key: cord-0862317-tstx42h7 authors: Bartsch, S. M.; O'Shea, K. J.; Ferguson, M. C.; Bottazzi, M.-E.; Cox, S. N.; Strych, U.; McKinnell, J. A.; Wedlock, P. T.; Siegmund, S. S.; Hotez, P. J.; Lee, B. Y. title: How Efficacious Must a COVID-19 Coronavirus Vaccine be to Prevent or Stop an Epidemic by Itself date: 2020-05-30 journal: medRxiv : the preprint server for health sciences DOI: 10.1101/2020.05.29.20117184 sha: 533b4820a382158dca79932cae1337fc1192e9e0 doc_id: 862317 cord_uid: tstx42h7 Background: Given the continuing coronavirus disease 2019 (COVID-19) pandemic and much of the U.S. implementing social distancing due to the lack of alternatives, there has been a push to develop a vaccine to eliminate the need for social distancing. Methods: In 2020, we developed a computational model of the U.S. simulating the spread of COVID-19 coronavirus and vaccination. Results: Simulation experiments revealed that when vaccine efficacy exceeded 70%, coverage exceeded 60%, and vaccination occurred on day 1, the attack rate dropped to 22% with daily cases not exceeding 3.2 million (reproductive rate, R0, 2.5). When R0 was 3.5, the attack rate dropped to 41% with daily cases not exceeding 14.4 million. Increasing coverage to 75% when vaccination occurred by day 90 resulted in 5% attack rate and daily cases not exceeding 258,029when R0 was 2.5 and a 26% attack rate and maximum daily cases of 22.6 million when R0 was 3.5. When vaccination did not occur until day 180, coverage (i.e., those vaccinated plus those otherwise immune) had to reach 100%. A vaccine with an efficacy between 40% and 70% could still obviate the need for other measures under certain circumstances such as much higher, and in some cases, potentially unachievable, vaccination coverages. Conclusion: Our study found that to either prevent or largely extinguish an epidemic without any other measures (e.g., social distancing), the vaccine has to have an efficacy of at least 70%. Introduction period duration). As individuals can transmit the virus prior to disease onset 8 , we assume they could transmit one day prior to the start of symptoms. Each individual moving to the 'I state' has a probability of being symptomatic, which governs if they are in the 'Is' or 'Ia state.' Each person in the model draws an infectious period duration from a distribution (range: 4-15 days, including the day prior to symptom onset). Infectious individuals remain in the 'I state' until they recover and are no longer infectious, moving from the 'I state' to the 'R state' (at the rate of 1/infectious period duration). Vaccination occurs on different days during the epidemic (varying with scenario) and protects in two different ways: by preventing infection and by preventing symptomatic disease (i.e., reducing viral shedding). Individuals who are vaccinated move into the 'V state'. Each day, vaccinated individuals mix with others and if they come in contact with an infectious person, they move to the 'E state'. When the vaccine prevents infection, individuals in the 'E state' will move to the 'R state' based on the vaccine efficacy. (Individuals for whom the vaccine is not effective move to either the 'Ia state' or 'Is state'.) When the vaccine reduces viral shedding, it reduces their probability of moving to the 'Is state' based on the vaccine efficacy. Effectively vaccinated individuals in the 'E state' move to the 'Ia state', where they transmit at a lower rate. We assume vaccination protection to be immediate (i.e., on day of vaccination) and that the vaccine had no impact on individuals who had already been infected and/or exposed to SARS-CoV-2. Each symptomatically infected person (i.e., COVID-19 case) travels through a probability tree of different sequential outcomes. 7 An infected person showing symptoms starts with a mild infection and has a probability of seeking ambulatory care or calling his/her physician (i.e., telephone consult). This person then has a probability of progressing to severe disease requiring hospitalization. If this person has only mild illness and is not hospitalized, he/she self-treats with over-the-counter medications and misses school or work for the duration of symptoms. If this person is hospitalized, he/she has a probability of developing severe pneumonia or severe nonpneumonia symptoms and has a probability of intensive care unit (ICU) admission. This patient then has a probability of having either sepsis or acute respiratory distress syndrome (ARDS), with or without sepsis. If this patient has ARDS, he/she requires the use of a ventilator. If the person is hospitalized, he/she has a probability of dying from COVID-19 complications. The person accrues relevant costs and health effects as he/she travels through the model. The third-party payer perspective includes direct costs (e.g., ambulatory care, hospitalization), while the societal perspective includes direct and indirect (i.e., productivity losses due to absenteeism and mortality) costs. Hourly wage across all occupations 9 serves as a proxy for productivity losses. Absenteeism results in productivity losses for the duration of symptoms. Death results in the net present value of productivity losses for missed lifetime earnings based on annual wage 9 for the years of life lost based on an individual's life expectancy. 10 Appendix Table 1 shows key model input parameters, values, and sources. All costs, clinical probabilities, and durations were age-specific when available and come from scientific literature or nationally representative data sources. Age-specific COVID-19 data are specific to the US context as of March 16, 2020. 11 We report all costs in 2020 values, discounting all past and future values using a 3% rate. We parameterized seeding SARS-CoV2-infected persons into the population for a given R0 such that simulated cases reflected case data reported as of March 24, 2020. 11 We evaluated the impact of introducing a vaccine with varying efficacies in the absence of other measures from the third-party payer and societal perspectives. Experiments consisted of 1,000 trial Monte Carlo simulations, varying parameters throughout their range (Appendix Table 1 ). Our initial scenario assumes an unmitigated epidemic with no vaccination. Experimental scenarios consisted of vaccinating individuals to either prevent infection or prevent symptomatic disease. Sensitivity analyses varied the vaccine efficacy (20%-100%), population coverage (those vaccinated and those otherwise immune, 50%-100%), timing of vaccination (0-180 days from the epidemic start), and R0 (2.5-3.5). [12] [13] [14] We report results as the median and 95% uncertainty interval (UI). Figure 1 ) show the resulting number of persons infected with SARS-CoV-2 over time (i.e., epidemic curves) when the R0 was 2.5 and 3.5 in the absence of vaccination, while Table 1 (and Appendix Table 2 ) show the number of cases, clinical outcomes, resources used, and costs over the duration of the epidemic. When the R0 of the virus was 2.5, the total attack rate was 86% with a maximum number of daily SARS-CoV-2 infections (i.e., epidemic peak) of 45.0 million. A higher R0 of 3.5, resulted in a total attack rate of 96% with 72.7 million cases at the epidemic peak. Even when R0 was 2.5, which is on the lower range reported, a 20% vaccine efficacy to prevent infection was not enough to substantially reduce the epidemic, even with 100% coverage. While reducing the attack rate (to 46%), there were still a total of 143.4 million (95% UI: 134.0-152.5 million) symptomatic cases and 29.7 million (95% UI: 27.8-31.6 million) hospitalizations. 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 May 30, 2020. . https://doi.org/10.1101/2020.05.29.20117184 doi: medRxiv preprint When the R0 was 2.5, a 40% efficacious vaccine that prevented infection reduced the number of cases at the peak to 2.3 million with 100% coverage (those vaccinated and those otherwise immune) and vaccinating within 90 days of the epidemic start ( Figure 1a ). When covering 100% of the population, vaccination resulted in 51.6 million (95% UI: 25.3-64.5 million) symptomatic cases. A 50% coverage reduced the number of cases at the peak by 59% but still resulted in a total of 154.5 million (95% UI: 147.5-161.9 million) symptomatic cases. Vaccinating on day 180 had no impact. When R0 was 3.5 ( Figure 2a) , a 40% efficacious vaccine reduced the number of cases at the peak by 67-83% with a coverage of 75% to 100% (Table 1) . If only reducing viral shedding (Appendix Figure 1a) , a 40% vaccine efficacy resulted in a total of 122.5 million (95% UI: 118.2-126.9 million) symptomatic cases even with 100% coverage. A 60% efficacious vaccine that prevented infection decreased the number of cases at the peak by 85-99% and stretch out the epidemic duration by ³2 months when R0 was 2.5-3.5 with coverages ³75% (Figures 1b and 2b ). For example, when R0 was 2.5, vaccination with 100% coverage on day 1 resulted in 2,449 (95% UI: 2,151-2,819) symptomatic cases, 89 (95% UI: 67-109) patients requiring a ventilator, and 2,815 (95% UI: 2,459-3,226) hospital bed days, costing $9.4 million in direct medical costs and $30.2 million in productivity losses. A 50% population coverage reduced the number of cases at the peak to 10.7 million (vs. 45.0 million), resulting in 115.5 million symptomatic cases ( Table 1 ). The impact was robust to changes in the timing of vaccination, such that vaccinating on day 90 shifted the curve earlier compared to vaccinating on day 1 but still resulted in a similar magnitude ( Figure 2a) . However, when vaccinating on day 180, near the peak, vaccination with 100% coverage served to shorten the epidemic duration ( Figure 1b ). When R0 was 3.5, vaccination with a 75% coverage (those vaccinated and otherwise immune) resulted in 103.1 million (95% UI: 95.6-78.9 million) symptomatic cases during the epidemic. With 100% coverage, vaccination reduced the number of cases at the peak to 1.4 million cases and resulted in an 11% overall attack rate and a total of 34.4 million symptomatic cases when vaccinating on day 90 ( Figure 2b ; Table 1 ). A vaccine that reduced viral shedding was not enough to effectively control the epidemic, even when coverage was 100% (Appendix Figure 1b) . While reducing the number of cases at the peak to 20.8 million, there were still a total of 72.6 million symptomatic cases (R0 2.5). When preventing infection, vaccinating with a 50% coverage on day 1 reduced the number of cases at the peak by 83%. A 60% coverage on day 1 largely extinguished the epidemic, reducing the number of cases at the peak by 93% (to 3.2 million infections), resulting in a 22% attack rate throughout the epidemic (Figure 2c ) and 6.1 million (95% UI: 51.2-68.1 million) symptomatic cases, costing $938.3 billion (95% UI: $633.3-$1,432 billion) in total costs. A 60% coverage 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 May 30, 2020. . https://doi.org/10.1101/2020.05.29.20117184 doi: medRxiv preprint when vaccinating on day 90, resulted in a 23% attack rate over the epidemic duration and daily cases did not exceed 3.4 million. When at least 75% of the population was covered on day 1, vaccination prevented the epidemic and resulted in a total of 90,760 (95% UI: 32,544-647,422) symptomatic cases, costing $1.4 billion (95% UI: $0.5-$11.0 billion) in total costs (R0 2.5). When covering at least 75% of the population on day 90, vaccination largely prevented the epidemic, resulting in an overall 5% attack rate and 13.0 million symptomatic cases ( Figure 2c ; Table 1 ). When R0 was 3.5, covering everyone prevented the epidemic and covering 75% of the population (those vaccinated and otherwise immune) reduced the number of cases at the peak by 93%, shifting it later into the year by 10 months when vaccinating on day 1 ( Figure 2c ; Table 1 ). With a 50% coverage, vaccinating on day 90 decreased the number of cases at the peak by 69%. If reducing viral shedding, vaccination reduced the number of cases at the peak by 63%, resulting in 50.1 million symptomatic cases over the epidemic duration with 100% coverage (R0 2.5; Appendix Figure 1c ). With an 80% efficacy, a vaccine that prevents infection was reduced the number of cases at the peak by ³36.6 million SARS-CoV-2 infections and shifted it to later in the year by more than 11 months when R0 was 2.5 with coverages of at least 50% and vaccinating within 90 days of the epidemic start ( Figure 1d ; Table 1 ). Covering 75% of the population resulted in a total of 3,346 (95% UI: 2,946-3,899) symptomatic cases and 3,848 (95% UI: 3,342-4,486) hospital bed days, costing $12.7 million in direct medical costs and $40.1 million in productivity losses. When R0 was 3.5, 50% coverage of the population reduced the number of cases at the peak by 76% but still resulted in 123.3 million symptomatic cases during the epidemic (Figure 2d , vaccinating at epidemic start). A 75% coverage decreased the total attack rate to 12%, resulting in 33.3 million symptomatic cases during the epidemic when vaccinating on day 90. If the vaccine only reduced viral shedding, a vaccine efficacy of 80% with 100% coverage reduced the number of cases at the peak by 32.4 million (R0 2.5), but still resulted in 30.6 million symptomatic cases, when vaccinating within 90 days of the epidemic start (Appendix Figure 1d ; Appendix Table 2 ). A 100% efficacious vaccine reduced the epidemic peak by ³44.1 million infections and shifted it ³12 months later, even if vaccination occurred 90 after the epidemic started and coverage was 50% (R0 2.5). With a 50% coverage, vaccinating resulted in 33.7 million symptomatic cases and 7.0 million hospitalizations, costing $538.1 billion in total costs during the epidemic. A 75% coverage at the start of the epidemic resulted in 928 (95% UI: 904-953) symptomatic cases and 1,067 (1,014-1,130) hospital bed days, costing $14.70 million (95% UI: $10.2-$22.0 million) in total costs. When R0 was 3.5, covering 50% of the population decreased the number of cases at 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 May 30, 2020. . https://doi.org/10.1101/2020.05.29.20117184 doi: medRxiv preprint the peak by 87%, resulting in 86.2 million symptomatic cases, even when vaccinating 90 days into the epidemic. A 75% coverage prevented the epidemic with 5.4 million symptomatic cases (2% attack rate) when vaccinating on day 90, and 246 (95% UI: 232-263) symptomatic cases when vaccinating on day 1. If the vaccine only reduced viral shedding, covering 100% of the population still resulted in 122.9 million SARS-CoV-2 infections when R0 was 2.5 and 239.8 million SARS-CoV-2 infections when R0 was 3.5 during the epidemic. Our study found that to either prevent or largely extinguish an epidemic without any other measures (e.g., social distancing), the vaccine has to have an efficacy (i.e., probability of preventing infection) of at least 70% when vaccination within 90 days of the epidemic start and covers at least 60% of the population. A 60% coverage would not be unreasonable since a recent poll by Reuters/Ipsos of 1,215 American adults found that 75% of respondents would get a coronavirus vaccine if assured it was safe. 15 The coverage threshold, which includes those people who are vaccinated and those people who are immune for other reasons, rises the later vaccination occurs, until the threshold reaches 100% at the peak of the epidemic. This is increase is due to the fact that as the epidemic processed the prevalence of infected people is greater and greater. There are circumstances that a vaccine with an efficacy between 40% and 70% can obviate the need for other measures but these require that 75% or more of the population be vaccinated. For example, when the vaccine efficacy is 60% at least 75% of the population would need to be vaccinated. If the R0 of the COVID-19 coronavirus goes from 2.5 to 3.5, then this coverage threshold would increase to 100%. Our study focused on identifying the efficacy thresholds required to eliminate the need for other measures (e.g., social distancing) in order for life to "return to normal", because that is a main concern of the general public. [16] [17] [18] [19] [20] A vaccine alone may not be able to do this. Nonetheless, a vaccine with an efficacies lower than the thresholds we identified, could still be useful (e.g., a vaccine efficacy of 40% could prevent ³2.8 million patients from requiring a ventilator and ³89.5 million hospital bed days). All models, by definition, are simplifications of real-life and cannot account for every possible outcome. 29 Our model inputs drew from various sources, and new data on SARS-CoV2 continues to emerge. The course of an actual COVID-19 coronavirus epidemic may not conform to our model data and assumptions. For example, we assumed individuals mixed equally with each other, and that individuals would be vaccinated by a certain day during the epidemic; 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 May 30, 2020. . https://doi.org/10.1101/2020.05.29.20117184 doi: medRxiv preprint however, in reality, individuals may be vaccinated at different time points over the epidemic. Moreover, assumptions about levels of herd immunity required may depend on differing individual susceptibility to SARS-CoV-2. Our study found that to either prevent or largely extinguish an epidemic without any other measures (e.g., social distancing), the vaccine has to have an efficacy of at least 70% when vaccination occurs within 90 days of the start of the epidemic and covers at least 60% of the population. This efficacy threshold increases the deeper one goes into the pandemic. A vaccine with an efficacy between 40% and 70% can still obviate the need for other measures under certain circumstances such as much higher, and potentially unachievable, vaccination coverages. The funders did not have any role in the study design, collection, analysis and interpretation of data, writing the report, and the decision to submit the report for publication. The authors of this manuscript are responsible for its content, including data analysis. Statements in the manuscript do not necessarily represent the official views of, or imply endorsement by, the National Institutes of Health, AHRQ, or the Department of Health and Human Services. 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. 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. (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 May 30, 2020. . https://doi.org/10.1101/2020.05.29.20117184 doi: medRxiv preprint Figure 2 . Epidemic curves for an epidemic with an R0 of 3.5 and the impact of a vaccine that prevents infection implemented at different points in the epidemic with a vaccine efficacy of A) 40%, B) 60%, C) 70%, and D) 80% 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. 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. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (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 May 30, 2020. . https://doi.org/10.1101/2020.05.29.20117184 doi: medRxiv preprint 9 Appendix Figure 1 . Epidemic curves for an epidemic with an R0 of 2.5 and the impact of a vaccine that reduces viral shedding implemented at different points in the epidemic with a vaccine efficacy of A) 40%, B) 60%, C) 70%, and D) 80% 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. Constructing target product profiles (TPPs) to help vaccines overcome postapproval obstacles A systems approach to vaccine decision making Infection control measures for norovirus: a systematic review of outbreaks in semi-enclosed settings Heidelberg Institute for Geoinformation Technology. Mapping COVI-19 Research The COVID-19 vaccine development landscape The Potential Health Care Costs And Resource Use Associated With COVID-19 In The United States Transmission of 2019-nCoV infection from an asymptomatic contact in Germany national occupational employment and wage estimates Early transmission dynamics in Wuhan, China, of novel coronavirus-infected pneumonia Preliminary estimation of the basic reproduction number of novel coronavirus (2019-nCoV) in China, from 2019 to 2020: A data-driven analysis in the early phase of the outbreak Preliminary prediction of the basic reproduction number of the Wuhan novel coronavirus 2019-nCoV Most Americans would take coronavirus vaccine if deemed safe: Reuters/Ipsos poll How Long Should Social Distancing Last? When Will COVID-19 Coronavirus End? In: Forbes What's It Going To Take To End The Shutdown? 5 Keys To Containing Coronavirus Social distancing until 2022?! Hopefully not Constructing target product profiles (TPPs) to help vaccines overcome postapproval obstacles A systems approach to vaccine decision making Infection control measures for norovirus: a systematic review of outbreaks in semi-enclosed settings Heidelberg Institute for Geoinformation Technology. Mapping COVI-19 Research The COVID-19 vaccine development landscape The Potential Health Care Costs And Resource Use Associated With COVID-19 In The United States Transmission of 2019-nCoV infection from an asymptomatic contact in Germany national occupational employment and wage estimates Early transmission dynamics in Wuhan, China, of novel coronavirus-infected pneumonia Preliminary estimation of the basic reproduction number of novel coronavirus (2019-nCoV) in China, from 2019 to 2020: A data-driven analysis in the early phase of the outbreak Preliminary prediction of the basic reproduction number of the Wuhan novel coronavirus 2019-nCoV Most Americans would take coronavirus vaccine if deemed safe: Reuters/Ipsos poll How Long Should Social Distancing Last? When Will COVID-19 Coronavirus End? In: Forbes What's It Going To Take To End The Shutdown? 5 Keys To Containing Coronavirus Social distancing until 2022?! Hopefully not Testing Could Unlock A Return To Normal Life, But Obstacles Persist When will the outbreak end and life get back to normal? In: BBC News The persistence and clearance of viral RNA in 2019 novel coronavirus disease survivors Virological assessment of hospitalized patients with COVID-2019 Temporal profiles of viral load in posterior oropharyngeal saliva samples and serum antibody responses during infection by SARS-CoV-2: an observational cohort study United States Department of Health & Human Services. HCUP facts and figures: statistics on hospital-based care in the United States Estimated the asymptomatic proportion of coronaviru disease 2019 (COVID-19) cases on board the Diamond Princess curise ship The annual impact of seasonal influenza in the US: measuring disease burden and costs Clinical Characteristics of Coronavirus Disease 2019 in China Clinical Characteristics of 138 Hospitalized Patients With 2019 Novel Coronavirus-Infected Pneumonia in Wuhan, China Clinical features of patients infected with 2019 novel coronavirus in Wuhan, China Consultation times in emergency telemedicine using realtime videoconferencing COVID-19 in 2 Persons with Mild Upper Respiratory Symptoms on a Cruise Ship What Does It Mean to 'Recover From Coronavirus'? Here's What You Need to Know Hospitalization Rates and Characteristics of Patients Hospitalized with Laboratory-Confirmed Coronavirus Disease 2019 -COVID-NET, 14 States Cost-effectiveness of dual influenza and pneumococcal vaccination in 50-year-olds Cost-effectiveness of newer treatment strategies for influenza Vaccination versus treatment of influenza in working adults: a costeffective analysis Effectiveness and cost-effectiveness of vaccination against pandemic influenza (H1N1) 2009 Cost-effectiveness of treating influenzalike illness with oseltamivir in the United States Cost-effectiveness of live attenuated influenza vaccine versus inactivated influenza vaccine among children ages 24-59 months in the United States Health outcomes and costs of community mitigation strategies for an influenza pandemic in the United States Estimating the cost-effectiveness of a national program to eliminate disparities in influenza vaccination rates among elderly minority groups Cost effectiveness of zanamivir for the treatment of influenza in a high risk population in Australia Economics of influenza vaccine administration timing for children The cost-effectiveness of vaccinating pregnant women against seasonal influenza in England and Wales A cost-effectiveness analysis of revaccination and catch-up strategies with the 23-valent pneumococcal polysaccharide vaccine (PPV23) in older adults in Japan economic evaluation and Budget impact analysis of Vaccination against Haemophilus influenzae Type b infection in Thailand Cost-effectiveness analysis of universal screening for carbapenemase-producing Enterobacteriaceae in hospital inpatients Economic evaluation of pneumococcal vaccines for adults aged over 50 years in Belgium Cost-utility analysis of antiviral use under pandemic influenza using a novel approach-linking pharmacology, epidemiology and heath economics A global economic model to assess the cost-effectiveness of new treatments for advanced breast cancer in Canada Cost-effectiveness of recombinant human activated protein C and the influence of severity of illness in the treatment of patients with severe sepsis Implementation of early goal-directed therapy for severe sepsis and septic shock: a decision analysis Costeffectiveness of procalcitonin testing to guide antibiotic treatment duration in critically ill patients: results from a randomised controlled multicentre trial in the Netherlands Costeffectiveness of neoadjuvant chemotherapy before radical cystectomy for muscle-invasive bladder cancer Costutility analysis of antiviral use under pandemic influenza using a novel approach -linking pharmacology, epidemiology and heath economics (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.