A model for determining optimal frequency and fraction of population to be tested, for cost-effective surveillance testing for SARS-CoV2: algorithm development (Preprint)
BACKGROUND Because Covid19 can be asymptomatic, surveillance testing with isolation of those who test positive can reduce spread. But what fraction of a population should be tested, how often, and how long should surveillance testing continue as vaccination progresses? OBJECTIVE Develop a model that allows determination of extent and frequency of testing that optimize cost-effectiveness, and time after which surveillance testing ceases to be cost-effective. METHODS Free software from Illumina (Analytica 101) was used to make a model that runs on personal computers. Graphical interfaces demonstrate how the model reaches its conclusions and allow users with little programming experience to choose values for parameters that vary in different circumstances. RESULTS Optimal test frequency and fraction of a population to be tested depend on test cost, test sensitivity, false negative rate, delay (if any) between testing and isolation of those who test positive, and rate of compliance with isolation. Overall cost-effectiveness and optimal duration of testing depend critically on the value placed on saving a life, effective spread rate (R_eff) when surveillance testing is instituted, and fraction of a population already immune due to prior infection or vaccination. CONCLUSIONS For current spread rates (R_eff ca 1.2-1.5) and percent immune (ca 10%), overall cost-effectiveness generally requires value per life saved to be greater than $100,000. Optimal test frequency and fraction tested range from days to months, and ca .1 to 1, respectively, for different types of tests (antigen, pcr) currently on the market.