LAUSR

Background: Understanding the dynamics of the COVID-19 pandemic and evaluating the efficacy of control measures requires knowledge of the number of infections over time. This number, however, often differs from the number of confirmed cases due to a large fraction of asymptomatic infections and variable testing strategies. Methods: This study uses death count statistics, age-dependent infection fatality risks and stochastic modeling to estimate the prevalence of SARS-CoV-2 infections among adults (age ≥ 20 years) in 165 countries over time, from early 2020 until June 25, 2021. The accuracy of the approach is confirmed through comparison to previous nationwide seroprevalence surveys. Results: The presented estimates reveal that the fraction of infections that are detected vary widely over time and between countries, and hence confirmed cases alone often yield a false picture of the pandemic. As of June 25, 2021, the nationwide cumulative fraction of SARS-CoV-2 infections (cumulative infections relative to population size) is estimated at 98% (95%-CI 93–100) for Peru, 83% (61–94) for Brazil and 36% (23–61) for the US. Conclusions: The presented time-resolved estimates expand the possibilities to study the factors that influenced and still influence the pandemic’s progression in 165 countries.