Racial disparities in mortality among middle aged adults in the United States: Is health insurance the solution?

Objectives: Disparities in life expectancy between white and non-white populations in the US are marked. While infant mortality contributes to this difference, the majority of excess deaths occur in middle age. Our research investigates whether health insurance ameliorates racial disparities in mortality among adults aged 45 - 64.

Methods: Data were drawn from the 2004 National Health Interview Survey (NHIS) Linked Mortality File linking respondent information from the nationally representative NHIS to death records through 2002. Our analyses examined 217,748 respondents aged 45 and 64 during interview years 1986, 1989-2000. We applied Cox proportional hazard models to estimate mortality hazard ratios (HR) by race/ethnicity (black, white, Hispanic, other) and insurance status (privately insured, publicly insured, uninsured) at time of interview. Multivariable analyses adjusted for demographic, resource and health status characteristics. All analyses were stratified by sex and weighted to account for sample design.

Results: Overall, 77.1% of respondents reported private insurance, 11.1% public insurance, and 11.8%, uninsured. Minorities were more likely to be uninsured (18.2% black and 27.2% Hispanic, versus 9.5% of white respondents). In analyses holding age and interview year constant, black adults had an elevated risk of death (Female: HR 1.62, CI 1.52-1.72. Male: HR 1.57, CI 1.48-1.66) compared to white adults. Hispanics, regardless of sex, did not differ in risk from whites, whereas men from other racial origins had lower risks of death than white men (HR: 0.83, CI 0.73-0.95). Adjustment for insurance alone reduced the hazard ratio for blacks, while not fully eliminating disparities (Female: HR 1.31, CI 1.23- 1.40. Male: HR 1.32, CI 1.25-1.41), which remained after further adjustment for poverty, region, rurality, and health status (Female: HR 1.11, CI 1.04-1.19; Male HR 1.08 CI 1.02-1.16). Adjustment for insurance status resulted in significantly lower risk of death for Hispanic women compared to white women. Further adjustment for socio-economic status, region, rurality, and health status suggested significantly lower risk of death for Hispanics of either sex compared to white adults. Of note, uninsured individuals were at higher risk of death in all analyses.

Conclusions: Health insurance ameliorates some, but not all, of racial disparity in mortality risk. Health insurance reform expanding coverage may be insufficient to eliminate racial disparities in mortality. Research is needed into contextual factors which may influence health and survival. Research into the lowered risk of death experienced by Hispanic adults may provide insight for other minority populations.