Dynamics of SARS-CoV-2 seroassay sensitivity: a systematic review and modeling study

In this project, we did a systematic meta-analysis of sensitivity data of serological assay sensitivities for SARS-CoV-2 across time. This involved collecting a novel dataset of the proportion of positive tests when testing previously diagnosed individuals. A slightly pre-processed version of this unique dataset can be found here. Data and code.

Key public health message: What did you want to address in this study? Knowing how many people get infected with SARS-CoV-2 is important for determining the severity of the virus, herd immunity thresholds and groups at higher risk. To estimate this, results from antibody tests are used. However, antibody levels fall with time after infection, which can make these tests unreliable. We aimed to quantify the reliability of different tests, to understand possible biases in our understanding of COVID-19. What have we learnt from this study? The change of antibody test reliability through time is very variable, and it depends on assay characteristics. Some assays will give strongly biased estimates of infections a couple of months after an epidemic wave, while others will remain reliable for many months. We provide a tool for researchers to assess the risk that an assay will give biased results, and to quantitatively correct for this effect. What are the implications of your findings for public health? Because test reliability changes across time, some antibody tests have the potential to strongly bias our understanding of crucial aspects of COVID-19. As this effect varies depending on the test, the reliability of test results needs to be considered in a test-specific way. For future outbreaks and new infectious diseases, it is important that public health agencies provide guidelines and tools to account for this possible bias.

Recommended citation: Nana Owusu-Boaitey, Timothy W Russell, Gideon Meyerowitz-Katz, Andrew T Levin, Daniel Herrera-Esposito (2023). " Dynamics of SARS-CoV-2 seroassay sensitivity: a systematic review and modeling study ." Eurosurveillance. 28(21).
Download Paper