The COVID-19 Pandemic, Years of Life Lost, and Life Expectancy: Decomposition Using Individual-Level Mortality Data (WP-22-03)
Paula Natalia Barreto Parra, Vladimir Atanasov, Jeffrey Whittle, John Meurer, Qian (Eric) Luo, Ruohao Zhang, and Bernard Black
Background
It is essential to understand the mortality impact of the COVID pandemic. Published estimates of COVID’s impact on life-expectancy use period-based methods, which assume the mortality rates observed during 2020 will persist indefinitely, account for limited demographic characteristics, and can be updated only on an annual basis. Policymakers need analyses of the impact of COVID on life expectancy that vary with age, gender, race/ethnicity, and socio-economic status, and over time during the pandemic period, and use the more realistic cohort-based approach.
Methods
The researchers used individual-level death-certificate data from three Midwest states (Illinois, Wisconsin, Indiana) and CDC state-level COVID mortality data, plus data on population, life expectancy, and socio-economic status (SES) from the American Community Survey (ACS) and Medicare fee-for-service. They estimated the population fatality rate (PFR) and life expectancy without COVID by age, gender, race/ethnicity and SES quintile, then calculated years of life lost (YLL) and life-expectancy loss (LEL) for various populations, from the pandemic onset through September 30, 2021; and compared cohort-based to period-based estimates.
Findings
For the three Midwest areas, COVID PFR was 0.19%, with mean YLL per COVID decedent of 12.8 years. Population LEL was 0.025 years or 9 days for the Midwest areas and 11 days for the U.S., compared to period-based estimates of over one year. However, LEL exceeded 3 months for persons aged 75+ and was substantially higher for Blacks and Hispanics than for Whites at all ages. Mean age at death was around 80 during 2020, but fell sharply during 2021 to less than 69 overall (57 for Hispanic men).
Interpretation
The researchers’ LEL estimates are a small fraction of period-based estimates and allow decomposition of LEL by age and SES and over time during the pandemic. They confirm the disproportionate impact on Black and Hispanic populations. While the pandemic’s initial impact was concentrated in the elderly, younger individuals contribute an increasing fraction of COVID-related YLL.