Self-Directed Mikoto Simulator Training for Colonoscopy in Gastroenterology Trainees: An International Multicenter Randomized Trial

Objectives: Simulation-based colonoscopy training may support early acquisition of insertion skills, but its clinical effect remains uncertain. We evaluated whether a self-directed curriculum using the mikoto colonoscopy simulator improved insertion performance in gastroenterology trainees. Methods: In this multicenter cluster-randomized trial, 12 gastroenterology trainees from 11 institutions in Japan and Taiwan were assigned 1:1 to mikoto simulator training or standard training, with the trainee as the unit of randomization. Each trainee performed 35 colonoscopies; intervention-arm trainees additionally completed 100 self-directed simulator sessions over 2 weeks. The primary endpoint was post-training cecal intubation rate (CIR), compared using mixed-effects logistic regression with trainee as a random intercept and pre-training CIR as a covariate. Results: A total of 420 colonoscopies were analyzed. Post-training CIR was 80.0% (72/90) in the mikoto group versus 77.8% (70/90) in controls (adjusted OR, 1.34; 95% confidence interval, 0.37–4.84; p = 0.606). Within the mikoto group, CIR increased from 60.0% to 80.0% (p = 0.003) and cecal intubation time decreased, but the group-by-period interaction was not significant. Detection outcomes were unchanged. Conclusions: Although the between-group difference in post-training CIR did not reach statistical significance, self-directed mikoto simulator training produced a significant within-group improvement in CIR and a parallel reduction in cecal intubation time, supporting it as a feasible and educationally meaningful adjunct to conventional colonoscopy training for gastroenterology trainees. Trial Registration: UMIN Clinical Trials Registry: UMIN000054647. © 2026 The Author(s). Journal of Gastroenterology and Hepatology published by Journal of Gastroenterology and Hepatology Foundation and John Wiley & Sons Australia, Ltd.