This trial was conducted to compare the analgesic efficacy of administering variable-frequency automated boluses at a rate proportional to the patient's needs with fixed continuous basal infusion in patient-controlled epidural analgesia (PCEA) during labour and delivery. We recruited a total of 102 parturients in labour who were randomly assigned to receive either a novel PCEA with automated mandatory boluses of 5 ml administered once, twice, three or four times per hour depending on the history of the parturient's analgesic demands over the past hour (Automated bolus group), or a conventional PCEA with a basal infusion of 5 ml.h−1 (Infusion group). The incidence of breakthrough pain requiring supplementation by an anaesthetist was significantly lower in the Automated bolus group, three out of 51 (5.9%) compared with the Infusion group, 12 out of 51 (23.5%, p = 0.023). The time-weighted mean (SD) hourly consumption of ropivacaine was similar in both groups, 10.0 (3.0) mg in the Automated bolus group vs 11.1 (3.2) mg in the Infusion group (p = 0.06). Parturients from the Automated bolus group reported higher satisfaction scores compared with those in the Infusion group, 96.5 (5.0) vs 89.2 (9.4), respectively (p < 0.001). There was no difference in the incidence of maternal side-effects and obstetric and neonatal outcomes.