OBJECTIVE: To investigate whether procedure-specific skills acquired in a surgical lab, versus usual training, result in improved operative competence. DESIGN: randomised controlled trial. Didactic and procedural training occurred using three low fidelity vaginal surgery models: anterior repair (AR), posterior repair (PR), vaginal hysterectomy (VH). POPULATION/SETTING: Junior gynaecology residents at three academic centres. METHODS: The primary outcome was performance evaluated by attending staff blinded to group, via global rating scale (GRS) in the real operating room and for corresponding procedures. Prespecified secondary outcomes included procedural steps knowledge, overall performance impression, resident satisfaction, self-confidence and intraoperative parameters. A priori sample size estimated 50 residents (20% absolute difference in GRS score, 25% SD, 80% power, alpha 0.05). RESULTS: 83 residents were randomised to intervention or control and 55 completed the trial (2012-22). All characteristics were similar between groups. Adjusted GRS scores (by age, level and baseline knowledge) showed a significant group difference overall (mean difference 8.2; 95% CI 0.2,16.1; p=0.044) and for VH (mean difference 12.0; 95% CI 1.8, 22.3; p=0.02), but not for AR or PR. The intervention group also had significantly higher procedural steps knowledge, satisfaction and self-confidence for VH and PR (p<0.05 for all). Estimated blood loss, operative time and complications were similar between groups. CONCLUSIONS: Compared to usual training, surgical education modules using procedure-specific low fidelity models for vaginal surgery resulted in significant improvements in actual operative performance and several other skill parameters. CLINICAL TRIAL REGISTRATION: ClinicalTrials.gov, [www.clinicaltrials.gov](http://www.clinicaltrials.gov), NCT05887570
