CERGAS Webinar “Estimating the death toll of COVID-19 in the United States”

Abstract

Providing timely estimates of deaths associated with the COVID-19 pandemic is critical to provide policymakers and the public with an accurate picture  of severity. However, official tallies of deaths likely represent an undercount, due to variations in practices for recording causes of death, lack of testing, and delays in reporting. Evaluating changes in rates of death due to any cause can provide a more complete picture of the severity of the pandemic. We have analyzed state-level data from the United States since March 1, 2020 to evaluate increases in deaths due to any cause and deaths due to pneumonia, compared to data from previous years. We use a regression model that adjusts for seasonality, year-to-year variations in incidence, influenza activity, and reporting delays to obtain an estimate for how many deaths would have been reported in a typical year. In most states, the number of excess deaths exceeds the number of reported deaths; the magnitude of the discrepancy between reported deaths and excess deaths varied by state and declined over time and was partially related to variations in the intensity of testing.