Objective: to compare the outcome of care given to women ‘booking’ for delivery in a midwife-led maternity unit with that for comparable women ‘booking’ for care in a consultant obstetric unit. Design and method: prospective cohort study with a quasi-experimental design and data extracted from case notes. Setting: east Dorset, midwife-led maternity unit at Royal Bournemouth Hospital and consultant-led maternity unit at Poole General Hospital. Subjects: two cohorts of women who satisfied the criteria for ‘booking’ at the Royal Bournemouth Hospital. Of these 794 ‘booked’ at Bournemouth from I November 1992 to 30 June 1993 and 705 ‘booked’ at Poole over the same period. Main process and outcome measures: care given, morbidity in women and their babies, transfers during the antenatal period and in labour. Findings: of the women who initially ‘booked’ for Bournemouth, 62.3% actually delivered there, 27.1% transferred before labour and a further 9.2% transferred during labour. No differences were seen between those ‘booked’ for Bournemouth or Poole in the proportions of low birthweight babies, babies who were transferred to special care or babies who had congenital abnormality. Higher proportions of babies whose mothers ‘booked’ for delivery in Poole were resuscitated and had one minute Apgar scores below seven but there was no difference in the five minute scores. Similar proportions of women had perineal tears but fewer of the women ‘booked’ for delivery in Bournemouth had an episiotomy. ‘Booking’ for Poole was associated with higher rates of induction and augmentation of labour and greater use of anesthesia. ‘Booking’ for Bournemouth was associated with a shorter first stage and a longer third stage of labour. Women ‘booked’ for delivery in Bournemouth were no more likely to be delivered by a midwife than those ‘booked’ for Poole. Conclusions: there was very little difference between the groups of women who intially ‘booked’ for delivery at the two units. There were differences in the patterns of care received, but no major differences in the outcome for the women or their babies were detected.