Trends in Mortality Disparities by Area-Based Poverty in New York City, 1990-2010

Abstract

Residing in a high-poverty area has consistently been associated with higher mortality rates, but the association between poverty and mortality can change over time. We examine the association between neighborhood poverty and mortality in New York City (NYC) during 1990-2010 to document mortality disparity changes over time and determine causes of death for which disparities are greatest. We used NYC and New York state mortality data for years 1990, 2000, and 2010 to calculate all-cause and cause-specific age-adjusted death rates (AADRs) by census tract poverty (CTP), which is the proportion of persons in a census tract living below the federal poverty threshold. We calculated mortality disparities, measured as the difference in AADR between the lowest and highest CTP groups, within and across race/ethnicity, nativity, and sex categories by year. We observed higher all-cause AADRs with higher CTP for each year for all race/ethnicities, both sexes, and US-born persons. Mortality disparities decreased progressively during 1990-2010 for the NYC population overall, for each race/ethnic group, and for the majority of causes of death. The overall mortality disparity between the highest and lowest CTP groups during 2010 was 2.55 deaths/1000 population. The largest contributors to mortality disparities were heart disease (51.52 deaths/100,000 population), human immunodeficiency virus (19.96/100,000 population), and diabetes (19.59/100,000 population). We show that progress was made in narrowing socioeconomic disparities in mortality during 1990-2010, but substantial disparities remain. Future efforts toward achieving health equity in NYC mortality should focus on areas contributing most to disparities.

Keywords: Ethnic groups; Geography; Mortality; Poverty; Socioeconomic factors.

FIG. 1

a Age-adjusted all-cause mortality rate per 1000 population by census tract poverty^a and race/ethnicity, New York City, 1990. Trend from lowest to highest poverty level is statistically significant for each race/ethnic group (p < 0.05, chi-square for trend). b Age-adjusted all-cause mortality rate per 1000 population by census tract poverty^b and race/ethnicity, New York City, 2000. Trend from lowest to highest poverty level is statistically significant for each race/ethnic group (p < 0.05, chi-square for trend). c Age-adjusted all-cause mortality rate per 1000 population by census tract poverty^c and race/ethnicity, New York City, 2010. Trend from lowest to highest poverty level is statistically significant for each race/ethnic group (p < 0.05, chi-square for trend).

FIG. 2

Difference in cause-specific age-adjusted death rates (mortality disparity) between census tracts with the lowest (<10 %) and highest (≥30 %) proportion of the population living below the poverty threshold for selected leading causes of death, New York City, 1990, 2000, and 2010.