Analysis: Hispanic, Black, and Native Americans have carried the burden of COVID-19 pandemic

Reviewed by Emily Henderson, B.Sc.Apr 16 2021

A UCLA Fielding School of Public Health-led team has found that Hispanic, Black, and Native Americans have carried the burden of the pandemic, both in overall mortality and specifically in years of potential life lost, in an analysis of 45 U.S. states and the District of Columbia (D.C.).

"COVID-19 certainly didn't cause these racial/ethnic disparities in health outcomes, but it did highlight and bring unprecedented national attention to long-standing societal and health inequalities that many communities of color in the U.S. face," said Dr. Ron Brookmeyer, dean of the Fielding School and distinguished professor of biostatistics. "It is imperative that we rise to the challenge of addressing the health needs of communities of color, both during the U.S. COVID-19 epidemic and long after its conclusion."

The findings on years of potential life lost (YPLL) mirror how the pandemic has hit different communities across the United States, the researchers found.

By the end of 2020, the U.S. had more than 20 million confirmed cases of COVID-19 and more than 350,000 deaths. COVID-19 doesn't affect all segments of the population equally, however, with males and older individuals at higher risk of post-infection mortality, for example."

Dr. Christina Ramirez, Fielding School Professor of Biostatistics

The report, published in a special issue of the peer-reviewed International Journal of Environmental Research and Public Health, supports federal findings that COVID-19 mortality rates are 1.9 to 2.4 times higher for Black, Latino, and American Indian or Alaska Native Americans than non-Hispanic Whites.

"Data on COVID-19 outcomes have revealed substantial racial/ethnic disparities, with Blacks, Hispanics, and American Indian or Alaska Natives experiencing disproportionately high numbers of cases, hospitalizations, and deaths," said co-author Dr. Jarvis T. Chen, with the Harvard T.H. Chan School of Public Health.

While the overwhelming majority of COVID-19 deaths occur among individuals in older age groups – approximately two-thirds of COVID-19 deaths occur among individuals over 65 – younger individuals are also susceptible to death from COVID-19. These deaths represent greater unrealized years of life, economic productivity, and broader contributions to society compared to decedents of greater age. An alternative epidemiological measure of mortality burden, explicitly weighting deaths that occur at earlier ages more heavily than deaths that occur at later ages, is years of potential life lost.

According to public records, the total number of COVID-19 deaths in the U.S. as of December 30, 2020 was 342,577. Data from the National Center for Health Statistics (NCHS), from which information on decedent ages are recorded, as of December 30, 2020 were available for 88% (301,679) of those deaths, corresponding to over 926,058 years of potential life lost (YPLL) before the age of 75. Proportionally scaling up this lower bound of total YPLL to the true number of COVID-19 deaths as of December 30, 2020 results in an estimated lower bound of total YPLL in excess of 1 million.

"YPLL is an innovative epidemiological measure of mortality burden that has particular relevance to COVID-19, simultaneously capturing the effects of elevated mortality rates and extreme impacts at younger ages," said Dr. Tom Belin, Fielding School professor of biostatistics and David Geffen School of Medicine professor of psychiatry and biobehavioral sciences. "The estimated racial/ethnic disparities are generally more pronounced when measuring mortality in terms of YPLL, compared to death counts, because of the greater intensity of the disparities at younger ages."

As an example, the UCLA-Harvard research team found that in terms of YPLL, Hispanic Americans accounted for an estimated 30.3-31.0% of total YPLL, despite only totaling 18.5% of the U.S. population and an approximately commensurate 18.6-19.1% of total COVID-19 deaths.

American Indian or Alaska Native (AIAN)

AIANs comprise 0.84% of the U.S. population but 1.1-1.6% of total COVID-19 deaths and an even higher estimated 2.0-2.6% of total YPLL.

Hispanic

Hispanics comprise 18.5% of the U.S. population and an approximately commensurate 18.6-19.1% of total COVID-19 deaths but an estimated 30.3-31.0% of total YPLL.

Black

Blacks comprise 13.2% of the U.S. population and 16.9-17.4% of total COVID-19 deaths and an even higher estimated 23.0-23.6% of total YPLL.

Asian

Asians comprise 6.3% of the U.S. population but only 3.6-4.0% of total COVID-19 deaths and an estimated 3.8-4.4% of total YPLL.

White

Whites comprise 61.2% of the U.S. population but only 58.9-59.4% of total COVID-19 deaths and an even lower estimated 40.1-40.8% of total YPLL.

"There is substantial state-to-state variability in the magnitudes of the racial/ethnic disparities, which suggest that they are driven largely by social determinants of health, whose degree of association with race/ethnicity varies by state," said Jay Xu, lead author of the study and PhD candidate in biostatistics at the Fielding School. "Despite this current gap in understanding, however, the sheer magnitude of the state-level disparities speaks to the degree of urgency for communities of color in relation to the U.S. COVID-19 epidemic."

Selected state-level findings:

American Indian and Alaska Native (AIAN)

In Mississippi, the age-adjusted YPLL rate for AIANs is estimated to be 18.0 to 34.1 times higher than that of Whites. In Arizona, the age-adjusted YPLL rate for AIANs is estimated to be 14.8 to 23.7 times higher than that of Whites. In Montana, the age-adjusted YPLL rate for AIANs is estimated to be 11.1 to 22.5 times higher than that of Whites. In New Mexico, the age-adjusted YPLL rate for AIAN's is estimated to be 11.8 to 31.6 times higher than that of Whites.

Hispanics: In D.C., the age-adjusted YPLL rate for Hispanics is estimated to be 8.4 to 24.3 times higher than that of Whites. In Maryland, the age-adjusted YPLL rate for Hispanics is estimated to be 7.2 to 9.9 times higher than that of Whites.

Blacks: In Michigan, the age-adjusted YPLL rate for Blacks is estimated to be 5.8 to 6.7 times higher than that of Whites.

Asians: In Minnesota, the age-adjusted YPLL rate for non-Hispanic Asians is estimated to be 3.8 to 6.0 times higher than that of non-Hispanic Whites.

The researchers said there are potential strategies to address some of these disparities, including the immediate goal of increasing vaccination distribution in communities that are otherwise falling behind.

"For example, offering COVID-19 vaccinations in nonmedical settings may increase vaccine uptake; partnering with faith-based organizations could also help," said Dr. Marc Suchard, Fielding School Professor of Biostatistics and David Geffen School of Medicine Professor of Human Genetics and Computational Medicine. "Both vaccines and non-pharmaceutical interventions should be pursued to prevent further devastation from this epidemic, and to confront future public health crises."

Methods: Researchers examined U.S. national COVID-19 mortality data from the National Center for Health Statistics (NCHS) as of December 30, 2020-;the last data update for calendar year 2020-;summarized as cumulative death counts within age groups stratified by state (as well as D.C. and Puerto Rico) and race/ethnicity.

The following 5 racial/ethnic groups were considered: non-Hispanic White (NH White), non-Hispanic Black (NH Black), Hispanic, non-Hispanic Asian (NH Asian), and non-Hispanic American Indian or Alaska Native (NH AIAN). The set of mutually exclusive, collectively exhaustive, and chronologically ordered age groups used are <1, 1–4, 5–14, 15–24, 25–34, 35–44, 45–54, 55–64, 65–74, 75–84 and 85+. Death counts within individual age groups between 1 and 9 were suppressed in the NCHS data. The team characterized racial/ethnic disparities in COVID-19-attributable YPLL through the estimation of percentages of total YPLL by race/ethnicity, contrasting them with their respective percent population shares, as well as age-adjusted YPLL Rate Ratios, anchoring comparisons to NH Whites, using novel Monte Carlo simulation techniques to account for the administrative interval censoring of ages at death, suppression of low death counts, and unknown race/ethnicity for a subset of deaths in the NCHS data that complicate the analysis.