Group B streptococcus (GBS) is a leading cause of infectious neonatal morbidity and mortality. Timely and accurate identification of colonized mothers is imperative so that antibioprophylaxis can be implemented during labour to reduce the risk of neonatal sepsis. We planned our study to analyse the diagnostic accuracy of an intrapartum PCR assay to identify GBS-colonized women and to allow the implementation of correct (i.e. at least 4 h) intrapartum antibiotic prophylaxis based on the PCR results. We included 695 women in labour who were tested for rectovaginal GBS carriage by culture and PCR. Women were also screened at 35-37 weeks of gestation. Intrapartum GBS colonization was 19.3%. Assay sensitivity was 81.0% for antenatal culture and 85.0% for intrapartum PCR; p 0.72. GBS colonization (n = 107) was known at least 4 h before delivery in 68 (64%) and 73 (68%) women based on antenatal culture and intrapartum PCR, respectively. Among 43 women delivering preterm, correct status was known at least 4 h before delivery in 10 (23%) and 32 (74%) women according to antenatal culture and intrapartum PCR, respectively. These results support the concept that GBS screening can be performed routinely during labour in a clinical setting. The intrapartum approach is at least as accurate as the antenatal screening, with the additional advantage of identifying women delivering preterm or not followed during pregnancy.