key: cord-0682938-mdhfgo58 authors: Makhoul, Monia; Ayoub, Houssein H; Chemaitelly, Hiam; Seedat, Shaheen; Mumtaz, Ghina R; Al-Omari, Sarah; Abu-Raddad, Laith J title: Epidemiological impact of SARS-CoV-2 vaccination: mathematical modeling analyses date: 2020-04-23 journal: nan DOI: 10.1101/2020.04.19.20070805 sha: 6db33b86aad853b3e8f1abb20144551d6c5f1ca2 doc_id: 682938 cord_uid: mdhfgo58 Background: Several SARS-CoV-2 vaccine candidates are currently in the pipeline. This study aims to inform SARS-CoV-2 vaccine development, licensure, decision-making, and implementation by determining key preferred vaccine product characteristics and associated population-level impact. Methods: Vaccination impact was assessed at various efficacies using an age-structured mathematical model describing SARS-CoV-2 transmission and disease progression, with application for China. Results: A prophylactic vaccine with efficacy against acquisition (VEs) of ≥70% is needed to eliminate this infection. A vaccine with VEs <70% will still have a major impact, and may control the infection if it reduces infectiousness or infection duration among those vaccinated who acquire the infection, or alternatively if supplemented with a moderate social-distancing intervention (<20% reduction in contact rate), or complemented with herd immunity. Vaccination is cost-effective. For a vaccine with VEs of 50%, number of vaccinations needed to avert one infection is only 2.4, one severe disease case is 25.5, one critical disease case is 33.2, and one death is 65.1. Gains in effectiveness are achieved by initially prioritizing those ≥60 years. Probability of a major outbreak is virtually zero with a vaccine with VEs ≥70%, regardless of number of virus introductions. Yet, an increase in social contact rate among those vaccinated (behavior compensation) can undermine vaccine impact. Conclusions: Even a partially-efficacious vaccine can offer a fundamental solution to control SARS-CoV-2 infection and at high cost-effectiveness. In addition to the primary endpoint on infection acquisition, developers should assess natural history and disease progression outcomes and/or proxy biomarkers, since such secondary endpoints may prove critical in licensure, decision-making, and vaccine impact. Following the Severe Acute Respiratory Syndrome (SARS) epidemic in 2002 and the Middle East Respiratory Syndrome (MERS) epidemic in 2012 [1] , a novel coronavirus, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), emerged in late December 2019 in Wuhan, Hubei province, China [2, 3] . While the earlier coronavirus epidemics were rather limited in scope and scale [1] , SARS-CoV-2 rapidly spread [4] and evolved into a pandemic [5] . In absence of a vaccine, even if partially efficacious [6] , containment of the epidemic in China necessitated large-scale contact tracing and testing through deployment of thousands of healthcare fieldworkers along with severe quarantine measures [4] . The strain put on healthcare systems [7] , and the global human [8, 9] and economic [10] losses caused by the virus and resulting disease, designated as Coronavirus Disease 2019 (COVID-2019) [11] , accelerated efforts towards vaccine development [6, 12] . While multiple vaccine candidates are currently in the pipeline, they are still in early stage of development [6, 12, 13] . Assessment of the population-level impact of vaccine candidates through mathematical modeling is a critical component of the process of vaccine development, value proposition, licensure, decision-making, and pathways and costs of vaccine administration, and has been utilized for a wide range of infectious diseases [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] . In early stages of development, modeling is used to define the vaccine's key preferred product characteristics, by estimating levels of efficacy necessary to observe significant population-level impact, determining necessary duration of protection/immunity incurred by the vaccine, and identifying priority populations for optimal effectiveness [21, 29, 30]. These parameters provide early guidance to developers, manufacturers, regulators, and decision makers about candidates that are likely to be optimal through specifying vaccine characteristics that will maximize public health impact and cost-effectiveness [21, 28, 29, 31, 32] . Once key attributes are established, modeling plays an integral role in building the case for investment in vaccine development, and in ensuring rapid roll-out post-licensing, through assessment of risks, costs, and predicted returns associated with different immunization strategies [29, 33] . Post-vaccination, modeling is used to inform design and interpretation of surveillance studies [25] [26] [27] . We aimed in this study to provide the scientific evidence necessary to inform and accelerate SARS-CoV-2 vaccine development, licensure, decision-making, and implementation by determining key preferred vaccine product characteristics and associated population-level impact, at a critical time for such development [6, 12, 13] . For each of vaccinated and unvaccinated populations, nine age groups were considered, each representing a 10-year age band except for the last category (0-9, 10-19, …, ≥80 years). Susceptible individuals were at risk of being exposed to the infection at varying hazard rates depending on their age group and vaccination status. Following a latency period, infected individuals develop asymptomatic or mild infection followed by recovery, or severe infection followed by severe disease and then recovery, or critical infection followed by critical disease and either recovery or disease mortality. Mixing between individuals of different age groups was determined by an age mixing matrix that allows a range of mixing. Details on model structure are in SM. The model was coded, fitted, and analysed using MATLAB R2019a [35] . The model was parameterized and calibrated using empirical data on SARS-CoV-2 natural history and epidemiology. Age-specific distributions of infected individuals across the mild, severe, or critical infection stages were based on case-severity levels observed in China [4, 36, 37] . Critical disease cases were at risk of disease mortality, with the relative mortality rate in each age group informed by the age-specific crude case fatality rate observed in China [3, 38] . Tables S1 and S2 and Section B of SM. We assessed the impact of a prophylactic vaccine that reduces susceptibility to infection. However, since the first available vaccine may only be partially efficacious against infection acquisition, we also assessed the impact of the vaccine assuming additional "breakthrough" effects, that is effects that modulate natural history of infection for those vaccinated but still 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 23, 2020. . https://doi.org/10.1101/2020.04.19.20070805 doi: medRxiv preprint acquire the infection. Specifically, we assumed that vaccination may reduce infectiousness per one contact (by reducing viral load), infection duration (by faster clearance with vaccine-induced immunity), and likelihood of developing severe or critical disease (by rapid immune response that prevents disease progression). Definitions of these vaccine efficacies are summarized in Table 1 . Other relevant characteristics include duration of protection elicited by the vaccine and vaccination effect on adherence to social distancing-we investigated impact of increasing social contact rate following vaccination with the perception of protection. Population-level impact of SARS-CoV-2 vaccination was assessed by quantifying incidence, cumulative incidence, and reduction in incidence of each of infections, severe disease cases, critical disease cases, and deaths arising in presence of vaccination compared to the counterfactual scenario of no-vaccination. Vaccination impact was further assessed by quantifying effectiveness, that is number of vaccinations needed to avert one infection or one adverse disease outcome (ratio of number of vaccinations relative to that of averted outcomes). The latter measure is essentially cost-effectiveness with no costs included, as they are not yet available. Vaccine was assumed to elicit protection over 10 years. 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 23, 2020. Two vaccination program scenarios were considered. In both programs, it was assumed that vaccination is introduced in absence of social-distancing interventions, as the purpose of vaccination is to replace such interventions. The first program scenario assumes vaccine introduction and scale-up to 80% coverage before epidemic onset. This scenario is relevant for assessing impact of vaccination on future SARS-CoV-2 introductions in countries where the epidemic has been contained or at low level, such as in China. The scenario is also relevant to assess maximum potential impact of vaccination regardless of current epidemic status. The second program scenario assumes vaccine introduction during the epidemic's exponential growth phase, with scale-up to 80% coverage within one month. Incidence of new infections was assessed at various levels of S VE to determine the minimum efficacy needed to fully control the infection. Incidence was also assessed in a scenario where vaccination was introduced with a social-distancing intervention to estimate level of social distancing needed to complement vaccination to control the infection. Incidence was assessed in another scenario where those vaccinated increased their social contacts (behavior compensation), to assess consequences on vaccination impact. Lastly, we derived and estimated likelihood of occurrence of a major outbreak following infection introduction in a vaccinated but infectionfree population (Section D of SM). A multivariable uncertainty analysis was conducted to determine range of uncertainty around model predictions using five-hundred model runs. At each run, Latin Hypercube sampling [42, 43] was applied in selecting the natural history and disease progression parameter values 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. impact on infection incidence, but reduced peak incidence of each of severe and critical disease by 38.5%, and deaths by 40.0% (Figure 1 ), and the cumulative incidence of the latter three outcomes by ~39% ( Figures S2 and S3 of SM). In the second scenario ( Figure 2 and Figures S4 and S5 of SM), where vaccination was rapidly scaled up to 80% coverage during the exponential growth phase, the epidemic peaked earlier and 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 23, 2020. (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 23, 2020. . https://doi.org/10.1101/2020.04.19.20070805 doi: medRxiv preprint with large number of vaccinations needed to avert one infection or one adverse outcome. Of note that there are minor differences in effectiveness over time. For instance, in initial phases of the epidemic, prioritizing those 60-69 years of age was slightly more effective than prioritizing those 40-49 years of age ( Figure S8 of SM). Meanwhile, towards end of epidemic cycle, the inverse was true. Figure 5B illustrates the gains in effectiveness as While a vaccine with 50% S VE = cannot fully control the epidemic, Figure S9 of SM shows the impact when vaccination is supplemented with a social-distancing intervention that reduces the contact rate. Less than 20% reduction in contact rate would be sufficient to fully control the epidemic. Vaccinated individuals may increase their contact rate with the perception of protection. Figure 6 shows the consequences of behavior compensation. A 20% increase in contact rate among those vaccinated lowered the reduction in cumulative incidence from 52.8% to only 21.0%. A 41.8% increase in contact rate nullified impact of vaccination in reducing incidence. Figure 7 reports probability of occurrence of a major outbreak at varying levels of vaccine efficacy. A gradual decrease is noted with the increase in efficacy that accelerates close to S VE 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 23, 2020. (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 23, 2020. Effectiveness can be further enhanced by prioritizing vaccination for those ≥60 years of age, for an optimal reduction in disease cases and deaths ( Figure 4 ). Conversely, prioritizing children is least effective, with their lower risk for developing adverse outcomes ( Figure 4 ) [4, 44, 45] . This being said, prioritizing vaccination for any single age group, regardless of that age group, has overall lower effectiveness than extending vaccination to all age groups-vaccinating only one age group reduces the reproduction number ( 0 R ) only marginally, whereas vaccinating all age groups reduces 0 R to an epidemic domain where small reductions in 0 R can have more substantial impact on epidemic size ( Figure S12 of SM, also Figure 3 versus Figure 4 ). Consequently, roll-out strategies should initially prioritize individuals ≥60 years of age, but then incrementally cover younger age cohorts, and eventually the entire population. 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 23, 2020. . https://doi.org/10.1101/2020.04.19.20070805 doi: medRxiv preprint Vaccination will also reduce the likelihood of a major outbreak following virus introduction/reintroduction into a population (Figure 7) . With a vaccine of S VE ≥70%, infection transmission chains may not be sustainable, regardless of number of virus introductions. Of concern, however, is the potential increase in social contact rate among those vaccinated cycles. We assumed a long duration of vaccine protection (10 years), but this has limited impact on the predictions for one epidemic cycle, provided the duration of vaccine protection is greater than one year. Despite these limitations, our model was complex enough to factor the different 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 23, 2020. . key vaccine product characteristics, but also parsimonious enough to be tailored to the nature of available data. The model also generated results that are valid to a wide range of model assumptions. With most of the world's population remaining susceptible to SARS-CoV-2 and the need to impose disruptive social-distancing interventions, vaccination is a reliable intervention in the long-term. Findings show that even a partially-efficacious vaccine provides a fundamental solution to the SARS-CoV-2 pandemic and at high cost-effectiveness. Vaccine impact and costeffectiveness will not only depend on its efficacy in preventing infection, but can be enhanced if those vaccinated who still acquire the infection have reduced infectiousness, duration of infection, and disease severity. Vaccine developers should thus not only assess the primary endpoint of reduction in acquisition, but also other outcomes and/or proxy biomarkers including reductions in viral load and disease outcomes and speed of infection clearance for those vaccinated and unvaccinated. Totality of these primary and secondary endpoints may prove critical in the licensure process, decision-making, and vaccine impact once introduced into a population. 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 23, 2020. . https://doi.org/10.1101/2020.04.19.20070805 doi: medRxiv preprint Table 1 . Key vaccine product characteristics used to assess impact of a vaccine against SARS-CoV-2. Proportional increase in social contact rate (reduced social distancing) among those vaccinated compared to those unvaccinated 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 23, 2020. . https://doi.org/10.1101/2020.04.19.20070805 doi: medRxiv preprint new critical disease cases, and D) new deaths in the scenario assuming vaccine scale-up to 80% coverage before epidemic onset. Duration of vaccine protection is 10 years. Impact was assessed at 50% 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 23, 2020. . https://doi.org/10.1101/2020.04.19.20070805 doi: medRxiv preprint new critical disease cases, and D) new deaths in the scenario assuming vaccine introduction during the exponential growth phase of the epidemic, with scale-up to 80% coverage within one month. Duration of vaccine protection is 10 years. Impact was assessed at 50% 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 23, 2020. 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 23, 2020. . https://doi.org/10.1101/2020.04.19.20070805 doi: medRxiv preprint Figure 4 . Effectiveness of age-group prioritization using a SARS-CoV-2 vaccine with VEs of 50%. Number of vaccinations needed to avert A) one infection, B) one severe disease case, C) one critical disease case, and D) one death, by prioritizing different age groups for vaccination. Scenario assumes vaccine scale-up to 80% coverage before epidemic onset, and duration of vaccine protection of 10 years. Effectiveness is assessed at end of epidemic cycle. 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 23, 2020. . https://doi.org/10.1101/2020.04.19.20070805 doi: medRxiv preprint Scenario assumes vaccine scale-up to 80% coverage before epidemic onset. Duration of vaccine protection is 10 years. Measures are assessed at end of epidemic cycle. 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 23, 2020. . https://doi.org/10.1101/2020.04.19.20070805 doi: medRxiv preprint Figure 6 . Impact of vaccination with reduced adherence to social distancing for those vaccinated. Figure shows impact of varying levels of behavior compensation post-vaccination on the vaccine-induced reduction in cumulative number of new SARS-CoV-2 infections by end of epidemic cycle. Scenario assumes vaccine scale-up to 80% coverage before epidemic onset, VEs is 50%, and duration of vaccine protection is 10 years. 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 23, 2020. Scenario assumes vaccine scale-up to 80% coverage before epidemic onset. Duration of vaccine protection is 10 years. Figure does not include the result for 2 P VE , as this efficacy has no impact on probability of occurrence of a major 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 23, 2020. 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 23, 2020. 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 23, 2020. We declare no competing interests. This publication was made possible by NPRP grant number 9-040-3-008 and NPRP grant (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 April 23, 2020. Figure S1 . 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 23, 2020. Vaccinated population aged 0-9 years: η η µ ξ α = − + + + 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 23, 2020. . For subsequent age groups, the following set of equations was used: Unvaccinated populations aged 10+ years: (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 23, 2020. The definitions of population variables and symbols used in the equations are in Table S1 . (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 23, 2020. H . This matrix provides the probability that an individual in the a age group will mix with an individual in the a′ age group (regardless of vaccination status). The mixing matrix is given by 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 23, 2020. Age e = , the mixing is fully assortative, that is individuals mix only with members in their own age group. The input parameters of the model were chosen based on current empirical data for SARS-CoV-2 natural history and epidemiology. The parameter values are listed in Table S2 . (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 23, 2020. Using the second generation matrix method described by Heffernan et al. [12] , the basic reproduction number in absence of vaccination is given by where ( ) w a is the proportion of the population in each age group. 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 23, 2020. . https://doi.org /10.1101 Meanwhile, the basic reproduction number in presence of vaccination is given by Based on Whittle's method [13] , and by constructing Bailey's ratios [13] , the probability of a major outbreak was derived, that is the probability that the fraction of susceptible individuals that become infected is ≥ γ , where γ is a specific chosen level of the final attack rate. We define 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 23, 2020. . Meanwhile, in the case of vaccinating a fraction of the population, the probability of a major outbreak is given by the following three cases: 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 probability of a major outbreak is 0. 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 23, 2020. 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 23, 2020. 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 23, 2020. 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 23, 2020. 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 23, 2020. 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 23, 2020. 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 23, 2020. . https://doi.org/10.1101/2020.04.19.20070805 doi: medRxiv preprint Figure S8 . Temporal evolution of effectiveness of age-group prioritization using a SARS-CoV-2 vaccine with VEs of 50%. Number of vaccinations needed to avert A) one infection and B) one death, by prioritizing individuals aged 40-49 or 60-69 years. Scenario assumes vaccine scale-up to 80% coverage before epidemic onset. Duration of vaccine protection is 10 years. 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 23, 2020. . https://doi.org/10.1101/2020.04.19.20070805 doi: medRxiv preprint Figure S9 . Impact of a social-distancing intervention reducing the contact rate in the population on the cumulative number of new SARS-CoV-2 infections, when introduced to supplement the impact of a vaccine that has 50% efficacy in reducing susceptibility, VES. Scenario assumes vaccine scale-up to 80% coverage before epidemic onset. Duration of vaccine protection is 10 years. 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 23, 2020. 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 23, 2020. . https://doi.org/10.1101/2020.04.19.20070805 doi: medRxiv preprint Figure S11 . Uncertainty analysis. Model predictions for the mean cumulative number of new infections and associated 95% uncertainty interval (UI) at various levels of vaccine efficacy in reducing susceptibility (VES ) generated through 500 simulation runs. Scenario assumes vaccine scale-up to 80% coverage before epidemic onset. Duration of vaccine protection is 10 years. The solid black line, dashed lines, and shades show respectively, the mean, 95% uncertainty interval, and individual estimates for the cumulative number of new infections across all 500 uncertainty runs. 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 23, 2020. . https://doi.org/10.1101/2020.04.19.20070805 doi: medRxiv preprint Figure S12 . Vaccine effectiveness of age-group prioritization and the reproduction number R0. Model predictions for SARS CoV-2 attack rate at various levels of R0, indicating also the impact on R0 of prioritizing those 60-69 years of age for vaccination or vaccinating all age groups. The blue dashed-dotted line shows the modelpredicted attack rate at various levels of R0. The blue, red, and green stars show, respectively, the model-predicted attack rates in absence of vaccination, by prioritizing vaccination at 80% coverage for those 60-69 years of age, and by extending vaccination at 80% coverage to all age groups. The figure highlights how effectiveness of the vaccine (number of vaccinations needed to avert one infection) depends on the position on the R0 curve-prioritizing vaccination for any single age group, regardless of that age group, has overall lower effectiveness than extending vaccination to all age groups. The reason is that vaccinating one age group reduces R0 only marginally, whereas vaccinating all age groups reduces R0 to an epidemic domain where small reductions in R0 can have more substantial impact on epidemic size. 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 23, 2020. . https://doi.org/10.1101/2020.04.19.20070805 doi: medRxiv preprint Figure S13 . Impact of varying levels of vaccine efficacy in reducing susceptibility, VES, on the cumulative number of new SARS-CoV-2 infections when the reproduction number R0 is 3. Scenario assumes vaccine scaleup to 80% coverage before epidemic onset. Duration of vaccine protection is 10 years. 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 23, 2020. . https://doi.org/10.1101/2020.04.19.20070805 doi: medRxiv preprint The deadly coronaviruses: The 2003 SARS pandemic and the 2020 novel coronavirus epidemic in China The Incubation Period of Coronavirus Disease 2019 (COVID-19) From Publicly Reported Confirmed Cases: Estimation and Application Characteristics of and Important Lessons From the Coronavirus Disease 2019 (COVID-19) Outbreak in China: Summary of a Report of 72314 Cases From the Chinese Center for Disease Control and Prevention Report of the WHO-China Joint Mission on Coronavirus Disease 2019 (COVID-19). Available from WHO Director-General's opening remarks at the media briefing on COVID-19 -11 Timely development of vaccines against SARS-CoV-2. Emerg Microbes Infect Are high-performing health systems resilient against the COVID-19 epidemic? Real estimates of mortality following COVID-19 infection COVID-19 and Italy: what next? Lancet National Institute of Allergy and Infectious Diseases (NIH), NIH clinical trial of investigational vaccine for COVID-19 begins Pipeline: Investigational Therapies for COVID-19 Modelling HIV vaccination Prophylactic vaccines, risk behavior change, and the probability of eradicating HIV in San Francisco Imperfect vaccines and herd immunity to HIV Substantial undocumented infection facilitates the rapid dissemination of novel coronavirus (SARS-CoV2) The Incubation Period of Coronavirus Disease 2019 (COVID-19) From Publicly Reported Confirmed Cases: Estimation and Application SARS-CoV-2 Viral Load in Upper Respiratory Specimens of Infected Patients Transmission of 2019-nCoV Infection from an Asymptomatic Contact in Germany Report of the WHO-China Joint Mission on Coronavirus Disease 2019 (COVID-19). Available from United Nations Department of Economic and Social Affairs Population Dynamics, The 2019 Revision of World Population Prospects Characteristics of and Important Lessons From the Coronavirus Disease 2019 (COVID-19) Outbreak in China: Summary of a Report of 72314 Cases From the Chinese Center for Disease Control and Prevention Clinical Characteristics of Coronavirus Disease 2019 in China Clinical features of patients infected with 2019 novel coronavirus in Wuhan Characterizing key attributes of the epidemiology of COVID-19 in China: Model-based estimations. medRxiv Perspectives on the basic reproductive ratio The outcome of a stochastic epidemic-a note on Bailey's paper