Publication Type: Poster
Abstract: Introduction. Few residential racial residential segregation-health studies are multilevel, and examine the effect of metropolitan-area (MA) level segregation across the U.S.
Methods. We merged the 2000 national Natality dataset (n=1,699,202) with MA-level measures of hypersegregation derived from Census 2000 calculations of 5 measures of Black-White segregation (Dissimilarity, Isolation, Centralization, Spatial Proximity, Concentration), tapping 5 dimensions. Using threshold definitions from Massey & Denton, we identified hypersegregated MAs as highly segregated across all 5 dimensions. We restricted analyses to non-Hispanic (NH) white and NH black women aged 20-45, singleton births > 500 g in 237 large MAs. We executed a two-level hierarchical linear regression using HLM 6.0 to examine how hypersegregation was associated with individual birth weight, including racial slopes as outcomes, after controlling for region, MA confounders (median income, population size, altitude, % below poverty), and individual-level demographic, socioeconomic, health behavior and medical risk factors.
Results. Infants of black mothers were 250 grams lighter than their NH White counterparts. NH black infants in hypersegregated MAs were born 30 grams lighter on average (p=.037) than their NH black counterparts in non-hypersegregated MAs. Moreover, we found that older black women in hypersegregated MAs exhibited significantly worse birth outcomes with each increasing year of age (a 3-way segregation, race, age interaction).
Conclusion: Our analyses implicate the importance of situating black-white birth disparities within a metropolitan context, which may pattern opportunities and barriers for mothers within and between neighborhoods, and may shed light on causes of racial disparities in birth outcomes, which are largely unknown.