Tree-based risk factor analysis of preterm delivery and small-for-gestational-age birth.

Using data collected at the Yale-New Haven Hospital, New Haven, Connecticut, in 1980-1982, the authors conducted a tree-based statistical analysis using preterm delivery and small for gestational age as outcomes and the following variables as putative risk factors: maternal age, marital status, ethnicity, education, current employment, smoking, alcohol use, caffeine consumption, marijuana use, hormones/diethylstilbestrol used by the mother, gravidity, parity, and passive smoking. The authors' analyses indicate that ethnicity is a leading factor contributing to both outcomes: Black women are more likely to have preterm deliveries as well as to deliver small-for-gestational age infants. The tree-based procedure leads to generally consistent results with more traditional secondary analyses of the same data set, but also provides a more fully integrated picture of the important relation. The full utility of these procedures for hypothesis generation remains to be explored.