==== Front PLoS One PLoS One plos plosone PLoS ONE 1932-6203 Public Library of Science San Francisco, CA USA 10.1371/journal.pone.0243026 PONE-D-20-19975 Research Article Medicine and Health Sciences Epidemiology Medical Risk Factors Medicine and Health Sciences Epidemiology Infectious Disease Epidemiology Medicine and Health Sciences Medical Conditions Infectious Diseases Infectious Disease Epidemiology Medicine and Health Sciences Medical Conditions Infectious Diseases Viral Diseases Covid 19 Biology and life sciences Organisms Viruses RNA viruses Coronaviruses SARS coronavirus SARS CoV 2 Biology and life sciences Microbiology Medical microbiology Microbial pathogens Viral pathogens Coronaviruses SARS coronavirus SARS CoV 2 Medicine and health sciences Pathology and laboratory medicine Pathogens Microbial pathogens Viral pathogens Coronaviruses SARS coronavirus SARS CoV 2 Biology and life sciences Organisms Viruses Viral pathogens Coronaviruses SARS coronavirus SARS CoV 2 Medicine and Health Sciences Epidemiology Pandemics Medicine and Health Sciences Medical Conditions Infectious Diseases Respiratory Infections Medicine and Health Sciences Medical Conditions Respiratory Disorders Respiratory Infections Medicine and Health Sciences Pulmonology Respiratory Disorders Respiratory Infections Medicine and Health Sciences Epidemiology People and Places Population Groupings Age Groups Estimating individual risks of COVID-19-associated hospitalization and death using publicly available data Estimating individual risks of COVID-19-associated hospitalization and deathhttps://orcid.org/0000-0002-3371-4101Bhatia Rajiv ConceptualizationData curationFormal analysisMethodologyValidationWriting – original draft1* Klausner Jeffrey MethodologyWriting – review & editing2 1 Department of Medicine (Affiliated), Stanford University, Stanford, California, United States of America 2 Department of Medicine and Public Health, University of California Los Angeles, Los Angeles, California, United States of America Shaman Jeffrey Editor Columbia University, UNITED STATES Competing Interests: The authors have declared that no competing interests exist. * E-mail: drajiv@stanford.edu 7 12 2020 2020 7 12 2020 15 12 e024302629 6 2020 14 11 2020 This is an open access article, free of all copyright, and may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose. The work is made available under the Creative Commons CC0 public domain dedication.We describe a method to estimate individual risks of hospitalization and death attributable to non-household and household transmission of SARS-CoV-2 using available public data on confirmed-case incidence data along with estimates of the clinical fraction, timing of transmission, isolation adherence, secondary infection risks, contact rates, and case-hospitalization and case-fatality ratios. Using the method, we estimate that risks for a 90-day period at the median daily summertime U.S. county confirmed COVID-19 case incidence of 10.8 per 100,000 and pre-pandemic contact rates range from 0.4 to 8.9 per 100,000 for the four deciles of age between 20 and 60 years. The corresponding 90-day period risk of hospitalization ranges from 13.7 to 69.2 per 100,000. Assuming a non-household secondary infection risk of 4% and pre-pandemic contact rates, the share of transmissions attributable to household settings ranges from 73% to 78%. These estimates are sensitive to the parameter assumptions; nevertheless, they are comparable to the COVID-19 hospitalization and fatality rates observed over the time period. We conclude that individual risk of hospitalization and death from SARS-CoV-2 infection is calculable from publicly available data sources. Access to publicly reported infection incidence data by setting and other exposure characteristics along with setting specific estimates of secondary infection risk would allow for more precise individual risk estimation. The author(s) received no specific funding for this work. Data AvailabilityAll relevant data are within the manuscript and its Supporting information files except for COVID-19 case incidence data which can be accessed at the URL: https://github.com/nytimes/covid-19-data.OutbreaksCOVID-19Data Availability All relevant data are within the manuscript and its Supporting information files except for COVID-19 case incidence data which can be accessed at the URL: https://github.com/nytimes/covid-19-data. ==== Body Introduction Perceptions of personal risk can modify pandemic disease transmission and population health outcomes by modifying behaviors such as mask wearing and by influencing demand for and compliance with government recommendations. A recent rapid review of studies that examined adherence with “quarantine” found adherence varied from 0% up to 92.8% and was influenced by social norms, perceived benefits of quarantine, perceived risk of the disease, and financial and material needs [1]. Public perceptions of risk also may influence how political leaders act. Ideally, public leaders should communicate risk precisely and transparently. To be useful to individuals, estimates of risk need to be specific to people, times, places, and activities. Numerous reports have described clusters of SARS-CoV-2 infections in diverse settings, yet there has been little attention to estimating and communicating how personal risk varies by place, time, and population. Surveillance measures used to monitor and characterize the COVID-19 pandemic, including laboratory-based “case” counts of infection and mortality rates describe risk in aggregate. In the United States, public data on laboratory-confirmed infections, hospitalizations and deaths have not disaggregated information based on exposure factors even though case-reports include this information. No large-scale U.S. community transmission studies have comparatively evaluated risk by setting. Recognizing the limitations of available data sources, we estimate and compare the person-level risks of SARS-CoV-2 infection associated hospitalization and death attributable to household and non-household contacts using public confirmed case incidence data and published transmission parameters. We identify knowledge gaps that, if filled, could make individual risk assessment more precise and useful. Materials and methods We conceptualize risk as the product of susceptibility, infection incidence, the timing and efficiency of transmission from an infected individual, the contact setting, the contact rate, and the severity of infection. Because the true infection incidence rate is unknown, we assume a stable multiplicative relationship between the incidence rate of confirmed infections and the incidence of infections (I ~ I Confirmed). We consider the transmission potential of clinical and subclinical infection separately, assuming that all symptomatic infected individuals will contribute to transmission during a pre-symptomatic infectious period and a share of symptomatic individuals will voluntarily self-isolate, effectively limiting transmission to within their own households. We assume those with subclinical infection will contribute to transmission for the duration of their infectious period but have a lower likelihood of forward transmission than those with symptoms. Eq (1) estimates the age-specific probability of a confirmed infection attributable to contact within household settings, P (Infection | HH) i, for age-strata i, as a function of the proportion susceptible to infection, S, the confirmed case incidence rate, I, the household specific secondary infection risk, BHH, an age-specific household contact rate, CHH i, the fraction of symptomatic infections, FSX, and the complementary fraction of asymptomatic infections, 1- FSX, multiplied by an estimate of their relative infectiousness, r. P(Infection|HH)i=SICHHiBHH{FSX+r(1−FSX)}(1) Eq (2) estimates the age-specific probability of a confirmed infection attributable to contact in non-household settings, P (Infection | HH) i, for age-strata i. The share of pre-symptomatic transmission of the clinical fraction is denoted by p. The fraction of symptomatic infected cases that do not self-isolate, 1-q, also contribute to transmission to non-household contacts during the period of post-symptomatic transmission. P(Infection|NH)i=SICNHiBNH{p(FSX)+(1−q)(1−p)(FSX)+r(1−FSX)}(2) We estimate the age-specific probabilities of hospitalization and death in Eqs (3) and (4) from the age specific probabilities of confirmed infection, multiplying these quantities by the age-specific confirmed case hospitalization and fatality ratios (CFRs), respectively. P(Hospitalization)i=CHRi×{P(I|HH)i+P(I|NH)i}(3) P(Fatality)i=CFRi×{P(I|HH)i+P(I|NH)i}(4) We do not estimate risks for persons under 20 years due to the infrequent incidence of severe disease events in this age group. We also omit risk estimates for persons over 60 years as case hospital and fatality ratios cannot be reliably calculated for non-congregate settings using publicly available data. We list our parameters and their sources in Table 1 and discuss them further below. 10.1371/journal.pone.0243026.t001Table 1 Parameters used for risk estimation. Symbol Parameter Unit Value Source I Average daily incidence rate of confirmed (reported) infections Infections per day per person Computed from data Complied by the NYT from various sources S Proportion of the population susceptible Unitless proportion 91% Anand F SX Proportion of infections with clinical illness Unitless proportion 69% (95% CI, 63%–74%) Buitrago-Garcia p Percentage of transmission occurring prior to symptom onset Unitless proportion 50% He, Casey, Ferrriti, US CDC q Compliance with household self-isolation Unitless proportion 75% Unsourced r Relative Risk of infection transmitted from an asymptomatic individual relative to a symptomatic one Unitless proportion RR = 0.35 (95% CI, 0.10–1.27) Buitrago-Garcia Β HH Household Secondary infection risk Infections observed per contacta during a monitoring period 31.1% (95% CI 19.4% -42.7%) Madewell Β NH Non-household Secondary infection risk Infections observed per contact during a monitoring period 4.0% (95% CI: 2.8%, 5.2%) Koh C i HH Household contact rate for age group i Contactsb / day 20–29 years: 2.77 Prem 30–39: 3.20 40–49: 3.18 50–59 3.33 C i NH Non-household contact rate for age group i Contacts / day 20–29 y 11.88 Prem 30–39 11.10 40–49 10.49 50–59 10.53 CHR i Case Hospitalization ratio for age group i Unitless proportion 20–29 years: 1.65% US CDC surveillance case reports 30–39 years: 3.13% 40–49 years: 4.75% 50–59 years: 7.47% CFR i Age-specific Case Fatality ratio for age group i Unitless proportion 20–29 years: 0.05% US CDC surveillance case reports 30–39 years: 0.15% 40–49 years: 0.37% 50–59 years: 0.97% a Most studies estimating the SIR defined contact as but most commonly either as face to face unprotected contact for greater than 15 minutes, prolonged contact, or household membership b The basis of these estimates, the POLYMOD study, defined contact as “…either skin-to-skin contact such as a kiss or handshake (a physical contact), or a two-way conversation with three or more words in the physical presence of another person but no skin-to-skin contact (a nonphysical contact). We acquired data on confirmed SARS-CoV-2 infection incidence rates from publicly reported statistics compiled by The New York Times [2], computing average daily county-level confirmed infection incidence rates for the period of the “summer wave” in the U.S. (June 16th to Sept 15th 2020), then finding the median incidence rate for each quintile of incidence. Confirmed infection rates underestimate the true incidence of infection as not all people with symptomatic infections obtain tests, test produce false negative results and subclinical infections occur. The US CDC currently estimates that an average of 11 infections may occur for every confirmed case, but the source of this estimate indicates substantial regional variation [3]. We assume that confirmed infections and total infections will be proportional and related multiplicatively and that unconfirmed infections will have similar characteristics to confirmed ones. This assumption removes the need for application of this parameter in our methodology. The true susceptible proportion of the population is also unknown. Pandemic risk assessments at the onset of the pandemic assumed a 100% susceptible population. Seroprevalence surveys conducted in US populations using convenience samples demonstrated low (<10%) antibody prevalence after the spring wave but substantial variability among regions [3]. A systematic study of the U.S. dialysis population estimated that seroprevalence of SARS-CoV-2 was 9·3% in US adult population, ranging from 3·5% in the west to 27·2% in the northeast [4]. Understanding the limitations of antibody-based determinations of susceptibility, we assumed the prevalence of susceptibility to be 91% based on the seroprevalence of Anand et al. but did not apply region-specific parameters in our methodology. The parameters for secondary infection risk (SIR) (alternatively, secondary attack rates) come from reviews of community transmission studies. Studies included in reviews identify cases through active and passive surveillance and follow secondary contacts for a specified duration (typically 14 days), monitoring contacts for symptoms and, in most cases, testing contacts to confirm infection. Studies estimating the SIR define “contact” variously but commonly either as face to face unprotected (i.e., mask-less) contact for greater than 15 minutes, prolonged contact, or household membership. Studies compute the SIR as the number of infections that occur among the reported contacts of an ascertained case during the monitoring period. Studies typically do not ascertain the frequency of contact. Published reviews of the SIR for SARS-CoV-2 have focused on household transmission. Koh et al reviewed 20 studies published by May 15, 2020, estimating a pooled household SIR at 15.4% (95% CI: 12.2%, 18.7%) [5]. The authors reported a higher SIR for adults relative to children (RR 1.40, 95% CI: 1.00, 1.96). This review also reported a summary non-household SIR of 4.0% (95% CI, 2.8%, 5.2%) which included several cluster investigations with high SIR estimates. Lei et al. summarized 24 studies with data on household transmission risk published by July 1, 2020 including case reports with > 10 households, estimating a pooled SIR of 27% (95% CI, 21–32%) [6]. The review found a 3-fold higher risks for household transmission to adults. Data from ten studies with data on non-household contact, specifically, provided an SIR of 1.65% (95% CI, 1.43–1.87) based on 218 secondary cases among 13194 non-household contacts. Madewell et al. estimated a summary SIR for 40 studies reporting SIRs for household and family contacts published through 31 July 2020, estimating a mean SIR for household contacts of 19·0% (95% CI, 14·9%–23·1%) and for family contacts of 18·1% (95% CI: 12·9%–34·8%) [7]. The estimated mean SIR for contact to adult household members was 31.1% (95% CI, 19.4% -42.7%). The authors also estimated a mean SIR for “close-contacts,” which combined household and non-household contacts, as 4·3% (95% CI: 2·9%–5·6%). For our risk estimates, we used the 31% summary estimate of the adult household SIR from Madewell et al. and the 4% summary estimate of the non-household SIR from Koh et al. We used the 1.65% estimate of the non-household SIR from Lei et al in a sensitivity analysis. In the U.S., government agencies have not reported confirmed infection incidence by symptom status. Estimates of the clinical fraction from published studies have significant heterogeneity. A non-quantitative review of studies published from 19 April through 26 May 2020, reported asymptomatic fractions of 43–65% in community populations, 46–8% on cruise ships, 48–58% among personnel on aircraft carriers, 63–88% in the context of homeless shelter outbreaks, and 96% among inmates [8]. A meta-analysis of studies reported through 10 June estimated the asymptomatic fraction to be 20% (95% confidence interval [CI] 17–25) based on 79 studies with 6616 cases and 31% (95% CI 26%–37%) based on 7 studies with prospective follow-up [9]. Alternatively, a single large case-ascertained follow up study of 5,484 contacts which included both RT-PCR and serological examination, which limited under ascertainment from false negative RT-PCR tests, found that only 32% of test positive contacts developed symptoms of cough, shortness of breath with the clinical fraction varying from 18% for those under 20 to 65% for those over 80 years old [10]. Another meta-analysis of international studies also reported that clinical symptoms vary from 21% to 69% from the youngest to oldest subpopulations [11]. We used the 69% summary estimate of the clinical fraction reported by Buitrago-Garcia et al. from the subset of prospective studies and the lower 32% all-age-group clinical fraction reported by Polletti et al in a sensitivity analysis. We assume that subclinical infections will contribute to non-household exposure for the duration of their infectious period while a fraction of those with clinical infections will self-isolate in their households following symptoms. We found no empirical data to estimate actual compliance with self-isolation requirements for the U.S. during the COVID-19 period. A recent survey of public attitudes in Israel found that willingness to comply with self-quarantine rose from 57% without governmental financial compensation to 94% with compensation [12]. We assumed that 75% of symptomatic individuals would self-isolate after developing symptoms and tested the sensitivity of our estimate to an alternative assumption of 50%. Several transmission studies have estimated that roughly half of infection transmission occurs pre-symptomatically [13–15]. He et al. estimated that 44% (95% CI, 30–57%) of secondary cases were infected during the index cases’ pre-symptomatic stage [13]. Based on a review of 17 studies reporting serial intervals or generation times, Casey et al. estimated that 56.1% of transmission occurred in the pre-symptomatic period [14]. Ferretti et al. found that the peak of transmission occurred at the time of symptom onset with 41% of transmission events occurring before symptoms onset and another 35% on the day of and day after symptom onset [15]. We estimate that 50% of infectiousness occurs before symptom onset. Subclinical infections may be less infectious than clinical infections, because of differences in viral burden or the result of symptoms, such as coughing or sneezing. The review of Buitrago-Garcia reported that the secondary infection risk was lower (relative risk 0.35, 95% CI 0.10–1.27) among contacts of asymptomatic cases relative to symptomatic cases [9]. We considered infections in the subclinical fraction as being 35% as infectious as the clinical fraction. In a sensitivity analysis, we considered subclinical and clinical infections to be equally infectious. We used age specific pre-pandemic estimates of contact rates from the work of Prem et al. who modeled age and location specific contact rates in 140 countries re-applying data from the European POLYMOD study [16]. The POLYMOD study defined contact as “…either skin-to-skin contact such as a kiss or handshake (a physical contact), or a two-way conversation with three or more words in the physical presence of another person but no skin-to-skin contact (a nonphysical contact)” [17]. This definition differs from the one applied most commonly for studies estimating the SIR for SARS-CoV-2. Prem et al. reported contact rates for home settings as well as for work, school and other settings. We summed the rates for non-home settings as a non-household contact rate (S1 Table). As we were unable to obtain age-stratified case incidence data, we did not account for the age-structure of contacts. We estimated the period case-hospitalization and case-fatality ratios by decile of age directly from US CDC surveillance case reports using the total counts of cases and those flagged for hospitalization or death [18]. We treated missing value in fields for hospitalization and death as non-hospitalized, non-fatal cases (S2 Table). We computed risks for hospitalization and death for 90-day period using the average period confirmed case incidence and the parameters listed in Table 1. We utilized a Monte Carlo simulation to incorporate each equation parameter with an error ranges as a normally distributed variable, reporting risk as the mean of the resulting distribution. To validate our estimates, we summed household and non-household risk and compared it to cumulative national age-specific mortality and hospitalization incidence rates for the same time period [19, 20] (S3 and S4 Tables). Results Among US Counties, in the 90-day period from June 16th, 2020 to September 15th, 2020, the median daily confirmed case incidence was 10.8 per 100,000, varying by county from 3.1 per 100,000 in the first quartile to 28.5 per 100,000 in the fifth quartile. Fig 1 illustrates the 90-day period risks for hospitalization and death attributable to household and non-household contacts across the range of U.S. county case incidence values applying the primary assumptions in Table 1. 10.1371/journal.pone.0243026.g001Fig 1 Estimated risk of hospitalization and death over a 90-day period at pre-pandemic U.S. contact rates within the range of recent U.S. county case incidence under the following assumptions: Household SIR 31.1.%; non-household SIR, 4.0%; clinical fraction, 69%; relative infectiousness of asymptomatic infection, 35%; isolation adherence, 75%. Table 2 combines the household and non-household components of risk of hospitalization and death for each quintile of confirmed case incidence. The estimated 90-day period risk of death at the period overall median county confirmed case incidence of 10.8 per 100,000 ranges between 0.4 and 8.9 per 100,000 for each decile of age between 20 and 60 years. The corresponding 90-day period risk of hospitalization ranges from 13.7 to 69.2 per 100,000. 10.1371/journal.pone.0243026.t002Table 2 Estimated risk of death and hospitalization at pre-pandemic U.S. contact rates over a 90-day time period at the median county incidence of confirmed infections during the “summer wave’ of the COVID-19 pandemic (June 16 –September 15, 2020). Cumulative Period Risk of Death (X 100000) Cumulative Period Risk of Hospitalization (X 100000) Quintile 1 2 3 4 5 1 2 3 4 5 Average Daily Case Incidence 3.1 6.9 10.8 17.1 28.5 3.1 6.9 10.8 17.1 28.5 20–29 years 0.1 0.3 0.4 0.7 1.1 4 8.8 13.7 21.6 36.2 30–39 years 0.4 0.9 1.3 2.1 3.6 8.2 18.1 28.4 44.9 75 40–49 years 0.9 2.1 3.3 5.2 8.8 12.1 27.1 42.4 66.9 111.9 50–59 years 2.6 5.7 8.9 14.2 23.7 19.9 44.2 69.2 109.2 182.6 Though computed for different age-strata, our estimates are comparable to observed U.S. average age-specific mortality rates during the same time period of 0.48, 1.46, 4.39, 10.81, and 24.67 per 100,000, for 10-year age groups from 15–24 years to 55–64 years. Our estimates are also comparable to average age-specific hospitalization rates among the 98 counties in the USCDC COVID-Net hospital surveillance program (18–29 years, 44.4 per 100,000; 30–39 years, 59.2 per 100,000; 40–49 years 78.5 per 100,000; 50–64 years 109.1 per 100,000). Table 3 provides the 90-day period risk estimates disaggregated by setting of contact (household or non-household) at the national median county confirmed case incidence of 10.8 per 100,000 and illustrates the sensitivity of risk estimates to the alternative assumptions. Using the primary assumptions in Table 1 (scenario 1), for a person aged 40 to 49 years, the 90-day period risk of death attributable to household and non-household contacts is 2.56 and 0.74 events in 100,000 respectively. For hospitalizations, the risk is 32.9 and 9.5 in 100,000 for household and non-household contacts, respectively. Assuming a non-household SIR of 4.0% and pre-pandemic contact rates, the share of transmissions attributable to household settings ranges from 73% to 78% depending on age. 10.1371/journal.pone.0243026.t003Table 3 Estimated risk of death and hospitalization at pre-pandemic U.S. contact rates over a 90-day time period at the median county incidence of confirmed infections during the “summer wave’ of the COVID-19 pandemic (June 16 –September 15, 2020). Risk of Death (x 100,000) Risk of Hospitalization (x 100,000) Household Share Age Group Household Contacts Non-Household Contacts Household Contacts Non-Household Contacts Scenario 1 (Base): Household SIR 31.1.%; Non-household SIR 4.0%, Clinical fraction 69%; Relative infectiousness of asymptomatic infection, 0.35; Isolation compliance 75% 20–29 Years 0.3 0.11 9.96 3.74 73% 30–39 Years 1.04 0.32 21.78 6.62 77% 40–49 Years 2.56 0.74 32.89 9.5 78% 50–59 Years 7.04 1.95 54.18 15 78% Scenario 2: Household SIR 31.1.%; Non-household SIR 1.65%, Clinical fraction 69%; Relative infectiousness of asymptomatic infection, 0.35; Isolation compliance 75% 20–29 Years 0.3 0.05 9.89 1.55 86% 30–39 Years 1.04 0.13 21.62 2.75 89% 40–49 Years 2.54 0.31 32.65 3.95 89% 50–59 Years 6.98 0.81 53.77 6.23 90% Scenario 3: Household SIR 31.1.%; Non-household SIR 4.0%%, Clinical fraction 31%; Relative infectiousness of asymptomatic infection, 0.35; Isolation compliance 75% 20–29 Years 0.21 0.09 6.84 3.04 69% 30–39 Years 0.72 0.26 14.97 5.39 74% 40–49 Years 1.76 0.6 22.6 7.74 74% 50–59 Years 4.83 1.59 37.22 12.21 75% Scenario 4: Household SIR, 31.1.%; Non-household SIR, 4.0%%, Clinical fraction, 69%; Relative infectiousness of asymptomatic infection, 1; Isolation compliance 75% 20–29 Years 0.38 0.16 12.6 5.13 73% 30–39 Years 1.32 0.44 27.56 9.08 77% 40–49 Years 3.24 1.01 41.62 13.03 78% 50–59 Years 8.9 2.67 68.55 20.56 79% Scenario 5: Household SIR, 31.1.%; Non-household SIR, 4.0%%, Clinical fraction, 69%; Relative infectiousness of asymptomatic infection, 0.35; Isolation compliance 50% 20–29 Years 0.31 0.13 10.15 4.38 72% 30–39 Years 1.06 0.37 22.21 7.77 76% 40–49 Years 2.61 0.87 33.54 11.15 77% 50–59 Years 7.17 2.28 55.23 17.59 78% Non-household risk falls proportionally using the lower estimate of the SIR of 1.65% (scenario 2). Decreasing the estimated clinical fraction to 31% (scenario 3) decreases both household and non-household risk estimates given we assume subclinical infections to be less infectious. In the second scenario, the relative size of the non-household share of risk also decreases slightly fewer infectious individuals are isolated. Assuming sub-clinical infections are equally infectious as clinical infections, (scenario 4) risks rise for both household and non-household contacts. Assuming adherence with isolation drops to 50% (scenario 5), risk increases for non-household contacts. Discussion We demonstrate a straightforward method to estimate individual risks of COVID-19 associated hospitalization and deaths, using publicly available data on case incidence, the clinical fraction, transmission timing, secondary infection risk, contact rates, and case hospitalization and fatality ratios. These estimates of risk reflect the average risk across a wide range of exposure settings and do not account for individual risk factors for vulnerability to severe illness other than age. While many parameters have significant uncertainties, the comparability of estimated risks to observed hospitalization and fatality incidence rates validates the approach. Aggregated case incidence data, parameter uncertainty, and the lack of setting specific transmission risk estimates are limitations of the approach and suggest opportunities to improve individual risk estimation. We assumed the susceptible fraction of the population remains high based on serological findings. The prevalence of detected antibodies varies with region and will change with time. Furthermore, protective immunity may not be well estimated by antibody detection alone. Durable lymphocyte responses may persist following exposure [21]. Observed cellular immune response to COVID-19 among unexposed individuals suggest prior exposure to related coronaviruses may further contribute to immunity [22]. Our estimates assume that prevalent infections are dispersed homogeneously within a county’s geography. This does not account for clustering within chains of transmission among related or socially connected individuals. We might also expect higher infection incidence among those living in congregate living facilities, in neighborhoods with more density or larger households, or among service workers and those working together in close quarters. Disaggregated public reporting of confirmed infection incidence by age, symptom status, and neighborhood and congregate living status would allow for risk estimates to be setting and population specific. Reviews of the secondary infection risk find significant heterogeneity among estimates. Most published estimates of the SIR come from observations outside the US and at an earlier time period in the pandemic, prior to normalization of behaviors intended to reduce the risk of infection transmission, such as increased hand washing, mask use and observing physical distance. SIR estimates are dependent on both contextual and study characteristics. Setting specific characteristics include the characteristics of human contact. Most studies defined contacts similarly, yet studies varied in the number of secondary contacts per primary case, suggesting variable interpretations of contact definitions or variable ability or wiliness to recall and disclose contacts. Most studies used a combination of RT-PCR and symptom monitoring to identify infected contacts. RT-PCR methods will produce false negative results [23]. Studies that ascertained infection with RT-PCR only at the beginning of contact observation could have missed pre-clinical or subclinical infections. The large difference between estimates of household and non-household SIR is not unexpected given that frequent, prolonged and intimate contact is the norm in households. Consistent with the relative size of the SIR estimates, we found that the majority of SARS-CoV-2 transmission can be attributed to households. The WHO mission to China also concluded that most infection transmission occurred in household settings [24]. An analysis of 1,038 cases identified in Hong Kong up to 28 April observed transmission to occur most frequently in households although non-household social settings could involve a larger number of secondary infections [25]. Relatively fewer studies have examined risk from non-household contact in detail; most of these occurred in East Asian countries and used data from governmental surveillance systems. We found insufficient estimates to apply SIRs for particular settings including workplaces, schools, and public transportation. There has been little systematic characterization of contacts of confirmed cases of COVID-19 in the U.S. Reporting of the proportion confirmed cases with known contacts and the setting of contact or relationships among cases of contacts varies by jurisdiction; overall, few jurisdictions report this information. Among confirmed cases identified before ‘stay-at-home’ orders in nine Colorado counties participating in a retrospective survey, only 27% reported a contact with an infected individual [26]. Another survey found only 46% adults testing positive for SARS-CoV-2 infection could recall a contact with a known COVID-19 patient; of those who could recall a contact, most recalled contacts were family members (45%) or work colleagues (34%) [27]. Our estimates do not account for individual variation in the SIR. With respiratory viruses, the number of secondary cases generated by each index case can vary significantly [28]. Studies suggest that a large share COVID-19 infection might be due to a small fraction of particularly infectious individuals [25, 29]. Characteristics of COVID-19 individual “super-spreaders” are not described but these individuals may make a disproportionately larger contribution to the spread of infection in non-household, social settings [25]. Clusters of COVID-19 infections have occurred in diverse settings [30]. We applied the summary estimate of the non-household SIR by Koh et al which included estimates of the SIR from cluster investigations thus representing the potential contribution of “super-spreader” events. We applied age-specific contact rates without regard to the age structure of contacts and based on pre-pandemic estimates. While changes in mobility and consumption suggest that many individuals may have lower contact rates, the nature and distribution of current social contacts is unknown. Furthermore, limits on social and workplace contacts outside the home may have increased household contact rates. Overall, we did not have a satisfactory way to adjust contact rates for post-pandemic conditions. The definition of contact used for our contact rates, which includes physical touch or face to face speech, differs from contact definitions used in studies of secondary risk from SARS-CoV-2, which was usually defined as face to face “unprotected” contact for greater than 15 minutes. Neither definition may optimally characterize the mode of COVID-19 transmission. For example, neither definition explicitly consider transmission from contact via shared contaminated surfaces. Studies of the SIR that define contacts alternatively, for example, as physical work in close quarters, might produce useful findings. We estimated the case hospitalization and fatality ratio by decile of age directly from line-level CDC surveillance case report data during the same period used to assess case-incidence. We note that these ratios have varied significantly both with time and by location over the course of the pandemic; however, our examination of state-level ratios over time demonstrates that recent period estimates have more homogeneity (S1 Fig). Our estimates do not account for variation in the risk of hospitalization and death due to individual co-morbid conditions [31]. Adjusting for these factors would lower estimated risks for most adults without common chronic disease conditions. It is unclear how these risk estimates compare to of perceived individual risk among adults in the U.S as there are few published estimates of perceived risk. In one online survey conducted in March 2020, the median perceived risk was 10.0% for infection and 5.0% for infection fatality [32]. Another study conducted in March and April found that adults found that in March, 14% of U.S. adults perceived the fatality risk from SARS-CoV-2 infection to be greater than a 1% benchmark while 67% reported a lower risk; however, the perception of risk increased over time [33]. Daily media reports of counts of confirmed infections and the perceived lack of control over exposure all may be influencing risk perception. Public risk perceptions may not be reflecting the relatively larger risk from household contacts. U.S. government action on restricting non-household social activity and the publicity surrounding disease clusters in social settings may have amplified perceived risks from non-household settings. In contrast, other countries took steps to quickly physically isolate infectious family member to limit transmissions from household contact [34, 35]. Return to community workplace and social life will require individuals to be have a better understanding of the personal risks of COVID-19 infection. Accurate setting and population specific estimates on the individual probabilities hospitalization and death may contribute to a more accurate risk perception. Systematically collected and publicly reported data on infection incidence by, for example, the geographic setting of exposure, residence type, whether a case had a known exposure, would allow more precise estimation than those possible with currently available public data. Calculation of secondary infection risks by setting and a more precise knowledge of susceptibility would improve individual risk estimates. Supporting information S1 Table Daily contact rates by age and setting. (DOCX) Click here for additional data file. S2 Table Estimates of the case hospitalization and case fatality ratios. (DOCX) Click here for additional data file. S3 Table Cumulative COVID-19 associated hospital admissions rates per 100,000 people during the period June 16 to September 15, 2020. (DOCX) Click here for additional data file. S4 Table Cumulative COVID-19 associated mortality rates per 100,000 people during the period June 16 to September 15, 2020. (DOCX) Click here for additional data file. S1 Fig Estimates of the state-specific case fatality ratio by month. (DOCX) Click here for additional data file. 10.1371/journal.pone.0243026.r001 Decision Letter 0 Shaman Jeffrey Academic Editor © 2020 Jeffrey Shaman2020Jeffrey ShamanThis is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.Submission Version0 14 Aug 2020 PONE-D-20-19975 Estimation of Individual Probabilities of COVID-19 Infection, Hospitalization, and Death From A County-level Contact of Unknown infection Status PLOS ONE Dear Dr. Bhatia, Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process. Please respond to the reviewer comments on a point-by-point basis and revise the manuscript accordingly. Please submit your revised manuscript by Sep 28 2020 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file. Please include the following items when submitting your revised manuscript: A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'. A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'. An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'. If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter. If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols We look forward to receiving your revised manuscript. Kind regards, Jeffrey Shaman Academic Editor PLOS ONE Journal Requirements: When submitting your revision, we need you to address these additional requirements. 1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf 2.In your Data Availability statement, you have not specified where the minimal data set underlying the results described in your manuscript can be found. PLOS defines a study's minimal data set as the underlying data used to reach the conclusions drawn in the manuscript and any additional data required to replicate the reported study findings in their entirety. All PLOS journals require that the minimal data set be made fully available. For more information about our data policy, please see http://journals.plos.org/plosone/s/data-availability. Upon re-submitting your revised manuscript, please upload your study’s minimal underlying data set as either Supporting Information files or to a stable, public repository and include the relevant URLs, DOIs, or accession numbers within your revised cover letter. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories. Any potentially identifying patient information must be fully anonymized. Important: If there are ethical or legal restrictions to sharing your data publicly, please explain these restrictions in detail. Please see our guidelines for more information on what we consider unacceptable restrictions to publicly sharing data: http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions. Note that it is not acceptable for the authors to be the sole named individuals responsible for ensuring data access. We will update your Data Availability statement to reflect the information you provide in your cover letter. [Note: HTML markup is below. Please do not edit.] Reviewers' comments: Reviewer's Responses to Questions Comments to the Author 1. Is the manuscript technically sound, and do the data support the conclusions? The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented. Reviewer #1: Yes ********** 2. Has the statistical analysis been performed appropriately and rigorously? Reviewer #1: Yes ********** 3. Have the authors made all data underlying the findings in their manuscript fully available? The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified. Reviewer #1: Yes ********** 4. Is the manuscript presented in an intelligible fashion and written in standard English? PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here. Reviewer #1: Yes ********** 5. Review Comments to the Author Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters) Reviewer #1: # PONE-D-20-19975 ## Overall comments I would like to thank the editor for sending me an interesting paper to review with such a concise yet useful model for estimating the risk to residents of US cities of contracting SARS-CoV-2 and subsequent hospitalisation and death. Better informing the public about the risks of infection is vital to ensuring proportional public health responses and adherence to guidelines by the public. Overall I am satisfied that the authors have thought through their work and its implications. The requested changes to the model above are meant to help the reader understand how much variability there is likely to be in the results based on uncertainty in published estimates used by the authors. ## Methods In Table 1 there are no subscripts on the CHR and CFRs indicating which age group they belong to. Given that they are indeed age-dependent, it should be noted that equations (3) and (4) are age dependent. Additionally, the probability of an individual in any age group considered becoming infected is likely to be related to the total number of contacts that they have, the age group of those contacts (Prem et al., 2017), and the prevalence of current infection within each age group (although I understand this may be difficult to obtain). Contacts per day are given in the Results section (Table 3) but it is not clear in the Methods where these values are from. Line 107: "can be both symptomatic and asymptomatic" should be "can be either symptomatic or asymptomatic". Regarding the asymptomatic fraction, Buitrago-Garcia et al. (2020) provide a living systematic review of asymptomatic fraction that provides an estimate of 31% (95% prediction interval: 24%-38%), twice that which is used by the authors here. As asymptomatic infection may play an important role in ongoing transmission (e.g. Rivett et al., 2020) it would be worth the authors investigating the effect of a doubling of the asymptomatic fraction, particularly as asymptomatic infections will never have the chance to self-isolate on onset of symptoms. Issues in case reporting are a major source of uncertainty in the estimation of prevalence and therefore risk of transmission. Russell et al. (2020) provide estimates of the reporting rate in the United States of America of 99% based on the method of Golding et al. (2020). The authors should be more explicit in how they came to the value of 75% which represents both under-reporting in symptomatic cases and undetected asymptomatic cases. An assumption of 100% of infections having unknown contacts doesn't match the assumption of household transmission that the authors use. I appreciate that the authors have included this parameter in their model for future work. The discussion should highlight this as a limitation and it may be worth considering varying this assumption for sensitivity analysis. In terms of culture and socioeconomics, I am skeptical that Israel provides a reasonable comparison to the United States of America for adherence to isolation guidelines. While financial compensation has been found to improve adherence there, an assumption that 75% of confirmed cases will voluntarily self-isolate is akin to assuming half of cases are having their income supported, given the numbers the authors provide. In contrast to income support in the United Kingdom and Singapore, income support in the United States of America is both low and spotty. A review earlier this year (Webster et al., 2020) indicated that trust in government and other sociocultural factors may play a role, and while much of the studies cited are related to Ebola in Africa and H1N1 pandemic influenza in Australia, studies of SARS in Canada indicated that a sense of "civic duty" and a belief in the importance of "following the law" were associated with increased adherence. A recent survey indicates association between adherence and in a belief in a moral imperative to comply (van Rooij et al., 2020). There is evidence in the USA that a county's political demographics play a role in adherence (Painter and Qiu, 2020), particularly viewership of Fox News (Simonov et al., 2020), and poverty (Wright et al., 2020). Given that the authors are considering county-level reported cases it may be worth estimating county-level rates of adherence to self-isolation guidelines. ## Results I am satisfied, for the most part, with the presentation of results and their discussion. There are no uncertainties or sensitivity analyses presented in the results, due to the lack of uncertainty in the parameter estimates in the authors' model. I would suggest that estimates of incidence should have uncertainty due to uncertainty in under-reporting rates, CHR and CFR, and average days infectious - where the authors use 8 days without providing a source, despite indications that duration of infectivity may have a great deal of uncertainty due to differences in viral load (Wölfel et al., 2020). Figure 1 is useful but I would also appreciate if the authors provided a figure with faceting by Event and colouring the lines by age group (colours to be consistent with the current Figure 2) in order to more easily see how the probability of each event varies with age group. The x axis is difficult to interpret, and the authors may want to consider converting from fractions to cases per 100,000. ## Discussion The authors' discussion does a reasonable job of explaining the limitations of the data available to them and they compare their results appropriately. At line 309 the authors the authors discuss mediation of risk perceptions based on restrictions on community action. I suggest they consider Webster et al. (2020) and their discussion on social norms and perception of risk. ## References Prem et al. (2017) https://doi.org/10.1371/journal.pcbi.1005697 Rivett et al. (2020) https://doi.org/10.7554/eLife.58728 Russell et al. (2020) https://cmmid.github.io/topics/covid19/global_cfr_estimates.html Golding et al. (2020) https://doi.org/10.1101/2020.07.07.20148460 Buitrago-Garcia et al. (2020) https://doi.org/10.1101/2020.04.25.20079103 Webster et al. (2020) https://doi.org/10.1016/j.puhe.2020.03.007 van Rooij et al. (2020) https://dx.doi.org/10.2139/ssrn.3582626 Painter and Qiu (2020) https://dx.doi.org/10.2139/ssrn.3569098 Simonov et al. (2020) https://doi.org/10.3386/w27237 Wright et al. (2020) https://dx.doi.org/10.2139/ssrn.3573637 Wölfel et al. (2020) https://doi.org/10.1038/s41586-020-2196-x ********** 6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files. If you choose “no”, your identity will remain anonymous but your review may still be made public. Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy. Reviewer #1: Yes: Samuel Clifford [NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.] While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step. 10.1371/journal.pone.0243026.r002 Author response to Decision Letter 0 Submission Version1 26 Oct 2020 Authors’ responses to reviewers’ comments C1. I would like to thank the editor for sending me an interesting paper to review with such a concise yet useful model for estimating the risk to residents of US cities of contracting SARS-CoV-2 and subsequent hospitalization and death. Better informing the public about the risks of infection is vital to ensuring proportional public health responses and adherence to guidelines by the public. R1. No response indicated C2. Overall I am satisfied that the authors have thought through their work and its implications. The requested changes to the model above are meant to help the reader understand how much variability there is likely to be in the results based on uncertainty in published estimates used by the authors. R2. No response indicated C3. In Table 1 there are no subscripts on the CHR and CFRs indicating which age group they belong to. R3. We have specified the age-specific CHR and CFR estimates in the revision. C4. Given that they are indeed age-dependent, it should be noted that equations (3) and (4) are age dependent. R4. We have revised the description of formulae to note their age specificity. C5. Additionally, the probability of an individual in any age group considered becoming infected is likely to be related to the total number of contacts that they have, the age group of those contacts (Prem et al., 2017), and the prevalence of current infection within each age group (although I understand this may be difficult to obtain). Contacts per day are given in the Results section (Table 3) but it is not clear in the Methods where these values are from. R5. In the revision, we now use age and setting specific pre-pandemic contact rates from the analysis of Prem at al for our risk estimates. Given the lack of publicly available age-specific data on infection prevalence at the county level, we could not apply the age structure of contacts. We note the limitation and the potential value obtaining and applying age-structured incidence data in our revision C6. Line 107: "can be both symptomatic and asymptomatic" should be "can be either symptomatic or asymptomatic". R6. Corrected C7. Regarding the asymptomatic fraction, Buitrago-Garcia et al. (2020) provide a living systematic review of asymptomatic fraction that provides an estimate of 31% (95% prediction interval: 24%-38%), twice that which is used by the authors here. As asymptomatic infection may play an important role in ongoing transmission (e.g. Rivett et al., 2020) it would be worth the authors investigating the effect of a doubling of the asymptomatic fraction, particularly as asymptomatic infections will never have the chance to self-isolate on onset of symptoms. R7. The revision applies the estimate of the clinical fraction from Buitrago-Garcia et al and applies a lower estimate from the work of Poletti et al in a sensitivity analysis. C8. Issues in case reporting are a major source of uncertainty in the estimation of prevalence and therefore risk of transmission. Russell et al. (2020) provide estimates of the reporting rate in the United States of America of 99% based on the method of Golding et al. (2020). The authors should be more explicit in how they came to the value of 75% which represents both under-reporting in symptomatic cases and undetected asymptomatic cases. R8. We have modified the methodology in the revision no longer apply or require a ratio between confirmed and unconfirmed infections. We have made and noted an assumption that the relationship between confirmed infections and all infections will be simply multiplicative within a time and place. C9. An assumption of 100% of infections having unknown contacts doesn't match the assumption of household transmission that the authors use. I appreciate that the authors have included this parameter in their model for future work. The discussion should highlight this as a limitation and it may be worth considering varying this assumption for sensitivity analysis. R9. Changes to the methodology do not apply a parameter for unknown contacts. C10. In terms of culture and socioeconomics, I am skeptical that Israel provides a reasonable comparison to the United States of America for adherence to isolation guidelines. While financial compensation has been found to improve adherence there, an assumption that 75% of confirmed cases will voluntarily self-isolate is akin to assuming half of cases are having their income supported, given the numbers the authors provide. In contrast to income support in the United Kingdom and Singapore, income support in the United States of America is both low and spotty. A review earlier this year (Webster et al., 2020) indicated that trust in government and other sociocultural factors may play a role, and while much of the studies cited are related to Ebola in Africa and H1N1 pandemic influenza in Australia, studies of SARS in Canada indicated that a sense of "civic duty" and a belief in the importance of "following the law" were associated with increased adherence. A recent survey indicates association between adherence and in a belief in a moral imperative to comply (van Rooij et al., 2020). There is evidence in the USA that a county's political demographics play a role in adherence (Painter and Qiu, 2020), particularly viewership of Fox News (Simonov et al., 2020), and poverty (Wright et al., 2020). Given that the authors are considering county-level reported cases it may be worth estimating county-level rates of adherence to self-isolation guidelines. R 10. We agree that data on adherence is insufficient. We were not able to find a reliable source for this parameter for the United States. We test the sensitivity of our estimates to this parameter. C11. I am satisfied, for the most part, with the presentation of results and their discussion. There are no uncertainties or sensitivity analyses presented in the results, due to the lack of uncertainty in the parameter estimates in the authors' model. I would suggest that estimates of incidence should have uncertainty due to uncertainty in under-reporting rates, CHR and CFR, and average days infectious - where the authors use 8 days without providing a source, despite indications that duration of infectivity may have a great deal of uncertainty due to differences in viral load (Wölfel et al., 2020). R11. We have added sensitivity analyses varying individual parameter estimates. We removed the parameter for duration of infectiousness as the SIR estimates include follow up duration that encompasses this interval. C12. Figure 1 is useful but I would also appreciate if the authors provided a figure with faceting by Event and colouring the lines by age group (colours to be consistent with the current Figure 2) in order to more easily see how the probability of each event varies with age group. The x axis is difficult to interpret, and the authors may want to consider converting from fractions to cases per 100,000. R12. The figure has been revised. C13. The authors' discussion does a reasonable job of explaining the limitations of the data available to them and they compare their results appropriately. R13. No response indicated C14. At line 309 the authors the authors discuss mediation of risk perceptions based on restrictions on community action. I suggest they consider Webster et al. (2020) and their discussion on social norms and perception of risk. R14. We reference the review of Webster in the introduction to the revision. Attachment Submitted filename: covidrisks responsetoreviewers.docx Click here for additional data file. 10.1371/journal.pone.0243026.r003 Decision Letter 1 Shaman Jeffrey Academic Editor © 2020 Jeffrey Shaman2020Jeffrey ShamanThis is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.Submission Version1 16 Nov 2020 Estimating individual risks of COVID-19-associated hospitalization and death using publicly available data PONE-D-20-19975R1 Dear Dr. Bhatia, We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements. Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication. An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org. If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org. Kind regards, Jeffrey Shaman Academic Editor PLOS ONE Additional Editor Comments (optional): Reviewers' comments: Reviewer's Responses to Questions Comments to the Author 1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation. Reviewer #1: All comments have been addressed ********** 2. Is the manuscript technically sound, and do the data support the conclusions? The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented. Reviewer #1: Yes ********** 3. Has the statistical analysis been performed appropriately and rigorously? Reviewer #1: Yes ********** 4. Have the authors made all data underlying the findings in their manuscript fully available? The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified. Reviewer #1: Yes ********** 5. Is the manuscript presented in an intelligible fashion and written in standard English? PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here. Reviewer #1: (No Response) ********** 6. Review Comments to the Author Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters) Reviewer #1: All my comments have been addressed and I have no further changes to request to the manuscript. Thank you. ********** 7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files. If you choose “no”, your identity will remain anonymous but your review may still be made public. Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy. Reviewer #1: Yes: Sam Clifford 10.1371/journal.pone.0243026.r004 Acceptance letter Shaman Jeffrey Academic Editor © 2020 Jeffrey Shaman2020Jeffrey ShamanThis is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. 19 Nov 2020 PONE-D-20-19975R1 Estimating individual risks of COVID-19-associated hospitalization and death using publicly available data Dear Dr. Bhatia: I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department. If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org. If we can help with anything else, please email us at plosone@plos.org. Thank you for submitting your work to PLOS ONE and supporting open access. Kind regards, PLOS ONE Editorial Office Staff on behalf of Prof. Jeffrey Shaman Academic Editor PLOS ONE ==== Refs References 1 Webster RK , Brooks SK , Smith LE , Woodland L , Wessely S , Rubin GJ . How to improve adherence with quarantine: rapid review of the evidence . Public Health . 2020 5 ;182 :163 –169 . 10.1016/j.puhe.2020.03.007 32334182 2 On ongoing repository of data on coronavirus cases and deaths in the U.S. The New York Times Company. [Cited: 2020 Oct 18] https://github.com/nytimes/covid-19-data 3 Havers FP , Reed C , Lim T , Montgomery JM , Klena JD , Hall AJ , et al Seroprevalence of Antibodies to SARS-CoV-2 in 10 Sites in the United States, March 23-May 12, 2020 . JAMA Intern Med . 2020 7 21 10.1001/jamainternmed.2020.4130 32692365 4 Anand S , Montez-Rath M , Han J , Bozeman J , Kerschmann R , Beyer P , et al Prevalence of SARS-CoV-2 antibodies in a large nationwide sample of patients on dialysis in the USA: a cross-sectional study . Lancet . 2020 9 25 10.1016/S0140-6736(20)32009-2 32987007 5 Koh WC , Naing L , Chaw L , Rosledzana MA , Alikhan MF , Jamaludin SA , et al What do we know about SARS-CoV-2 transmission? A systematic review and meta-analysis of the secondary attack rate and associated risk factors . PLoS One . 2020 10 8 ;15 (10 ):e0240205 10.1371/journal.pone.0240205 33031427 6 Lei H , Xu X , Xiao S , Wu X , Shu Y . Household transmission of COVID-19-a systematic review and meta-analysis . J Infect . 2020 8 25 :S0163-4453(20)30571-5. 10.1016/j.jinf.2020.08.033 32858069 7 Madewell ZJ , Yang Y , Longini IM , Halloran ME , Dean NE . Household transmission of SARS-CoV-2: a systematic review and meta-analysis of secondary attack rate . medRxiv [Preprint]. 2020 7 31 :2020.07.29.20164590. 10.1101/2020.07.29.20164590 32766596 8 Oran DP , Topol EJ . Prevalence of Asymptomatic SARS-CoV-2 Infection: A Narrative Review . Ann Intern Med . 2020 9 1 ;173 (5 ):362 –367 . 10.7326/M20-3012 32491919 9 Buitrago-Garcia D , Egli-Gany D , Counotte MJ , Hossmann S , Imeri H , Ipekci AM , et al Occurrence and transmission potential of asymptomatic and presymptomatic SARS-CoV-2 infections: A living systematic review and meta-analysis . PLoS Med . 2020 9 22 ;17 (9 ):e1003346 10.1371/journal.pmed.1003346 32960881 10 Poletti P, Tirani M, Cereda D, et al. Probability of symptoms and critical disease after SARS-CoV-2 infection. [Preprint]. 2020 Jun 22. arXiv:2006.08471v2 11 Davies NG , Klepac P , Liu Y , Prem K , Jit M . Age-dependent effects in the transmission and control of COVID-19 epidemics . Nat Med . 2020 6 16 ;26 (8 ). 10.1038/s41591-020-0962-9 32546824 12 Bodas M , Peleg K . Self-Isolation Compliance In The COVID-19 Era Influenced By Compensation: Findings From A Recent Survey In Israel . Health Aff (Millwood) . 2020 4 9 ; 39 (6 ). 10.1377/hlthaff.2020.00382 32271627 13 He X , Lau EHY , Wu P , Deng X , Wang J , Hao X , et al Temporal dynamics in viral shedding and transmissibility of COVID-19 . Nat Med . 2020 5 ;26 (5 ):672 –675 . 10.1038/s41591-020-0869-5 Erratum in: Nat Med. 2020 Sep;26(9):1491–1493. 32296168 14 Casey M. , Griffin J. , McAloon C.G. , Byrne A.W. , Madden J.M. , McEvoy D. , et al (2020 ). Estimating pre-symptomatic transmission of COVID-19: a secondary analysis using published data . [Preprint] medRxiv . 2020 6 11 10.1101/2020.05.08.20094870 15 Ferretti L , Ledda A , Wymant C , Zhao L , Ledda V , Abeler-Dorner L , et al The Timing of COVID-19 Transmission . [Preprint] MedRxiv . 2020 9 16 10.1101/2020.09.04.20188516 . 16 Prem K , Cook AR , Jit M . Projecting social contact matrices in 152 countries using contact surveys and demographic data . PLoS Comput Biol . 2017 9 12 ;13 (9 ):e1005697 10.1371/journal.pcbi.1005697 28898249 17 Mossong J , Hens N , Jit M , Beutels P , Auranen K , Mikolajczyk R , et al Social contacts and mixing patterns relevant to the spread of infectious diseases . PLoS Med . 2008 3 25 ;5 (3 ):e74 10.1371/journal.pmed.0050074 18366252 18 COVID-19 Case Surveillance Public Data Access, Summary, and Limitations. U.S. Centers for Disease Control and Prevention. [Cited: 2020 Oct 18] https://data.cdc.gov/Case-Surveillance/COVID-19-Case-Surveillance-Public-Use-Data/vbim-akqf 19 Provisional COVID-19 Death Counts by Sex, Age, and Week. U.S. Centers for Disease Control and Prevention. [Cited: 2020 Oct 18] https://data.cdc.gov/NCHS/Provisional-COVID-19-Death-Counts-by-Sex-Age-and-W/vsak-wrfu 20 Laboratory-confirmed Covid-19- Associated Hospitalizations. U.S. Centers for Disease Control and Prevention. [Cited: 2020 Oct 18] https://gis.cdc.gov/grasp/COVIDNet/COVID19_3.html 21 Robbiani DF , Gaebler C , Muecksch F , Lorenzi JCC , Wang Z , Cho A , et al Convergent antibody responses to SARS-CoV-2 in convalescent individuals . Nature . 2020 8 ;584 (7821 ):437 –442 . 10.1038/s41586-020-2456-9 32555388 22 Grifoni A , Weiskopf D , Ramirez SI , Mateus J , Dan JM , Moderbacher CR , et al Targets of T Cell Responses to SARS-CoV-2 Coronavirus in Humans with COVID-19 Disease and Unexposed Individuals . Cell . 2020 6 25 ;181 (7 ):1489 –1501.e15 . 10.1016/j.cell.2020.05.015 32473127 23 Kucirka LM , Lauer SA , Laeyendecker O , Boon D , Lessler J . Variation in False-Negative Rate of Reverse Transcriptase Polymerase Chain Reaction-Based SARS-CoV-2 Tests by Time Since Exposure . Ann Intern Med . 2020 8 18 ;173 (4 ):262 –267 . 10.7326/M20-1495 32422057 24 World Health Organization. Report of the WHO-China Joint Mission of Coronavirus Disease 2019 (COVID-19). 2020 Feb 16–24. 25 Adam DC , Wu P , Wong JY , Lau EHY , Tsang TK , Cauchemez S , et al Clustering and superspreading potential of SARS-CoV-2 infections in Hong Kong . Nat Med . 2020 9 17 10.1038/s41591-020-1092-0 32943787 26 Marshall K , Vahey GM , McDonald E , Tate JE , Herlihy R , Midgley CM , et al Exposures Before Issuance of Stay-at-Home Orders Among Persons with Laboratory-Confirmed COVID-19—Colorado, March 2020 . MMWR Morb Mortal Wkly Rep . 2020 7 3 ;69 (26 ):847 –849 . 10.15585/mmwr.mm6926e4 32614809 27 Tenforde MW , Billig Rose E , Lindsell CJ , Shapiro NI , Files DC , et al Characteristics of Adult Outpatients and Inpatients with COVID-19–11 Academic Medical Centers, United States, March-May 2020 . MMWR Morb Mortal Wkly Rep . 2020 7 3 ;69 (26 ):841 –846 . 10.15585/mmwr.mm6926e3 32614810 28 Lloyd-Smith JO , Schreiber SJ , Kopp PE , Getz WM . Superspreading and the effect of individual variation on disease emergence . Nature . 2005 11 17 ;438 (7066 ):355 –9 . 10.1038/nature04153 16292310 29 Endo A ; Centre for the Mathematical Modelling of Infectious Diseases COVID-19 Working Group , Abbott S , Kucharski AJ , Funk S . Estimating the overdispersion in COVID-19 transmission using outbreak sizes outside China . Wellcome Open Res . 2020 7 10 ;5 :67 10.12688/wellcomeopenres.15842.3 32685698 30 Leclerc QJ , Fuller NM , Knight LE ; CMMID COVID-19 Working Group , Funk S , Knight GM . What settings have been linked to SARS-CoV-2 transmission clusters? Wellcome Open Res . 2020 6 5 ;5 :83 10.12688/wellcomeopenres.15889.2 32656368 31 Kim L , Garg S , O’Halloran A , Whitaker M , Pham H , Anderson EJ , et al Risk Factors for Intensive Care Unit Admission and In-hospital Mortality among Hospitalized Adults Identified through the U.S. Coronavirus Disease 2019 (COVID-19)-Associated Hospitalization Surveillance Network (COVID-NET) . Clin Infect Dis . 2020 7 16 :ciaa1012. 10.1093/cid/ciaa1012 32674114 32 Bruine de Bruin W , Bennett D . Relationships Between Initial COVID-19 Risk Perceptions and Protective Health Behaviors: A National Survey . Am J Prev Med . 2020 8 ;59 (2 ):157 –167 . 10.1016/j.amepre.2020.05.001 32576418 33 Niepel C , Kranz D , Borgonovi F , Emslander V , Greiff S . The coronavirus (COVID-19) fatality risk perception of US adult residents in March and April 2020 . Br J Health Psychol . 2020 6 10 : 10.1111/bjhp.12438 32519364 34 Pan A , Liu L , Wang C , Guo H , Hao X , Wang Q , et al Association of Public Health Interventions With the Epidemiology of the COVID-19 Outbreak in Wuhan, China . JAMA . 2020 5 19 ;323 (19 ):1915 –1923 . 10.1001/jama.2020.6130 .32275295 35 Thompson D . What’s Behind South Korea’s COVID-19 Exceptionalism? The Atlantic . 2020 5 6 .