Multiple risk factors do not identify cocaine use in rural obstetrical patients

This nonconcurrent, cohort study of consecutive admissions to one of three hospital units: labor and delivery (n = 474), well-born nursery (n = 100), and the neonatal intensive care unit (n = 100), was designed to determine the prevalence of cocaine exposure in a rural obstetrical sample and to determine the relationship between exposure and perinatal variables. Urines were analyzed for benzoylecgonine, and the Obstetrical Complications Scale was completed for each mother-infant pair. Elementary comparisons were made using chi 2 analyses and Student's t test. Stepwise discriminant and discriminant function analyses were performed. The prevalence of exposure in the three groups of subjects ranged from 5%-7%. No significant differences in perinatal variables were found between users and nonusers in either of the newborn samples. In the maternal sample the groups differed on twelve mother or infant factors. However, no single variable or set of variables predicted use versus nonuse in any of the groups.