For this study, we recruited women admitted to our labor and delivery service, enrolling all consenting patients with a history of prenatal cocaine use and the next admission with no recorded use. During the immediate postpartum period, experienced researchers conducted private, structured interviews to obtain details of prenatal cocaine use and to identify a priori exclusion criteria (other illicit drug use, high alcohol use and chronic illnesses and medications). Specific protocols were used to collect amniotic fluid, cord blood, infant urine, meconium and maternal hair. All specimens were analyzed blind with respect to exposure using gas chromatography/mass spectrometry. Of 115 subjects, 46 had one or more biologic specimens positive for cocaine metabolites and five admitted prenatal use, but had negative specimens. Of these 51 identified as users by any method, 38 admitted, 32 were positive for urine, 28 for hair and 25 for meconium. Of the 38 admitters, 87% had positive specimens; of the 77 denying use, 17% were positive. Urine was most frequently positive in identified users, 67% overall and 62% of users who denied. Hair was next, positive in 65% of all users and 50% of users who denied. Of the 13 subjects who denied use but were positive on at least one specimen, four were identified solely by urine, two only by hair and one only by meconium. Self-report identified five users with all negative specimens. Although no one method identified all users, the single method that maximally identified users was detailed history taken by experienced interviewers.