The 131st Annual Meeting (November 15-19, 2003) of APHA

Abstract #73191

Jay S. Kaufman, PhD, Department of Epidemiology, UNC School of Public Health, 2104C McGavran-Greenberg Hall, Pittsboro Road, CB#7400, Chapel Hill, NC 275997400, Barbara A. Laraia, PhD, MPH, RD, Department of Nutrition, The University of North Carolina at Chapel Hill, Carolina Population Center, CB #8120, Chapel Hill, NC 27599-8120, (919)966-5969, blaraia@email.unc.edu, Paul A. Buescher, PhD, State Center for Health Statistics - North Carolina, 1908 Mail Service Center, Raleigh, NC 27699-1908, and Lynne C. Messer, MPH, Department of Epidemiology, The University of North Carolina at Chapel Hill, CB # 7435, School of Public Health, Chapel Hill, NC 27599.

Objective: To examine the association between neighborhood-level characteristics and preterm delivery (< 37 weeks estimated gestational age at birth) using random-effects binary-event regression models. Methods: We use three consecutive years (1999-2001) of birth certificate records for Wake County, North Carolina, linked to community characteristics at the census block group level. Race-stratified models (non-hispanic whites and blacks) are fit with adjustment for individual maternal characteristics and a random coefficient specified for the block group intercept. Block group level characterisitcs include poverty, as represented by the proportion of individuals in the block group falling below the federal poverty line. We fit random-effects generalized linear models with binomial distribution in order to estimate exposure effects associated with block-group-level characteristics while accounting for unmeasured between-groups heterogeneity in baseline risk and individual-level maternal characteristics. Results: There were 17,796 non-hispanic white births and 5426 non-hispanic black births recorded in the 3-year period, within 262 block groups in Wake County. The crude preterm delivery proportions were 9.6% and 14.4% in white and black women, respectively. Level of aggregation of risk was modest, especially so among white women. Complete modeling results will be presented and discussed, including estimation of individual and community-level effects on preterm delivery, and conditional variances for the distribution of random effects, representing unexplained between-block-group heterogeneity in baseline risk after accounting for compositional effects of measured maternal characteristics.

Learning Objectives:

• To examine the association between neighborhood-level characteristics and preterm delivery (< 37 weeks estimated gestational age at birth) using random-effects binary-event regression models

Presenting author's disclosure statement:
Disclosure not received
Relationship: Not Received.