We compared results of urine drug analysis with clinical data and history to test the usefulness of peripartum drug screening and to establish guidelines for optimal testing. Urine from 28 mothers and 52 babies was analysed. Drugs not suspected by history were found in 10 mothers and six babies. Results assisted in the management of neonatal withdrawal in three babies. Drugs suspected by history were not found in 11/22 mothers and 23/35 babies. About half of these results were associated with delayed urine collection. In 12/28 mothers, drugs administered in hospital could have confused interpretation of screen results. We conclude that urine drug screening without strict protocols for specimen collection is of limited usefulness for management of drug abuse in pregnancy and neonatal drug withdrawal. We favour testing of maternal urine obtained before drugs are administered in hospital. Neonatal urine, if used, should be collected in the first day of life.