Explaining variation in health status across space and time: implications for racial and ethnic disparities in self-rated health.

We use the Metropolitan Community Information Center-Metro Survey-a serial cross section of adults residing in the City of Chicago, USA, conducted from 1991 through 1999-in combination with 1990 census data to simultaneously examine the extent to which self-rated health varies across Chicago neighborhoods and across time. Three-level hierarchical logit models are employed to decompose individual, spatial, and temporal variance in self-rated health. Results indicate that variation in self-rated health across neighborhoods is explained, in part, by variation in the level of neighborhood affluence. Neighborhood level poverty, however, is not a significant predictor of self-rated health. Community level affluence, moreover, accounts for a substantial proportion of the residual health deficit experienced by African-Americans when compared with Whites (after controlling for individual level SES). The effects of affluence hold when controlling for spatial autocorrelation and when considered in primarily African-American neighborhoods. Findings also indicate that individuals living in the City of Chicago became significantly healthier over the decade of the 1990s, and that this improvement in health is explained largely by the increasing education and income levels of Chicago residents.