Few studies have investigated the relative ability of different biologic specimens to identify fetal cocaine exposure. As part of a federally funded, prospective study, we identified subjects through chart review of all patients admitted to Labor and Delivery. Experienced interviewers enrolled all consenting patients with history of prenatal cocaine use and the next admission with no recorded use. Extensive interviews postpartum: 1). verified history of use; 2). helped eliminate women with chronic illnesses and medications, other illicit drug use, >1.5 oz of absolute alcohol per day; and 3). detailed amount, frequency and timing of drug use. Amniotic fluid, cord blood, infant urine and meconium, and maternal hair were collected using specific protocols. Analysis of biologic samples was performed blindly using immunoassay verified by gas chromatography/mass spectrometry. Of 65 subjects whose specimens have been analyzed to date, 45 had one or more biologic specimens positive for cocaine metabolites or admitted prenatal use: 33 were identified by interview, 26 by meconium, 21 (of 38 analyzed) by hair, 21 by urine, 16 (of 42) by cord blood, and 9 (of 25) by amniotic fluid. The best combination of two methods, interview and meconium identified, 39 (87%) of the 45; interview, meconium and hair together identified all but 7%. Of the 73% identified by interview, 9 admitted only first trimester use and 9 admitted use in each trimester. As hypothesized, urine and meconium were generally not positive for admitted early use. However, hair analysis did detect over 70% of those admitting first trimester use, but, disappointingly, only identified about half of those positive by another method. No one biologic specimen identified all the users detected by other methods. The single method that maximumly identified users was detailed history taken by experienced interviewers. Further analyses will compare type and quantity of metabolites detected in maternal versus infant specimens within each trimester.