Air Pollution and Mortality at the Intersection of Race and Social Class

Abstract

Background: Black Americans are exposed to higher annual levels of air pollution containing fine particulate matter (particles with an aerodynamic diameter of ≤2.5 μm [PM_2.5]) than White Americans and may be more susceptible to its health effects. Low-income Americans may also be more susceptible to PM_2.5 pollution than high-income Americans. Because information is lacking on exposure-response curves for PM_2.5 exposure and mortality among marginalized subpopulations categorized according to both race and socioeconomic position, the Environmental Protection Agency lacks important evidence to inform its regulatory rulemaking for PM_2.5 standards.

Methods: We analyzed 623 million person-years of Medicare data from 73 million persons 65 years of age or older from 2000 through 2016 to estimate associations between annual PM_2.5 exposure and mortality in subpopulations defined simultaneously by racial identity (Black vs. White) and income level (Medicaid eligible vs. ineligible).

Results: Lower PM_2.5 exposure was associated with lower mortality in the full population, but marginalized subpopulations appeared to benefit more as PM_2.5 levels decreased. For example, the hazard ratio associated with decreasing PM_2.5 from 12 μg per cubic meter to 8 μg per cubic meter for the White higher-income subpopulation was 0.963 (95% confidence interval [CI], 0.955 to 0.970), whereas equivalent hazard ratios for marginalized subpopulations were lower: 0.931 (95% CI, 0.909 to 0.953) for the Black higher-income subpopulation, 0.940 (95% CI, 0.931 to 0.948) for the White low-income subpopulation, and 0.939 (95% CI, 0.921 to 0.957) for the Black low-income subpopulation.

Conclusions: Higher-income Black persons, low-income White persons, and low-income Black persons may benefit more from lower PM_2.5 levels than higher-income White persons. These findings underscore the importance of considering racial identity and income together when assessing health inequities. (Funded by the National Institutes of Health and the Alfred P. Sloan Foundation.).

Figure 1.. PM_2.5 Exposure during the Study Period.

Shown is the annual average exposure to fine particulate matter (particles with an aerodynamic diameter of ≤2.5 μm per cubic meter [PM_2.5]) for Black higher-income persons (blue solid line), Black low-income persons (blue dashed line), White higher-income persons (orange solid line), and White low-income persons (orange dashed line) from 2000 through 2016. The orange dashed and solid lines overlap substantially, which indicates that the White higher-income and low-income persons were exposed to similar levels of PM_2.5 pollution. The inset shows the same data on an enlarged y axis.

Figure 2.. Exposure–Response Curve for PM_2.5 Exposure and Mortality among the Full Medicare Population.

Shown are the point estimate (solid line) and 95% confidence interval (gray shaded area) of the hazard ratio for death corresponding to decreases in annual average PM_2.5 exposure (to 6 to 11 μg per cubic meter) with respect to 12 μg per cubic meter on average for the full population. Estimates below 6 μg per cubic meter are not shown in order to focus attention on plausible ranges for PM_2.5 pollution policy. Confidence intervals were not adjusted for multiplicity; therefore, they should not be used in place of hypothesis testing.

Figure 3.. Exposure–Response Curves for PM_2.5 Exposure and Mortality among Marginalized Subpopulations.

Shown are point estimates (solid lines) and 95% confidence intervals (gray shaded areas) of the hazard ratio for death corresponding to decreases in annual average PM_2.5 exposure (to 6 to 11 μg per cubic meter) with respect to 12 μg per cubic meter on average for subpopulations defined in selected ways. Estimates below 6 μg per cubic meter are not shown in order to focus attention on plausible ranges for PM_2.5 pollution policy. In all panels, curves for Black persons are blue and White persons are orange. Panel A defines persons according to racial identity only without regard to income. Panel B includes only higher-income persons. Panel C includes only low-income persons. Low income was defined as dual eligibility for both Medicare and Medicaid. Confidence intervals were not adjusted for multiplicity; therefore, they should not be used in place of hypothesis testing.

Figure 4.. Differences in Mortality with Decreasing PM_2.5 Exposure among Marginalized Subpopulations.

Shown are point estimates and 95% confidence intervals of the hazard ratio for death comparing different levels of annual average PM_2.5 exposure (12 μg per cubic meter vs. 11, 10, 9, or 8 μg per cubic meter) on average for subpopulations defined in selected ways. Low income was defined as dual eligibility for both Medicare and Medicaid. Confidence intervals were not adjusted for multiplicity; therefore, they should not be used in place of hypothesis testing.