Background Birth associated perineal trauma affects millions of women worldwide. The aim of the Perineal Assessment and Repair Longitudinal Study (PEARLS) was to evaluate if an enhanced, cascaded training programme improved implementation of evidence-based practise in perineal assessment and repair and reduced subsequent maternal morbidity. Methods PEARLS was a pragmatic matched pair cluster randomised controlled trial with 22 participating UK maternity units. Within each of the 11 matched pairs one unit was randomised to receive the intervention early (cluster A) and the other late (cluster B). Women sustaining a second-degree tear or episiotomy were eligible. Outcomes included pain on activity at 10–12 days postnatal, clinically reported outcomes by women and implementation of evidence-based surgical repair. Analysis was based on summary statistics at cluster level, using paired t-tests. Results 1470 and 2211 women were recruited in groups A and B respectively. No significant difference in mean primary outcome was noted between clusters that had received the intervention and those who had not (0.7% 95% CI (–10.1%, 11.4%), p = 0.89), with the overall percentage of women being 77% and 74% respectively. Improvement was seen in implementation of evidence-based perineal management. A significant reduction was noted in mean percentages of women reporting wound infections and needing suture removal in the early intervention clusters. Conclusion PEARLS is the first RCT to assess the impact of a ‘hands-on’ training package on implementation of evidence-based perineal trauma management and clinical outcomes for women. Findings will support improvements in clinical practise and women’s longer-term health.