A study estimates the underreporting of COVID-19 cases in the United States, finding that a potentially significant percentage of infections may have gone unreported and that the United States is likely far from achieving herd immunity through infection alone. Despite various sources providing information about the number of SARS-CoV-2 infections in the United States, biases in testing likely contribute to an undercount of the true number of cases. Adrian Raftery and colleagues developed a Bayesian framework to estimate viral prevalence based on available data sources, including deaths, number of confirmed cases, and number of tests performed. The model was anchored in data from previous random sample testing surveys in Indiana and Ohio, yielding estimated infection fatality rates of 0.84% and 0.83%, respectively. Model results suggested that approximately 60% of infections may have gone unreported, with around 20% of the population having likely been infected by March 2021. According to the authors, the results show the value of random sample testing during a pandemic and suggest that the estimated true number of infections is far below potential herd immunity thresholds and that immunization may achieve herd immunity without the fatalities that accompany infections.
###
Article #21-03272: "Estimating SARS-CoV-2 infections from deaths, confirmed cases, tests, and random surveys," by Nicholas J. Irons and Adrian E. Raftery.
MEDIA CONTACT: Adrian E. Raftery, University of Washington, Seattle, WA; email: raftery@uw.edu
Proceedings of the National Academy of Sciences