Relationship of Socioeconomic and Racial Factors, Both Individual and Community-Level, to Infant Birth Weight

Background:  Large, urban studies have documented a neighborhood effect for population concentration of minority race (typically black) residents upon birth outcomes.   But rarely are individual-level effects and socioeconomic status (SES), primary confounders, considered in these studies.   The goal of this study was to examine the relative association of maternal race and SES, at the individual-level and at the neighborhood-level, upon infant birthweight within a mid-size community whose maternal demographic profile was comparable to the nation as a whole.

Methods:  The study was a cross-sectional design utilizing two secondary datasets of Kalamazoo County, Michigan: (1)Y2010 birth certificate records(N=2,861)and (2)Y2010 census tracts(N=57).  Infant birthweight (measured in grams) was the outcome measure.  Predictors at both levels were dichotomized race and socioeconomic status.  Individual-level birth records were geo-coded, then linked to census tract through an ArcGIS10.0 spatial join.  Birthweight was modeled using Bayesian regression with spatial random CT effect.  A Proper CAR prior was specified for the spatial random effect.Bayesian credible intervals for the regression coefficients were used to assess this effect.

Results:  Across the county, mean birthweight was 3,330 grams, with 6.6% of births falling into the low-birthweight category (<2,500 grams). LBW prevalence ranged from 0 to 18%, with a mean of 6.7% across the census tracts.   Within the maternal-birth sample, 19.2% were black and 47.7% were low-income.  Across the county, 33.3% of census tracts met criteria for concentrated poverty and 31.6% met study criteria for concentrated black residents.

Individual-level factors consistently outweighed community-level factors predicting birthweight.  Race and SES were each significant predictors of lower birthweight:  being black was associated with a 1.7 increase in odds of having an LBW infant, and, among the over-LBW births, with an 82 gram decrease in birthweight (CI -104.7,-54.4grams); being poor was associated with a 1.7 increase in odds of having an LBW infant, and, among the over-LBW births, with an 115 gram decrease in birthweight (CI -144.2,-76.6grams).  Living with concentrated black residents or concentrated poverty did not add predictive power.

Conclusions: 

Maternal race and SES equally and independently predict birthweight, regardless of neighborhood factors regarding concentrated black residents or concentrated poverty.