key: cord-0837684-nd9shlvx authors: Jevdjevic, Milica; Listl, Stefan; Beeson, Morgan; Rovers, Maroeska; Matsuyama, Yusuke title: Forecasting future dental health expenditures: Development of a framework using data from 32 OECD countries date: 2020-11-30 journal: Community Dent Oral Epidemiol DOI: 10.1111/cdoe.12597 sha: 29745c2b5a31b4ad19c109e383c87dc47bfc84fb doc_id: 837684 cord_uid: nd9shlvx OBJECTIVE: To (1) develop a framework for forecasting future dental expenditures, using currently available information, and (2) identify relevant research and data gaps such that dental expenditure predictions can continuously be improved in the future. METHODS: Our analyses focused on 32 OECD countries. Dependent on the number of predictors, we employed dynamic univariate and multivariate modelling approaches with various model specifications. For univariate modelling, an auto‐regressive (AR) dynamic model was employed to incorporate historical trends in dental expenditures. Multivariate modelling took account of historical trends, as well as of relationships between dental expenditures, dental morbidity, economic growth in terms of gross domestic product and demographic changes. RESULTS: Estimates of dental expenditures varied substantially across different model specifications. Models relying on dental morbidity as one of the predictors performed worst regardless of their specification. Using the best‐fitted model specification, that is the univariate second‐order autoregression [AR(2)], the forecasted dental expenditures across 32 OECD countries amounted to US$316bn (95% forecasted interval, FI: 258‐387) in 2020, US$434bn (95%FI: 354‐532) in 2030 and US$594bn (95%FI: 485‐728) in 2040. Per capita spending in 2040 was forecasted to be highest in Germany (US$889, 95%FI: 726‐1090) and lowest in Mexico (US$52, 95%FI: 42‐64). CONCLUSIONS: The present study demonstrates the feasibility and challenges in predicting dental expenditures and can serve as a basis for improvement towards more sustainable and resilient health policy and resource planning. Within the limitations of available data sources, our findings suggest that dental expenditures in OECD countries could increase substantially over the next two decades and vary considerably across countries. For more accurate estimation and a better understanding of determinants of dental expenditures, more comprehensive data on dental spending and dental morbidity are urgently needed. The policy debate on health systems has been dominated in recent decades by concerns about growing cost pressures and nonsustainability of health systems' financing. If healthcare budgets are limited, increased expenditures for one type of care may mean that other types of care are no longer fundable to previous extents. 1, 2 Hence, monitoring the dynamics of health expenditures is relevant to inform healthcare decision-makers about potential challenges in the future financing of health care and to take timely action in the interest of population wellbeing. Up-to-date and reliable information on health expenditures is particularly important for decision-makers who aim at allocating healthcare resources efficiently and equitably. 3 Over recent decades, per capita global spending on health has almost doubled, from US$472 in 2000 to US$1007 in 2015. 4 There has also been an increase in healthcare expenditures relative to the global economy: spending on health care amounted to 8.6% of the gross domestic product (GDP) in 2000, and it increased to 9.9% of GDP in 2015. 4 Dental diseases are among the most prevalent and persistent diseases worldwide and impose a substantial economic burden to humankind. [6] [7] [8] [9] Dental expenditures currently account for the third highest proportion of health spending in the European Union, with €90bn compared to €111bn spent on diabetes and €119bn spent on cardiovascular diseases. 9, 10 Considerable proportions of dental expenditures are attributable to out-of-pocket payments in many countries. In addition to a greater financial hardship, future increases in dental expenditures could imply a substantial increase in unmet needs for dental care among less affluent populations. Dental expenditures are therefore particularly relevant to the United Nation (UN) and World Health Organization (WHO) goals of Universal Health Coverage. [11] [12] [13] Despite recent attempts to better estimate the economic impacts of dental diseases, considerable room for improvement has been documented for the international reporting of dental expenditures, with mixed levels of availability of information across countries. The reporting of dental expenditures for member countries of the Organization for Economic Cooperation and Development (OECD) appeared to be more consistent than for several other countries. 7, 8 Also, there is very little insight into the future economic implications of dental diseases, exceptions being work on dental expenditure in the US from The Centers for Medicare and Medicaid Services 14 and by Nasseh, Vujicic. 15 The aims of this study, therefore, are to (1) develop a framework for forecasting future dental expenditures, using currently available information, and (2) identify relevant research and data gaps such that dental expenditure predictions can continuously be improved in the future. For the development of the forecasting framework, we collated publicly available data on the predictors of interest (dental expenditure, morbidity, demographic changes and economic growth [GDP] ). Dependent on the number of predictors, we employed both, dynamic univariate and multivariate modelling approaches with varying model specifications. The best-performing model was identified based on performance indicators. Further details are provided below. Annual dental health expenditures (expressed as % of GDP) were derived from the OECD online platform-OECD Statistics (Health Expenditure and Financing) that relies on the Joint OECD, Eurostat and WHO National Health Accounts reports. 16 bootstrapped expectation-maximization algorithm and has been commonly employed for multiple imputation of time-series data. 17 As a re- Further details about the data sources used are shown in Table 1 . We estimated future dental morbidity (prevalence of dental diseases) for consideration as potential predictor variables in multivariate models. Firstly, following the equation described in formula 1, the 2017 to 2040 incidence per 100 000 persons was forecasted for dental caries in permanent teeth, periodontal disease (ie, having a gingival pocket depth equal or more than 6 mm, or Community Periodontal Index of Treatment Needs score of 4, or a clinical attachment loss more than 6 mm), and severe tooth loss (ie, having less than nine remaining teeth). All variables were log-scaled before deriving first difference. We built projection models for disease incidence using data for The model with the smallest RMSE was rerun using 1990-2016 data, and the parameters were used to extrapolate the disease incidence through the year 2040. Then, as shown in formula 2, the same procedure was applied to forecast disease prevalence until 2040, using the forecasted disease incidence as an additional predictor. where Δ: first difference, c: country, y: year, Country: country dummy variables. To forecast future dental health expenditures, a dynamic modelling approach was applied as recommended by previous literature. 21, 22 For the univariate modelling approach, we used existing and imputed dental health expenditure values from 2000 to 2016. In the multivariate approach and in addition to previous trends in dental health expenditures, dental morbidity, demographic changes and GDP were included as potential predictors. The primary condition to produce reliable statistical inferences in time-series analysis is data stationarity. 23 The statistical properties (1) Morbidity forecast formula 1: ΔIncidence c,y = ΔIncidence c,y − 1 + ΔIncidence c,y − 2 . . . + ΔIncidence c,y − n + Country c . (2) Morbidity forecast formula 2: ΔPrevalence c,y = ΔIncidence c,y + ΔPrevalence c,y − 1 + ΔPrevalence c,y − 2 . . . + ΔPrevalence c,y − n + Country. Expenditure forecast model 3: ΔDHE c,y = ΔDHE c,y − 1 + ΔDHE c,y − 2 … + ΔDHE c,y − n + ΔPop c,a,y − 1 + ΔGDP c,y − 1 + Country c . Expenditure forecast model 4: ΔDHE c,y = ΔDHE c,y − 1 + ΔDHE c,y − 2 … + ΔDHE c,y − n + ΔPrevalence c,a,d,y − 1 + ΔGDP c,y − 1 + Country c . where Δ: first difference, DHE: dental health expenditure, Country: country dummy variables, Prevalence: the prevalence of dental diseases, Pop: the size of population, GDP: gross domestic product, the subscripts c, y, a and d, indicate country, year, age group (<15 yearsold, 15-64 years-old and ≥ 65 years-old) and dental diseases (caries, periodontitis, and severe tooth loss), respectively. Note that all variables were log-scaled before deriving first differences. Similar to morbidity predictions (see above), we ran models using To consider the uncertainty for the forecasting models, we esti- Auxiliary analyses were conducted to assess the potential implications of additional dental care systems characteristics. In addition to population needs (expressed as morbidity, see above) and the number of practicing dentists per 1000 people, dental care utilization was captured by the number of yearly dental visits per capita. Further details are provided in the Appendix. Table 2 reports the performance of the various forecasting models. According to the RMSE and MAPE criterions, the best-performing model was yielded by the univariate second-order autoregression [AR (2) ]. Among the multivariate models considered, a specification according to formula 5 performed best. Models relying on dental morbidity as one of the predictors performed worst regardless of the number of lags in their specification. The auxiliary analyses revealed that additional factors such as utilization of dental services (eg. number of yearly dental consultations per capita) and density of dental providers (eg. number of practicing dentists per 1000) might also be important to account for when quantifying future dental expenditure (see Appendix). Regression analyses indicated that the density of dental providers is significantly associated with the amount of dental expenditure in countries with lower economic growth (GDP per capita lower than the OECD mean). However, the extent to which additional analysis could be performed was restricted by nonavailable or incomplete data. F r a n c e G e r m a n y G r e e c e H u n g a r y I c e l a n d I r e l a n d I s r a e l J a p a n nancing. Moreover, the number of (dental) healthcare providers 30, 31 may be an important driver of health expenditures, as corroborated by our auxiliary analysis. In addition, potential implications of unpredictable events like the recent COVID-19 outbreak which has caused suspension of dental care cannot be ruled out. This may result in significant changes in dental spending, as suggested by Nasseh, Vujicic. 32 All the more, forecasts of dental expenditures are highly important to closely monitor and predict the dynamics of dental expenditure to equip oral health systems when dealing with the consequences of unanticipated events (external shocks). For a more robust assessment of the future economic trajectories, additional inputs such as the share of out-of-pocket payments, private insurance coverage and government spending on dental health are necessary. Countries with greater spending on dental health do not automatically have better oral health outcomes or financial risk protection. 33 Therefore, economic forecasting can also be a useful tool to compare and improve the performance of oral healthcare systems. Moreover, the pay-off of resources dedicated to the current preventive services could be evaluated and provide a guidance for future investments in oral health care. Our findings revealed promising new avenues for future research. With investing in better data collection, processing and predictive modelling, policy makers would be able to anticipate future needs and identify gaps with available resources. This is particularly relevant at the time of the increasing support to integrate oral health into universal health coverage. 34 The present study substantiates both the societal relevance and the methodological challenges involved in providing robust and reliable predictions of future dental expenditures. Health systems and resource planning could benefit from these findings, as they emphasize the critical importance of more comprehensive health economic monitoring as a key information source for sustainable and resilient health policy and resource allocation. In conclusion, our findings show that given the expected future dynamics of dental expenditures and continuing uncertainties in oral health systems planning (including due to unexpected events such as the COVID-19 pandemic), coordinated health policy action is needed to attenuate the predicted economic burden and to warrant efficiency and sustainability of dental care systems. The authors did not receive any financial support and declare no potential conflicts of interest with respect to the publication of this manuscript. MJ contributed to conception, design, data acquisition and interpretation, performed analyses, drafted and critically revised the manuscript. SL contributed to conception, design, interpretation and critically revised the manuscript. MB contributed to interpretation, performed analyses and critically revised the manuscript. MR contributed to interpretation and critically revised the manuscript. YM contributed to conception, design, interpretation, performed analyses and critically revised the manuscript. All authors gave their final approval and agreed to be accountable for all aspects of the work. The data that support the findings of this study are available from the corresponding author upon reasonable request. We have searched for information on additional factors that could potentially drive dental health expenditures (eg, utilization, the percentage of the population with a yearly dental visit, the number of dentists per population, out-of-pocket payments). We have found a high rate of discrepancy in data availability between the countries. Data on the number of yearly dental consultations per capita and the number of dentists per 1000 people were the most consistent. 16 However, information on dental utilization was completely missing for Iceland, Israel, Latvia, Norway, Slovenia and Switzerland. On the other hand, data on providers density was not available for Greece, Ireland, New Zealand, Slovak Republic, South Korea, Spain and the United States. Taking into account already limited data on dental expenditures, the scarcity of information on all variables of interest across multiple time-points deterred more comprehensive analysis. On all available observations, a multiple linear regression was performed to explore the correlation between per capita dental expenditures, GDP per capita, dental utilization as well as the dentist density. As shown in Table A1 , when applied on the entire sample only GDP per capita appeared to be a significant determinant of dental health expenditures (β = 1.147, P = .000), whereas dental utilization (β = 0.084, P = .245) and dentist density (β = 0.252, P = .067) were not significant. If the analysis was performed exclusively on the countries whose GDP per capita was lower than the OECD mean in the respective year, in addition to economic growth (GDP per capita; β = 1.668, P = .000), dentist density was positively correlated with dental expenditures (β = 0.817, P = .007) while dental utilization remained nonsignificant (β = 0.031, P = .780). However, that was not the case among the countries with GDP per capita higher than the OECD mean for which utilization shown the strongest relationship, although not statistically significant. Countries with GDP per capita a OECD mean β (GDP per capita) β (dental utilization) β (dentist density) Economics and the evaluation of health care programmes: generalisability of methods and implications for generalisability of results Methods for the economic evaluation of health care programmes Fiscal sustainability of health systems: Bridging health and finance perspectives World Health Organization. Global spending on health. Global Health Expenditure database Global Burden of Disease Health Financing Collaborator Network. Future and potential spending on health 2015-40: development assistance for health, and government, prepaid private, and out-of-pocket health spending in 184 countries Global, regional, and national prevalence, incidence, and disability-adjusted life years for oral conditions for 195 Countries, 1990-2015: a systematic analysis for the global burden of diseases, injuries, and risk factors Global economic impact of dental diseases Global-, regional-, and country-level economic impacts of dental diseases in 2015 Oral diseases: a global public health challenge What is health economics? Community Dent Health Can people afford to pay for health care? New evidence on financial protection in Europe The impact of out-of-pocket payments for dental care on household finances in low and middle income countries Ending the neglect of global oral health: time for radical action Resea rch-Stati stics -Data-and-Syste ms/Stati stics -Trend s-and-Repor ts/Natio nalHe althE xpend Data/Natio nalHe althA ccoun tsPro jected Dental expenditure expected to grow at a much lower rate in the coming years. Health Policy Institute Research Brief Amelia II: A program for missing data Real GDP long-term forecast (indicator) Global health data exchange, Institute for Health Metrics and Evaluation World Population Prospects: The 2017 Revision, custom data acquired via website. UN Population Division Forecasting health expenditure: Methods and applications to international databases. Centre for Health Economics and Policy Analysis (CHEPA) Advances in panel data analysis in applied eco Introduction to econometrics Aggregation and the measurement of health care costs Once bitten, twice shy? Lessons learned from an experiment to liberalize price regulations for dental care Is skill mix profitable in the current NHS dental contract in England? Cost-effective workforce planning: optimising the dental team skill-mix for England Payment systems and incentives in dentistry The effect of different methods of remuneration on the behaviour of primary care dentists Determinants of healthcare expenditure in Economic Cooperation Organization (ECO) countries: evidence from panel cointegration tests Is health care a luxury? New evidence from OECD data Modeling the impact of COVID-19 on US dental spending. Health Policy Institute Research Brief Waste not, want not: the efficiency of health expenditure in emerging and developing economies Strengthening oral health for universal health coverage Forecasting future dental health expenditures: Development of a framework using data from 32 OECD countries Germany 361 Germany 465 (380, 570) Germany 641 (523, 786) Germany 889 (726, 1090) United States 342 United States 433 (353, 530) United States 558 (456, 684) United States 729 (595, 894) Switzerland 337 Switzerland 412 (337, 505) Switzerland 527 (430, 645) Switzerland 684 (558, 838) Norway 303 Canada 347 (283, 425) Canada 438 (357, 536) Canada 563 (459, 690) Canada 332) Spain 363 (296, 445) Austria 485 (395, 594) Iceland 241 Iceland 271 (221, 332) Austria 360 (293, 441) Sweden 481 (393, 590) Spain 220 Spain 264 305) France 316 (258, 387) Luxembourg 410 (335, 503) France 182 France 239 (195, 293) Australia 308 (251, 377) Australia 383 (313, 470) Belgium 162 United Kingdom 223 (182, 273) United Kingdom 291 (238, 357) United Kingdom 383 (313, 470) Israel Estonia 185 (151, 227) Israel 249 (203, 305) Ireland 126 Czech Republic 132 (108, 162) Czech Republic 180 (147, 220) Czech Republic 248 (202, 304) Czech Republic 113 Finland 112 (92, 137) Finland 146 (120, 179) Finland 194 (158, 238) Finland 103 Slovenia 104 (85, 128) Slovenia 141 (115, 173) Slovenia 193 (158, 237) Slovakia 93 New Zealand 92 (75, 113) Hungary 125 (102, 154) Hungary 178 (145, 218) Greece 92 Hungary 90 (73, 110) Slovakia 118 (96, 145) Slovakia 164 (134, 201) Lithuania 85 South Korea 88 (72, 107) South Korea 115 (94, 140) South Korea 156 (127, 191) Hungary 78 Slovakia 87 (71, 107) New Zealand 114 (93, 140) New Zealand 145 (118, 177) Slovenia 72 Poland 72 (59, 89) Poland 101 (83, 124) Poland 144 (118, 177) Poland 64 Lithuania 66 (54, 81) Lithuania 91 (75, 112) Lithuania 129 (106, 159) Latvia 56 Latvia 62 (51, 76) Latvia 89 (73, 110) Latvia 128