A Working Model to Inform Risk-Based Back to Work Strategies [article]

Kristen Meier, Kirsten J. Curnow, Darcy Vavrek, John Moon, Kyle Farh, Martin Chian, Robert Ragusa, Eileen de Feo, Phillip G. Febbo
2020 medRxiv   pre-print
The coronavirus disease 2019 (COVID-19) pandemic has forced many businesses to close or move to remote work to reduce the potential spread of disease. Employers desiring a return to onsite work want to understand their risk for having an infected employee on site and how best to mitigate this risk. Here, we modelled a range of key metrics to help inform return to work policies and procedures, including evaluating the benefit and optimal design of a SARS-CoV-2 employee screening program.
more » ... ng program. Methods: We modeled a range of input variables including prevalence of COVID-19, time infected, number of employees, test sensitivity and specificity, test turnaround time, number of times tested within the infectious period, and sample pooling. We modeled the impact of these input variables on several output variables: number of healthy employees; number of infected employees; number of test positive and test negative employees; number of true positive, false positive, true negative, and false negative employees; positive and negative predictive values; and time an infected, potentially contagious employee is on site. Results: We show that an employee screening program can reduce the risk for onsite transmission across different prevalence values and group sizes. For example, at a pre-test asymptomatic community prevalence of 0.5% (5 in 1000) with an employee group size of 500, the risk for at least one infected employee on site is 91.8%, with 3 asymptomatic infected employees predicted within those 500 employees. Implementing a SARS-CoV-2 baseline screen with an 80% sensitivity and 99.5% specificity would reduce the risk of at least one infected employee on site to 39.4% and the predicted number of infected employees onsite (false negatives) to 1. Repetitive testing is required for ongoing vigilance of onsite employees. The expected number of days an infected employee is on site depends on test sensitivity, testing interval, and turnaround time. If the test interval is longer than the infectious period (~14 days for COVID-19), testing will not detect the infected employee. Sample pooling reduces the number of tests performed, thereby reducing testing costs. However, the pooling methodology (eg, 1-stage vs 2-stage pooling, pool size) will impact the number of employees that screen positive, thereby affected the number of employees eligible to return to onsite work. Conclusions: The modeling presented here can be used to help employers understand their risk for having an infected employee on site. Further, it details how an employee screening program can reduce this risk and shows how screening performance and frequency impact the effectiveness of a screening program. The primary factors determining the effectiveness of a screening program are test sensitivity and frequency of testing.
doi:10.1101/2020.12.18.20248512 fatcat:ihwlnuikanaalke4tw3ayuf5ti