Using repeated antibody testing to minimize bias in estimates of prevalence and incidence of SARS-CoV-2 infection

Abstract Objectives The prevalence and incidence of SARS-CoV-2, the virus which causes COVID-19, at any given time remains controversial, and is an essential piece in understanding the dynamics of the epidemic. Cross-sectional studies and single time point testing approaches continue to struggle with appropriate adjustment methods for the high false positive rates in low prevalence settings or high false negative rates in high prevalence settings, and post-hoc adjustment at the group level does not fully address this issue for incidence even at the population level. Methods In this study, we use seroprevalence as an illustrative example of the benefits of using a case definition using a combined parallel and serial testing framework to confirm antibody-positive status. In a simulation study, we show that our proposed approach reduces bias and improves positive and negative predictive value across the range of prevalence compared with cross-sectional testing even with gold standard tests and post-hoc adjustment. Using data from the North Carolina COVID-19 Community Research Partnership, we applied the proposed case definition to the estimation of SARS-CoV-2 seroprevalence and incidence early in the pandemic. Results The proposed approach is not always feasible given the cost and time required to administer repeated tests; however, it reduces bias in both low and high prevalence settings and addresses misclassification at the individual level. This approach can be applied to almost all testing contexts and platforms. Conclusions This systematic approach offers better estimation of both prevalence and incidence, which is important to improve understanding and facilitate controlling the pandemic.