Racial residential segregation impacts on low birth weight using improved neighborhood boundary definitions

Abstract

Research on local racial residential segregation and health generally utilize census tract boundaries as a proxy from within which to estimate individual exposures. Census tracts however, may not accurately reflect the neighborhood environments in which people live and interact. Census tract geography may also capture non-exchangeable populations in socially stratified cities, impacting statistical assumptions of independence. To address these concerns, this study assessed the impact of racial residential segregation (i.e., racial isolation and racial clusters) on low birth weight (i.e., intrauterine growth retardation (IUGR) and preterm birth) in the Detroit Metropolitan Area using optimized neighborhood boundary definitions. Automated zone-matching (AZM) methodology was applied to redefine neighborhood (zones). Maternal and infant health information was obtained from Michigan’s vital statistics birth registry (n=137,965) for the years 2004–2006. Multilevel models were estimated to assess the effect of high racial isolation and high racial clusters on IUGR and preterm birth, controlling for maternal race, single marital status, smoking and area-level poverty. The results from this study showed that high racial isolation had a significant effect on IUGR, while the odds of preterm birth were higher in racially clustered zones. African American mothers were at increased odds of having IUGR or preterm infants than other mothers; however, these disparities reduced in highly segregated zones. The predicted incidence of IUGR across racially isolated zones and census tracts differed indicating a modifiable area unit problem (MAUP). MAUP effects were not observed in models predicting preterm incidence in high racial clusters or IUGR or preterm incidence in high poverty areas, demonstrating the stability-reliability of these estimates. Future research should continue to optimize neighborhood boundary definitions, while assessing the sensitivity of segregation measures to changes in scale, to improve our understanding of segregation impacts on racial disparities in low birth weight.