key: cord-0825656-q5dcvb04 authors: Paul, Prabasaj; Mosites, Emily; Laws, Rebecca L.; Scobie, Heather; Slayton, Rachel B.; Clarke, Kristie E. N. title: Modeling the frequency and number of persons to test to detect and control COVID-19 outbreaks in congregate settings date: 2020-11-20 journal: bioRxiv DOI: 10.1101/2020.11.20.391011 sha: 1ac1c15c69e0c1c95f71229a6ef2bbf131a0f66a doc_id: 825656 cord_uid: q5dcvb04 Background Congregate settings are at risk for coronavirus disease 2019 (COVID-19) outbreaks. Diagnostic testing can be used as a tool in these settings to identify outbreaks and to control transmission. Methods We used transmission modeling to estimate the minimum number of persons to test and the optimal frequency to detect small outbreaks of COVID-19 in a congregate facility. We also estimated the frequency of testing needed to interrupt transmission within a facility. Results The number of people to test and frequency of testing needed depended on turnaround time, facility size, and test characteristics. Parameters are calculated for a variety of scenarios. In a facility of 100 people, 26 randomly selected individuals would need to be tested at least every 6 days to identify a true underlying prevalence of at least 5%, with test sensitivity of 85%, and greater than 95% outbreak detection sensitivity. Disease transmission could be interrupted with universal, facility-wide testing with rapid turnaround every three days. Conclusions Testing a subset of individuals in congregate settings can improve early detection of small outbreaks of COVID-19. Frequent universal diagnostic testing can be used to interrupt transmission within a facility, but its efficacy is reliant on rapid turnaround of results for isolation of infected individuals. 23 Background: Congregate settings are at risk for coronavirus disease 2019 (COVID-19) outbreaks. Diagnostic testing can be used as a tool in these settings to identify outbreaks and to control 25 transmission. 26 Methods: We used transmission modeling to estimate the minimum number of persons to test and the 27 optimal frequency to detect small outbreaks of COVID-19 in a congregate facility. We also estimated the 28 frequency of testing needed to interrupt transmission within a facility. Testing for early detection of outbreaks 56 Outbreak detection was defined as identifying at least one positive test, regardless of true 57 underlying prevalence. We defined "early" detection to mean identifying at least one positive result 58 before case counts surpass a set number; here we chose 2, 5, and 10 true cases. For this purpose, the 59 number of people that need to be tested in a congregate setting is dependent on the true underlying 60 prevalence of infection, and the sensitivity and the specificity of the test. We estimated the number of randomly-selected individuals (n) needed to test for detection of an outbreak based on an expected minimum positive predictive value. When the required exceeded the facility size , then the outbreak 64 was considered too small to be detected. size where the prevalence of infections is , and presence of an outbreak is indicated by at least one The true underlying prevalence of infection, , was determined by the rate of introduction, ( = 76 community incidence × facility size) and the doubling time ( ) within the facility as: Solving for the expected number of infections at time after first introduction (i.e., with n(0)=1) gives: At a facility where the rate of introduction is and the doubling time is , the expected number of 81 infections at time after first introduction is: The frequency of testing for detection at a facility should not exceed the time from first 124 introduction of an infected person from the community to the maximum threshold of allowable Using testing alone without additional isolation of symptomatic individuals ( Figure 1B ) would 143 require more frequent testing to achieve a 60% reduction in transmission. Daily testing would be required if results were available in 24 hours and testing every 2 days required if there was a 12-hour 145 delay in test results. Figure 1 (14). Ultimately, the frequency of testing at a facility will 162 depend on the balance between risk tolerance, frequency of introducing infections, and resource 163 availability. One limitation of this analysis is that we applied general formulas that did not account for 165 specific characteristics of individuals residing in and working in these congregate settings. Furthermore, 166 uncertainties in the modeling parameters introduce imprecision in derived estimates. These values can 167 provide starting points for consideration for testing strategies but should not be considered definitive. the best available data. morbidity and mortality among individuals in congregate settings, prevent further spread into the 175 community, and decrease strain on healthcare systems. 176 177 High Contagiousness and Rapid 182 Spread of Severe Acute Respiratory Syndrome Coronavirus 2 COVID-19 in a Long-Term Care Facility 185 CoV-2 Infections in Residents of a Long-Term Care Skilled Nursing Facility -189 COVID-19 in Correctional 199 and Detention Facilities -United States Serial Laboratory Testing 202 for SARS-CoV-2 Infection Among Incarcerated and Detained Persons in a COVID-19 Among 205 Workers in Meat and Poultry Processing Facilities -19 States Temporal dynamics in viral shedding and 210 transmissibility of COVID-19 The authors would like to thank Dr. Eric Mooring for his thoughtful review of the manuscript.