Mortality by Race and Ethnicity

Definition

Systematic differences in age-adjusted mortality rates across racial and Hispanic-origin (RE) groups, documented by cause of death, age, and sex. In the United States, four groups are typically analyzed — non-Hispanic White (WNH), Hispanic, Black, and Asian/Pacific Islander (API) — using Centers for Disease Control and Prevention (CDC) WONDER death records and International Classification of Diseases, 10th Revision (ICD-10) cause-of-death classifications. Age-adjustment uses the 2000 U.S. standard population to control for differences in age structure across groups.

Key Ideas

All-cause ranking: Black males have the highest all-cause age-adjusted mortality rate; API females the lowest. This ordering holds for most cause-of-death categories.
Hispanic mortality paradox: Hispanics have lower all-cause mortality than WNH despite socioeconomic disadvantage. The advantage is strongest among foreign-born Hispanics and at older ages, and holds for every major cause except digestive diseases (alcoholic liver disease), endocrine/metabolic diseases (diabetes), and perinatal conditions.
Post-2010 reversal: Midlife mortality (ages 25–64) was higher in 2019 than in 2010 for every RE/sex group except Black and API women — reflecting stalled cardiovascular improvement and rising deaths of despair.
Racial universalization of deaths of despair: Poisoning mortality quintupled for WNH males and rose $6.5\times$ for WNH females 1999–2019. Among Black males it more than tripled. In 2019 Black males overtook WNH males in poisoning mortality ( $33.9$ vs. $32.8$ per $100{,}000$ ), driven by the fentanyl third wave.
Category exceptions: Respiratory disease (chronic obstructive pulmonary disease (COPD)/emphysema) at all ages is higher for WNH males than Black males — a reversal driven by cohort smoking history. Nervous system/dementia mortality is highest for WNH. Digestive disease is highest for Hispanic males in parts of the period.
Converging cancer disparities: The Black–WNH male cancer mortality gap narrowed from $94.8$ to $23.8$ per $100{,}000$ excess deaths 1999–2019.
Persistent human immunodeficiency virus (HIV) disparity: Black male HIV mortality fell $>75\%$ with antiretroviral therapy but remained $7.4\times$ WNH male rates in 2019.
American Indian/Alaska Native (AIAN) exclusion and mortality flip: American Indian/Alaska Native people are typically excluded from systematic analyses due to severe death-certificate misclassification (sensitivity $\approx 55\%$ ; Classification Ratio $= 1.30$ ). Crucially, this misclassification reverses AIAN's apparent mortality ranking: raw age-adjusted death rate (AADR) ratio vs. White $= 0.85$ (apparent advantage); corrected ratio $= 1.11$ (real disadvantage, Arias et al. 2008). AIAN thus have the worst true mortality trajectory of any RE group, but are systematically invisibilized in uncorrected data. See Death Certificate Race-Ethnicity Misclassification.

How It Works

Cause-specific age-adjusted mortality rates are computed from CDC WONDER data (universe of death certificates) combined with ICD-10 chapter-level classifications. Direct standardization using the 2000 U.S. standard population removes the effect of differing age distributions across RE groups, enabling valid cross-group comparisons. The Hispanic paradox's mechanism remains contested: leading explanations include the healthy-immigrant effect (self-selection of healthier migrants), the salmon-bias hypothesis (return migration of the ill to country of origin), and cohort differences in smoking prevalence.

Standard-choice effects on racial ratios: The black/white all-cause AADR ratio depends on which standard population is used. Under the 1940 standard, the ratio was $\approx 1.60$ ; under the 2000 standard (in effect since 1999), it falls to $\approx 1.41$ . The compression occurs because the black/white mortality ratio is largest at young ages ( $\approx 2.0$ for ages $0$ – $24$ ) and smallest at old ages ( $\approx 1.1$ for ages $65+$ ) — and the 2000 standard places greater weight on older ages where racial disparities are narrowest. Any time-series comparison of the black/white AADR ratio spanning the 1999 standard switch will show a spurious narrowing of $\approx 0.19$ that does not reflect a real convergence in age-specific mortality. See Age Standardization.

Why It Matters

RE-group mortality differentials directly affect Social Security program outcomes: retirement benefit collection periods (higher mortality → shorter receipt), Disability Insurance (DI) incidence rates, and survivor benefit timing vary systematically across groups.
The 2019 Black–WNH poisoning mortality crossover is clinically and politically underreported — it signals that the opioid/fentanyl epidemic, historically framed as a "white crisis," has fully diffused into the Black community.
AIAN exclusion from standard analyses likely causes systematic understatement of the severity of U.S. mortality disparities at the national level.
Cancer mortality convergence suggests that targeted screening and treatment investments can close large racial gaps within two decades.

Education-Stratified Life Expectancy by Race (Olshansky et al. 2012)

Cross-sectional 2008 analysis by Olshansky et al. (2012), using Multiple Cause of Death records matched to American Community Survey data, reveals that education stratifies life expectancy (LE) within race groups at least as strongly as race stratifies LE across groups.

Within-race education gradients (2008)

Group	LE gap: highest vs. lowest education
White men	$12.9$ years
White women	$10.4$ years
Black men	$9.7$ years
Black women	$6.5$ years
Hispanic men	$5.5$ years
Hispanic women	$2.9$ years

The headline cross-group disparity: $14.2$ years between white men with $16+$ years of education and black men with $<12$ years; $10.3$ years for the equivalent women. This rivals the most extreme international comparisons.

Education dominates race at the credential extremes

Highly educated blacks and Hispanics ( $16+$ years) outlive low-education whites ( $<12$ years) by $7.5$ – $13.6$ years — a complete reversal of the typical racial ranking. Education, as a proxy for access to health-protecting knowledge, income, and social capital, can dominate racial disadvantage when the educational contrast is sufficiently large.

Residual racial gap conditional on education

Even among the college-educated ( $16+$ years), black men live approximately $4.2$ fewer years than white men. Education narrows but does not eliminate the racial LE gap, consistent with weathering, discrimination-related chronic stress, and residential segregation operating through channels independent of schooling attainment. See Arline Geronimus.

Trend 1990–2008 (credential-based measure, subject to compositional bias)

Low-education white men and women experienced declining LE 1990–2008, while low-education blacks and Hispanics improved. Bound et al. (2015) subsequently showed the apparent decline for whites is largely a compositional artifact of rising attainment — the $<12$ -year group shrank from $\approx 55\%$ to $\approx 37\%$ of the population, becoming increasingly adversely selected. A real but modest stagnation for low-socioeconomic status (SES) white women survives after correction; the $3.4$ – $5.3$ year decline figures overstate the magnitude. See Income-Mortality Gradient for the full quantitative critique.

Historical benchmark

Low-education whites in 2008 had LE equivalent to all Americans in the early 1970s — a lag of $\approx 35$ years relative to the national average.

Age-Period-Cohort (APC) Analysis of Long-Run Black-White Trends, 1959–2009 (Masters et al. 2014)

Masters, Hummer, Powers, Beck, Lin, and Finch (2014) use APC intrinsic estimator models on National Center for Health Statistics (NCHS) death data 1959–2009, ages 15–74, to decompose black-white mortality trends into period and cohort components. Data source: NCHS Multiple Cause of Death Files; denominator: Census-based population estimates; causes: all-cause plus heart disease, stroke, lung cancer, breast cancer, other cancers, homicide, accidents, infectious diseases.

Cohort effects dominate period effects for chronic disease. Both black and white adult mortality changes 1959–2009 show more cohort variation than period variation for all-cause and chronic disease mortality — consistent with Yang (2008). This means cumulative life-course exposures (early-life conditions, education, healthcare access) are the primary mechanism, not contemporaneous shocks.

Pre-WWII cohorts: white cohort gains exceeded black cohort gains. The black-white relative rate ratio in adult mortality rose from ~1.25 (cohorts born before 1900) to ~1.75 (1930s/1940s male cohorts) — institutionalized racism, segregation, and legal exclusion prevented blacks from benefiting equally from social and public health advances.

Mid-1960s period shock: dramatic reductions in black mortality following the Civil Rights Act, Medicaid/Medicare (1965), Food Stamps (1964) — especially large for black women, rapidly closing the period-based black-white gap.

Post-WWII cohorts: black cohort-based reductions accelerated. Expanded access to education, wages, and medical care meant post-WWII black cohorts made faster cohort-based mortality gains than pre-WWII cohorts — narrowing the black-white cohort gap (male relative rate ratio: ~1.75 → ~1.25 across 1940s → 1960s cohorts). Driven primarily by chronic disease (heart disease, stroke, cancers).

Troubling sign: heart disease stalling for recent cohorts of BOTH races. Recent (post-1960s) cohorts of black and white men and women show cohort-based stalling or increases in heart disease mortality — likely the obesity epidemic embedding higher cardiac risk that will manifest decades later. This warning is invisible in period-based analyses.

Breast cancer: widening on both period and cohort dimensions. White women had rapid period-based breast cancer mortality reductions since 1985 (mammography, improved treatment). Black women had period-based increases through the early 1990s (obesity epidemic, differential screening access). Black women's cohort reductions are slower. A preventable cancer where racial disparities are growing on all temporal dimensions — a public health emergency.

Young adult mortality: period effects dominate. HIV/AIDS, homicide, and accidents (ages 15–34) are driven primarily by period changes. Recent period reductions in HIV and homicide have been larger for blacks than whites, narrowing the gap. White accident mortality shows recent period increases (drug overdoses) — an early signal of the Deaths of Despair pattern subsequently documented by Case and Deaton (2015).

Open Questions

Why did Black and API women avoid the post-2010 midlife mortality reversal that affected every other RE/sex group?
To what extent does the Hispanic mortality paradox reflect true health advantage vs. statistical artifacts (salmon bias, data linkage errors in CDC WONDER)? Arias et al. (2008) confirm the paradox survives death certificate (DC) misclassification correction (corrected AADR $= 83\%$ of WNH vs. $79\%$ raw), but salmon bias and cohort smoking differences remain live explanations.
How will the fentanyl third wave reshape RE-group mortality rankings over the next decade, given that Black male poisoning mortality already overtook WNH in 2019?
Can the cancer convergence be sustained or replicated for cardiovascular disease?