medication information between dentistry and medicine [and vice versa] including current medications, contraindications, tolerances, etc. marshfi eld clinic cattails software has created a dashboard that readily identifi es this for both the medical and dental providers. not is time saved but chances for complications or escalation of conditions is reduced [both of which impact cost]. for example an integrated record allows medical providers treating respiratory infections to include or exclude oral fl ora as the possible source of the infection which would lead to more knowledgeable prescribing decision on the antibiotic used. coordination of care has a direct impact on cost for the system and the patient. • for example, in . % of the us population aged years and older that was diagnosed diabetes had been to the dentist in the past year (healthy people ( )). the us government's program healthy people includes an initiative to increase the proportion of people with diagnosed diabetes who have at least an annual dental examination. the american diabetes association recommends that diabetic patients be seen semi-annually and more if bleeding gums or other oral issues are present. the american diabetes association also recommends the consultation between the dentist and doctor to decide about possible adjustments to diabetes medicines, or to decide if an antibiotic is needed before surgery to prevent infection. the target from the healthy people is a % improvement at . %. integrated medical/dental records could allow for the coordination of efforts between providers to include communication of treatment plan and services leading to quicker resolution, increased patient compliance, and less patient time away from work or home and potentially less travel. similarly, integrated records also create a platform to integrate clinical appointing • between medicine and dentistry. as such, combative patients or severely disabled patients needing anesthesia in order for care to be delivered can be treated with one hospital sedation vs. multiple sedations. family health center of marshfi eld, inc. (fhc) dental clinics shares an iehr with marshfi eld clinic and uses it integrated scheduling feature to complete dental care, lab work, ent care, woman's health, preventive studies, all in one visit. follow up care management can be more focused and coordinated. for example, • without the knowledge or dental conditions, medical providers could spend months attempting to control diabetes with periodontal disease. however, with access to an iehr, the practitioner or allied care manager can determine patient's oral health status immediately to determine possible infl uence of periodontal disease. similarly an iehr with a shared patient data dashboard brings to light history • and physical examination data without having to have patients be the historian to their physician on their last dental visit or for the dentist to have to rely on the patient's recall of medications or medical diagnosis. for example, if an integrated record saved providers min per hour of patient care, that would be min per day. imagine giving a physician or dentist min more a day. in a capitated system, this allows for more patients to be seen in a day for roughly the same amount of expenditure. in a production based clinic this allows more patients to be seen and more charges per day. in either case, the investment into informatics is covered. in an underserved area, more patients get care quicker, which creates the opportunity for quicker resolution, which can lead to a healthier society, which in turn may lead them back to a productive livelihood sooner. an iehr results in one system for acquisition, orientation, training and support. • pc based owners who also own a mac and mac owners who also have to operate a pc can relate. need we say more? imagine if your pc function just like a mac [or your mac function just a pc]. no cross learning of software quirks. not having to purchase two separate units to begin with. reduced costs, increased space. not having to jump from one computer to the other computer to get data from one data from another to create a report. not having to call two separate computer companies for service or updates. third party coordination. having an iehr creates a platform for interfacing • with third party payers. a common system and language for timely reimbursement. in part, the result of an iehr is driving the diagnostic coding for dentistry. such an integrated interface provides a tool to bridge with healthcare payors that historically kept payment as segregated as the oral and medical health professions. the iehr overcomes that limitation. timely payment, consolidation of payment, expansion of covered patient and provider benefi ts based on clinical integration, and a viable system for interfacing are all potential economic benefi ts of iehr clinical data. the iehr creates new horizons for research that will lead to cost saving discov-• eries. as example, knowing the benefi ts of research, marshfi eld clinic research foundation (mcrf) has created an oral and systemic health research project (oshrp). the creation of oshrp, led by dr. murray brilliant, will allow mcrf to capitalize on its existing and growing strengths in the areas of complex disease interactions and personalized health care (phc) to advance oral health and the health of the rest of the body. the oshrp has three specifi c goals: understand the connections between oral and systemic health (diabetes, heart disease, pre-term births) understand the causes of oral diseases and determine the effect of genetics, diet, water source (well/city + fl uoridation) and microbiome. understand how improving oral health aids systemic health (comparative effectiveness) and bring personalized health care (phc) to the dental arena. the oshrp research resource will be unique in the nation. as mcrf has done • with other projects, it will share this resource with qualifi ed investigators at other academic institutions both within and outside of wisconsin. oshrp will advance scientifi c knowledge, improve healthcare and prevention, reduce the cost of oral healthcare, and create new economic opportunities. such knowledge will have a direct economic impact on the cost of care and care management. the iehr creates an ability to have an integrated patient portal to comprehen-• sively maintain their health. portals are becoming more and more popular in the healthcare industry as a means to helping maintain compliance with care management recommendations and preventative procedures. portals provide patents a tool to stay up to date on their care and recommendations. portals can take iehr clinical data, adapt it through programming, and provide creative visual reinforcement for patients as they monitor their health status. the more patients engage in owning their health status, the more preventative services are followed through with. the more medicine and dentistry can leverage the prevention potential [which insurance companies have come to realize] the more likely costly conditions can be avoided. the link between oral health and systemic health is well documented. the separation of dental and medical is not a sustainable model in modern healthcare delivery. a new model of integrated care is necessary. aristotle said, "the whole is greater than the sum of its parts." increased access to combined medical and dental histories and diagnosis at the providers' fi ngertips makes vital information available. shared diagnosis between physicians and dentists could aid in formulating interventions and to accelerate decision making abilities by allowing for prioritizing of medical/ dental procedures. clinical management and treatment of the patient would be expedited with immediate access to both records. quality could be improved through a complete picture of the patient through the dashboard. all of which have a direct or indirect economic benefi t. the iehr will be the tool that facilitates such delivery and the studies and scenarios described in these pages point to signifi cant economic benefi ts to patients, payors, and providers. if increased access, multi-provider monitoring, shared problems lists with enhanced decision making abilities from iehr could reduce healthcare costs. the greatest cost reduction will be with using the iehr to manage chronic disease. a combined dental-medical electronic record with a shared data informatics platform is most likely to yield the best long-term economic solution while maintaining or enhancing positive patient outcomes. this section reveals viewpoints from a variety of medical and dental providers. one section focuses on optimal use of ophthalmic imaging, which should show how that the challenges of clinical data integration go beyond those encountered in the effort to bring oral health and systemic health together. wendy e. mouradian , suzanne boulter , paul casamassimo , and valerie j. harvey powell oral health is an important but often neglected part of overall health. historically separate systems of education, financing and practice in medicine and dentistry fuel this neglect, contributing to poorer health outcomes for vulnerable populations such as children, while increasing costs and chances for medical error for all patients. advances in understanding the impact of oral health on children's overall health, changing disease patterns and demographic trends strengthen the mandate for greater integration of oral and overall healthcare, as reviewed in two recent institute of medicine reports (iom a, b ) . the pediatric population could realize substantial benefit from oral disease prevention strategies under a coordinated system of care enhanced by integrated electronic health records (ehr). this approach would benefit all children but especially young children and those from low socioeconomic, minority and other disadvantaged groups who are at higher risk for oral disease and difficulties accessing dental care. this section focuses on the pediatric population and the need for close collaboration of pediatric medical and dental providers. first we consider how a child's developmental position and their parents' level of understanding might affect oral health outcomes. next we address the importance of children's oral health and the urgency of seizing missed opportunities to prevent disease. we then briefl y outlines some steps to preventing early childhood oral disease utilizing some of the many health providers that interact with families. finally we examine one pediatric hospital's approach to choosing an integrated ehr technology. children have unique characteristics which distinguish their needs from those of adults. children's developmental immaturities may increase their risks for poor oral health outcomes ( fig. all children, but especially young children, are limited in their ability to care for their own health and must depend upon adults. a child's parent/caregiver may also lack basic oral health knowledge and an awareness of their child's oral health needs, and/or suffer from poor oral health themselves. low oral health literacy is prevalent among patients and health professionals alike in america; individuals of low socioeconomic status or from ethnically diverse backgrounds may be at particular risk for low oral health literacy (iom a ) . without appropriate education, a parent…. may not correctly interpret a child's symptoms or signs of oral disease • may not know that caries is an infectious disease that can be spread to a child by • sharing spoons, for example, may not know the potential value of chewing gum with xylitol, • may not fully grasp the importance of good oral health hygiene habits, • may not grasp the consequences of a child consuming quantities of sugared • foods or beverages, may have diffi culty controlling the child's consumption of sugared foods or bev-• erages in or out of the home, may not realize the consequences of chronic use of sugared medications, • may not know the potential for systemic spread of disease from a toothache, or • for liver damage due to overuse of acetaminophen or other analgesics, may not grasp the long-term consequences of early childhood caries, • may live in a community without fl uoride in the tap water and not know about • alternative sources of fl uoride, may overlook oral health due to the stress of living in poverty, • may be fearful of dentists or oral health care due to their own experiences, • may have diffi culty locating a dental provider accepting public insurance, or • have other problems navigating the health care system. parents in turn depend on access to medical and dental providers with current understanding of the most effective ways to prevent caries and promote the child's oral and overall health. an important element in helping families is the provision of culturally-sensitive care to a diverse population. children are the most diverse segment of the population with % from minority backgrounds compared with % of the overall population (us census bureau ) . the separation of medical and dental systems and the lack of shared information can create additional barriers for families, especially for those with low health literacy or facing linguistic or cultural barriers. all pediatric health professionals have increased ethical and legal responsibilities to promote children's health, including advocacy for them at the system level (mouradian ) . although many factors can infl uence children's oral health outcomes, caries is largely a preventable disease. despite this, national trends and other data on broader considerations of medical and dental data integration children's oral health attest to this persistent national problem ) . some important facts include the following…. caries is the most prevalent chronic disease of childhood, • caries is a preventable disease unlike many chronic diseases of childhood, • yet according to (nicdr ) % of children - have had dental caries in • their primary teeth; % of children - have untreated dental caries. further, " % of children - have had dental caries in their permanent teeth; % of children - have untreated decay." overall "[c]hildren - have an average of . decayed primary teeth and . decayed primary surfaces," the latest epidemiologic evidence shows increasing rates of caries for young-• est children, reverse from the healthy people goal of decreasing caries. according to (nicdr ), overall "dental caries in the baby teeth of children - declined from the early s until the mid s. from the mid s until the most recent ( ) ( ) ( ) ( ) ( ) ( ) ) national health and nutrition examination survey, this trend has reversed: a small but signifi cant increase in primary decay was found. this trend reversal was more severe in younger children." disparities in children's oral health and access to care persist by age, income • level, race and ethnicity, and parental education level (edelstein and chinn ) . of concern, the latest increase was actually in a traditionally low-risk group of young children (dye and thornton-evans ) . the human and economic costs of early childhood caries are substantial • (casamassimo et al. ) . according to catalanotto ( ) , health consequences include… extreme pain, spread of infection/facial cellulitis, even death (otto - ) diffi culty chewing, poor weight gain falling off the growth curve (acs et al. - ) risk of dental decay in adult teeth (broadbent et al. - ; li and wang ) crooked bite (malocclusion) -children with special health care needs (cshcn) may be at higher risk for oral • disease and diffi culties accessing care. analyzing data from the national survey of children with special health care needs, (lewis ) found that "cshcn are more likely to be insured and to receive preventive dental care at equal or higher rates than children without special health care needs. nevertheless, cshcn, particularly lower income and severely affected, are more likely to report unmet dental care need compared with unaffected children." children who were both low-income and severely affected had . times the likelihood of unmet dental care needs, dental care is the highest unmet health care need of children; . million children • had unmet dental care needs because families could not afford care compared with . million with unmet medical needs for the same reasons (cdc ) , according to the national survey of children's health, children are . times as • likely to lack dental as medical insurance (lewis et al. ) , there is evidence that children who get referred to a dentist early may have lower • costs of care and disease. savage et al. ( ) reported that children "who had their fi rst preventive visit by age were more likely to have subsequent preventive visits but were not more likely to have subsequent restorative or emergency visits" and concluded that preschool "children who used early preventive dental care incurred fewer dentally related costs," ramos- gomez and shepherd ( • ) , in their "cost-effectiveness model for prevention of early childhood caries," conclude that preventive ecc interventions could reduce ecc by - % for a particularly vulnerable population of children, and that part of the costs of interventions will be offset by savings in treatment costs. as these facts convey, and the deaths of more than one child from consequences of untreated caries make painfully clear, there is an urgent need for more attention to the oral health needs of children. a more coordinated system for oral health care including integrated ehr would be an important advance. a glance at table . , an ideal model, reveals that intervention should begin before birth and that a range of medical and oral health professionals can contribute to the child's oral health. early intervention is necessary because of the transmissibility of cariogenic bacteria from mother/caregiver to infant, and importance of oral health practice in preventing disease. the following professionals may be involved: • pediatric medical provider family physician pediatrician pediatric nurse nurse practitioner in pediatric/family practice physician assistant in pediatric /family practice - other appropriate allied health professionals • the availability of some of these professionals can be affected socioeconomic status, health insurance, place of residence, or by a child's special health care need. one obvious limitation on developing a "relay" as in table . , with a "hand-off" from family care to obstetric care to pediatric care is the education of the medical providers. as part of pre-conception and perinatal healthcare, providers should address oral health, but may lack the knowledge to do so. additionally, as noted by ressler-maerlaender et al. ( ) , "some women may believe that they or their table . timeline of some oral health interventions to prevent early childhood caries (ecc) -birth to years age (marrs et al. ; lannon et al. ; han et al. ; ezer et al. ; aap ; mouradian et al. ) child's age intervening professional(s) planning conception, prenatal and perinatal family physician the physician and/or obstetric provider educates mother-to-be about good maternal oral hygiene and infant oral health issues, including transmissibility of caries. mother's dentist assesses and treats caries, gingivitis or other oral health problems and educates the mother-to-be obstetrician/nurse midwife obstetric nurse general dentist obstetric nurse obstetric nurse advises new mother to chew xylitol gum, limit salivary contact between mother and infant, and help child avoid sugar intake (exposure) while asleep and from common sugar sources (medicines, sugared water, bottle feeding on demand at night with fl uid other than water -following tooth eruption, certain foods), and to schedule dental exam at year age months pediatric medical provider first dental examination recommended by aapd when the fi rst tooth comes in, usually between to months pediatric/general dentist educate mother about optimal fl uoride levels . assess the woman's oral health status, oral health practices, and access to a dental home; . discuss with the woman how oral health affects general health; . offer referrals to oral health professionals for treatment; . educate the woman about oral health during pregnancy, including expected physiological changes in the mouth and interventions to prevent and relieve discomfort; and . educate the woman about diet and oral hygiene for infants and children and encourage breastfeeding a combination of anticipatory guidance, with continuity from prenatal and perinatal care to pediatric care, can help move infant oral health from "missed opportunities" to "seized opportunities." others who may be of assistance to families in closing these gaps are professionals at the women, infants and children's (wic) supplemental nutrition program, early head start/head start and neurodevelopmental/birth to three programs. together medical, dental and community professionals can help create a system of care to improve maternal and child oral health. for the envisioned model in table . to be realized, the mother requires access to a general dentist with accurate information on her oral health during pregnancy and on her infant's oral health, including the need for an early dental visit. the mother and child then need access to a pediatric medical provider who will provide oral health screening/counseling, and who will guide the family to establishing the child's dental home by age . success in dental referral requires access to a pediatric or general dentist willing and able to provide infant oral health. (dela cruz et al. ) , in a discussion of the referral process mentioned that among the factors in assessing the likelihood of a dental referral were the medical providers' "level of oral health knowledge, and their opinions about the importance of oral health and preventive dental care." since young children are much more likely to access medical than dental care, the medical provider plays an important role in promoting children's oral health. (catalanotto ) recommends, as part of a pediatric well child checkup: an oral screening examination, • a risk assessment, including assessment of the mother's/caregiver's oral health, • application of fl uoride varnish • anticipatory guidance (parental education) including dietary and oral hygiene • information, attempted referral to a dental home. • the aap recommends that child healthcare providers be trained to perform an oral health risk assessment and triage all infants and children beginning by months of age to identify known risk factors for early childhood caries (ecc). the oral health component of pediatric care is integrated into the aap's "recommendations for preventive pediatric health care (periodicity schedule)" (aap ) . to what extent are medical and dental and providers aware of recommendations for a fi rst dental visit for a child by age one, as recommended by the aap, the american academy of pediatric dentistry (aapd), and the american dental association? (wolfe et al. ) reported that % of licensed general dentists in iowa were familiar with the aapd age dental visit recommendation and that most obtained the information through continuing education; % believed that the fi rst dental visit should occur between and months of age. however, according to (caspary et al. ) , when pediatric medical residents were asked the age for the fi rst dental visit, the average response was . years, while % reported received no oral health training during residency. in a national survey of pediatricians ) reported that less than % of had received oral health education in medical school, residency, or continuing education. finally (ferullo et al. ) surveyed allopathic and osteopathic schools of medicine and found that . % reported offering less than h of oral health curriculum, while . % offered no curriculum at all. other workforce considerations relevant to preventing early childhood caries include the training of dentists in pediatric oral health (seale et al. ) , the number and diversity of the dental workforce, the number of pediatric dentists, and the use of alternative providers such as dental therapists, expanded function dental assistants and dental hygienists (mertz and mouradian ; nash ) . examples of integrated care models do exist, such as that presented by (heuer ) involving school-linked and school-based clinics with an "innovative health infrastructure." according to heuer, "neighborhood outreach action for health (noah)" is staffed by two nurse practitioners and a part-time physician to provide "primary medical services to more than , uninsured patients each year" in scottsdale, arizona. heuer counts caries among the "top ten" diagnoses every year. mabry and mosca ( ) described community public health training of dental hygiene students for children with neurodevelopmental/intellectual disabilities. they mentioned that the dental hygiene students had worked together with school nurses and "felt they had impacted the school nurses' knowledge of oral disease and care." the decision to acquire an integrated ehr as pediatric clinicians (both medical and dental) work more closely together, they require appropriate ehr systems that integrate a patient's medical and dental records. following is a set of local "best practices" from nationwide children's hospital in columbus, ohio, which may help other children's hospitals in planning acquisition of an integrated pediatric ehr system. integrated (medical-dental) ehr technologies are becoming more widely available outside the federal government sector (see integrated models e and e in fig. . ). nationwide children's 'drivers' for the acquisition process were, in : . minimize registration and dual databases . patient registration takes time and requiring both a stand-alone dental and a medical patient registration inhibits cost-effective fl ow of services. integration allows for the use of single demographics information for all clinics in the comprehensive care system serving the patient. clinicians always have an updated health history on patients, if they have been a patient of record. if not, and for a dental clinic that sees walk-ins, a brief "critical" dental health history can be completed on paper by a parent and scanned into the emr. in designing an integrated medical-dental record for patients of record, the system can sort essential health history elements into a brief focused dental history without the detail needed by other medical specialty clinics. kioskdriven electronic health histories for those children who are new to clinic similar to those used in airline travel could be considered if feasible in busy clinics. . for charting, no more key/mouse strokes than with paper . some commercial dental record products try to accomplish too much. moving from paper to electronics should be driven in part by effi ciencies. the tooth chart, which is an essential part of any dental record, must be such that examination fi ndings can be transferred quickly and accurately to either paper or electronic capture. a helpful exercise is visualization of the functionality of the charting process, including both the different types of entries (caries, existing restorations and pathology) and how these are entered in the paper world. if charting will be able to be used for research the system should be able to translate pictures to numerical values, often a complex programming function. dental practitioners and faculty may want to use drawings of teeth or graphics of surfaces because that is their current comfort level. a true digital charting is possible with no images of teeth, but some habits are hard to change. . maximizing drop downs with drop down building possible . duplication of paper chart entries using drop downs which can be upgraded as more clinical entities are found is a staple of an emr. the paper process usually relies on a clinician's wealth of medical-dental terms since inclusion of every possible, or even the most common fi ndings, is prohibitive on a paper chart. the emr drop down requires front-end loading of the most common clinical fi ndings with opportunity for free-hand additions. being able to add terms to any drop down is a needed capability. . don't design a system for uncommon contingencies, but for your bulk of work . a pediatric dental record should be primarily designed around dental caries, with secondary emphases on oral-facial development (orthodontics) and a lesser capability to record traumatic injuries and periodontal fi ndings. these second and third level characteristics can be hot-buttoned and should not drive the design of the basic system which is caries charting for % of our patients. sadly in most dental schools, the chart is slave to every teaching form, few of which ever exit with the dds into practice! these forms may have little relationship to patient care and only create "signature black holes" that need to be addressed, usually after treatment is completed. . progress notes should be designed for the routine entries with free-hand modification possible . student learners tend to write too much and a carefully crafted progress note format with standard entries in required fi elds helps patient fl ow and record completion. in federally funded clinics and residencies, attending reconciliation of student/resident service delivery is a compliance requirement. a well-designed emr system can "stack" required co-signing tasks on a computer screen, offer standard entries as well as free-hand options, and create a process far faster than paper records for an attending's validation (same as reconciliation?). . tie examination results to treatment planning and treatment planning into billing . a good system allows easy transfer of clinical fi ndings needing treatment into some problem "basket" and ideally in a tabulated format. an alternative is a split screen that allows a clinician to visualize clinical fi ndings, radiographic fi ndings while compiling a treatment plan. again, in clinical settings where compliance to medicaid/medicare regulations is required, the design of the record should give attention to auditing principles and security. a good emr system allows portals of entry for billing and compliance personnel. . plan for users of different skill levels and different periods of exposure . the teaching hospital or dental school environment often involves learners and attendings with varying skill levels and computer experience who may be there for brief periods of time. this reality adds signifi cant security and userfriendliness issues. some medical record systems are far too complex for shortterm or casual users. a well-integrated medical-dental emr allows navigation of the depths of the medical side should a user want to explore, but should focus on the dental portion. some suggestions in design: initial opening or logging into the dental portion for dental users, rather than • opening into the medical portion, clearly indicated options for exploration of medical portions, • orientation of major dental component (examination, radiographs, treatment • plan) in a logical dental treatment fl ow to replicate the way dentistry works rather than trying to reshape dentistry's normal fl ow to the record, minimization of seldom-used functions on the main dental screen, such as • specialty medical clinics, old laboratory tests and hyperfunctionalities like letter writing, clear identifi cation of existing non-caries dental portions like orthodontics or • trauma, so a novice user need not randomly search to see if a patient has any of these records. unfortunately, many pediatric hospitals do not yet have an ehr system that supports convenient communication among a pediatric patient's medical and dental providers. evidence of this state of affairs was provided unintentionally by (fiks et al. ) . some pediatric hospitals may have an awkward mix of systems serving physicians, dentists, and orthodontists and their shared patients. this section demonstrates how closely medical and dental professionals must collaborate to deliver appropriate oral health care for infants and children. such collaboration is especially important given the developmental vulnerabilities of children and the urgency of the oral health needs of many children, especially those from underserved populations. collaboration is made more diffi cult by the long-standing separation of medical and dental systems and poor oral health literacy of parents and medical professionals alike. teamwork in the delivery of pediatric care requires appropriate electronic patient record technology to facilitate sharing of patient information, to avoid patient record discrepancies between systems, and to create effi ciencies by maintaining only a single repository for patient demographics. only comparatively recently have appropriate integrated systems become available to support a range of clinical sites from pediatric special needs clinics to the largest children's hospitals. nationwide children's has given practical examples of effi cient decision-making in identifying an integrated system to acquire. much more work will be needed to develop the means to move towards integrating offi ce and community-based care for children through the sharing of electronic health records. oral health is an oft neglected area in the care of patients who have chronic kidney disease. furthermore, the provision of care by dentists and physicians to the same patient is fragmented as communication between the two health care providers is scant. emerging data suggesting the periodontal disease is closely linked to chronic kidney disease highlights the importance of proper oral health and the importance of communication between dentists and physicians in the care of the patient. investigators used data from nhanes iii, including information on , adults who had an oral examination by a dentist who categorized each patient as having no periodontal disease, periodontal disease or edentulous to examine the relationship between numerous risk factors for moderate to severe chronic kidney disease, as determined by calculation of estimated gfr through use of the mdrd formula (fisher et al. ) . no chronic kidney disease was defi ned as an estimated gfr of ml per min per . m . three percent of the patients had ckd, . % were hypertension and . % had diabetes ( . % with glycated hemoglobin of % or higher). four models were constructed to examine the potential relationship between periodontal disease and ckd. in model one adults with either periodontal disease or edentulous had an adjusted odds ratio of . (with % confi dence intervals of . - . ) of having ckd, independent of the other risk factors for ckd including of age above years, ethnicity, hypertension, smoking status, female gender and c-reactive protein elevation. the fourth model contained potential risk factors including the periodontal disease score and for every -unit increase in the score, the risk of having ckd increased by % controlling for the other risk factors. the authors hypothesized from their results that the relationship between periodontal disease and ckd was bidirectional in that ckd may increase the risk of periodontal disease which in turn increases the risk of ckd. grubbs et al. ( ) also used nhanes data to look more closely at the relationship between periodontal disease and ckd, using dental examinations obtained from to (n = , adults, - years) (grubbs v, et al. ) . in this analysis edentulous subjects were excluded and those with albuminuria were included in the defi nition of ckd. in the entire population ckd was present in . %, but in those with moderate to severe periodontal disease this increased to . %. other associations with moderate to severe periodontal disease were being older, male, nonwhite, less educated and poor. there was a strong relationship between periodontal disease and ckd ( . unadjusted odds ratio). when adjusted for age, gender, tobacco use, hypertension, diabetes, ethnicity, poverty and educational attainment, the odds ratio for the association of periodontal disease and ckd was still signifi cant ( . ). in some groups (mexican american, poor, and poorly educated) dental care was not received on an annual basis in the majority of this segment of the population. periodontal disease has been associated with an increased risk of death in hemodialysis patients (kshirsagar et al. ). this relationship has been poorly studied in peritoneal dialysis patients. this requires further study but it appears possible that periodontal disease might hasten loss of residual kidney function and perhaps contribute to atherosclerosis in dialysis patients and therefore, contribute to the high mortality in this population. patients who desire a kidney transplant are required to undergo a thorough evaluation beforehand including an oral examination by a dentist. some patients on dialysis have inadequate insurance which does not cover dental care, leading to a situation in which a kidney transplant is denied because the patient cannot afford the dental examination. communications between dentists and physicians in the care of the patient is scant. if oral surgery is required in a dialysis patient, the surgeon generally requires a brief summary from the nephrologist with recommendations. these might include suggestions for prophylactic antibiotics, avoidance of vasoconstrictor agents to an excess locally (which can elevate blood pressure) and the increased risk of bleeding of a dialysis patient. for more routine dental examinations no information is requested which could potentially lead to drug interactions or a dangerous situation. most nephrologists and health care providers in the dialysis unit do not inquire of the patient concerning dental health and examination of the mouth is quite uncommon. although the dialysis patient is seen monthly at a minimum, there is little conversation or documentation of oral health. connecting the electronic health records of in-patient care, the out-patient dialysis unit and the dentists' offi ce could potentially have a large impact in improving the care of those with end stage kidney disease. integrating medical and dental records in ehr's may or may not be the "golden ring." first, we need to integrate the clinical thinking…something we both realize is important, but not likely to be solved by an inert computer. i also think that integrated records will be very cumbersome, given the fact that the language used by the separate disciplines is so different, and the kind of detail required to support good decisions and good work is so different. it could be done…but for many professionals on either "side," they would never open the other module. to me, a more sensible solution may be to have a condensed "nugget" of information that could cross populate. "moderate periodontal disease" may be what the medical doctor needs to know, plus know what a treatment plan may include. she won't need to know the number of the teeth with the deepest pockets and erosions but will need to support the patient's determination to follow through. on the other hand, if the patient has shown remarkable initiative in gum care and has successfully migrated to a lower severity index, that would be important for congratulation and reinforcement…and also to encourage similar diligence in managing, let's say, the hypertension that is not optimally controlled. in the other direction, the dentist should know that a patient has been erratic in clinical follow up, does not self-test blood glucose, uses hypoglycemic drugs only intermittently, and has failed several appointments for eye exams. this would lead to a rather different set of approaches from a highly motivated grandmother who is enrolled in a community cultural center's senior exercise club, and is learning to become a lay community teacher for diabetes. right now, i don't think even this superfi cial degree of information is exchanged. we need to support each other's efforts, but we probably do not need to share minute details. the benefi ts of an electronic health record are well described. ehrs allow for legible standardized documentation and easier sharing of patient data between providers at multiple locations. they are less prone to loss and require much less space to store. they have the potential to result in a reduction in the cost of health care. a distinct disadvantage of the ehr, in its current confi guration, is the problem of information overload. simply put, there is often too much information presented in a way that is diffi cult to review and digest. the ehr equivalent of thumbing through a chart quickly is not yet available. as a result we frequently see practitioners look only at the last note or two as they review a patient's history. we require a way to communicate information directly relevant to patient diagnosis, treatment and prognosis among subspecialists and primary care providers. we require a way to identify subclinical cerebrovascular disease in a patient, independent of blood pressure and other traditional risk factors. we require a way to recognize which patients with cerebrovascular disease are two to four times more likely than average to develop a stroke in the next years. we have a way -retinal imaging. the eye is the one place in the body we can directly observe arteries, veins and a cranial nerve in a noninvasive manner. routine imaging of the retina and optic nerve could allow primary care providers to assess retinal, and by proxy systemic, end organ damage from atherosclerosis in an effi cient manner. the key to optimal use of the medical record and effi cient yet effective communication among providers may lie with the familiar adage; a picture is worth a , words. traditionally, when ophthalmologists communicate with primary care providers they send brief letters regarding the fi ndings seen during a yearly dilated examination and the presence, absence or progression of diabetic retinopathy. these letters end by exhorting the virtues of improved blood sugar, blood pressure and lipid control, a sentiment that the primary care provider likely shares. this system of communication does not provide particularly useful information for the primary care provider, except to serve as a notice that the standard of care screening guidelines have been met. the box has been checked. if primary care providers, cardiologists, nephrologists had access to routine ophthalmic imaging, they would be able to directly visualize the effect that suboptimal blood sugar control is having on their diabetic patients. as importantly, they would be equipped with information directly predictive of congestive heart failure, stroke, and cardiovascular mortality for their patient with hypertension, hyperlipidemia and for those who smoke. large clinical studies have shown that assessment of retinal vascular changes such as retinal hemorrhages, microaneurysms and cotton wool spots provides important information for vasculopathy risk stratifi cation. as an example, wong et al. showed that the presence of retinopathy indicates susceptibility to and onset of preclinical systemic vascular disease, independent of and qualitatively different from measuring blood pressure or lipids (wong and mcintosh ) . in the atherosclerosis risk in communities (aric) study, individuals with hypertensive retinopathy signs such as cotton wool spots, retinal hemorrhages and microaneurysms were two to four times more likely to develop a stroke within years, even when controlling for the effects of blood pressure, hyperlipidemia, cigarette smoking and other risk factors (wong et al. ) . in a recent study by werther et al., patients with retinal vein occlusions were found to have a two-fold increased risk of stroke compared to controls (werther et al. ) . in addition, the aric study group reported that individuals with retinopathy were twice as likely to develop congestive heart failure as individuals without retinopathy, even after controlling for pre-existing risk factors (wong et al. a ) . interestingly, even among individuals without pre-existing coronary artery disease, diabetes or hypertension, the presence of hypertensive retinopathy was associated with a three-fold increased risk of congestive heart failure events (wong et al. a ) . in the beaver dam eye study, cardiovascular mortality was almost twice as high among individuals with retinal microaneurysms and retinal hemorrhages as those without these signs ( wong et al. a, b ) . the aric and beaver dam eye studies have also shown that, independent of other risk factors, generalized retinal arteriolar narrowing predicts the incidence of type ii diabetes among individuals initially free of the disease (wong et al. a (wong et al. , b . a primary care provider with access to patients' retinal photographs may therefore have the evidence needed to suggest which patient with either established systemic vascular disease or preclinical systemic vascular disease requires a more aggressive treatment and risk factor modifi cation. they could do this without wading through the electronic equivalent of piles of records. one photograph could refl ect both acute changes in blood pressure (retinal hemorrhages, microaneurysms and cotton wool spots) and chronic changes resulting from cumulative damage from hypertension (av nicking and generalized arteriolar narrowing) (sharrett et al. ; wong et al. a ; leung et al. ) . in brown et al. out of patients, excluding those with known diabetes, that presented with a single cotton wool spot or a predominance of cotton wool spots on examination of the retina were found to have underlying systemic disease (brown et al. ) . systemic work-up revealed diagnoses including previously undiagnosed diabetes, hypertension, cardiac valvular disease, severe carotid artery obstruction, leukemia, metastatic carcinoma, systemic lupus erythematosus, aids and giant cell arteritis (brown et al. ) . these fi ndings illustrate the importance of retinal fi ndings on a systemic level. the utilization and integration of ophthalmic imaging may serve to achieve more effective communication among subspecialists and primary care providers and ultimately to provide improved diagnosis and treatment for delivery of optimal quality of patient care. moreover, the improved integration and maximal use of resources may serve to reduce overall health care cost and perhaps decrease provider frustration with the electronic health record (fig. . ). there are cotton wool spots, exudates, intraretinal dot-blot hemorrhages and microaneurysms. av nicking is also present especially along the superior arcade just as the vessel leaves the optic nerve ( fig. . ) . av nicking, tortuosity of vessels, intraretinal hemorrhages and dry exudates are seen ( fig. . ) . there is edema of the optic nerve head, with cotton wool spots and fl ame shaped hemorrhage along the disc margin. there are several cotton wool spots along the vascular arcades and scattered dot hemorrhages throughout the posterior pole and periphery ( fig. . ) . notice the cholesterol plaque in the vessel just as it exits the optic nerve head and the pallor in the superior macula corresponding to retinal ischemia and edema ( fig. . ). the cholesterol embolus has resulted in lack of blood fl ow to the superior arcade ( fig. . ) . there is pooling of subretinal blood just superior to the optic disc with a central fi brin clot and associated vitreous hemorrhage (fig. . ) . optic disc edema, fl ame hemorrhages and venous congestion are seen in a patient with severe hypertension. biju cheriyan in clinical practice, an otolaryngologist often needs a dental consult not only because of the topographically adjacent nature of the structures but also because most structures are supplied by the same neurovascular bundle and therefore there is overlapping of symptoms. the converse scenario can also apply. apart from this, there are many systemic medical conditions (for example: bleeding diatheses, diabetes) a hypertensive optic neuropathy dentist encounters throughout his or her practice which can determine the outcome of a successful treatment. sometimes, providers may observe a cluster of diagnostic criteria which may have to a single source. in the sections below, i will explore a few of these scenarios and conditions, and indicate where and how an integrated electronic health record (ehr) could optimize delivery of health care by dentists and otolaryngologists. cleft palate/cleft lip : cleft lip and cleft palate (cl/cp) are congenital conditions that require multidisciplinary management by dentists, oral and maxillofacial surgeons, orthodontists, otolaryngologists, speech pathologists and plastic surgeons a number of studies report that a multidisciplinary approach is essential for better treatment outcomes (wangsrimongkol and jansawang ) and for post operative rehabilitation (furr et al. ). these multidisciplinary approaches may lead to new ways to manage and treat cl/cp patients (salyer et al. ). hutchinson's teeth : notching of the upper two incisors is typically seen in individuals infl icted with congenital syphilis. macroglossia refers to enlarged tongue in relation to oral cavity. macroglossia is an important sign. it can indicate important systemic diseases like systemic amyloidosis, congenital hypothyroidism, acromegaly, or down syndrome. a common complaint that dentists and otolaryngologists encounter in their practice is the common headache. because of the special nature of the neurovascular bundle of the head and neck this symptom can be presented to both dentists and otolaryngologists (ram et al. ). any sinus pathology can present as a headache to an otolaryngology practice. since the maxillary sinus fl oor is in close proximity to the maxillary premolars and molars, it is imperative to obtain a dental evaluation in persistent cases of headache. there are a number of causes for headache from the dental and otolaryngology perspective. a mal-aligned denture patient with chronic headache, whom i saw in my practice was shuttled between departments and an array of investigations only to fi nd at the end that an ill-fi tting denture caused the intractable headache. in these cases, an integration of fi ndings is extremely important in providing quality treatment to the patient and also saves money and time for the whole health care system. hence it is important to have an integrated patient record for this particular symptom alone. trigeminal neuralgia is facial pain of neurogenic origin experienced along the distribution of the trigeminal nerve(fi fth cranial nerve). it can present as a dental pain and can also be triggered by brushing teeth among other trigger factors. as a result, patients with dental pain without obvious causes are required to have a physicians' consultation to rule out this obscure condition. sometimes it is diagnosed by omission (aggarwal et al. ; rodriguez-lozano et al. ; spencer et al. ). any tumor of the nasal sinuses (specifi cally maxillary and ethmoids) can erode the lower bony wall and present in the oral cavity (usually the maxillary arch) as dental pain, loose tooth, etc. therefore, these are areas of interest to both dentists and otolaryngologists. such tumors most commonly present fi rst to a dentist or could also be an accidental fi nding. cancers of the naso/oro/laryngo pharynx can also present as toothache to a dentist as these structures have a common nerve supply from cranial nerves , and . therefore, an integration of the patient record may even help in early diagnosis of the tumor. the same principle applies to all oral tumors, tumors of the nasopharynx, the oropharynx etc. this is especially true of malignant lesions of the oral cavity as these may help in early detection and treatment of cancer. in these cases, an early biopsy and histopathology can save the life of the patient. therefore, it is imperative to say that a collaborative patient record can save patients' lives. ulcers of the oral cavity from aphthous ulcers to carcinomas can present both to a dentist and an otolaryngologist. oral ulcers can be of dental origin. contact ulcers from sharp edges of a mal-aligned tooth can result in intractable ulcers, where a simple smoothing of sharp edges may eradicate the ulcer and terminate it as a chronic condition and can even prevent the ulcer turning into a malignancy. if you have an integrated electronic health record (ehr) these problems are immediately addressed and managed. otherwise, the condition will consume valuable time of both the patient and the physician concerned. in addition to this, there are a few conditions which require special attention: aphthous stomatitis (canker sore), which may indicate oral manifestation of defi ciencies of iron, vitamin b , folate deficiency and oral candidiasis, which can be a sign of diabetes mellitus or of an immunocompromised patient (e.g. aids). temperomandibular joint (tmj) disorders can present in a variety of symptoms to both dentists and otolaryngologists. they can present as a headache, earache, toothache, or as facial pain. there can be a number of causes for this including osteoarthritis of the tmj, recurrent dislocation, bruxism, or even an ill fi tting denture. there have been cases where patients have been subjected to removal of teeth for chronic toothache only to discover at the end that the symptom was a referred pain from tmj! therefore, an integrated ehr can prevent misdiagnoses and resulting impairment or disability to patients. trismus (lock jaw) can indicate important diagnoses such as tetanus and rabies.it is due to a spasm of muscles of mastication, which is an important oral manifestation of widespread muscle spasm. apart from these conditions, other causes of trismus are peritonsillar abscesses, and scleroderma. other problems dentists and otolaryngologists encounter in clinical practice are concurrent systemic diseases (patients with multiple problems): patients with bleeding diatheses, diabetes mellitus and a hidden primary malignancy. a non-healing ulcer in the oral cavity may hide a primary malignancy behind it. in these cases, you have to look for it specifi cally. similarly, one has to be aware of oral manifestations of internal pathology. some of them are crohn's disease, ulcerative colitis and gastro-intestinal tract malignancies. often dentists see patients after a tooth extraction with intractable bleeding to fi nd that they have a bleeding diathesis. so, this may be the fi rst presentation of these patients' bleeding disorder. when this patient undergoes any elective procedure in future, it will be a great help to surgeons to be aware of this information to prevent any inadvertent complications. therefore an integrated ehr can prevent unwanted complications where a patient's life may be in jeopardy. the source of otalgia or earache can be from a number of sites other than ear itself. technically ear lobe and ear canal are supplied by four different cranial nerve branches ( th, th, th, th). therefore, an area with a common nerve supply can present as earache. common dental problems which present as referred otalgia are ( ) dental caries ( ) oro-dental diseases or abscesses ( ) an impacted molar tooth (which is a common cause) ( ) malocclusion ( ) benign and malignant lesions of oral cavity and tongue (kim et al. ) . therefore, it is essential these two departments collaborate with each other in diagnosing and treating these diseases, and one way of facilitating it is through an integrated ehr system. there is a lot of overlap between dentists and otolaryngologists in the diagnosis and treatment of patients with halitosis (delanghe et al. ; bollen et al. ) . poor oral hygiene is the most common cause for this common complaint. oral causes include tooth caries, oral ulcers, periodontal diseases, unhealthy mucosa of the oral cavity. it is interesting to note that a simple oral ulcer can form an abcess eroding the fl oor of mouth and becoming a life-threatening oral cellulitis (ludwig angina). once the cellulitis has developed, it becomes a medical emergency. therefore, it is essential to prevent it before it can progress into a life-threatening condition, which of course is possible. causes pertaining to otolaryngologists include: chronic sinusitis or mucociliary disorder, chronic laryngitis or pharyngitis, pharyngeal pouches-related pathology, tumors or ulcers of naso/oro/laryngopharynx, diseases or conditions that impair normal fl ow of saliva such as salivary gland diseases or stones preventing fl ow of saliva, medications which cause dryness of mouth: antihistamines, antidepressants; local manifestation of systemic disorders: auto immune disorders, sjögren syndrome, dehydration from any cause, diabetes mellitus and gastro esophageal refl ux disorder (gerd). gerd is caused by improper neuro-autonomy of the lower esophageal sphincter (les). the les does not close tightly after food intake which causes gastric content to enter the esophagus. over time this can erode mucosa and cause various diseases even becoming cancerous (friedenberg et al. ). this disorder is attributed to life style. fast food consumption habits (oily fried foods) and eating habits (swallowing food without properly chewing) are partly responsible for this disorder (lukic et al. ; al-humayed et al. ) . here again an early diagnosis can manage the disease process before it is fully developed. at present there are no integrated ehr systems serving these specialties (dentistry and otolaryngology). an integrated ehr would facilitate effi cient communication between a dentist and an otolaryngologist who are providing care to the same patient and addressing a problem with a shared focus between the two disciplines. such integrated communication, may only require consulting the available medical or dental record of the patient, based on the particular circumstance. even enabling this simple communication would avoid duplication of effort, clarify the context of certain symptoms and reduce stress endured by the patient. it also has the potential to reduce healthcare delivery costs, and in some cases, even contribute to saving the patient's life. henry hood, allan g. farman, and matthew holder in this chapter, the authors attempt to put forth a justifi cation for precisely this kind of collaborative approach through a summary and discussion of a series of actual clinical cases. the protocols discussed in the management of each of these clinical cases illustrate the value in providing whole-person, interdisciplinary health care to this complex patient population. there is arguably no single patient population for whom the provision of collaborative, interdisciplinary health care is more challenging than for patients with neurodevelopmental disorders and intellectual disabilities (nd/id). in planning and delivering the generally-accepted standard of health care to this unique population, myriad biomedical, psychosocial and sociopolitical realities converge to create a landscape that is, at best, daunting for patients with these disorders, and for the clinicians who are charged with their care. anecdotal and scientifi c evidence suggest that this landscape has produced a paucity of physicians and dentists who are willing and able to provide care to patients with nd/id, and that american medical and dental schools are providing little training focused on their care (holder et al. ; wolff et al. ) . in february of , th surgeon general david satcher issued a report, which documented that americans with nd/id experience great diffi culty accessing quality health care (thompson ) . in that same report, former health and human services secretary tommy thompson said, "americans with mental retardation and their families face enormous obstacles in seeking the kind of basic health care that many of us take for granted." (thompson ) the disparities identifi ed by dr. satcher and secretary thompson require that physicians and dentists approach this population in a spirit of collaboration, compassion, and teamwork in order to produce positive health outcomes for them. perhaps, an even greater imperative driving the need for collaboration between medicine and dentistry in this arena is the fact that many patients with intellectual disabilities have developed this cognitive impairment as the result of an underlying neurodevelopmental disorder that is often undiagnosed. and it is this neurodevelopmental illness and the constellation of potentially devastating complications associated with that illness that create a biomedical fragility and a vulnerability that neither begins nor ends at the oral cavity, and that leaves these patients at risk in almost every aspect of their daily lives. when, for example, patients with nd/id are dependent upon publicly-funded programs for their health care, and when these systems fail to provide the health services that biomedically complex cases require because they fail to account for and accommodate the link between medical and dental pathologies, the risk of a negative outcome is greatly enhanced. such was the case for an intellectually disabled woman in michigan who, in october of , was unable to access dental services through the state's public medical assistance program, and who fatally succumbed to a systemic bacteremia resulting from an untreated periodontal disease (mich. dent. assoc. ). the american academy of developmental medicine and dentistry (aadmd) defi nes a neurodevelopmental disorder as a disorder involving injury to the brain that occurs at some point between the time of conception and neurological maturationapproximately age or (zelenski et al. ). examples of frequently-encountered neurodevelopmental disorders would include fragile x syndrome, a genetically acquired neurodevelopmental disorder caused by a mutation at the distal end of the long arm of the x chromosome (see fig. . ), trisomy , another genetic disorder, which features extra genetic material at the chromosome site (see fig. . ), and cerebral palsy, a prenatal or perinatal, acquired neurodevelopmental disorder (see fig. . ). patients with neurodevelopmental disorders tend to present clinically with one or more of fi ve frequently-encountered, objective symptom complexes or primary complications. these fi ve, classic primary complications include intellectual disability (aka: mental retardation), neuromotor impairment, seizure disorders, behavioral disturbances, and sensory impairment (aadmd). additionally, multiple secondary health consequences can derive from the fi ve primary complications; and any one of these secondary health consequences, or a combination of them, can produce profound morbidity. an example of a common secondary health consequence seen in patients with nd/id, which is derived from intellectual disability and / or neuromotor impairment, is the patient who is unable to care for his or her own mouth, and who develops ubiquitous caries and advanced periodontal disease as a result (see: fig. . ). another example would be the patient who suffers from the secondary health consequence of gastroesophageal refl ux disease (gerd) as a result of the neuromotor impairment associated with multiple neurodevelopmental disorders; and whose tooth enamel and dentinal tissues become chemically eroded as a result of the chronic intraoral acidity produced by gerd (see: fig. . ) . the diagnosis and management of these secondary health consequences provide dentists and physicians with a unique opportunity to work together to improve the quality of health and quality of life for their patients by implementing a team approach, which crosses the traditional interdisciplinary lines of communication, and which expands each clinician's ability to make meaningful treatment options available. indeed, it is often the case that quality primary care provided in one discipline will provide potentially valuable information to an attending clinician from another discipline. such is the case with the patients featured in figs. . and . . the patient whose intraoral photograph is featured in fig. . is a year-old male patient who presented to a special needs dental clinic accompanied by his mother. the mother indicated that her son was exhibiting hand-mouthing behaviors that she believed suggested he was experiencing mouth pain. a comprehensive radiographic and intraoral exam revealed, among other maladies, notched incisors, multiple diastemas, grossly decayed mulberry molars, and advanced periodontal disease. the patient also exhibited moderate to severe intellectual disability. these fi ndings were all consistent with a diagnosis of congenital syphilis. however, in developing the medical history with the mother, it was learned that no previous diagnosis of syphilis had been discussed with the mother, nor was it included in the health history. in cases like this, a comprehensive dental treatment plan should always include consultation with the primary care physician for purposes of moving forward with confi rmation of the clinical diagnosis by serologic testing, and consultation with a cardiologist to assist in the management of potential cardiovascular sequelae. as the dental treatment plan is being developed, consideration should also be given to human immunodefi ciency virus (hiv) testing for this patient, as coinfection is a common fi nding . this issue could easily be attended to by a primary care physician, an internist or an infectious disease specialist. in the absence of any of these team members, the dentist should feel entirely comfortable ordering hiv testing. the primary care physician and the developmental dentist should continue to advise each other and their respective consultant specialists of any signifi cant developments or new information, which could in any way impact either the medical or the dental treatment plan. as treatment progresses, both the physician and the dentist should expect improvement in the patient's periodontal status, which will likely be refl ected in a decrease in the frequency of immune-related illnesses, and in the maladaptive behaviors produced by chronic oral pain. it is quite often the case in this patient population that, with a reduction in maladaptive behaviors, comes a reduction of the use of psychotropic medications prescribed in a frequently futile attempt to manage behaviors that were born of an undiagnosed medical or dental illness. gerd is defi ned as the refl ux of gastric contents into the esophagus. gerd is primarily associated with incompetence of the lower esophageal sphincter; however there are numerous co-contributors, which may predispose a patient to gerd or exacerbate an existing refl ux problem. these co-contributors include a diet high in fat, neuromotor impairment associated with functional abnormalities such as dysphagia, neuromotor impairment associated with impaired ambulation and prolonged periods of recumbence, and the use of multiple medications including anxiolytics, calcium channel blockers, and anticholinergics. gerd is thought to affect approximately - % of the general us population. it has been established in the literature that the incidence of gerd in patients with intellectual disabilities is signifi cantly higher than in the neurotypical population, and that the relative number of unreported cases of gerd is much higher in patients with a neurodevelopmental diagnosis, as well. patients who have gastric refl ux as a function of a neurodevelopmentally-derived neuromotor impairment and a coexisting intellectual disability are impaired in their ability to voice the complaint that would, in the neurotypical patient, commonly lead to an encounter with either a family physician or a gastroenterologist and, ultimately, to a diagnosis. this inability to voice a complaint can be problematic in that, left untreated, gerd can produce maladaptive and sometimes aggressive behaviors in this population. and, of even greater concern, is the fact that undiagnosed esophageal refl ux can lead to more complex conditions that can produce signifi cant morbidity or even mortality -maladies such as barrett's esophagus or adenocarcinoma of the esophagus. chronic gerd can also produce an acidic intraoral environment, which can lead to the chemical erosion of the enamel and dentinal tissues of the teeth. ali et al. have established a link between erosion of the enamel and dentinal tissues of the teeth and gerd. there is additional anecdotal evidence suggesting a link between tooth enamel erosion and gerd, and related maladies. a special needs dental clinic in the eastern united states serving , patients with nd/id, has reported that, of nine patients referred to gastroenterology who presented for dental exam with a fi nding of either tooth enamel erosion or ubiquitous caries, two cases were diagnosed with gerd, two with barrett's esophagus, three with gastritis, and one with duodenitis. in all cases, medical treatment was required. in light of all that is known about the incidence of gerd and of the gerdrelated risks unique to this patient population; and in light of the link between tooth enamel erosion and gerd, it is incumbent upon any dentist encountering tooth enamel erosion in a patient with an intellectual disability to immediately refer that patient to gastroenterology for a work up, which should include esophagogastroduodenoscopy (egd) and ph monitoring. a dentist encountering gerd in a patient with an intellectual disability must be aware that he or she may be the fi rst and only link between that patient and the diagnosis of a potentially life-threatening illness. phenytoin-induced gingival enlargement can appear as either an infl ammatory lesion or a more dense, fi brotic hyperplastic lesion. the infl ammatory lesion is one in which the gingival tissues are swollen and bleeding, and in which pain is often a component. this type of gingival enlargement is the more acute lesion, frequently seen in patients who are currently taking phenytoin. in advanced cases of infl ammatory gingival enlargement, the tissues can appear botryoid, with a characteristic grape-cluster appearance. in advanced cases of phenytoin-induced gingival enlargement, the lesion can sometimes shroud entire sections of the dentition. phenytoin has long been a common medication used to treat seizure disorders in patients with neurodevelopmental disorders and intellectual disabilities. however, the gingival enlargement it produces, and the obstacle this lesion can pose to effective oral hygiene -especially in a population in which oral hygiene is typically compromised -can, over time, lead to periodontal disease, edentulism, and in advanced cases, systemic bacteremias. gingivectomy performed to reduce phenytoin-induced gingival enlargement will typically fail unless the patient is weaned off the offending medication, and another anti-seizure medication is titrated to effect. multiple alternative anti-seizure medications are currently available, which do not have the side effect profi le of phenytoin, and most patients who are weaned off phenytoin will demonstrate a virtual % resolution of the infl ammatory lesion within a matter of or months. the image in fig. . is of a year-old, microcephalic african-american male with intellectual disability, neuromotor impairment, and a seizure disorder. figure . illustrates the appearance of this patient's gingival tissues while he was currently on phenytoin. figure . features the same patient months after being weaned off phenytoin and placed on topiramate. these images illustrate the dramatic result that can be achieved when a dentist and a physician work in collaboration in the best interests of the patient. it is worth noting that this particular collaboration required only one intervention to achieve this result: the patient was weaned off phenytoin and was placed on a safer alternate anti-seizure medication. any dentist caring for a patient with an intellectual disability who presents with phenytoin induced gingival enlargement should immediately contact either the primary care physician or neurologist managing the patient's seizure disorder, and strongly urge that the patient be weaned off phenytoin and placed on a safer alternative anti-seizure medication. edentulism and bacteremia need not be a side-effect of a seizure management protocol. the patient seen in fig. . is a year old male patient with idiopathic intellectual disability who presented to an outpatient dental clinic for comprehensive dental evaluation and treatment. he was accompanied by his father. his father was referred to the clinic by the staff at his son's day program workshop. the day program staff had observed hand-mouthing behaviors, and they had voiced concern that the patient may be in pain. in the waiting room, the patient exhibited behaviors consistent with neurodevelopmental dysfunction. he was non-communicative, and his gaze aversion and tactile defensiveness were suggestive of autism. he was resistant and somewhat combative when directed to the dental chair, and effective behavior management in both the waiting room and operatory required the combined efforts of his father and two staff fig. . the adult patient suspected of having fragile x syndrome members. the patient's health history was positive for attention defi cit hyperactivity disorder (adhd), and there was no history of seizure or neuromotor impairment. the father indicated that, at age ten, the patient was admitted to an inpatient psychiatric unit for evaluation of his uncontrollable behavior. the following day, the parents were told that managing the patient's behavior was beyond the ability of the psychiatric unit staff, and the parents were asked to take the child home. the father also indicated that the psychiatric unit staff described the child's behavior as overwhelming. the patient was last seen by a dentist years prior to presentation; examination and treatment at that time were carried out in the operating room under general anesthesia. effective oral examination of this patient required utilization of papoose board and molt mouth prop. multiple options for behavior management, including utilization of general anesthesia in the operating room, were discussed with the father, and informed consent to utilize medical immobilization techniques for purposes of this examination was obtained and documented prior to taking the patient into the operatory. in the operatory a dental examination was performed, and a baseline panel of digital radiographs was obtained. the head and facial features of this patient were suggestive of fragile x syndrome (see: fig. . ) . the body of the mandible was somewhat elongated; the nose was prominent; the head had somewhat of a triangular shape, and the patient readily averted his gaze. upon further inquiry, the father reported that the patient also exhibited macroorchidism, although he indicated that no physician or dentist had ever suggested a work up for fragile x. fragile x syndrome is a disorder with which many clinicians are unfamiliar. yet it is the second leading genetic cause of intellectual disability in the united states, and it is the leading known cause of autism in the u.s. in addition to the phenotypic fi ndings noted in this case, there are other frequently-encountered physical characteristics consistent with fragile x that may move a clinician toward this diagnosis. they include pectus excavatum or funnel chest (see fig. . ) and joint laxity (see fig. . ) . gaze aversion, as previously mentioned, is a typical fi nding in autism and in fragile x syndrome. indeed, in conjunction with non-verbal behaviors, gaze aversion is often the fi nding that initially alerts the clinician to the possibility of a neurodevelopmental diagnosis featuring autism as a complication. figure . features a photograph of fi ve children at a school for children with special needs. four of the children have been diagnosed with autism, and a fi fth child is a neurotypical child who was visiting his brother on the day the photograph was taken. the reader is left to decide which child is the neurotypical child. any physician or dentist who encounters a patient with an obvious intellectual disability, who does not have an established underlying neurodevelopmental diagnosis, and who presents with additional fi ndings, which may include gaze aversion, shyness, a prominent chin, pectus excavatum, a large nose or large ears, should suspect a possible fragile x diagnosis. the primary care clinician -physician or dentist -should discuss with the guardian or family member the importance of establishing a neurodevelopmental diagnosis. the family member or guardian should be informed that genetic counseling should be made available to all members of the extended family, since fragile x syndrome is a genetic disorder that can be passed from parents to offspring. once this discussion has taken place, a referral to a geneticist for a complete genetic work up is indicated. both the dentist and physician should feel entirely comfortable making this referral. in remote areas where the services of a geneticist may not be available, the attending physician or dentist may order a high resolution chromosomal analysis and a fragile x dna test, and have those results sent to a remote location for interpretation by a geneticist. consultation with a psychiatrist or a clinical psychologist may also be advisable, as patients with fragile x can sometimes experience enhanced social integration as a benefi t of behavioral therapy. the healthcare access problem for americans with neurodevelopmental disorders and intellectual disabilities is, at its core, a healthcare education problem -an education problem resulting from a long-standing defi ciency in professional training focused on the care of this patient population. and it is clear that the medical and dental professions share equally in responsibility for these defi ciencies. eighty-one percent of america's medical students will graduate without ever having rendered clinical care to a single patient with a neurodevelopmental disorder or intellectual disability; and the graduates of % of america's medical residency programs will graduate from those residencies having had no formal training whatsoever -didactic or clinical -in the care of this patient population. additionally, % of graduating dentists have never treated a single patient with a disability. it is no wonder that patients like those whose cases were discussed in earlier sections of this chapter have such diffi culty accessing quality health care. as robert uchin, dean of nova southeastern university college of dental medicine observed in a speech in to his faculty, "not only do we not have enough doctors to care for these patients; we don't have enough teachers to teach them how to care for them." as a result of these defi ciencies in professional education, few clinicians with any expertise in developmental medicine or developmental dentistry are to be found in communities across america. the experts in developmental medicine and dentistry, for the most part, tend to be physicians and dentists who work at the few remaining intermediate care facilities, and at special needs outpatient clinics, psychiatric hospitals, and nursing homes. these physicians and dentists possess the knowledge and expertise in these disciplines because they are the physicians and dentists with the clinical experience. unfortunately for the patients with neurodevelopmental disorders who are clamoring for quality care, there are too few of these clinicians. national experts in developmental medicine and dentistry, however, have begun to collaborate in the creation of patient care protocols; and they have produced multidisciplinary curricula in both dvd and online format. the aadmd has made available hours of online curriculum in developmental medicine, developmental dentistry, and developmental psychiatry (see: list of urls). the curriculum program is entitled, the continuum of quality care , and it teaches collaborative patient care in three disciplines through an interdisciplinary format. the aadmd, through a grant from the wal mart foundation and the north carolina developmental disabilities council, and in collaboration with the north carolina mountain area health education center and the family medicine education consortium, has also established the national curriculum initiative in developmental medicine. this initiative, which is scheduled for completion in , will develop curriculum standards for physicians in the primary care of adults with nd/id. the curriculum stresses the importance of a collaborative approach, which includes medicine, dentistry, podiatry, optometry, and multiple ancillary health professions. if the disparities in access to healthcare for americans with nd/id are to be resolved, physicians and dentists must be willing to cross professional boundaries and work together to plan and deliver whole-person healthcare to their patients with nd/id. interdisciplinary protocols in the diagnosis of neurodevelopmental disorders and in the management of the secondary health consequences associated with these disorders must be established. additionally, clinicians with expertise in these arenas must be willing to work and teach in our nation's medical and dental schools. the clinicians with expertise must be willing to develop predoctoral and postdoctoral curricula, and the deans of america's professional schools must be willing to include these curricula as part of their larger programs in primary and specialized care. the clinicians with expertise in developmental medicine and dentistry must also be willing to conduct patient-focused, interdisciplinary, clinical research in an effort to solve the myriad problems that create obstacles to the delivery of the standard of care for patients with nd/id. they must be willing to obtain institutional review board approval for this research, and they must be willing to make this research available to their colleagues through publication in peer-reviewed journals and text books, and in professional lecture forums. the patient featured in figs. . and . is a man named james. he is a year old patient with idiopathic intellectual disability who presented to a dental clinic for evaluation of a painful facial swelling. a comprehensive intraoral exam revealed a cellulitis resulting from multiple grossly decayed teeth, and a generalized advanced periodontitis. no fewer than fi ve clinicians became involved in this patient's care. they included a general dentist, two oral surgeons, a family practice physician, and a geneticist. over the course of several months, as the treatment plan was completed, and as the chronic dental and periodontal infections were eliminated, james experienced signifi cant improvement in his overall state of health. a comparison of these two photographs reveals not only signifi cant improvement in his aesthetic appearance, but also in his skin turgor and color. these improvements in the patient's health translated to improvements in his daily life. he found gainful employment, and his caregivers now report that he smiles constantly -at work and at home. these photographs were entered into evidence in before a congressional subcommittee investigating the death of a young african-american boy who died as a result of an untreated dental abscess. the photographs were intended to make the point that patients with intellectual disabilities need not die as a result of medical illnesses derived from untreated dental disease. this patient's case illustrates that, when physicians and dentists are willing to work together toward a common goal of whole-person health for their patients, profoundly positive outcomes can be achieved. in a larger context, if our nation's medical and dental professions are willing to commit to a shared agenda, one which promotes the idea of collaborative, interdisciplinary care as a foundational concept, signifi cant improvements in quality of health and quality of life can be realized, not just for americans with neurodevelopmental disorders, but for every patient seeking quality care. in light of the events of , bioterrorism has become subject of increased attention from all members of society. government agencies, professional associations, academia, etc. have expressed their determination to wage war on such threats by all means available. dentists can also participate in this effort by providing assistance at interested groups and the general public (flores et al. ) . in this chapter we will examine the elements and components that may play a role in the establishment of an electronic network for the dental profession for supporting the fi ght against bioterrorism. in this section we review the threats, the public health system, current electronic surveillance systems, regulations and ethical issues, the computerization of dentistry, and how dentistry can serve in improving biosurveillance efforts. the aftermath of september and the anthrax incidents in october ( lane and fauci ) , made the us government reorganize its priorities and reform its current structure (white house offi ce of the press secretary ) . in response to these incidents, president bush proposed the "health security initiative" (white house letter ) in february nd of . this effort labeled the "bioshield initiative," (white house letter ) has the purpose to stimulate research and development of medical countermeasures against bioterrorism attacks. however, despite all these efforts, terrorist attacks are likely to happen in the future and even the best work from intelligence and security agencies will be unable to prevent such events (betts and richard ; council on foreign relations ; baker and koplan ) . to cope with this threat, a report published by an independent task force sponsored by the council on foreign relations "america-still unprepared, still in danger" (council on foreign relations ) , suggested a series of steps to assist the government in preparing to better protect the country. one of these suggestions is the bolstering of the "public health systems". baker et al. defi ne the u.s. public health system as a system that consists of a broad range of organizations and partnerships needed to carry out the essential public health services, such as hospitals, voluntary health organizations, other non-governmental organizations and the business community (baker and koplan ) which can collaborate with local, state and federal public health entities. after the unfortunate incidents in the public health system was revisited and the realization that "the nation's public health infrastructure is not fully prepared to meet this growing challenge" (frist ) became clear. to address this need, congress and president bush enacted the public law (p.l.) - titled "public health security and bioterrorism preparedness and response act of " (frist ; th congress ) . the main purpose of this law was to improve the public health capacity by means of increasing funding and fostering other measures. frist ( ) , described the law as a "good start" and that "to be prepared for bioterrorism, it is imperative that we develop a cohesive and comprehensive system of ongoing surveillance and case investigations for early detection". in this way, several early detection systems have been implemented with different levels of success among different geographic regions in the us. one of the most important initiatives over the years has been the establishment of the national electronic disease surveillance system (nedss) (baker and koplan ; nedss ) . the national electronic disease surveillance working group establishes that the "nedss is a broad initiative focused on the use of data and information systems standards to advance the development of effi cient, integrated, and interoperable surveillance systems at the state and local levels. the long-term objectives for nedss are the ongoing automatic capture and analyses of data needed for public health surveillance". the purpose of this system is to take into consideration and integrate the information of current public health systems implemented at different health department levels: county, state and fi nally at the centers for disease control and prevention (cdc). another initiative spearheaded by the cdc is biosense (looks ) . the purpose of this program is to develop advance detection capabilities of health related events including disease outbreaks. in addition, its emphasis is to improve situational awareness by integrating advanced analytics to process data generated by different health providers and other entities in the us. now that we have examined the general aspects, we will continue our background review focusing on the aspects that pertain to the specifi cs of the dental profession. this section will provide some perspective of the structure of the dental profession in comparison with its medical counterpart. "there are approximately , active dentists in the united states" (mertz and o'neil ) . in the dentistto-population ratio was of - , . and it is expected that by the year the ratio will be . , which translates into one dentist for every , people. " in contrast, the physician-to-population ratio has been increasing for the past years and now stands at per , , about one physician for every people." eighty percent of the dentists are in general practice. during march and of , the american dental association and the us public health service sponsored the conference "dentistry's role in responding to bioterrorism and other catastrophic events" (palmer ; national institute of dental and craniofacial research ) . this meeting reviewed several aspects of bioterrorism and the dental profession: the nature of biological pathogens and its oral manifestations, what needed to be communicated, how dentists should participate, etc. dr. michael c. alfano described the diffi culties that biological pathogens create for clinicians because "they are so insidious." while discussing the anthrax mailings after september th he pointed out that: "… early symptoms appeared so they resembled the aches, fever, and malaise of fl u so those affected delayed seeking treatment, a delay that has proven fatal in some cases". lieutenant colonel ross h. pastel of the us army medical research institute of infectious disease (usamriid) listed the "category a" pathogens as defi ned by the centers for disease control and prevention, and those are: smallpox, anthrax, plague, botulinum toxin, tularemia and viral hemorrhagic fever. he also described an outbreak of smallpox in yugoslavia in and the measure that had to be taken to control it. dr. michael glick described the oral manifestations of smallpox showing "signs hours before skin rash. these oral signs include tongue swelling, multiple mucosa vesicles, ulceration, and mucosal hemorrhaging. oral signs are also evident in inhalation and gastro-intestinal anthrax. in oropharyngeal anthrax the mucosa appears edematous and congested; there may be neck swelling, fever, and sore throat" . dr. ed thompson, deputy director of the centers for disease control and prevention mentioned that "none of the new counter-bioterrorism measures can be effective unless local health practitioners are vigilant in observing and reporting a possible disease outbreak. such surveillance-knowing what to look for and whom to report to-is critical and applies not only to suspected bioterrorist agents, but to a list of reportable diseases which has grown to include such entities as west nile virus and sever acute respiratory syndrome (sars)." dr. sigurs o. krolls presented the response at the local level and he "stressed the importance of communication and the need for redundant systems", "to keep all the parties informed". he also posed the question "can dentists recognize signs and systems of contagious diseases?", and emphasized that education can be essential. dr. louis depaola made several connotations that can be key in the scope of this paper by saying "dentists can contribute to bioterrorism surveillance by being alert to clues that might indicate a bioterrorism attack. such surveillance would note if there is an infl ux of people seeking medical attention with non-traumatic conditions and fl ulike or possibly neurological or paralytic symptoms… or even specifi c signs of a bioterrorist agent. patterns of school of work absence, appointment cancellations or failures to appear, could also be indicators." dr. depaola made clear that in cases of limited release of bioterrorist agents, dentists "have little to offer" but "a widespread attack can certainly tap into dental professional skills in recognition, isolation and management". in addition, dr. guay ( ) lists all the possible roles in which dentists can participate including "education, risk communication, diagnosis, surveillance and notifi cation, treatment, distribution of medications, decontamination, sample collection and forensic dentistry." dental informatics must pay attention to these and other recommendations, in order to develop integrated systems that take these recommendations into consideration. it is also important to understand that informatics has to work with technologies already in place like the computer-based oral health record and current standards. the fi nal recommendation from the meeting stated that to play an important role in biodefense, a serious amount of coordination and preparation will be required, not only from dentists but from other groups, most likely requiring medical and dental data integration. the cohr as described by rhodes ( ) "can provide a structure for documentation that goes beyond the concept of a blank form on a page, it includes a glossary of dental terminology for the entire content of the form as well as knowledge bases and expert systems that can enhance the practitioner's diagnostic and treatment planning decisions". he also acknowledges that one of the advantages of this type of documentation is that it "is much more transportable". he also recognizes the need for standardized methods for collecting information from dentists. schleyer and eisner ( ) defi ned several scenarios where the cohr is used in a "shared" environment where several healthcare providers interact and information is seamless communicated, improving the decisions made by clinicians. delrose and steinberg ( ) discuss how the "digital patient record" enhances clinical practice by providing "better quality information" to the clinician. although all of these benefi ts sound promising and encouraging some still express concern of the lack of standards among different information systems, which translates in communication breakdowns (schleyer ) . on the other hand, heid and colleagues ( ) mention a list all the steps that are currently being taken by different organizations such as the ada in order to produce a standardized cohr. other examples of standardization can be found in a paper presented by narcisi ( ) where ada's participation as a voting member in the american national standards institute has allowed edi or the cohr to be discussed and improved at a national level. additional infl uences in the standardization of the cohr are the security regulations mandated by hipaa, the health insurance portability and accountability act of . dentists are required to "adopt practices necessary for compliance" (sfi kas ; chasteen et al. ) . these and other regulations (szekely et al. ) will encourage the homogeny among different system vendors. computer ownership, on the other hand, has increased steadily during the last years. according to schleyer et al. ( ) in only % of dental professionals used computers in their practices compared to % in the year . additionally similar trends in internet connectivity where described. the issues mentioned above describe the issues that have to be considered in order to create surveillance system against bioterrorism for the dental profession. this review has tried to be inclusive by covering different aspects starting with the current state of affairs and environment, treats, technology, law, etc. next we present a blueprint for developing a biosurveillance system. the purpose of developing an electronic health surveillance system is to gather information from patients directly ( wagner et al. ) by detecting signs and/or symptoms, or indirectly by obtaining other types of information such as over the counter medication sales, patients' no-shows, usage of internet search engines keywords, etc. in this particular case, the proximity of contact between the dentist and the patient is equivalent to a medical inspection in terms of immediacy and/or closeness. such signs and symptoms can be easily detected if the dentist is properly prompted to search for them. this is just one example of ways how a system could provide assistance in the detection of a bioterrorism incident. but, before describing our proposed system, it would be important to address the fact that current syndromic surveillance systems have certain advantages in terms of its particular technological implementation . the rods laboratory obtains data directly from chief complains in the emergency departments from hospitals. the advantage of this surveillance system is that the implementation has to be made with only a limited number of parties (hospitals, clinics, health systems, etc.). on the other hand, our system would have to deal with thousands of different implementations (one in each dental offi ce). this and other challenges have to be considered when designing the proposed system: the proposed system should work at multiple levels: the system would have to provide a mechanism to alert the dentist if there is • suspicion that a bioterrorist attack may be happening. the mechanism would increase the dentist's awareness in case of fi nding suspicious signs or symptoms in a patient. this can be triggered by the patient's characteristics such as geographic location of residence, etc. automated collection of information from the patient's oral health record. the • system would report to a central database signs or symptoms of interest. the aggregation of this data could generate information that would eventually identify the presence of patterns that may lead to the early detection of such events. collection of additional information, which combined with other sources, can be • useful in terms of detecting or tracing some incident. patients' "no-shows" is the primary example, that, if combined with others such as work or school absenteeism can provide a relevant pattern for public health offi cials. dr. x, who practices in a community min away from capitol city, installed a new clinical management system months ago. among the features that were included in this new clinical management system (cms), a bioterrorism detection module was added. she felt curious because of recent news she read in the newspaper about possible attacks against the us and decided to install such feature. he read about how the module would work in combination with the cms she just bought. the educational information provided with the software instructed dr. x, that in case that a patient victim of a bioterrorism attack happens to be seen in her practice, the software would collect information and would send it to public health offi cials. when installing the software, dr. x was asked if she agreed to share such information with authorities. she was provided the option to receive notifi cation in case some information was sent but she decided not to enforce it. during the last week a patient walked into dr. x's dental offi ce. the patient presented some signs that indicated the presence of a disease; still its origin was not clear. an epidemiologic study later would show that the patient was present at the football stadium when an infectious agent was released (fig. . ) . although, at that time his medical history showed no indication of a systemic disease, the presence of multiple oral vesicles prompted the dentist to make an annotation into the cohr. the system, by using a natural language processing engine, detected such sign and sent this information to a central database. the patient was discharged and instructed to take some support medication to treat the oral ulcers. the next day, the central database pinpointed the presence of an out of the ordinary increase in the number of cases with the same signs and symptoms around that region. when the presence of this peak was detected, the central server sent a request to the dentist computer for additional information. one of the requested elements was if there was any use of medication for treating oral ulcers. fortunately this information was available. the central database crossed this with the information of other surveillance systems together with the information from other patients that happen to have similar clinical signs and/or symptoms. dr. x received an email from a public health offi cial asking her to communicate to the local health department to discuss information about one her patients. the case depicted above simulates the release of smallpox during a football game. in the case of smallpox oral symptoms include tongue swelling, multiple oral mucosal vesicles, ulceration, and mucosal hemorrhaging (national institute of dental and craniofacial research ) . dentists could be alerted by an electronic system to search for such signs or they can be detected automatically. in case of a high incidence within a group of patients, in a confi ned area, public health offi cials get to be notifi ed. in our hypothetical case there are issues that need to be addressed in order to make such detection system feasible: as described by schleyer et al. ( ) , % of dentists in the us use a computer in their practices. this fi gure would generate an estimate of , computers in dental practices. this prevalence of computers represents an opportunity for public health data collection. the creation of a software application for surveillance purposes must rely on existing technology. currently there are approximately major clinical management software packages in the market (dentistry today ) . out of these , clearly permit direct database manipulation. this characteristic can easily allow the creation of a "querying" application that would look for specifi c information within the data stored by those packages. additionally, a natural language processing engine could be embedded into the application in order to detect variations in data input on the computer oral health record. nevertheless, it is necessary to obtain a detailed list of the oral manifestations of diseases that are likely to be found on patients. successful implementations of similar systems have been shown to work successfully (chapman et al. ; ivanov et al. ) and using the same approach for our system seems technically feasible. this collected information later would be send to a central server in order to be analyzed and interpreted. the components of our system would be as follows (fig. . ) : thin client: a software application distributed for data collection. it would be • conformed of a "querying" mechanism, combined with a natural language processing engine and a communication module. this software client should be as thin as possible to reduce the work load on the dentist's equipment and should be embedded as a plug-in for current clinical management systems. vendors should be contacted to ask for their collaboration in the development of such application to ensure maximum compatibility and integrity of data collection. central servers: server software in charge of integrating all the data collected • from dental offi ces. it has to be capable of handling simultaneous requests from multiple users. this server would integrate all the data and would perform an analysis with the intention of detecting anomalies. it would be recommended that redundant servers should be located in different data centers with mirroring capabilities to guarantee their survivability in case of technical diffi culties. communication network: the transmission of information should be done using • the internet. this, of course, would essentially depend on the practitioner's current connectivity. if that is not available, backup connection to the central servers should be established. dentistry uses several standards for transmission of health related information. clinical management systems use standard-based technology to transmit information (narcisi ; chasteen et al. ; szekely et al. ; dentrix dental systems ) . dentists are aware of these standards and use them in a day-to-day basis to transmit information to insurers. additionally, in order to interact with other surveillance systems such as the nedss, our application should rely on the same standards. the software both client and server should be thoroughly verifi ed to be secure in terms of being safe against hacker attacks. on the server side, redundancy should be provided so downtime is reduced from design. the system should be developed so mirrored servers are always up and running. data integrity mechanism should also be considered. privacy and confi dentiality are important issues that need to be incorporated as part of a robust biosurveillance system and distinct regulations such as hipaa require protecting patient information (frist ; chasteen et al. ; bayer and colgrove ; etzioni ; ivanov et al. ) . in our hypothetical case we describe the use of several sources of information for detecting a bioterrorist attack. we described how syndromic information is transmitted to a central database which initially should be de-identifi ed. later, after the suspicion a bioterrorist attack more information is requested (medications) and more inferences are made. this, although technically possible, would require changing our processes and also the will to share clinical information. this leads to the discussion mentioned in the background section about "individual rights" vs. "common good". although hipaa addresses public health , some other implications may arise and the health professionals including dentists, physicians, public health offi cials and patients should discuss and address such issues. as discussed earlier, legislators face a diffi cult task in terms of determining what is best on behalf of the individuals they were asked to represent. legislation may have to be passed in order to guarantee the functioning of such a system. individual freedom and privacy are important values which may pose a confl ict when collecting individuals' information even for their own good. in any case, careful consideration has to be given to which information is required to detect a bioterrorist attack and also, by keeping in mind that it is always important to reduce, as much as possible, the collection and transmission of patients' information over the internet or any other network. a detection algorithm has to be created or adapted in order to determine the presence of a bioterrorist attack. some algorithms have proven their effectiveness (wong et al. a, b ) and it is likely that from these, a new analysis should be done in order to select or create one that addresses the particular needs of our system. a study was conducted to assess the feasibility of using oral manifestations in order to detect disease outbreaks (torres-urquidy et al. ) . it was found that for diseases such as botulism and smallpox it would be feasible to gather data that contains oral manifestations that would allow creating a detection signal using natural language processing followed by the use of statistical methods such as moving average to serve as part of a detection algorithm. the system should also be thoroughly evaluated, before and after implementation. to perform the evaluation before the system implementation computer simulation can be used to assess the effectiveness and likelihood of detection. simulation and modeling techniques (reshetin and regens ) have been used to estimate the effects of a bioterrorist attack. the same techniques can be used to evaluate our system. in case of the study by torres-urquidy ( ) , the investigators utilized synthetic outbreaks to test the performance of different signals. from their evaluation process, they learned, for instance, how many cases would be necessary to occur for the system to reach certain detection thresholds. several dental organizations have shown publicly their support of measures against bioterrorism. the american dental association and the national institute of dental and craniofacial research are two organizations who could play an important role in the development, deployment and ongoing support for our system. local dental societies also would also play an important role in the deployment of the proposed system. similarly, local, state and federal public health agencies should engage in activities that could make these mechanisms for health surveillance feasible. if dentists want to play an active role in the fi ght against bioterrorism, they should commit to collaborate with public health entities as well as to seek a way to integrate their information with the rest of electronic biosurveillance systems. professional organizations such as the american dental association can also participate by endorsing such efforts and by collaborating in the educational process of the dental professionals and their patients. as mentioned by dr. depaola (national institute of dental and craniofacial research ) dentists "have little to offer" in the current biosurveillance state of affairs. however, the integration of different technologies can change this perception. goldenberg et al. ( ) described over-the-counter medication sales as a technique for discovering disease outbreaks and stated that their approach may be "more timely" than traditional medical or public health approaches. medical cases that result from bioterrorism attacks do not produce symptoms until they have fully developed, so it is likely that different patterns can be detected before the patients start reaching the emergency department. as stated earlier (torres-urquidy et al. ) , it may be possible to have dentists participating of biosurveillance efforts, if we solve the proper organizational and technical challenges. dr. john r. lumpkin ( ) states that "hippocrates noted the health of the community was dependent on characteristics of a community and the habits of the people who lived there." dr. krolls (nidcr ) in his fi nal remarks during his presentation at the dentistry's role conference against bioterrorism, said, "dentists may pick up telltale information about what is happening in the community. after all, dentists spend more time with their patients than any other health specialty". kass-hout t, zhang x. biosurveillance: methods and case studies. muhammad f. walji maintaining patient records are essential for both clinical care and research. clinical research often occurs in the context of also providing patient care, yet the systems that are used for each are different and often cannot exchange data. the lack of data exchange between systems pose signifi cant barriers to effi ciently treating patient and conducting clinical research in dentistry. the purpose of this section is to review the benefi ts and challenges of integrating electronic health record (ehr) used for patient care and electronic data capture (edc) which is used for clinical research such as clinical trials. an increasing number of dentists routinely use ehrs (schleyer et al. ) . most dental schools in north america also use ehrs. benefi ts of using ehrs include increased legibility, portability, and improved patient safety (buntin et al. ) . recent federal incentives, although not directly benefi cial to dentists, will also likely spur the adoption of ehr (blumenthal and tavenner ) . clinical researchers, especially those conducting clinical trials, are also discovering benefi ts of using electronic data capture compared to paper. a clinical trial is a process in which new treatments, medications and other innovations are tested to evaluate safety and effi cacy. a standard part of health care, clinical trials are often lengthy and costly due to myriads of regulatory oversight. recent estimates set the cost of drug development in excess of $ million (grabowski et al. ) . accurately documenting data with suffi cient detail is critical for providing patient care and conducting clinical research. while the medical record is the foundation for patient care, the case report form is the foundation in a clinical trial. not all clinical research is clinical trials. clinical trials whose data will be submitted to the fda as a new therapy or device have additional requirements relating to the collection and transmission of the data. similarly for patient care data, ehrs need to meet the privacy and security requirements of hipaa. case report forms (crf) are a medium in which research study sites collect subject data in pre-defi ned formats for communication with clinical trial sponsors (rondel and webb ) many clinical trials data are collected on paper (rondel and webb ) . data measurement, collection, and recording are considered the "most crucial stage" in the data management process (hosking et al. ) . traditionally, study coordinators often record information in a case report form and subsequently mail or fax the crf to the centralized coordinating center. there, data entry staff, sometimes with the aid of optical character recognition systems, input crf data into a computer. errors made during this second data entry process are diffi cult to detect and correct (hosking et al. ) . lengthy guidelines in literature discuss methods for developing paper case report forms to reduce data entry mistakes (hosking ) . a well-designed crf may allow a user to effi ciently collect and record pertinent data. however, forms are often revised and redesigned during a clinical trial due to changes in protocol, unforeseen outcomes, or oversight (singer and meinert ) . there has been a recent drive to use electronic case report forms (ecrf). direct data entry at a study site shortens time to analysis and provides opportunities to audit data at time of entry. this could reduce data errors that might otherwise be caught weeks after submission. for quality control purposes, some studies require double data entry using computers and paper (day et al. ) , though alternative solutions have been explored including the use of data sampling (king and lashley ) and probability statistics to select only those forms likely to contain errors (kleinman ) . ecrfs may also facilitate data collection from existing electronic information systems such as lab systems. however, ecrfs are almost always reside in a separate system that is not linked to a patients record. although many clinical research studies are still being conducted using paper, an increasing number of studies are using ecrfs and electronic data capture (edc). for example, a review of canadian clinical trials found that % use edc (el emam et al. ). studies that are sponsored by a pharmaceutical company and are multicenter appear to use edc at a higher rate than those sponsored by government or a university. the cost of a commercial edc is substantial. recently a freely available edc has become popular amongst universities called redcap. a tool originally developed at vanderbilt university, it is now being used at over a institutions worldwide (harris et al. ). however, such tools are generally not integrated with the institutions ehr. although moving from paper to electronic will afford benefi ts there is a great need to allow data exchange between the patient care and clinical research components of information systems. although ehr and edc are similar, several challenges remain unresolved that prevent integration. one of the major barriers is likely to be different workfl ows for patient care purposes and to collect data for research. research is needed in defi ning an optimal workfl ow that can streamline the tasks associated with patient care and research, while at the same time providing a unifi ed information system that support these activities. also, the data that are collected for care and research are likely to differ. a researcher may require far more granularity of an oral health measurement than a clinician seeking to provide care. in cases when conducting a double blind placebo controlled clinical trial, the investigator may not even know the type of treatment that has been delivered to the patient. due to complexities of each domain, and large differences in goals, to date mutually exclusive workfl ows have arisen. a clinician investigator who sees a patient for both care and research, will likely need to enter data on this same patient twice; once in the ehr and once in the edc system. despite the availability of electronic systems, a major barrier is the integration and compatibility of disparate health information systems to converse with one another. the languages are important because they can help data sharing. clinical trials are not usually conducted in isolation, but are part of conventional medical care. therefore sharing data by clinical trials, patient care and laboratory systems becomes especially important with the adoption of ehrs in dentistry. in biomedical informatics, standardized terminologies are recognized as a critically important area to help better represent and share data for use in electronic systems (cimino ) . the systematized nomenclature of medicine clinical terms (snomed-ct), developed by the college of american pathologists, is the most comprehensive medical terminology (strang et al. ; chute et al. ) and is used in a number of health informatics applications. the us department of health and human services ( ) has also licensed snomed-ct, allowing access throughout the us at no charge. therefore snomed-ct is even more likely to be the vocabulary used in electronic formats of patient records in the future. the medical dictionary for regulatory activities (meddra) is terminology used by the fda and drug development industry to classify, retrieve, present, and communicate medical information throughout the medical product regulatory cycle (brown et al. ) . in particular it is used to record and report adverse drug event data. therefore standard languages are essential in sharing clinical trials data between sites, and also with regulatory agencies. no one single terminology is suited for all tasks. snomed-ct is likely to be more comprehensive to code clinical encounters, while meddra is more suited to help adverse event reporting. however, it is important that terminologies are widely adopted and used for similar purposes. even when standard terminologies are agreed upon, such information needs to be interchanged in standard formats. health level (hl ) is an important organization whose standards are widely adopted in healthcare to exchange information between computer systems. the clinical data interchange standards consortium (cdisc) is also an important group that helps to defi ne different data standards specifi cally for clinical trials research, such as clinical trials or regulatory submissions. one particular challenge in oral health has been the lack of a standardized terminology to describe diagnoses. although icd contains oral health concepts, they are often not granular enough to be useful for some patient care or research purposes. recently a dental diagnostic terminology has been developed by a group of dental schools, and has already been adopted by several institutions and used within dental ehrs (kalenderian et al. ) . the american dental association (ada) has also been developing snodent, but is not yet publically available for clinical use (goldberg et al. ). another link between ehr data and clinical research is the potential to fi nd human subjects. recruiting suffi cient numbers of patients that meet eligibility requirements within an allotted time frame for clinical trials is challenging. as ehrs contain detailed information about patients, they can be used to fi nd patients that meet specifi c inclusion and exclusion criteria. informatics for integrating biology and the bedside (i b ), an open source data warehousing platform, has been found to be a useful tool for cohort selection especially if the source data from an ehr is represented in a structured format (deshmukh et al. ). further, with health information increasingly available to patients through the internet, it is possible interested patients will be more effective in fi nding clinical trials than investigators looking for patients. many clinical trial registers are now available online. the national institutes of health (nih) have made available their database of nih funded research (mccray ) . there is currently no single repository for patients to fi nd all trials studying a health condition. a recent study assessed the comprehensiveness of online trial databases concerning prostate and colon cancer and found that online trial registries are incomplete, especially for industry-sponsored trials (manheimer and anderson ) . a more collaborative effort between government and industry-sponsored research groups to compile and standardize information may be a mutually benefi cial effort. it is not clear how many patients now enroll in clinical trials through online discovery. ehr data originally collected for patient purposes can be potentially used for research. aggregating data from multiple sources can provide a large dataset that could otherwise not be available. electronic health records (ehr) contain a wealth of information and are a promising source to conduct research. data extracted from ehrs differ from other sources such as population surveys or data obtained from payers, as they provide a more detailed and longitudinal view of patients, symptoms, diseases, treatments, outcomes, and differences among providers. therefore ehr data in dentistry can potentially provide valuable insight into oral health diseases, and treatments performed on a large cohort of subjects. ehrs also play an important role in enhancing evidence-based decision-making in dentistry (ebd) and improving clinical effectiveness through decision support (atkinson et al. ; walji et al. ; valenza and walji ; taylor et al. ; spence et al. ; chambers et al. ; langabeer nd et al. ; walji mf et al. ). the consortium of oral health related informatics (cohri) provides an example of how dental ehrs are used for research purposes (schleyer et al. ; stark et al. ) . cohri was formed in by a group of dental schools who used the same ehr platform and who are interested in sharing clinical and education data. through funding from the national library of medicine, four dental schools are participating in a pilot project to develop an inter-university oral health research database by extracting and integrating data from ehrs. one promising area where data repositories derived from ehr data can be used for new discoveries is in the area of comparative effectiveness research. comparative effectiveness research is defi ned as "a rigorous evaluation of the impact of different options that are available for treating a given medical condition for a particular set of patients." (congressional budget offi ce ) further, such research includes focusing on the clinical benefi ts and risks of each option (clinical effectiveness), and an analysis on the costs and benefi ts (cost effectiveness analysis). comparative effectiveness research (cer) is also likely to reduce costs of dental care and increase access to the majority of the population who currently receive no dental care. unfortunately many recent systematic reviews focusing on cer questions in dentistry have been inconclusive due to the lack of existing evidence in the scientifi c literature. secondary analysis of the data that reside in dental electronic health records (ehr) is a particularly appealing approach to facilitate cer and generate new knowledge. ehr data has the potential to provide a comprehensive picture of patients' histories, treatments, and outcomes, and if integrated with similar data from other dental clinics can include a large and diverse set of patients. however, numerous challenges must be solved before ehrs can be used for cer. first, data suitable for cer must actually be collected from ehr systems. second, this data, which often resides in proprietary systems, must be accessible and retrievable. and lastly, this data should be structured in a format that can be integrated with data from other sources or institutions. practice-based research networks (pbrn) are groups of primary care clinicians and practices working together to answer community-based health care questions and translate research fi ndings into practice. pbrns engage clinicians in quality improvement activities and an evidence-based culture in primary care practice to improve the health of all americans. in , the national institute of dental craniofacial research funded three such research networks. the dental pbrn's to date have been conducting both prospective and retrospective research. for example, barasch et al. conducted a case controlled study to investigate risk factors associated with osteonecrosis of the jaws . many prospective studies conducted as part of pbrns still require separate data collection systems for the research data. ehr data contained in practices as part of pbrns are beginning to be used for secondary purposes. for example fellows et al. conducted a retrospective analysis of data contained in electronic health records to estimate incidence rates of osteonecrosis of the jaws ( fellows et al. ) . pbrns provide great promise of how ehr and clinical research data can be used effectively to promote both patient care and new discoveries. another area that intersects both the patient care and research realm are patient registries. patient registries are ways to track groups of patients who have had specifi c diseases or have had certain treatments. while ehr data would contain information on all types of patients, their diseases, and treatments, registries would allow focus on specifi c diseases or treatments of interest. registries would not be as costly in terms of resource requirements like a traditional clinical trial, but would require specifi c eligibility criteria, informed consents, and collection addition to that collected as part of routing care. dentistry has lagged far behind in forming data registries, primarily because dentistry is practiced in small offi ces and not in large hospitals making the process of integrating data very diffi culty. however, dental schools which themselves house large clinical operations are ideally positioned to create disease specifi c registries that can potentially use data collected for patient care and extend for research purposes. there is great potential for providing new insight in oral health by the integration of patient records and clinical research from both a workfl ow and information systems perspective. the technology challenges of developing systems that can exchange data, and use standardized terminologies appear solvable. however, the socio-technical issues such as determining how to incorporate optimal workfl ows for conducing both patient care and research with minimal additional overhead appear to be the greatest challenge before widespread adoption. similarly, there appears to be great potential in using ehr data originally collected for patient care for the secondary use of research and discovery. this will require collaboration between patients, providers and researchers from all healthcare disciples, and institutions with friendly policies for sharing data to improve both patient care and drive new discoveries. amit acharya , andrea mahnke , po-huang chyou , and franklin m. din more recently there has been a strong push from the united states federal government for the adoption of the electronic health record (ehr) within the healthcare industry. as a result, $ . billion is made available to incentivize the physicians, dentists and hospitals for the adoption of the ehr through the health information technology for economic and clinical health (hitech) act. as the nation head towards adoption of the ehrs, there has also been a growing interest with the majority of the u.s. dental schools to implement ehrs within the educational setting. fifty of the fi fty-six u.s. dental schools, as well as dental schools in canada and europe, are either using or in the process of adopting some aspects of a common dental ehr framework (white et al. ) . a group of dental schools known as consortium for oral health-related informatics (cohri) was formed in which used this common dental ehr framework -axium (stark et al. ) . currently there are about dental schools within cohri. the ehr will not only support clinical care, but will also result in training the next generation dental students and to conduct innovative research that was not possible earlier. however, not much is known about how many of these dental schools' electronic dental records are integrated with their respective university's electronic medical record. a common medical-dental ehr model at healthcare universities would enable a holistic approach to providing patient care and provide the much needed electronic infrastructure to study interrelationship between the various oral-systemic diseases. recently a group of researchers from marshfi eld clinic in wisconsin, us conducted a survey to investigate the current states of health information technology and informatics within the dental school in the us. list of us dental schools were identifi ed through the american dental education association (adea) web site. dental schools were contacted to determine who the most appropriate person to take the survey would be. once the list of contact was developed from each dental school, an email was sent to us dental schools with a link to a survey created in surveymonkey. the survey was administered on tuesday march , . reminder survey emails were sent to all recipients on march and march . the survey was closed on march . the anonymous survey was at most questions, depending on how questions were answered. the survey focused on topics such as presence of dental informaticians within the dental schools, use of fi nancial and clinical information systems, interest in federal stimulus support for ehr adoption provided through american recovery and reinvestment act and meaningful use of ehr, relationships with health care entities and bidirectional nature of the dental and medical ehrs. the study was approved as exempt from the marshfi eld clinic institutional review board under section cfr . (b) and waived requirement for an authorization. thirty out of the fi fty fi ve dental schools responded to the survey (response rate of %). however, fi ve of the thirty dental schools representative did not complete the survey and hence their response was not included in the analysis. regarding the question about the presence of a dedicated department or a center for information technology (it) or informatics within the dental school in us, % (n = ) of the responding dental schools had a dedicated it/informatics department or center (p-value of . ). the it or the informatics department size (in terms of the number of personnel) at the dental schools is illustrated in fig. . . thirty fi ve percent (n = ) of the us dental schools that housed an it / informatics departments had personnel with not only it training but also dental informatics training. while % (n = ) of the dental schools were considering integration of dental informatics personnel within their department or center. twenty fi ve percent (n = ) of the dental schools did not have any plans of integrating personnel with dental informatics personnel within their department or center (see fig. . ). partial responses to additional questions in the section of the survey is provided under table . . the majority of the responding dental schools were currently using financial electronic systems (fes) (p-value of < . ) and electronic dental records (edr) (p-value of . ). the use of fes outnumbered the use of edrs in the dental schools (see fig . ) . about % of the dental schools that were currently utilizing the edrs used it in all the clinical modules (p-value of . ), while % of the dental schools used the edrs in some of the clinical modules. when asked about the commercial edr system that the dental schools were using, axium (exan group, canada) was by far the most implemented edr system. two dental schools had salud (two-ten health limited, ireland) implemented and two dental schools had gsd academic (general systems design group, iowa, us) implemented. combinations of two ehr systems (home grown and dentrix) were implemented at two dental schools. one school had a dentrix only implementation, while another had developed its own edr system (home grown) (see fig. . ). there were dental schools which had implemented an edr fi ve or more years ago, dental schools - years ago, dental schools - years ago and dental schools less than a year ago (see fig. . ) (p value of . ). when the dental schools were asked the question as to whether they were expecting to apply for the medicaid meaningful use incentive program, majority ( %) of the dental schools did not know and only % of the dental schools were expecting to apply within the next years (fig. . ) (p-value of . ). challenges or barriers identifi ed by some of the dental schools in complying with the meaningful use objectives were (a). lack of certifi ed edr and information regarding it, (b). issues with getting auxium certifi ed and (c). qualifi cations of the edr as many of the meaningful use objectives do not apply to dentistry and lack of specifi c information about it. only % of the responded dental schools were part of a health system. fifty two percent (n = ) of the responded dental schools had a formal relationship with other health care delivery entities in terms of sharing facilities, patient transfer, training programs. some of the types of relationship mentioned by the dental schools that had a formal relationship with other health care delivery entities included: (a). a gpr program and an emergency dental unit in the hospital, (b). affi liated hospital, (c). affi liation agreements, (d). oral and maxillofacial surgery (omfs), anesthesia and pedodontics all have some portion of education in medical health center, (e). omfs residents are also residents of medical health center, (f). residents providing care under contract with area hospitals, (g). sharing patients wand facilities with the health center, (h). students rotating in the community health centers and (i) collaborative grand programs. eighty fi ve percent of the dental schools that had a formal relationship with the health care delivery entities had routine interaction with them because of their existing relationship (p-value of . ). their usual method for exchanging information was through informal medium such as phones, emails and faxes and formal medium such as memorandums, letters and contracts. when the dental schools were asked about the communication between the health systems' emr and the school's edr, majority of the dental schools did not have any communication ( %) or did not know is such a communication existed ( %) (p-value of < . ) (see fig. . ). out of the % (n = ) of the responded dental schools who's edr did not communicate with the health system's emr, % (n = ) of the dental schools stated that they did not need to exchange patient information electronically as a reason for the non-communication, while % (n = ) dental schools states that they would like to exchange patient information electronically but there were barriers that prevent them from doing so. some of the barriers identifi ed by these dental schools were (a). the hospitals and the dental school are not part of the same medial system and hipaa concerns prevent sharing data, (b). the dental school currently neither did have an edr nor the infrastructure to support one and (c). hospital is not interested and has high and perhaps unrealistic security standards. the remaining % (n = ) of the dental schools expected to exchange patient information electronically in the near future (next years). some of the information categories that were shared between the edr and emr in the small number of dental schools are illustrated in fig . . finally when asked about any research projects under way in their dental school to investigate discrepancies between medical and dental records for the same patient, only ( %) dental school was currently undertaking such project. in all common diseases, including those that affect the oral cavity, both the environment and genetics are pathogenic conspirators. unfortunately, we currently know little about the specifi c mechanisms underlying any common disease; and oral diseases are among the least understood. elucidating the etiology of chronic oral diseases will involve a synthesis of results from careful experiments of environmental exposures such as diet and tobacco use, the oral microbiome, co-morbidities, largescale, well-designed genetic studies, and the various interaction effects. with regard to genetics, the past few decades have witnessed transformative developments in our ability to interrogate the entire genome for genes that contribute to disease. while dramatic advances in experimental designs, statistical approaches, and clinical insights have greatly aided this scientifi c campaign, the central driver of this progress has been the development of high-throughput, inexpensive genetic technologies. following initial molecular studies using variant forms of enzymes, or allozymes, a major breakthrough was the use of highly informative dna-based markers throughout the genome (botstein et al. ) . this idea of directly assaying existing dna variation to conduct linkage and association studies in genetics began a revolution in disease gene mapping. recent interest from commercial entities has produced a feverish pace of technological innovation, markedly reducing cost and expanding the depth of inquiry. previously unfathomable, the testing of over one million single nucleotide polymorphisms (snps) in thousands of patients and controls is now commonplace (wellcome trust case control consortium ; schaefer et al. ) ; and very recently, next generation sequencing technologies have progressed to the point where sequencing of the entire protein-coding portion of the genome (exome) or even the entire genome is a costeffective method to examine disease traits across the entire spectrum of genetic variants in small numbers of affected individuals (ng et al. ) . there is little doubt that soon whole genome sequencing will be applied to nuclear family-based designs, studies among distantly-related affected individuals in extended pedigrees, and case/control studies involving thousands of individuals. this unprecedented scope of inquiry made possible by large-scale genetics, has begun to yield fascinating resulting into predisposition to oral cancers, caries, and periodontal disease that will molecularly redefi ne these pathologies, explicate unique biological connections with related diseases, give impetus to the development of directed therapeutics, and indeed personalize medicine. still, much more genetic focus on oral disease phenotypes is required if we are to realize this medical impact in a timely fashion. as genetic technologies have allowed the progression of interrogating single protein variants to single dna markers to entire genes to markers across the genome, and now to the entire genome sequence, the promises of these large-scale genetic studies have understandably undergone monumental expansion. it may be reasonable to expect the results from whole genome sequencing to decidedly revolutionize medicine within the next two decades. however, this new scientifi c capacity comes at a cost. as genetics, and biology in general, transitions to a data-rich science, practitioners have found themselves woefully unprepared to store and analyze the volume of data generated. once analyzed, interpretation and integration of these abundant and multifaceted results into medical practice will also be an appreciable challenge. insuffi cient assimilation of genetic fi ndings into merged dental and medical records will severely limit the ability of clinicians to appropriately treat patients. inadequately addressing these informatics issues will severely derail efforts in the basic sciences efforts as well as the translational and clinical sciences. this chapter explores the current state of genomics studies, what we have learned from genetic investigations into oral diseases, and where we may be headed. genetic studies have much to offer investigations of disease etiology. why do some acquire diseases and others do not? for those affected, why do some progress more rapidly than others? what causes some patients to respond to therapies, while others suffer from adverse reactions? these are all fundamental questions in both biology and medicine, whether the focus is on the gastrointestinal tract, the hippocampus, the lymphatic system, metabolic disorders, or oral diseases. speaking generally across disease areas, a portion of the answers to these questions often lies in described environmental effects. in numerous chronic diseases, infectious agents are likely contributors to the disease process -periodontitis, for example, is initiated by gram negative anaerobes in susceptible individuals (holt and ebersole ) . surely, unique and latent environmental exposures provide a random component to common disease susceptibility and progression. through twin studies, studies of risk in close relatives, and quantitative traits experiments, it is well-understood that heritable factors, including but not limited to dna variation, are typically responsible for - % of the phenotype variability for common diseases. this section will attempt to cover, at least at a cursory level, the major salient developments affecting genetic insights into chronic and aggressive periodontitis, with some comment on genetic factors infl uencing susceptibility to caries and oral cancers. while it would be extremely naïve to view genetic studies as an immediate panacea for our ills, the discovery of disease-causing genes does illuminate hitherto unknown biological pathways and molecular mechanisms, draws unforeseen connections with other traits, may improve prognostic models applicable for individuals, and suggests specifi c therapeutics. industrialization has brought forth increased lifespan and wellness through vaccination, modern sanitation practices, public health policies, and advances in medical science translated into practice. however, the accompanying physical inactivity coupled with a high calorie diet are probable contributors to an extremely common, chronically infl amed metabolic syndrome (hotamisligil ) that is thought to give rise to a multitude of intimately related disease traits: insulin resistance, compromised insulin signaling, hyperglycemia, obesity, dyslipidemia, hypertension, impaired kidney function, elevated liver enzymes and steatohepatitis, poor wound healing, neurodegeneration, vascular disease, pregnancy complications, accelerated immunosenescence, and periodontal disease (ford et al. ; ferrannini et al. ; eaton et al. ; holvoet et al. ; speliotes et al. ; eckel et al. ; d'aiuto et al. ) . these diseases often co-occur within the same patient and could be considered variable expression complications arising from a state of aberrant caloric fl ux that induces metabolic dysfunction and chronic, systemic infl ammation. these features constitute a disruption in a fundamental homeostatic mechanism with intensifying pathogenic consequences. the rapidly increasing incidence and decreasing age of onset for this pathophysiological state have generated a major source of mortality and morbidity in modern cultures (ford et al. ; ferrannini et al. ; weiss et al. ) . it is becoming increasing clear that many chronic diseases have an infectious component. there is relatively convincing evidence that many systemic, t-cell mediated autoimmune disorders may be initiated by infections. for example, from archaeological data, it is believed that an infectious agent -currently unknown -is necessary for rheumatoid arthritis (firestein ) , and both guillain-barre syndrome and rheumatic fever have well-described pathogeneses triggered by specifi c infections in susceptible individuals (bach ) . in many instances, oncogenesis and tumor progression can be traced to pro-infl ammatory responses at the site of chronic infection (coussens and werb ) , although it is not known whether these effects are mediated through the actions of the immune system, the infectious agents, or a combination thereof. several cancers fall into this category including gastric adenocarcinoma (uemura et al. ) , cervical cancer (walboomers et al. ) , hepatocellular carcinoma (saito et al. ) , and kaposi's sarcoma (dictor ) , all having unequivocal infectious agent etiologies. recent fi ndings of antiinfl ammatory pharmaceuticals, particularly those that inhibit cox- and cox- , reduce the incidence of certain classes of cancers are consistent with this view (dannenberg and subbaramaiah , rothwell rothwell et al. ) . in addition, there is moderate evidence that several bacteria -the most studied is chlamydia pneumoniae -play a role in atherosclerosis and myocardial infarction (saikku et al. ; watson and alp ) , however the studies are not conclusive and antibiotic treatment does not appear to be effective (andraws et al. ) . chronic periodontal disease is fi rmly footed at the intersection of infection, chronic infl ammation, and metabolic dysfunction. chronic periodontitis is characterized by infl ammation of the periodontal membrane, slowly causing gingival recession and eventual bone loss. the proximate cause of periodontitis is the virulent oral microbiome. the involvement of gram negative anaerobes has been fi rmly established for the disease. aside from the known oral pathogenic species p. gingivalis , t. denticola , and t. forsythensis , the so-called "red complex" (holt and ebersole ) , new bacterial species associated with chronic periodontitis have also been described (kumar et al. ) . the advent of an extensive database covering the oral microbiome will surely propel such investigations . numerous studies have shown that periodontal disease covaries with many diseases, presumably due to overlapping molecular etiologies. compelling meta-analyses demonstrate a highly signifi cant synchronicity of obesity and periodontal disease (chaffee and weston ) . in addition, the correlation between periodontal diseases/ alveolar bone loss and frank metabolic syndrome is repetitively observed (nesbitt et al. ; andriankaja et al. ) . extensive work has also shown a strong role for both infl ammation-related genes and circulating infl ammatory markers in periodontal disease (nikolopoulos et al. ; bretz et al. a, b ) . treatment studies further support the link between periodontal disease and immuno-metabolic syndrome. these experiments have demonstrated a signifi cant improvement in intermediate molecular markers of infl ammation when chronic periodontitis in the presence of metabolic syndrome (acharya et al. ) or type diabetes (iwamoto et al. ) was treated. conversely, treatment of periodontal disease with reduction of bacterial load leads to greater glycemic control among diabetic patients (simpson et al. ; stewart et al. ) . given the high prevalence of periodontitis and the co-morbidity of metabolic syndrome with periodontal disease, these treatment experiments appear to suggest that the virulent oral microbiome could play an important role in the pathogenesis of systemic infl ammatory metabolic syndrome, and is exacerbated by the syndrome. certainly, further studies are needed to defi nitively answer this question. as chronic periodontal disease seems to be a critical feature of sustained, systemic dysfunction of both metabolic and infl ammatory networks, uncovering the genetic variants carried by susceptible individuals would not only provide much needed insight into the molecular pathogenesis of chronic periodontal disease, but would also markedly aid our understanding of the infl ammatory metabolic syndrome and how it drives related co-morbidities. such genetic studies may also shed light on the specifi c mechanisms that appear to improve cardiovascular, infl ammatory, and diabetic outcomes when periodontal disease is treated, potentially leading to therapies and medical/dental intervention with greater effectiveness. such studies may also provide clues to which subsets of individuals respond more effectively than others and why they do so. periodontal disease can also present in a rapid manner with aggressive bone loss and early-onset. this is termed aggressive periodontitis (lang et al. ) . in contrast to chronic periodontitis, there is often a greater degree of familial aggregation with aggressive periodontitis, and it is hypothesized that most aggressive cases may affl ict individuals with one or more defective immune genes (zhang et al. ; amer et al. ; machulla et al. ; carvalho et al. ; toomes et al. ; hart et al. ; hewitt et al. ) . mutations in the lysosomal protease, cathepsin c, have been shown to be responsible for some forms of aggressive periodontitis, along with complications associated with other infl ammatory diseases (laine and busch-petersen ) . the specifi c hla variants thought to play a role in aggressive periodontitis, are also involved in infectious disease susceptibility and autoimmunity; and, interestingly, two of the non-mhc-linked regions, fam c and a locus on chromosome p , have been implicated in myocardial infarction (connelly et al. ) and may have action as a tumor suppressor in tongue squamous cell carcinoma (kuroiwa et al. ) . as with chronic periodontal disease, an infectious microbiome is heavily involved. however, in general, microbiome differences could not explain the presence of chronic versus aggressive forms of the disease, although in some aggressive periodontitis patients, a highly leukotoxic a. actinomycetemcomitans strain may contribute to the disease process (mombelli et al. ) . we currently do not fully know the differences between the genetic susceptibility factors for the chronic and aggressive forms of the disease. the most prevalent chronic disease in both children and adults is dental caries (national institute of dental and craniofacial research) . caries formation is a complex disease with several interacting components form the environment and host genetics. similar to gingivitis and periodontitis, caries have an infection-initiating etiology with acidifi cation leading to localized demineralization. epidemiological studies have long shown that diet is a strong predictor of caries formation; and the reduction in ph is exacerbated by high consumption of carbohydrates. the principal pathobacterial species are streptococcus mutans and lactobacillus ( van houte ) . there are also several reports of positive correlations of caries with infl ammatory diseases, although the association is not always repeatable. it is also not clear what proportion of the putative association with infl ammatory disease is due to innate upregulation of immune networks in contrast to the immuno-modulating pharmaceuticals prescribed to those with infl ammatory disease (steinbacher and glick ) . much of the effect is reported to result from lack of saliva volume (steinbacher and glick ) . interestingly, the presence of epilepsy may be associated with higher caries rates (anjomshoaa et al. ) . fluoride is an effective antimicrobial agent that interferes with bacterial growth and metabolism (wiegand et al. ) . hence, topical fl uoride administration as well as ingestion of fl uoridated water inhibits cariogenesis and caries progression (ripa ) . amelogenesis is a key process involved in modifying the rate of caries formation. both common variation and rare mutations in enamel formation genes such as amelogenin and enamelin are involved in caries rates (patir et al. ; kim et al. ; crawford et al. ) , the molecular actions of which are beginning to be revealed (lakshminarayanan et al. ) . over , new cases of cancers affecting the oral cavity and pharynx were expected in the united states for , with deaths numbering , (jemal et al. ) . the majority of these malignancies involved solid tumors originating from cancerous changes in squamous cells of the mouth. again, oral cancers have a complex etiology existing of entangled genetic, epigenetic, infectious, and dietary causes, further modifi ed by tobacco, alcohol and other environmental exposures. as with most cancers, it is reasonable to expect that both germline and somatic genetic changes will be involved in carcinogenesis, tumor growth, and metastasis. promoter hypermethylation of genes central to cellular growth, differentiation, dna fi delity, apoptosis, and metabolic stability is an important facet of these cancers (poage et al. ) . indeed, methylation-mediated silencing of genes involved in tumor suppression (e.g. the cyclin-dependent kinase inhibitor a), detoxifi cation (e.g. mgmt ), and apoptosis (e.g. the death-associated protein kinase- ) are commonly found in oral squamous cell carcinoma samples (ha and califano ) . to quantify the proportion of the variance in a phenotypic trait that is due to variance in genetic factors, population geneticists defi ned the concept of heritability (visscher et al. ; falconer and mackay ) . researchers subsequently developed several methods for estimating heritabilities using the measure of a trait (e.g. occurrence of disease/not-disease) in combinations of relatives (e.g. parentoffspring, or monozygotic-dizygotic twins). in general, the higher the measured heritability of a variable phenotype, the larger the contribution of genetic factors is in comparison to environmental effects. it is fallacious to assume that the heritable variation is composed entirely of alleles residing in the dna sequence, for heritability studies simply examine the covariance between relatives without comment on specifi c molecular mechanisms. hence, any heritable variation such as methylation patterns, vertically-transmitted infectious agents, as well as dna variation can contribute to the heritability measure. heritability results are important because they not only give a rough estimate of the collective effects of heritable factors, but also can provide a measure to quantify how much of the total genetic effect is accounted for by specifi c loci examined. for periodontal disease, four twin-based studies of heritability have been performed (michalowicz et al. ; corey et al. ; michalowicz et al. ; mucci et al. ) . although varying in sample size and methodological details, all four arrived at consistent results, with - % of the variance in periodontal disease being attributed to genetic variability for chronic periodontitis. given the segregation patterns described in the literature, it is reasonable to assume that aggressive periodontal disease exhibits a higher heritability. therefore, given the prevalence of periodontal disease, heritable factors within the population at large are likely appreciable. using twin pairs, bretz and colleagues reported substantial heritability values for multiple traits related to caries ranging from % to % (bretz et al. a, b ) . lastly, mutagen sensitivity studies of head and neck cancer patients suggest a signifi cant effect of genetic factors for the carcinogenesis of oral cancers (cloos et al. ) . hence, there is every reason to believe that a sizable pool of genetic and/or epigenetic factors await discovery for oral diseases. once the development of pcr (saiki et al. ) was applied to the idea of using naturally-occurring dna variation (botstein et al. ) , large-scale dna-based studies of disease underwent a substantial acceleration (schlotterer ) . genotyping of short, tandem repeated sequences (weber and may ) -microsatellites -spurred on a wave of genome-wide linkage studies, which evaluate the co-segregation of disease state with microsatellite markers, for both rare mendelian disorders as well as more common diseases with complex inheritance patterns. while the rarer traits with more coherent transmission patterns generally relinquished their genetic secrets to linkage analysis, more common diseases did not. in the mid-to late s, several theoretical studies had shown that the power to detect disease-causing alleles is higher with association-based designs such as a case/control experiment or association in the presence of a linkage signal as in the transmission/disequilibrium test if the frequency of those alleles is high and the effects are moderate (kaplan et al. ; risch and merikangas ; jones ; long and langley ) . however, to conduct genomewide association studies presented an ominous obstacle for the genetic technologies at the time. the number of markers required to effectively cover the genome was prohibitively large as the chromosomal blocks in population-based samples used in association designs were expected to be small. even within large extended families, the limited number of recombination events generates substantial chromosomal blocks passed through the pedigree, but researchers had both theoretical and empirical evidence that the blocks in population-based samples were on the order of k base pairs for most large human populations. as the reader can imagine, the mean length of blocks that are shared by descent is inversely related to the product of recombination rate, the number of affected individuals and the number of meioses separating the affected individuals. in practice, even very large extended families segregate regions shared by affected members on the order of several million base pairs in length. however, once geneticists seriously considered large-scale studies using a case/control design where individuals are separated by say , meioses, it became clear that to adequately cover the much smaller shared regions across the entire genome, hundreds of thousands of markers would be required (kruglyak ) . utilizing the human genome sequence (venter et al. ; lander et al. ) , a number of studies at celera diagnostics provided an intermediate solution, where approximately , putative functional snps primarily located in genes were assayed through allelespecifi c pcr in a number of common diseases using a staged case/control design. these studies were successful in identifying several gene-centric polymorphisms associated with common diseases (begovich et al. ; cargill et al. ) (fig. . ) . concurrently, several groups had performed extensive sequencing and genotyping across the genome to produce a genome-wide map of haplotype structure (hinds et al. ) , useful in linkage disequilibrium mapping. within years, technology for snp hybridization arrays had advanced so as to enable genome-wide association studies capable of capturing most of the common genetic variation in the genome either through direct genotyping or indirect interrogation using linkage disequilibrium -the term linkage disequilibrium is a measure of the correlation of alleles at closely-linked sites (see fig. . ). these investigations were met with numerous successes (klein et al. ; kathiresan et al. ; graham et al. ; gudmundsson et al. ) . inexpensive genotyping platforms and urging from theoreticians ensured that these genome-wide association studies were, in general, highly powered to detect all but very mild effects from high frequency alleles. these efforts, led by large academic consortia such as the wellcome trust, the international multiple sclerosis genetics consortium, and the broad institute and commercial entities such as decode genetics and perlegen have greatly expanded our understanding of the basic biology of common diseases: we now know, for example, that (i) autophagy-related genes are involved in crohn's disease (rioux et al. ) , (ii) there are a number of genes such as the protein tyrosine phosphatase, ptpn and the interleukin- receptor, il r , that exhibit ample pleiotropic effects among autoimmune conditions (lopez-escamez ; safrany and melegh ) , (iii) in the case of age-related macular degeneration, predictive models using the genetic results enable fairly accurate prognosis of individuals who are at high risk of disease (seddon et al. transcription factor tcf l plays a role in type diabetes (grant et al. ) , and (v) aberrant il- signaling likely contributes to multiple sclerosis susceptibility (gregory et al. ) . the plot shows the tremendous progress in genotyping technology where, a decade ago, very little of the genome was accessible for disease studies using association designs through the current wave of viable sequencing-based whole exome studies ( ) ( ) and whole genome studies ( ) ( ) . in fig. . , the average distance between adjacent genetic markers is plotted as a function of year of introduction to the disease mapping community. impressively, the total number of genetic markers has increased a million-fold over the past decade. although successful in uncovering numerous pathogenic pathways for common diseases, results from the current wave of genome-wide association studies, with a few exceptions, explain little of existing disease heritability. the reasons for this are cryptic and the subject of heavy debate (manolio et al. ) . multiple rare sequence variants generating high levels of allelic heterogeneity, functional de novo mutations, structural mutations such as copy number variants and large deletions, and epigenetic effects constitute four of several possible disease models that could account for the heritability discrepancy. the answer will almost certainly consist of a conglomeration of these and other effects. bringing forth the new genome-wide technologies that illuminate these previously non-or under-interrogated properties of the genome to bear on this enigma is a reasonable next step for all complex traits including oral diseases. the most commonly used measure of ld in a sample of chromosomes is linkage disequilibrium (ld) is a measure of the correlation between alleles at two sites in a sample of chromosomes. for two biallelic sites, if the a allele is always paired with the b allele, and the a allele is always on the same haplotype as the b allele, then the two sites are said to be in perfect ld. successive recombination diminishes ld. interrogating one site for disease association allows investigators to indirectly interrogate other sites in sufficiently high ld with the interrogated site. a key feature explicitly studied in molecular population genetics and implicitly used in disease gene mapping studies is the site frequency spectrum; that is, the distribution of allele frequencies at single sites in the genome that vary in the human population studied. from both diffusion models (kimura ) and coalescent theory (hudson ) in theoretical population genetics, we know that the vast majority of realistic models generate many more rare variants compared to common polymorphisms. this is particularly true for expanding populations. are these rare variants the source of much of the missing heritability? recently, with the application of high-throughput sequencing technology to human studies over the past decade, empirical studies have clearly verifi ed these predictions -the large majority of variants have low frequencies (the international hapmap consortium ) . the distribution of deletions appears to be skewed toward more rare frequencies, presumably due to the deleterious effects of such variants. individual mutations appearing de novo typically are extremely rare events per locus, but collectively are numerous. other types of genetic variability, such as copy number repeats, span both ends of the frequency spectrum with the preponderance of the markers being rare. thus, there is a sizable pool of low-frequency variants in human populations that have yet to be thoroughly investigated. over the past few years it has become increasingly clear that structural variants exist in the human genome at a far higher rate than previously thought. structural variants can exist in a multitude of forms including deletions, copy number variants, and inversions among others. due to the nature of these genetic changes, many are considered to be highly disruptive of molecular function if they lie in functional motifs. indeed, there are several mendelian diseases are caused by fully-penetrant structural variants impacting a chromosomal region (lupski ) . numerous structural variants have recently been reported to be associated with common diseases, particularly in the neurological fi eld (sebat et al. ; stefansson et al. ; elia et al. ) , infectious disease susceptibility (gonzalez et al. ) , and drug metabolism (zackrisson et al. ) . although they have improved dramatically over the past few years, algorithms using snp-based data from hybridization arrays to infer copy number variants have had high error rates, perhaps explaining the rather low rates of replication of structural variation association results for common diseases. nevertheless, given the high frequency of structural variants, their pathogenic potential, and that we are on the precipice of a sequencing revolution in genome-wide studies, examination of these variants should be a high priority for new sequencing-based studies in oral disease susceptibility, progression, and related pharmacogenetic applications. as different technologies examine different portions of the site frequency spectrum (i.e. genome-wide snp scans interrogate variation that is common in the hapmap populations, whereas sequencing-based studies typically interrogate the entire frequency spectrum), where one believes genetic causation is harbored should infl uence the selection of genotyping technology. if common genetic variation contains the vast majority of heritable effects on disease phenotypes, then an investigator would be wise to employ a snp-based experimental design. if, however, there is reason to believe that a signifi cant portion of the genetic load of the disease studied exists in the highly populated portion of the distribution -the rare variants -then a sequencing-based study may be better suited to unravel causative alleles. the studies of heritability discussed previously show that there is heritable variation underlying a substantial portion of the variance observed in oral diseases. as discussed above, sequencing technologies may address many aspects of dna variation including copy number loci, rare haplotypes, inversions, and insertions/deletions, but it is also worthwhile to repeat that the molecular mechanisms for disease heritability are not necessarily limited to variation at the dna level. for a disease state, the covariance between relatives could be driven by co-inherited chromosomal regions or other phenomena. chief alternative heritable mechanisms include dna methylation (hammoud et al. ) , modifi cations to the histones (bestor ) , complex rna zygotic transfer (rassoulzadegan et al. ) , and vertical transmission of infectious agents. additionally, transgenerational effects offer an intriguing class of epigenetic mechanisms (nadeau ) . in a thorough review on epigenetics and periodontitis, gomez et al. make a strong argument for consideration of both cpg dinucleotide methylation and deacetylation actions on cytokine expression as a credible avenue for further investigation in periodontal disease etiology (gomez et al. ) . genome-wide epigenetic studies have been successfully conducted for oral cancers (poage et al. ) . the scale of this study on head and neck squamous cell carcinomas allowed these researchers to show a global pattern of tumor copy number changes signifi cantly correlated with methylation profi les that was not detectable at the individual gene promoter level. with advanced chromatin immunoprecipitation and new methods to study dna methylation, efforts to apply highthroughput epigenetic methods to oral diseases should be accelerated. numerous studies have been conducted in oral disease traits using a candidate gene approach. there are two large reviews of the existing candidate gene results (nikolopoulos et al. ; ) . laine and colleagues have recently put together a comprehensive review article covering gene polymorphisms. there are some suggestive fi ndings for cyclooxygenase- gene, cox- , the cytokineencoding genes, il and il b , the vitamin d receptor, vdr , a polymorphism immediately upstream of cd , and the matrix metalloproteinase- gene, mmp . however, these initial results will require further confi rmation, for the association patterns are inconsistent across independent studies, the statistical signifi cance is moderate, and the posterior probability of disease is decidedly bland. the striking pattern that emerges from the laine et al. summary data is the lack of coherent replication of genetic association for the vast majority of polymorphisms examined. the situation is reminiscent of genetic association studies prior to large-scale snp studies where poor repeatability of results plagued the fi eld. in a pivotal study from , hirschhorn and colleagues (hirschhorn et al. ) examined the state of genetic association studies, fi nding that "of the putative associations that had been studied three or more times, only six have been consistently replicated." the dearth of robust results was largely remedied when large-scale genetic studies were applied to very substantial numbers of well-characterized patients and geneticallymatched controls and stringent statistical criteria enforced. one can only suspect that a similar state of affairs is operating in genetic studies of chronic periodontitis. perhaps efforts to ( ) reduce the heterogeneity of the disease state through detailed clinical and laboratory assessments, ( ) drastically increase sample sizes, and ( ) expand the scope of inquiry to larger numbers of genes/regions, and examine a more comprehensive set of variants/epigenetic effects will improve the current situation. the second large study is a meta-analysis of studies, where nikolopoulos and colleagues analyzed six cytokine polymorphisms linked to il a , il b , il , and tnf-alpha (nikolopoulos et al. ) . two of these, an upstream snp in il a and a snp in il b , exhibited signifi cant association with chronic periodontal disease risk. although the results were not particularly strong, as is typical with complex diseases, the results do suggest the importance of infl ammation-response variability in chronic periodontitis predisposition. perhaps the strongest, most replicable genetic association fi nding with coronary heart disease and myocardial infarction is centered on the short arm of chromosome ( p . ) (mcpherson et al. ; helgadottir et al. ) . two studies of periodontal disease showed that the same alleles at the p . locus confer risk for aggressive periodontitis (schaefer et al. ; ernst et al. ) . the discovery of such a pleiotropic locus may explain a portion of the aggregation of periodontal disease with other co-morbid conditions. further studies investigating overlapping genetic susceptibility factors between periodontitis and cardiovascular disease, diabetes mellitus, metabolic syndrome, rheumatoid arthritis, and other related diseases may be a fruitful strategy for honing in on shared genes affecting these immuno-metabolic disorders. using patients from families from the philippines, the fi rst genome-wide linkage study for caries was completed in (vieira et al. ) . the study identifi ed fi ve loci which exhibit suggestive statistical evidence (lod scores exceeding . ): q . , q . , xq . , q . , and q . . the latter of which overlapped with a quantitative trait locus discovered from mapping work in the mouse. further work is necessary to refi ne these signals and localize the variants that may be driving these linkage signals. aggressive periodontal disease and rarer dental diseases have also been subjected to linkage analysis. results from linkage studies for dentinogenesis imperfecta type i, for example, have gone on to produce the novel gene fi ndings of the dentin sialophosphoprotein-encoding gene on q . being responsible (song et al. ; crosby et al. ) . a linkage study in african american families examining localized aggressive periodontitis found a strong linkage signal in a region covering approximately megabases on chromosome (li et al. ) . several interesting genes are in this region. in a study earlier this year further mapping from carvalho et al. in brazilian families identifi ed haplotypes in this region on q in fam c which were associated with aggressive periodontitis (carvalho et al. ) . the function of the fam c protein is not fully understood. fam c is localized in the mitochondria and it appears to play a role in vascular plaque dynamics and risk of myocardial infarction (laass et al. ) . it should also be noted here that other types of mapping analyses such as homozygosity mapping to identify have yielded gene discoveries. for example, the lysosomal protease cathepsin c gene for the recessively-inherited papillon-lefevre syndrome which is characterized by aggressive and progressive periodontitis was effectively mapped using homozygosity mapping (fischer et al. ; connelly et al. ) . cathepsin c is highly expressed in leukocytes and macrophages and is a key coordinating molecule in natural killer cells (rao et al. ; meade et al. ) . although sparse, these linkage results are undoubtedly encouraging. employing very large extended families subjected to genome-wide genotyping or sequencing will surely shed much needed light on chromosomal regions and genes relevant to oral disease research (fig. . ). for periodontitis, a single study has employed a genome-wide association design in an effort to uncover aggressive periodontal variants (schaefer et al. ) . this study by schaefer and colleagues discovered and replicated an intronic snp, rs , in the glycosyltransferase glt d which is signifi cantly correlated with aggressive periodontal disease in both german and dutch samples. often, seemingly signifi cant results from large studies are due to the effect of reporting the top result from a great many statistical tests -this is called the multiple testing problem. in this situation, the strength of the fi nding, along with the replication across three case/control studies, argues for true association with aggressive periodontal susceptibility. the snp may modulate the binding affi nity of gata- . the association with glt d is currently one of strongest genetic associations for aggressive periodontal disease, testifying to the power of genome-wide studies to generate novel, relevant molecular pathophysiology for complex diseases. it seems unlikely that glt d would be extremely high on a candidate gene list, and it was only through a genome-wide scan that it appeared. like many excellent studies, the fi nding by schaefer et al. raises more questions than it answers and will undoubtedly provide fertile ground for ensuing molecular work. after a somewhat sluggish start, due to a lack of critical mass of investigators aiming to collect large numbers of patient samples and bring high throughput genetic technologies to caries susceptibility, gingivitis, and periodontal disease traits, the future of genetic studies in oral health is bright. scientifi c progress in revealing the molecular pathogenesis of oral diseases is dependent on genome-wide genetic studies; and i have argued that progress in related immuno-metabolic diseases is also dependent on these large-scale genetic studies in periodontal disease. to study sporadic disease, substantial patient collection efforts are required for the application of these technologies. this may involve a combination of new recruitment and consortiumrelationships with existing collections. the beginning of such a collection for sporadic aggressive periodontitis in europe has shown extremely intriguing initial results, but more patients are needed to examine rare variants of moderate effect. both the german/dutch collection of aggressive periodontitis and the brazilian collection have begun to revolutionize the study of periodontal disease susceptibility with the discovery of glt d snps and fam c -linked haplotypes. there is little doubt that subsequent molecular work on these two genes will uncover novel mechanisms for the predisposition to aggressive periodontal disease. focus should also be placed on the collection of extended families segregating these diseases. applying sequencing technologies to large pedigrees can be an effective method of identifying rare variants and structural variants in a highly-refi ned phenotype. furthermore, applying these methods to the entire genome would make for a comprehensive genetic study. several trends in large-scale genomics science hold promise to signifi cantly advance our understanding of oral disease pathogenesis: the sociology of biological sciences has changed over the past years so as to • become more collaborative. essential for association-based designs, consortiumbased genetic research has blossomed over that time period, increasing sample sizes and therefore the power to detect disease-causing variants. there currently is consortium-based research in periodontal disease and oral cancer. further expanding these efforts will enhance subsequent studies, particularly those investigating rare alleles and/or rare epigenetic effects. through over a century of laboratory work, the collective knowledge of bio-• chemical pathways, signal transduction, cell physiology, regulatory mechanisms, and structural biochemistry is weighty. incorporation of this information into etiological models may substantially advance oral disease work as well as the fi eld of complex disease genetics in general. sophisticated analysis techniques are needed to perform this task. recent advances merging results from network science with probability theory within the context of computer science have produced the fi eld of machine learning. this rigorous framework can be used to identify those factors responsible for disease status and can also be used to develop robust predictive models using known biological networks and genetic data. the output from such models, typically the probability of disease, an estimate of disease progression rate, or a probability of adverse reaction, can be used by physicians and dentists to personalize medical care. until relatively recently, population genetics did not contribute a great deal to • human genetics research. that has changed in the past decade where effort spent on association studies surpassed that spent on family-based studies. those investigating disease gene mapping began to collaborate with population geneticists and population geneticists took up a wide-spread interest in fi nding disease alleles. incorporation of population genetics theory into such studies markedly improved association studies on several levels: confounding by population stratifi cation was effectively treated using population genetics, linkage disequilibrium patterns. use of population genetics theory in large-scale oral disease mapping studies may accelerate discoveries. sequencing technology has rapidly progressed over the past decade. currently, • sequencing studies across the exome can be accomplished at reasonable cost and yield data for all known genes in the genome. within the next few years, sequencing costs will depreciate to a point where whole-genome sequencing studies will be commonplace, using both family-based and population designs. application of these technologies to oral disease studies is imperative for comprehensive studies of etiology. high-throughput dna methylation and chromatin immunoprecipitation studies • will enable large-scale epigenetic studies in oral diseases (meade et al. ; ehrich et al. ; bibikova et al. ; ren et al. ; pokholok et al. ) . these have already started to play an important role in delineating mechanisms responsible for oral cancers (poage et al. ) . additional application of these techniques to studies of gingivitis, caries, and periodontal diseases may generate novel fi ndings. molecular biologists and pharmacologists have increasingly become able to • develop and evaluate highly targeted pharmaceuticals based on genetic discoveries. the use of such genetic information may improve the chances of developing effi cacious therapies. geneticists and disease researchers are beginning to realize that oral diseases • both impact and are intrinsically tied to susceptibility and progression of other common diseases. a synthesis of genetic fi ndings from immuno-metaboliclinked disorders would seem to greatly increase the knowledge of these diseases and better pinpoint their respective etiologies. as the new high-throughput genomics and epigenomics technologies become • implemented in oral disease research, the storage, management, analysis, and interpretation of the ensuing colossal amounts of data will be critical to enable clinicians to use these results in daily practice. advances in dental and medical informatics will facilitate these steps. we are in exciting times where advances in genetic technologies will uncover the genetic causes of diseases, including those that affect the oral cavity. with more focus in the area of oral disease genomics and the harnessing of new high-throughput sequencing and epigenetic technologies, novel insights into the pathways driving these diseases are imminent. these discoveries will, in turn, motivate directed therapies, aid in illuminating the molecular etiology of related disorders such as diabetes, and increase the level of personalized medicine. joseph kilsdonk the title of this section reinforces a institute of medicine (iom) report titled "dental education: at the crossroads." to quote yogi berra, a baseball sage: "when you come to a fork in the road, take it." the implication being that dental education must take action and move beyond its crossroads. these crossroads are described in the fi rst third of the section. it includes a summary and recommendations of the iom report and three transitional reports that followed: the surgeon general's report identifying oral health as a silent epidemic, the josiah macy foundation report, and a "pipeline" study funded by both the robert wood johnson and the california foundations. having been at the crossroads for a decade or so, the middle portion of the section highlights educational models that may lead to a more promising future. the later third of this section describes an alternative path of action for dental education which emphasizes the central roles of clinic-based education and dental informatics in dental education curriculum. it is unknown how traditional dental educators may view this model; however, it is effectively a logical conclusion and responsive to the reports. in the institute of medicine (iom) published "dental education at the crossroads" (field ) . the title was apropos as the authors' analysis concluded: ( ) economics surrounding dental education were unsustainable ; ( ) student service learning opportunities and access to care for patients were limited; and ( ) new dental schools were not replacing those forced to close due to the economic climate. the iom report additionally proposed key recommendations to reform dental education and service delivery. fifteen years later, we remain at "the crossroads" as these issues remain largely unresolved. furthermore, these recommendations have retained their validity. their implementation would directly impact structures and services for contemporary models of dental education in the future. the following iom recommendations (field ) are intrinsic to the proposed dental education reform: recommendation : to increase access to care and improve the oral health status of underserved populations… recommendation : to improve the availability of dental care in underserved areas and to limit the negative effects of high student debt… recommendation : to prepare future practitioners for more medically based modes of oral health care and more medically complicated patients, dental educators should work with their colleagues in medical schools and academic health centers to: move toward integrated basic science education for dental and medical • students; require and provide for dental students at least one rotation, clerkship or • equivalent experience in relevant areas of medicine and offer opportunities for additional elective experience in hospitals, nursing homes, ambulatory care clinics and other settings; continue and expand experiments with combined md-dds programs and • similar programs for interested students and residents; increase the experience of dental faculty in clinical medicine so that they, • and not just physicians, can impart medical knowledge to dental students and serve as role models for them. recommendation : to prepare students and faculty for an environment that will demand increasing effi ciency, accountability, and evidence of effectiveness, the committee recommends that dental students and faculty participate in effi ciently managed clinics and faculty practices in which the following occurs: patient-centered, comprehensive care is the norm; • patients' preferences and their social, economic, and emotional circumstances • are sensitively considered; teamwork and cost-effective use of well-trained allied dental personnel are • stressed; evaluations of practice patterns and of the outcomes of care guide actions to • improve both the quality and the effi ciency of such care; general dentists serve as role models in the appropriate treatment and referral • of patients needing advanced therapies; larger numbers of patients, including those with more diverse characteristics • and clinical problems, are served. recommendation : because no single fi nancing strategy exists, the committee recommends that dental schools individually and, when appropriate collectively evaluate and implement a mix of actions to reduce costs and increase revenues. potential strategies, each of which needs to be guided by solid fi nancial information and projections as well as educational and other considerations, include the following: increasing the productivity, quality, effi ciency, and profi tability of faculty • practice plans, student clinics, and other patient care activities; pursuing fi nancial support at the federal, state, and local levels for patient-• centered predoctoral and postdoctoral dental education, including adequate reimbursement of services for medicaid and indigent populations and contractual or other arrangements for states without dental schools to support the education of some of their students in states with dental schools; rethinking basic models of dental education and experimenting with less • costly alternatives; raising tuition for in or out-of-state students if current tuition and fees are low • compared to similar schools; developing high-quality, competitive research and continuing education • programs; consolidating or merging courses, departments, programs, and even entire • schools. in summary, the iom report identifi ed that: ( ) an outdated curriculum continues to be retained which refl ects past dental practice rather than current and emerging practice and knowledge; ( ) clinical education does not suffi ciently incorporate the goal of comprehensive care, with instruction focusing too heavily on procedures; ( ) medical care and dentistry are not integrated; and ( ) the curriculum is crowded with redundant material, often taught in disciplinary silos. the iom's report was followed by the surgeon general's report on oral health in and a subsequent supplement by the surgeon general in called "the national call to action" (u.s. department of health and human services ) . five signifi cant fi ndings and recommendations from the surgeon general's report(s) that have implications pertaining to the envisioned structure and services of new models for dental education include: changing the perception of oral health so that it will no longer be considered • separate from general health; improving oral health care delivery by reducing disparities associated with popu-• lations whose access to dental treatment is compromised by poverty, limited education or language skills, geographic isolation, age, gender, disability, or an existing medical condition; encouraging oral health research, expanding preventive and early detection pro-• grams, and facilitating the transfer of knowledge about them to the general population; increasing oral health workforce diversity, capacity, and fl exibility to overcome • the underrepresentation of specifi c racial and ethnic groups in the dental profession. in this regard, the national call to action urged the development of dental school recruitment programs to correct these disparities and to encourage parttime dental service in community clinics in areas of oral health shortage; increasing collaboration between the private sector and the public sector to cre-• ate the kind of cross-disciplinary, culturally sensitive, community-based, and community-wide efforts to expand initiatives for oral health promotion and dental disease prevention. spurred by the iom report and the surgeon general's report, the josiah macy foundation ( ) conducted a study entitled "new models of dental education." the study was prompted by concerns about declines in dental school budgets and the diffi culties experienced by schools in meeting their educational, research, and service missions. the macy study concluded that: financial problems of dental schools are real and certain to increase. • current responses of schools to these economic challenges are not adequate. • most promising solutions require new models of clinical dental education. • macy study lead researcher dr howard bailit, and his team recently concluded in reference to points one and two above, that: "if current trends (to aforementioned) continue for the next years, there is little doubt that the term crisis will describe the situation faced by dental schools. further, assuming that it will take at least ten or even more years to address and resolve these fi nancial problems, now is the time for dental educators, practitioners, and other interested parties from the private and public sectors to come to a consensus on how to deal with the coming crisis. clearly, these fi nancial problems will not be solved by minor adjustments to the curriculum, modest improvements in the clinical productivity of students or faculty, or even signifi cant increases in contributions from alumni. the solutions 'must involve basic structural changes in the way dental education is fi nanced and organized' (bailit et al. ) ." this statement is supported by the fact that in the past years more dental schools have closed than opened. specifi cally eight schools have closed, whereas to date a couple has opened and a handful is pending. curriculum relevance was also a focus of the study. findings concluded that "changing the curriculums in dental schools to allow students to spend more time in community venues would be highly benefi cial to both society and student. society benefi ted from having underserved patients cared for while students were assessed as being fi ve to ten times more productive, more profi cient, more confi dent, more technically skilled and more competent in treating and interacting with minority patients" (brodeur ) . macy study (formicola et al. ) outcomes represented signifi cant and foundational guideposts for assessing and planning any future models for dental education. their report led to the robert wood johnson foundation pipeline study ( ) , a major research study funded by the robert wood johnson foundation and the california endowment (tce). the goal of the dental pipeline program was to reduce disparities in access to dental care. the pipeline study provided over $ million for the start up or expansion of schools and student clinical programs that incorporated services to underserved extramural clinical settings (primarily community health centers). the following recommendations from the surgeon general's report structured • the goals of the pipeline's initiative: increase the number of under-represented minority and low-income students enrolled in the dental schools participating in the pipeline program so that there would be a voice of minority and low-income students at all the funded schools. provide dental students with courses and clinical experience that would prepare • them for treating disadvantaged patients in community sites. have senior dental students spend an average of days in community clinics • and practices treating underserved patients. increasing the community experience of dental students was expected to have an immediate impact on increasing care to underserved patients (brodeur ) . this third point is pivotal to future success of dental curricula and dental education economics. recently published in a supplemental volume to the journal of dental education , february, , the pipeline study reported the following outcomes: minority recruitment of low-income students increased by %; • the rate of recruitment for under-represented populations was almost twice that • of non-pipeline schools; the length of time dental students spent in extramural rotations increased from a • mean of days to a mean of days over a period of years. procedural profi ciency increased compared to that of their non-extramural peers. of the pipeline-funded programs, only four schools achieved the goal of • days of extramural rotations; through extra funding from tce, the four schools extended extramural rotations to an average of days; based on this publication, it appears that only a handful of pipeline schools defi -• nitely plan to sustain their extended extramural rotations. financial concerns were highlighted as the major problem in sustaining future recruitment and placement of students beyond the timeframe of the study; a survey of program seniors indicated a mean of % [range of - % by • school] were planning to devote ³ % of their practice to serving minority patients. only % [range of - % by school] were planning to practice at community clinics. in the context of these outcomes, discussion indicated that the unwillingness of students to practice in underserved settings was based on several factors: students that participated were already enrolled in traditional programs and were • not necessarily seeking a pipeline experience or a future in community service. concern over future reimbursement as a provider in a community setting; • limited time spent in underserved settings; • limited loan forgiveness scholarship opportunities. • the fact that the large majority of pipelines were unsustainable was attributed to lack of productivity in the school clinics while the students were on rotation at community based clinics. schools generate meager, yet necessary revenue streams on intramural student clinical activity to support the costly clinical and faculty infrastructure. currently, similar economic constraints involved with outsourcing students to serving rural and underserved populations impacts the ability of tradition dental schools to participate in sustained outreach programs. most recently, the pew center on the states national academy for health policy ( ) released "help wanted: a policy makers guide to new dental providers". this report provided an excellent summary outlining workforce needs, access issues, and strategies for dental-related services to help states and institutions develop creative ways to solve oral health access and care issues. the guide proposes the following relevant components and trends for consideration in development of future sustainable school models: dental colleges are willing to bear a large and disproportionate share of the burden • in terms of access to care, particularly during a time of incredibly scarce resources. expanded, extensive, and/or creative extramural rotations have been developed • in recent years under the conceptual umbrella of service-learning. these often involve clinics providing direct or indirect payment to dental schools or clinics managed in some way by dental schools. dental education has certain obligations. first, education must adhere to accreditation standards with the goal of producing competent practitioners. second, education must remain responsive and impact the societal need for care. lastly, the delivery of dental education must be economically sustainable. the macy, rwjf, and iom reports note that improved oral health, sustainable dental education economic models, and competent workforce pipelines converge around community health centers (chc). university of michigan researchers fitzgerald and piskorowski ( ) reaffi rm this conclusion in an evaluation of an ongoing -year program, stating that: (the chc model) is self-sustaining and can be used to increase service to the underserved and increase the value of students' clinical educational experiences without requiring grant or school funding, thus improving the value of dental education without increased cost. self-sustaining contracts with seven federally qualifi ed health centers (fqhcs) have resulted in win-win-win-win outcomes: win for the underserved communities, which experienced increased access to care; win for the fqhcs, which experienced increased and more consistent productivity; win for the students, who increased their clinical skills and broadened their experience base; and win for the school in the form of predictable and continuing full coverage of all program costs (fitzgerald and piskorowski ) (fig. . ). however, unlike medicine that outsources their students to clinical sites, dental education programs retain the majority of the student time within their own "clinical laboratories" as documented by the aforementioned studies, this limits students' exposure to extramural experiences. costs to operate such intramural clinical programs are ever increasing and many schools' clinical operations run defi cits. if that component can be outsourced to community-based resources such as a chc, then the burden of cost is shifted away from the school. an example would be a.t. still university's arizona school of dentistry and oral health (asdoh) which matriculated its fi rst class in . at the prompting of the state's community based clinics, asdoh designed a program that placed students into community-based settings for up to months, an unprecedented length of time for an extramural rotation. they also saw this as an opportunity to use an adjunct centric faculty that signifi cantly reduced traditional education overhead. through this innovation, the school was able to develop a program that was sustained by "fair market" value tuition and trained students where community needs were greatest for up to months (which was then unprecedented). conversely, if the chc can rely on student service-learning to care for patients, the cost of care is reduced. other schools are also advancing with innovative education and care delivery. adea's charting progress (valachovic ) fig. . the synergy between access to care, student competency, and fi nancially sustainable dental education converge around chc/fqhcs little rock, arkansas; and the university of southern nevada in south jordan, utah. western university is planning placement of % of their fourth year class in community health centers, while east carolina is seeking to set up rural clinical campuses as well as clinical partnerships with the state's fqhc. at the time of this publication, several existing schools are expanding or looking to expand including the university of north carolina, marquette university, midwestern university in downer's grove, il. such expansions will contribute to solving the existent access supply and demand issues. however, it was observed even with all the start ups and expansions, graduation numbers will not approach the output of schools in the late s and early s. these creative models establish the foundation for a sustainable clinic structure by generating self-sustaining revenue through student service-learning, which, unlike medical student services, are billable. simultaneously these new models provide access to care for the needy while student exposure to clinical experiences that are often not available in academic patient pools. these models also shift some of the cost of providing clinical education from the dental college to community-based clinics. however, this innovation is not without criticism. schools are dependent on the success of their clinics and clinic partnerships. one author cautions: "however, these creative models also may present potential political strategic risk or confl ict: private practitioners may organize and protest higher than normal reimbursement schemes. potentially, such protests could even jeopardize the very existence of such models (dunning et al. ) ." notably, community health centers have historically received strong bipartisan support. for example, during the bush administration, fqhc funding was doubled and most recently expanded through health reform legislation by the obama administration. according to the institute for oral health, "the group practice of the future is the dentist working with the physician" (ryan ) . the ada reported "multidisciplinary education must become the norm and represent the meaning and purposes of primary care as it applies to dentistry. educational sequences should include rotation strategies across discipline specialties in medicine and dentistry, clerkships and hospital rotations, and experience in faculty and residency clinics." (barnett and brown ) the models alluded to, were school-based attempts at improving educational outcomes. perhaps the proverbial fork in the road regarding the future of dental education leaves two paths for consideration. is it better to travel down a road that leads a school to develop and operate a clinic? or is the road less traveled, where a clinic becomes a school, the better of the two options? the answer, perhaps, is that a combination of both will accomplish the desired outcome. for example, didactic knowledge is measured by examination whereas competency as a practitioner is measured by clinical demonstration. at a minimum, the result must achieve learner competency, quality, and sustainability. however, the road less traveled has not been taken yet. william gies, in his revered report written years ago on the state of american dental education, wrote "dental faculties should show the need…. for integrated instruction in the general principles of clinical dentistry and in its correlations with clinical medicine" (gies ) . basic sciences aside, could a clinicalbased educational training center have an advantage over a school-based clinical center? soon-to-be-implemented new commission on dental accreditation (coda) standards will require schools to demonstrate competency in patient-centered care (valachovic ) . might an enterprise profi cient at running a successful clinical business model have an advantage running a professional, patient-centered clinical training program as compared to a pedagogical business model attempting to run a clinical training model? these questions should challenge us to reexamine why our thinking about educational models should be limited to schools being the starting point for the development of a profession that demands clinical competency, patientcenteredness, and integration as outcomes. the clinic based model may serve as an equivalent starting point and, have some distinct advantages for achieving responsiveness to recommendations and directions cited in this section. beginning in november through august , the family health center (fhc) of marshfi eld, inc, marshfi eld, wisconsin, launched of a broad network of developing dental clinics, targeting dental professional shortage areas with the provision of dental services to the underserved communities whose dental needs were not being adequately met by the existing infrastructure. fhc-marshfi eld is a federally qualifi ed health center (fqhc). as an fqhc, fhc receives cost-based reimbursement for its dental services to medicaid populations. along with the cost-based reimbursement, fqhcs are obligated to provide care to anyone regardless of their ability to pay. presently, fhc is the nation's largest federally qualifi ed dental health center. to date, this network of dental clinics has served over , unique patients, % of whom were under % of poverty. notably, service was provided to a signifi cant number of cognitively and developmentally disabled patients in special stations developed for serving patients with special needs. these patients frequently travel the furthest to get to our dental centers for care. beginning in , fhc stepped up the pace of dental clinic expansion, constructing two new dental centers in , two in , and two more are slated to open in . when fully operational, this will establish capacity to serve , patients annually. each site has proactively included dedicated clinical and classroom training space for dental residents or students, thus laying the framework for clinic-based training of new dental professionals. the plan is to continue to stand up new dental centers until they have the capacity to serve , patients annually or approximately % of the , underserved patients in the rural service area. in addition to the capacity for training residents and students, a dental post-baccalaureate program is being considered in partnership with regional year under graduate campuses. the post-baccalaureate program is aimed at preparing students from rural and underserved areas who desire to practice in rural and underserved areas for acceptance and success in dental schools. presently fhc in partnership with marshfi eld clinic is moving forward with plans to develop dental residencies at these sites and a dental post baccalaureate training program to better prepare pre-doctoral students from rural and/or underserved backgrounds to be successful in dental school as a means to create a dental academic infrastructure responsive to rural environments which have been classically underserved. marshfi eld clinic has a long-standing history in medical student education and multiple medical residency programs. creating access for the underserved population was the major motivational force driving the establishment of the dental clinic network back in . the fi ndings of the iom, macy, and rwjf reports became the foundational framework for developing the vision of a dental education model that would realize the major recommendations found in the reports. by establishing clinical campuses in regional underserved dental health professional shortage areas, access to care where care is needed most was provided. sustainment of a work force for provision of care across the dental clinic network is accomplished by schools contracting with fqhc's for service learning, thus circumventing challenges associated with releasing dental students at traditional dental schools to distant extramural training sites as discussed previously. this model is however not without its own set of challenges including calibration of faculty, supervision and evaluation of students in training, and achieving accreditation acceptance. however, through video connectivity and iehr technology curriculum, learning plans, competency assessment, progression, performance, faculty development, and learner evaluations can be centrally calibrated. additionally, this dental service-learning model based in a community health center setting offers students unique state-of-the-art exposures to alternative access models, cutting-edge informatics (including access to a combined dental-medical record) and a quality-based outcomes-driven practice. given the novelty of such an extended extramural dental clinical training model, there is limited data on the success of rural placement leading to retention to practice in a rural setting. the pipeline study piloted a model for getting students into underserved communities. however, that experiment was limited to -day rotations. outcome driven programs may provide a predictive surrogate for purposes of comparative analysis. for example, the rural medical education "rmed" program of the university of illinois medical school at rockford, has sustained a longstanding program in illinois. over years in duration with over student participants of whom % have been retained as primary care medicine practitioners in rural illinois. rabinowitz et al. ( a ) further reinforced that medical school rural programs have been highly successful in increasing the supply of rural physicians, with an average of - % of graduates choosing to practice in rural areas. they also noted rural retention rates of - % among the programs (rabinowitz et al. a) . recently, the university of wisconsin school of medicine and public health (uwsmph) launched the wisconsin academy for rural medicine (warm program). the warm program places medical students in rural academic medical centers during their third and fourth years in medical school. marshfi eld clinic is one of those sites. warm students affi liating with marshfi eld clinic's system would ultimately share learning experiences with dental students, clinical rotations, team-based rounding, lectures, and exposure to a combined medical-dental patient record. in an analogous manner, the marshfi eld clinic dental education model will incorporate a curriculum that embeds students in rural clinical practice for up to years. a secondary but not insignifi cant outcome of placing residents and students in clinical campuses focused on developing competency and providing care where needs are often greatest is the cost savings to taxpayers associated with the public care of patients. these savings are accomplished through the "service-learning" of the student. for example, in the model described where clinical training is embedded within the fhc clinics, the stipend resident or unpaid student learner provides the patient care as part of their service learning training while requiring oversight from one paid faculty per four to six learners. as a result, an academic based clinical partnership creates a model that reduces the cost for care provided to underserved patients. an additional benefi t to the community based clinic might be realized through tuition assistance by the academic program to help support patient procedures that develop learner competencies. in educational quality and infl uence, dental schools should equal medical schools, for their responsibilities are similar and their tasks analogous (william gies ) . the commission on dental accreditation (coda) notes that one of the learning objectives of an advanced education general dentistry (aegd) residency is to have the graduate function as a "primary care provider". to function competently in this role, the graduate needs to have a strong academic linkage to primary care medicine. at a dental deans forum, years after the gies report, dr polverini made the statement "dentistry has never been linked to the medical network but unless dentistry becomes part of the solution to the challenge of providing comprehensive patient care, it will be looked on as part of the problem, and ultimately, all dental schools will be called into question." (polverini ) the use of dental informatics and an integrated record are elements essential to this competency. on april , , fhc and marshfi eld clinic successfully transitioned all of their dental centers to a new practice management and electronic health record system that fully integrates medical and dental; one of the fi rst such systems in the nation. along with the benefi ts derived in fig. . , chc placement also exposes students to an integrated medical-dental care setting where learners can develop skills in system-based practice to include the interdependence of health professionals, systems, and the coordination of care. on the administrative side, dental and medical appointments can be coordinated to enhance convenience for patients and improve compliance with preventive dental visits. in , marshfi eld clinic's research foundation biomedical informatics research center hired their fi rst dental informatician, dr. amit acharya, bds, ms, phd. with dedicated biomedical informatics and research resource centers, the marshfi eld clinic has laid the groundwork for true medical/dental integration with appropriate electronic health record decision support and is positioned to develop a dental education curriculum capable of implementing the iom recommendations. downstream benefi ts of using such a curriculum are the ability of future practitioners to use informatics to improve quality of care and reduce the burden of disease. according to an institute of oral health report ( ) it is widely accepted across the dental profession that oral health has a direct impact on systemic health, and increasingly, medical and dental care providers are building to bridge relationships to create treatment solutions. as early as , william gies recognized that "the frequency of periodic examination gives dentists exceptional opportunity to note early signs of many types of illnesses outside the domain of dentistry" (gies ) . the following examples show how integration of dental and medical care can impact patient outcomes, underlining the importance of this concept in dental curriculum design. a study of , blue cross blue shield of michigan (bcbs) members with diabetes, who had access to dental care lead researchers, and bcbs executives to conclude that treatment of periodontal disease signifi cantly impacts outcomes related to diabetes care and related costs (blue cross blue shield of michigan ) . another example is found in the context of preterm delivery and miscarriage. according to research cited by cigna ( ) , expecting mothers with chronic periodontal disease during the second trimester are seven times more likely to deliver preterm (before th week), and have dramatically more healthcare challenges throughout their life. cigna also cites the correlation between periodontal disease and low birth weights, pre-eclampsia, gestational diabetes. equally important is the opportunity to develop and implement the team-based curriculum that trains future dentists and physicians in the management of chronic disease as an accountable care organization (aco) in a patient-centered environment. as an example, joseph errante, d.d.s., vice president, blue cross blue shield of ma, reported that medical costs for diabetics who accessed dental care for prevention and periodontal services were signifi cantly lower than those who didn't get dental care (errante ) . these data suggest that team based case management of prevalent chronic health conditions have considerable cost savings opportunities for government payers, third party payers, employers and employees (errante ) . these economic benefi ts to integration as it relates to the iehr are discussed elsewhere in this book, but begin with the ability of providers to function in a team based environment and as such, underscore the importance of training in such an environment. dentists trained in a fqhc iehr integrated educational model will be well positioned to function successfully within an aco model. an aco is a system where providers are accountable for the outcomes and expenditures of the insured population of patients they serve. the providers within the system are charged with collectively improving care around cost and quality targets set by the payor. within this system, care must be delivered in a patient-centered environment. the patient-centered environment according to the national committee for quality assurance (ncqa), is a health care setting that cultivates partnerships between individual patients and their personal physicians and, when appropriate, the patient's family. care is facilitated by registries, information technology, health information exchange and other means to assure that patients receive defi ned, timely and appropriate care while remaining cognizant of cultural, linguistic and literacy needs of the patient being served. the model includes the opportunity to deliver patient care that is patient-centric, incorporates the patient in the care planning, considers the patient's beliefs and views, and incorporates the patient's families as needed. the model allows providers to deliver care that is inclusive of needs, attentive, and accessible. the model equips payers to purchase high quality and coordinated care among teams of providers across healthcare settings. while this describes the medical home, most dental practices also follow this process. many dental practices function in this regard with insured populations and refl ect elements of the model that medicine is creating. william gies would be proud. training in the delivery of accountable and patient-centered medical-dental care must be done purposefully. commenting on the inadequate training relative to the integration of medical and dental education, baum ( ) stated that "we need to design new curricula with meaningful core competencies for the next generation of dentists rather than apply patches to our existing ones." while this statement was made in reference to the basic sciences, the same holds true for patient-centered system-based practice competencies. utilizing state-of-the-art electronic medical records as a tool and the fhc infrastructure as the service venue, meaningful patient-centered system-based practice core competencies achievement becomes possible in a manner highly responsive to societal needs. by defi nition, fqhcs must provide primary medical care, dental care, and behavioral health. fqhc have also historically been utilized as healthcare workforce training centers and the affordable care act of reinforced their role as healthcare training centers. specifi cally, this legislation serves to promote fqhcs as the entity through which the primary care workforce (including dental) will be developed and expanded. in combination, fqhcs and primary care centers are positioned to be the front runners in a medical/-dental home training model which will be essential to preparing future practitioners for practice in an aco. critical to this success is the ability to train these practitioners on an integrated medical-dental record and informatics platform. use of this platform imprints most strongly during the learner's formative years of training; instructing and guiding disease management, decision making, patient care coordination, prevention, and both outcome-based and comprehensive care. training in this hybrid academically orientated clinically integrated setting moves dental education off its crossroads and creates the highway to its future. concerns with the new models extend to their ability to integrate medical and dental disciplines at the clinical and informatics level. while the iom report identifi ed the need to integrate medical and dental curriculum, success at the curricular and technological level within schools, has been limited. three major factors have contributed to the limited progress: access priorities. creating access to care has outranked the need to integrate • care. in part, this refl ects societal need for care and public demand to reduce the burden of the "silent epidemic." schools play an important role as a safety net to care for the uninsured and underinsured through intramural clinical service learning. even though "dental colleges seem to be willing to bearing a large and disproportionate share of the burden in terms of access to care" (dunning et al. ) , schools were challenged as part of iom, surgeon general, and macy reports, to expand that role. while these reports have prompted creative educational solutions to increase access, the reports understate the tremendous opportunity, quality and cost benefi ts that could result from an integration of medicine and dentistry. it is diffi cult to change the culture and structure of existing schools. this is not • unique to dentistry. however, the iom report specifi cally recommended that schools "eliminate marginally useful and redundant courses and design an integrated basic and clinical science curriculum". the challenges with this are many. examples include: some schools may not have other disciplines to draw from to create an inte-grated curriculum; a number of schools use a faculty senate to determine curriculum. this can result in curriculum that preserves the current faculty structure; changing curriculum is associated with expense and can be fi nancially pro-hibitive to some schools physical changes may be needed and represent an expense and/or may, in some instances, may not be practicable based on structure of existing facilities. public school programs may direct the fi nal curriculum, as boards or regent's one or two steps removed from the curriculum often have fi nal authority conversely, private schools may specify business or mission objectives that determine fi nal design. perhaps most germane to this text is the lack of a common technology plat-• form between disciplines in a learning environment. an integrated curriculum requires an integrated platform to accomplish delivery and evaluation. this is particularly essential to clinical management of the patient by professionals in training as part of a healthcare delivery team. some progress in establishing shared basic sciences curricula has been documented in the literature. to date, no single integrated electronic health (medical-dental) record has been meaningfully adapted for educational purposes, including incorporation of assessment of the learner relative to integrated competencies, integrated case-based and problem-based curriculum, and integrated evaluation and assessment. another concern with new educational programs emerging in response to these reports and relative to creating a transformational integrated curriculum is that some of the programs are focusing primarily on creating clinicians with no value or emphasis on integrating training with research and/or scholarly activity. integrated training models counter such concerns. research will be fundamental to measuring the relative benefi ts and outcomes associated with treatment of patients in a shared curriculum setting and will be the catalyst for the development of integrated medical-dental informatics incorporating educational capabilities. additionally, accreditation will also need to evaluate its response to such models. presently it is unclear how accrediting bodies will view an integrated crossdisciplinary curriculum. further, due to its integrated nature, such a curriculum would lie outside of the expertise of a single traditional accrediting body focused on one particular discipline. it has yet to be determined how accrediting bodies will review and appraise such cross-disciplinary competencies. lastly, it is important to recognize that a successful education model with innovative informatics is only successful if its focus is patient care. graduating learners with competency only in the use of informatics will be limited unless adapted to training and delivery programs that result in patient centric care. research and reports over the past years support the need to reform dental education. first steps have been taken and lead the way for continued innovation around clinic-based education and integrated curriculum. the models identifi ed point to a strong partnership and interrelationship with chcs for creative, cost saving, effective and sustainable delivery methods. moreover, chc's must be more involved in a training curriculum integrated with informatics. chcs, in turn, benefi t from residents and students through service-learning to help meet a societal and workforce need, while the learners benefi t from increased competency. in order to train an evidence-based, patient-centered, medical-dental workforce, it is imperative that medical and dental data and record accessibility be incorporated into these training and care delivery initiatives. in order to keep moving away from the crossroads, such integration must become the pathway on which curriculum is developed and implemented. public law - cardioprotective effect of periodontal therapy in metabolic syndrome: a pilot study in indian subjects effect of professional oral health care on the elderly living in nursing homes. oral surg oral med oral pathol study shows preventive dental care may help reduce risk of preterm birth dentists' and specialists' knowledge of chronic orofacial pain: results from a continuing professional development survey. prim dent care an examination of periodontal treatment and per member per month (pmpm) medical costs in an insured population an examination of periodontal treatment, dental care, and pregnancy outcomes in an insured population in the united states the changing pattern of upper gastro-intestinal lesions in southern saudi arabia: an endoscopic study association between hla antigens and periodontal disease chronic care: making the case for ongoing care effects of antibiotic therapy on outcomes of patients with coronary artery disease: a meta-analysis of randomized controlled trials association between metabolic syndrome and periodontal disease caries is associated with asthma and epilepsy medicare coverage for patients with diabetes infections and autoimmune diseases state-supported dental schools: fi nancial projections and implications strengthening the nation's public health infrastructure: historic challenge, unprecedented opportunity dental health policy analysis series: issues in children's access to dental care under medicaid. chicago: american dental association inadequate training in the biological sciences and medicine for dental students a missense single-nucleotide polymorphism in a gene encoding a protein tyrosine phosphatase ( ptpn ) is associated with rheumatoid arthritis the dna methyltransferases of mammals fixing intelligence, foreign affairs high-throughput dna methylation profi ling using universal bead arrays blue cross blue shield of michigan. study links good oral care to lower diabetes care costs halitosis: a multidisciplinary problem construction of a genetic linkage map in man using restriction fragment length polymorphisms longitudinal analysis of heritability for dental caries traits systemic infl ammatory markers, periodontal diseases, and periodontal infections in an elderly population attention to dental coverage lacking in health-care debate. from st. petersburg times community based dental education: the pipeline, profession and practice program cotton-wool spots a large-scale genetic association study confi rms il b and leads to the identifi cation of il r as psoriasis-risk genes mother's gum disease linked to infant's death chronic disease management. health reform gps: navigating the implementation process fam c contributes to aggressive periodontitis us government printing offi ce centers for disease control. dental service use and dental insurance coverage-united states, behavioral risk factor surveillance system association between chronic periodontal disease and obesity: a systematic review and meta-analysis evaluation of negation phrases in narrative clinical reports the health insurance and portability and accountability act: practice of dentistry in the united states: privacy and confi dentiality the human oral microbiome database: a web accessible resource for investigating oral microbiome taxonomic and genomic information prevalence of doctor diagnosed arthritis and arthritis attributable activity limitation -united states cigna dental oral health maternity program trends in medical care expenditures of us adults with arthritis and other rheumatic conditions to genetic susceptibility to head and neck squamous cell carcinoma genetic and functional association of fam c with myocardial infarction self-reported periodontal disease in a virginia twin population infl ammation and cancer genetic mapping of the dentinogenesis imperfecta type ii locus association of the metabolic syndrome with severe periodontitis in a large u.s. population-based survey the economic burden of diabetes targeting cyclooxygenase- in human neoplasia: rationale and promise an inventory of patients' response to treatment at a multidisciplinary breath odor clinic the clinical signifi cance of the digital patient record evidence-based integrated care plan oral and craniofacial data resource center of the national institute of dental and craniofacial research hipaa department of veterans affairs. va health information technology improves quality of health care while reducing costs dental disease and risk of coronary heart disease and mortality human herpesvirus and kaposi's sarcoma strategic management and organizational behavior in dental education: refl ections on key issues in an environment of change prevalence of hypertension, dyslipidemia, and dyslipidemic hypertension the metabolic syndrome quantitative high-throughput analysis of dna methylation patterns by base-specifi c cleavage and mass spectrometry rare structural variants found in attention-defi cit hyperactivity disorder are preferentially associated with neurodevelopmental genes replication of the association of chromosomal region p . with generalized aggressive periodontitis (gagp) using an independent casecontrol cohort integration of medical and dental benefi ts for improved overall health. institute for oral health public health law: a communitarian perspective introduction to quantitative genetics hyperinsulinaemia: the key feature of a cardiovascular and metabolic syndrome dental education at the crossroads: challenges and change. an institute of medicine report evolving concepts of rheumatoid arthritis mapping of papillon-lefevre syndrome to the chromosome q region building a medical neighborhood for the medical home bidirectional relationship between chronic kidney and periodontal disease: a study using structural equation modeling clinical outreach: seven years of a self-sustaining model dentistry and bioterrorism prevalence of the metabolic syndrome among us adults the macy study: a framework for consensus cms projects focus on chronic illnesses: three demonstration projects seek to improve effi ciency and care. caring for ages prevalence and risk factors for gastroesophageal refl ux disease in an impoverished minority population public health and national security: the critical role of increased federal support extending multidisciplinary management of cleft palate to the developing world automated syndromic surveillance for the winter olympics systemic diseases and their treatment in the elderly: impact on oral health dental education in the united states and canada early statistical detection of anthrax outbreaks by tracking over-the-counter medication sales epigenetics and periodontal disease: future perspectives the infl uence of ccl l gene-containing segmental duplications on hiv- /aids susceptibility genetic variants near tnfaip on q are associated with systemic lupus erythematosus variant of transcription factor -like ( tcf l ) gene confers risk of type diabetes interleukin receptor a chain ( il r ) shows allelic and functional association with multiple sclerosis powe nr, for the centers for disease control and prevention ckd surveillance team. vulnerable populations and the association between periodontal and chronic kidney disease dentistry's response to bioterrorism: a report of a consensus workshop genome-wide association and replication studies identify four variants associated with prostate cancer susceptibility promoter methylation and inactivation of tumour-suppressor genes in oral squamous-cell carcinoma distinctive chromatin in human sperm packages genes for embryo development localisation of a gene for prepubertal periodontitis to chromosome q and identifi cation of a cathepsin c gene mutation the electronic oral health record a common variant on chromosome p affects the risk of myocardial infarction the role of cathepsin c in papillon-lefevre syndrome, prepubertal periodontitis, and aggressive periodontitis whole-genome patterns of common dna variation in three human populations a comprehensive review of genetic association studies economic cost of babies in the us in preparing health professionals to provide care to individuals with disabilities porphyromonas gingivalis, treponema denticola, and tennerella forsythia: the "red complex", a prototype polybacterial pathogenic consortium in periodontitis association between circulating oxidized low-density lipoprotein and incidence of the metabolic syndrome infl ammation and metabolic disorders gene genealogies and the coalescent process oral health and a referral arterial disease the effect of periodontal disease on medical and dental costs in a middle-aged japanese population: a longitudinal worksite study institute for oral health. focus groups on oral health in healthcare reform -focus group # accuracy of three classifi ers of acute gastrointestinal syndrome for syndromic surveillance the effect of antimicrobial periodontal treatment on circulating tumor necrosis factor-alpha and glycated hemoglobin level in patients with type diabetes the relative power of linkage and association studies for the detection of genes involved in hypertension periodontal disease, tooth loss and incidence of ischemic stroke new models of dental education kaiser family foundation. the uninsured likelihood methods for locating disease genes in nonequilibrium populations six new loci associated with blood lowdensity lipoprotein cholesterol, high-density lipoprotein cholesterol or triglycerides in humans amelogenin p.m t and p.w s mutations underlying hypoplastic x-linked amelogenesis imperfecta stochastic processes in population genetics, with special reference to distribution of gene frequencies and probability of gene fi xation complement factor h polymorphism in age-related macular degeneration prospects for whole-genome linkage disequilibrium mapping of common disease genes periodontal disease adversely affects the survival of patients with end-stage renal disease new bacterial species associated with chronic periodontitis expression of the fam c in tongue squamous cell carcinoma localisation of a gene for papillon-lefevre syndrome to chromosome q -q by homozygosity mapping inhibitors of cathepsin c (dipeptidyl peptidase i) gene polymorphisms in chronic periodontitis perturbed amelogenin secondary structure leads to uncontrolled aggregation in amelogenesis imperfecta mutant proteins initial sequencing and analysis of the human genome bioterrorism on the home front a new challenge for american medicine consensus report: aggressive periodontitis impact of current and past blood pressure on retinal arteriolar diameter in an older population localized aggressive periodontitis is linked to human chromosome q heart disease & stroke statistics - update the power of association studies to detect the contribution of candidate genetic loci to variation in complex traits biosense -a national initiative for early detection and quantifi cation of public health emergencies periodontal therapy reduces the rate of preterm low birth weight in women with pregnancy-associated gingivitis a variant of ptpn gene conferring risk to autoimmune diseases may protect against tuberculosis air, water, places, and data-public health in the information age the role of the nutrition in the pathogenesis of gastroesophageal refl ux disease, barrett's oesophagus and oesophageal adenocarcinoma structural variation in the human genome / , dqb in german patients suffering from rapidly progressive periodontitis (rpp) and adult periodontitis (ap) finding the missing heritability of complex diseases dental insurance: design, need and public policy births: fi nal data for va uses integrated health informatics to produce $ billion in savings. daily break a common allele on chromosome associated with coronary heart disease a family with papillon-lefevre syndrome reveals a requirement for cathepsin c in granzyme b activation and nk cell cytolytic activity diabetes mellitius and infl ammatory periodontal diseases the growing challenge of providing oral health care services to all americans periodontal fi ndings in adult twins evidence of a substantial genetic basis for risk of adult periodontitis the relationship between reduction in periodontal infl ammation and diabetes control: a report of cases can presence or absence of periodontal pathogens distinguish between subjects with chronic and aggressive periodontitis? a systematic review employer survey on health insurance in montana environmental and heritable factors in the etiology of oral diseases -a population-based study of swedish twins medicaid coverage and utilization of adult dental services transgenerational genetic effects on phenotypic variation and disease risk the convergence of dental and medical care the american dental association's commitment to electronic data interchange dentistry's role in responding to bioterrorism and other catastrophic events association of periodontitis and metabolic syndrome in the baltimore longitudinal study of aging exome sequencing identifi es the cause of a mendelian disorder cytokine gene polymorphisms in periodontal disease: a meta-analysis of studies including cases and controls keep america smiling: oral health in america. chicago: oral health america oral biofi lms, periodontitis and pulmonary infections partnership for solutions national program offi ce. chronic conditions: making the case for ongoing care enamel formation genes are associated with high caries experience in turkish children genetic and epigenetic somatic alterations in head and neck squamous cell carcinomas are globally coordinated but not locally targeted genome-wide map of nucleosome acetylation and methylation in yeast the ann arbor dental deans forum: crafting a response to the emerging tiered system of dental education global. health nexus, college of dentistry medical school programs to increase the rural physician supply: a systematic review and projected impact of widespread replication human dipeptidyl-peptidase i clinical characteristics and diagnosis of atypical odontalgia: implications for dentists rna-mediated non-mendelian inheritance of an epigenetic change in the mouse association of periodontitis with rheumatoid arthritis: a pilot study genome-wide location and function of dna binding proteins simulation modeling of anthrax spore dispersion in a bioterrorism incident the computer-based oral health record genome-wide association study identifi es new susceptibility loci for crohn disease and implicates autophagy in disease pathogenesis a half-century of community water fl uoridation in the united states: review and commentary the future of genetic studies of complex human diseases foundation pipeline study. evaluation of pipeline, profession and practice: community-based dental education neuropathic orofacial pain after dental implant placement: review of the literature and case report. oral surg oral med oral pathol oral radiol endod oral hygiene can reduce the incidence of and death resulting from pneumonia and respiratory tract infection effect of daily asprin on long-term risk of death due to cancer: analysis of individual patient data from randomized trails access to oral health care for elders: mere words or action? reconnecting the head to the body. dental considerations for the optimal medical management of people with diabetes. seattle: institute for oral health functional variants of the interleukin- receptor gene in non-gastrointestinal autoimmune diseases enzymatic amplifi cation of b -globin genomic sequences and restriction site analysis for diagnosis of sickle cell anemia serological evidence of an association of a novel chlamydia , twar, with chronic coronary heart disease and acute myocardial infarction hepatitis c virus infection is associated with the development of hepatocellular carcinoma the severity of periodontal disease is associated with the element of glucose intolerance in non-diabetics skeletal facial balance and harmony in the cleft patient: principles and techniques in orthognathic surgery periodontal disease and mortality in type diabetes pneumonia in nonambulatory patients identifi cation of a shared genetic susceptibility locus for coronary heart disease and periodontitis a genome-wide association study identifi es glt d as a susceptibility locus for periodontitis integrating dental offi ce technology -the next frontier the computer-based oral health record: an essential tool for cross-provider quality management the technologically well-equipped dental offi ce clinical computing in general dentistry the evolution of molecular markers -just a matter of fashion? strong association of de novo copy number mutations with autism prediction model for prevalence and incidence of advanced age-related macular degeneration based on genetic, demographic, and environmental variables hipaa security regulations protecting patients' electronic health information retinal arteriolar diameters and elevated blood pressure: the atherosclerosis risk in communities study treatment of periodontal disease for glycaemic control in people with diabetes relationship between periodontal disease and c-reactive protein among adults in the atherosclerosis risk in communities study social security and medicare boards of trustees. actuarial publications, status of the social security and medicare programs: a summary of the annual reports dentin phosphoprotein frameshift mutations in hereditary dentin disorders and their variation patterns in normal human populations fatty liver is associated with dyslipidemia and dysglycemia independent of visceral fat: the framingham heart study toothache or trigeminal neuralgia: treatment dilemmas consortium for oral health-related informatics: improving dental research, education, and treatment large recurrent microdeletions associated with schizophrenia the dental patient with asthma: an update and oral health considerations the effect of periodontal treatment on glycemic control in patients with type diabetes mellitus projected costs of chronic disease. health care cost monitor legal issues of the electronic dental record: security and confi dentiality the international hapmap consortium. integrating common and rare genetic variation in diverse human populations national electronic disease surveillance system (nedss): a standards-based approach to connect public health and clinical medicine help wanted: a policy maker's guide to new dental providers. national academy for state health policy, wk kellogg foundation chronic conditions account for rise in medicare spending from medical status and complications in relation to periodontal disease experience in insulin-dependent diabetics loos-of-function mutations in the cathepsin c gene result in periodontal disease and palmoplantar keratosis detection of disease outbreaks by the use of oral manifestations technical description of rods: a real-time public health surveillance system department of health and human services. oral health in american; surgeon general's call to action. rockville: usdhhs, national institute of dental and craniofacial research, nation institutes of health helicobacter pylori infection and the development of gastric cancer unitedhealth center for health reform & modernization. the united states of diabetes: challenges and opportunities in the decade ahead. minnetonka: unitedhealth group healthy people how will we seize them? american dental education association charting progress holding ourselves to the highest standard: doing what's best for patients. american dental education association charting progress role of micro-organisms in caries etiology the sequence of the human genome genome-wide scan fi nd suggestive caries loci heritability in the genomics era -concepts and misconceptions handbook of biosurveillance human papillomavirus is a necessary cause of invasive cervical cancer worldwide the assessment of treatment outcome by evaluation of dental arch relationships in cleft lip/palate role of chlamydia pneumoniae in atherosclerosis abundant class of human dna polymorphisms which can be typed using the polymerase chain reaction obesity and the metabolic syndrome in children and adolescents wellcome trust case control consortium. genome-wide association study of , cases of seven common diseases and , shared controls myocardial infarction and cerebrovascular accident in patients with retinal vein occlusion fact sheet project bioshield evaluating a dental diagnostic terminology in an electronic health record review on fl uoride-releasing restorative materials -fl uoride release and uptake characteristics, antibacterial activity and infl uence on caries formation dental students' experiences with and attitudes toward people with mental retardation hypertensive retinopathy signs as risk indicators of cardiovascular morbidity and morality retinal microvascular abnormalities and incident stroke: the atherosclerosis risk in communities study retinal microvascular abnormalities and blood pressure in older people: the cardiovascular health study retinal arteriolar narrowing and risk of diabetes in middle-aged persons retinal microvascular abnormalities and ten-year cardiovascular abnormality: a population-based casecontrol study wsare: what's strange about recent events? retinopathy and risk of congestive heart failure retinal arteriolar narrowing, hypertension and subsequent risk of diabetes mellitus periodontal disease and risk of cerebrovascular disease national and state medical expenditures and lost earnings attributable to arthritis and other rheumatic conditions -united states high frequency of occurrence of cyp d gene duplication/multiduplication indicating ultrarapid metabolism along suicide cases the digital pioneer evaluation of human leukocyte n-formylpeptide receptor (fpr ) snps in aggressive periodontitis patients integration of pediatric medical and pediatric dental care the effect of dental rehabilitation on the body weight of children with early childhood caries recommendations for preventive pediatric health care (periodicity schedule) american academy of pediatric dentistry. dental care for your baby does caries in primary teeth predict enamel defects in permanent teeth? a longitudinal study california dental association. oral health during pregnancy and early childhood: evidence-based guidelines for health professionals. sacramento: california dental association beyond the dmft: the human and economic cost of early childhood caries perceptions of oral health training and attitudes toward performing oral health screenings among graduating pediatric residents role of the medical team in preventing early childhood caries trends in oral health by poverty status as measured by hp objectives update on disparities in oral health and access to dental care for america's children early childhood caries: the dental disease of infants teaching oral health in u.s. medical schools: results of a national survey electronic medical record use in pediatric primary care term stillbirth caused by oral fusobacterium nucleatum integrated medical and dental health in primary care advancing oral health in america improving access to oral health care for vulnerable and underserved populations the bright futures training project: implementing systems to support preventive and developmental services in practice dental care and children with special health care needs: a population-based perspective preventive dental care for children in the us: a national perspective oral health and pediatricians: results of a national survey predicting caries in permanent teeth from caries in primary teeth: an eight-year cohort study interprofessional educational partnerships in school health for children with special oral health needs early childhood caries: determining the risk factors and assessing the prevention strategies for nursing intervention addressing children's oral health in the new millennium: trends in the dental workforce making decisions for children disparities in children's oral health and access to dental care executive summary: progress in children's oral health since the surgeon general's report adding dental therapists to the health care team to improve access to oral health care for children new york state department of health. oral health care during pregnancy and childhood cost-effectiveness model for the prevention of early childhood caries oral health during pregnancy: current research at the division of oral health, centers for disease control and prevention early preventive dental visits: effects on subsequent utilization and costs dental education's role in improving children's oral health and access to care survey of iowa general dentists regarding the age dental visit integration of patient records and clinical research: lack of integration as a barrier to needed research electronic patient records for dental school clinics: more than paperless systems risk factors for osteonecrosis of the jaws: a case-control study from the condor dental pbrn the "meaningful use" regulation for electronic health records the medical dictionary for regulatory activities (meddra) the benefi ts of health information technology: a review of the recent literature shows predominantly positive results assessing user perceptions of an electronic patient record (epr) before implementation the content coverage of clinical classifi cations. for the computer-based patient record institute's work group on codes & structures from data to knowledge through concept-oriented terminologies: experience with the medical entities dictionary research on the comparative effectiveness of medical treatments: issues and options for an expanded federal role pub. no. double data entry: what value, what price? evaluating the informatics for integrating biology and the bedside system for clinical research the use of electronic data capture tools in clinical trials: web-survey of canadian trials the signifi cance of snodent. stud health technol inform returns on research and development for s new drug introductions research electronic data capture (redcap)-a metadata-driven methodology and workfl ow process for providing translational research informatics support data collection and transcription the development of a dental diagnostic terminology a quantifi able alternative to double data entry adaptive double data entry: a probabilistic tool for choosing which forms to reenter economic outcomes of a dental electronic patient record survey of public information about ongoing clinical trials funded by industry: evaluation of completeness and accessibility better access to information about clinical trials clinical data management clinical computing in general dentistry format-independent data collection forms documentation of clinical workfl ow: a key step in a plan to facilitate implementation of an electronic patient record consortium for oral health-related informatics: improving dental research, education, and treatment which coding system for therapeutic information in evidencebased medicine integrating electronic patient records into a multimedia clinic-based simulation center using a pc blade platform: a foundation for a new pedagogy in dentistry creating a searchable digital dental radiograph repository for patient care, teaching and research using an online photo management and sharing application a human-centered design of a dental discharge summary (dds) for patients factors infl uencing implementation and outcomes of a dental electronic patient record system continuum of quality care series, american academy of developmental medicine and dentistry