Evaluation of Group Testing for SARS-CoV-2 RNA

Nasa Sinnott-Armstrong; Daniel Klein; Brendan Hickey

. CC-BY 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Pooling dimension Fraction of positives among re−tests Figure S1 : Comparison of random (black) versus case-probability-ordered (red) modes of pooling. Simulations done with known case probability drawn from Beta(0.5, 10) (5% prevalence). Pooling dimension k is the size of the k × k plate of individuals to be tested (rows and columns). In both the total number of tests relative to naïve (left) and the fraction of positives in the resulting secondary testing scheme (right), the ordered method results in slight improvements around the optimum pooling dimension. We suggest these are not sufficient to warrant the additional complexity this method entails. Prevalence overestimation in current testing Relative cases discovered per test Figure S2 : Expected improvements relative to current testing. Across global tests performed so far, we choose the best grouped testing strategy per country. In order to double the current tests performed (left), grouped testing requires progressively fewer total tests at lower prevalence. Prevalence scaled by observed prevalence within each country (so 2-fold overestimation in a country with 3% observed prevalence = 1.5% tested prevalence). Testing these lower-prevalence individuals using group testing finds more cases per test than existing strategies did, even at lower overall prevalence than existing screening (right).