Using GIS to Examine the Relationship of Maternal Race and Neighborhood Racial Segregation with Birth Outcomes

Background: Racial disparities in birth outcomes have been shown to vary widely, and are thought to be closely linked to a community's racial segregation. This study utilized GIS to define racially segregated neighborhoods and examine the association of race and racial segregation with birth outcomes at the individual maternal level as well as the neighborhood,tract level. Methods: This was a secondary data analysis of Kalamazoo County 2008 birth certificate data. Birth records were geo-coded and spatially linked to census tracts. Pearson Chi Square and correlation tests were conducted for maternal race (Black/White) and racial segregation (below/above the county mean Black residents). Regression analyses was then completed for each of the four birth outcome variables (infant death, VLBW, LBW and prematurity). Results: Kalamazoo County census tracts varied widely in racial composition, from 0.1% to 85.4% Black residents, with twenty meeting the racial segregation criteria. Racially segregated tracts had statistically significantly higher vacant housing, lower home-owner-occupied housing, higher poverty and lower high-school-graduation rates. Births in these neighborhoods, regardless of maternal race, were less likely to have early or adequate prenatal care. Regression models identified the following factors associated with poor birth outcomes: HS graduation at the tract level, and maternal race, prenatal smoking, STI and inadequate maternal weight gain at the individual level. Conclusions: Being a Black individual was strongly and independently related to poor birth outcomes; while at the tract level, results suggested an indirect relationship mediated by neighborhood impoverishment, educational attainment and prenatal healthcare resources.