Methods for Counting COVID-19 Deaths in US States and Territories

Abstract

Purpose: The US Centers for Disease Control and Prevention (CDC) never issued guidelines on how states should publicly report COVID-19 deaths, and as a result, states ended up settling on different methods. Critics have claimed that some methods, such as counting “deaths with COVID-19” rather than “deaths from COVID-19”, resulted in overcounts of COVID-19 deaths. While it is clear from alternate measures like excess mortality that all states undercounted COVID-19 deaths, we investigated whether different methods had any effect on state death reporting. Methods & Materials: We identified the methods states use to report COVID-19 deaths on their dashboards, then compared state-reported data collected by CDC in its COVID Data Tracker against data from the CDC's National Center for Health Statistics (NCHS)’s national review of death certificates measuring deaths from COVID-19. Comparing state data against this standardized national dataset allowed us to see if state definitions affected their reporting. Results: Contrary to critics’ claims, few states have definitions that could systematically include individuals who died of causes unrelated to COVID-19. State death definitions instead vary along two lines: whether positive PCR tests were required to count COVID-19 deaths—with states requiring tests having cumulative totals 9.03% lower than NCHS data, while states not requiring them had totals 10.21% higher (p=.06)—and whether states exclusively used death certificates to measure COVID-19 deaths, with these states’ totals coming in 27.63% higher than NCHS data, compared to 3.26% higher for states using additional information (p=.11). Conclusion: Rather than certain methods leading to systematic overcounts of COVID-19 deaths, it appears some methods could be especially prone to undercounts. Only states using death certificate reviews had counts substantially higher than NCHS data, and since NCHS also uses death certificates, the difference appears to be a matter of the faster speed of state review. Meanwhile, many states using different methods have discovered they missed large quantities of COVID-19 deaths in their counts. The CDC should have provided leadership to states on how they should report COVID-19 deaths to avoid any preventable undercounts, and public communication to quell misinformation about overcounts.