The COVID Excess Mortality Percentage and Racial/Ethnic Disparities in COVID Mortality: Evidence from Indiana and Wisconsin

Importance: COVID-19 mortality rates increase with age, are higher among men than women, and vary across racial/ethnic groups, but this is also true for other natural causes of death. We develop a new measure of COVID-19 mortality burden, the COVID Excess Mortality Percentage (CEMP), defined as COVID-19 deaths as a fraction of all deaths from natural causes other than COVID-19. This measure can control for the effects of underlying population characteristics, including general population health, age, gender, race/ethnicity, and zip-code-level socioeconomic status (zip-SES) in predicting the COVID-19 mortality burden.Objective: We use CEMP to study how COVID-19 mortality varies by age, gender, race/ethnicity, and zip-SES and between the pre-vaccine and vaccine-available periods. Design. Retrospective analysis of all deaths from natural causes.Setting: Indiana and Wisconsin.Participants: All adult decedents from natural causes over the pandemic period from April 2020-March 2022.Exposure: Demographic factors and vaccine availability. Main Outcome and Measures: We report CEMP within sub-populations defined by age, gender, and race/ethnicity during the pre-vaccine (April 2020-March 2021) and vaccine-available (April 2021-March 2022) periods, and odds ratios from multivariable logistic regression.Results: CEMP is broadly similar for men and women and rises gradually with age during the pre-vaccine period, but peaks at age 40-49 during the vaccine-available period. Racial/ethnic disparities can be very high, especially for Hispanics in the pre-vaccine period, with CEMP ratios for Hispanics to non-Hispanic Whites as high as 9:1 for men aged 50-59, and higher for men than for women. CEMP disparities were smaller but substantial for other minorities and declined with age after 60+. Differences in zip-SES and education explain only a small part of these disparities. National results for 2020 are consistent with our Indiana-Wisconsin findings.Conclusions and Relevance: We studied COVID-19 mortality using a new measure that controls for non-COVID natural mortality rates. This approach is important in understanding racial/ethnic disparities in COVID-19 mortality. Disparities have been observed before, but not the very high Hispanic/White ratios we find for younger and middle-aged persons, especially men. Explanations for these disparities must account for age, gender, and time variation.The Online Appendix is available at: http://ssrn.com/abstract=4244054