Probability Samples or Clinic Populations to Study Pregnancy and Children's Health? Contrasting Approaches of Demography and Epidemiology

David A. Savitz, University of North Carolina at Chapel Hill
Nancy Dole, University of North Carolina at Chapel Hill
Diane Kaczor, University of North Carolina at Chapel Hill
Amy H. Herring, University of North Carolina at Chapel Hill
Anna Maria Siega-Riz, University of North Carolina at Chapel Hill
Jay S. Kaufman, University of North Carolina at Chapel Hill
John M. Thorp, Jr., University of North Carolina at Chapel Hill

Research on pregnancy and children's health has two distinct research traditions, the study of geographically defined populations and clinical populations, such as prenatal care sites or hospitals, each bringing distinctive strengths and weaknesses. To make more informed choices for programs such as the National Children's' Study, we examined influences on pregnancy outcome in a clinic population enrolled for the Pregnancy, Infection, and Nutrition (PIN) Study (N=2,289) compared to births in counties served by participating clinics (N=59,979) in 1996-2000. Pregnancy outcomes were preterm birth (<37 weeks' completed gestation) and term low birth weight (<2500 grams and gestation 37+ weeks). The demographic profile of women differed and in some cases, the relations between predictors and outcomes differed, particularly Black/White comparisons for preterm birth (relative risk = 1.8 for area women, 1.0 for PIN women). Given compensating strengths and weaknesses, clear priorities are needed to match the optimal strategy to specific research goals.