You are here:
ESTIMATING THE INFLUENCE OF INDIVIDUAL POVERTY-ADJUSTED EDUCATIONAL ATTAINMENT ON TERM BIRTH WEIGHT USING CONDITIONAL MODELS
Citation:
MESSER, L. C., J. S. KAUFMAN, AND N. DOLE. ESTIMATING THE INFLUENCE OF INDIVIDUAL POVERTY-ADJUSTED EDUCATIONAL ATTAINMENT ON TERM BIRTH WEIGHT USING CONDITIONAL MODELS. Presented at Society for Epidemiologic Research Annual Meeting, Seattle, WA, June 22 - 24, 2006.
Description:
Reported maternal education is an important predictor of pregnancy outcomes. Like income, it is believed to allow women to locate in more favorable conditions than less educated or affluent peers. We examine the effect of reported educational attainment on term birth weight (birth weight for women delivering between 37-42 weeks gestation) for white and black non-Hispanic (NH) women living in Wake County NC (1995-2005) using conditional (fixed effects) models. Geocoded Pregnancy, Infection and Nutrition (PIN) cohort data (n=1088) were analyzed. Maternal education categories (> 12, 12, < 12 years) differed by race. Approximately 36 percent of black NH women (n=612) had > 12 years of education compared with more than 43 percent of white NH women (N=436). At equal levels of education, black and white women differed in poverty level. At > 12 years of education, mean poverty for white women was 292 percent, compared with 165 percent for black women. Race-stratified conditional linear models with tract fixed effects were fit. Conditioning on neighborhood changed black NH education estimates by 31.1 percent; white education estimates were unchanged. Reporting < 12 years of education was associated with lower term
birth weight, compared with > 12 years, for white NH [~b~=-69.0 (95% Confidence Interval [CI]: -201.1, 63.1)] and for black NH women [~b~=-110.3 (-217.3, -3.3)]. Adjusting for poverty did not change the education effects meaningfully, but education effects were attenuated following adjustment for maternal age (white NH women [~b~=-16.2 (95% CI: -139.3, 181.8)], black NH women [~b~=85.3 (95% CI: -240.0, 69.7)]. Neighborhood fixed effects models are an especially useful analytic approach because they condition on all measured and unmeasured community attributes shared by area residents that may influence birth weight. Educational attainment differentiates term birth weight for NH black women, even after adjusting for poverty and the tract environment.