Using the Index of Concentration at the Extremes at multiple geographical levels to monitor health inequities in an era of growing spatial social polarization: Massachusetts, USA (2010-14).

BACKGROUND: Metrics that quantify economic and social spatial polarization at multiple geographical levels are not routinely used by health agencies, despite rising inequalities.
METHODS: We employed the Index of Concentration at the Extremes (ICE), which quantifies how persons in a specified area are concentrated into the top vs bottom of a specified societal distribution, to examine associations with Massachusetts mortality data (2010-14). Our a priori hypotheses were that these associations would: be greater at the local [census tract (CT)] compared with city/town level; vary by race/ethnicity but not gender; and be greatest for our new ICE for racialized economic segregation. Mortality outcomes comprised: child (< 5 years); premature (< 65 years); and cause-specific (cancer; cardiovascular; diabetes; suicide; HIV/AIDS; accidental poisoning; smoking-attributable).
RESULTS: As illustrated by child mortality, in multilevel models jointly including CT and city/town metrics, the rate ratio comparing the worst to best-off ICE quintile for the total population ranged from 2.2 [95% confidence interval (CI) 1.6, 3.0] for the CT-level ICE for racialized economic segregation down to 1.1 (95% CI 0.8, 1.7) for the city/town-level ICE for racial segregation; similar patterns occurred by gender and for the non-Hispanic White population. Larger associations for the ICE for racialized economic segregation were at the CT-level for the Black non-Hispanic population (6.9; 95% CI 1.3, 36.9) and at the city/town level for the Hispanic population (6.4; 95% CI 1.2, 35.4).
CONCLUSIONS: Results indicate that health agencies should employ measures of spatial social polarization at multiple levels to monitor health inequities.