Multicenter, randomized, controlled trial of virtual-reality simulator training in acquisition of competency in colonoscopy

Background: The GI Mentor is a virtual reality simulator that uses force feedback technology to create a realistic training experience. Objective: To define the benefit of training on the GI Mentor on competency acquisition in colonoscopy. Design: Randomized, controlled, blinded, multicenter trial. Setting: Academic medical centers with accredited gastroenterology training programs. Patients: First-year GI fellows. Interventions: Subjects were randomized to receive 10 hours of unsupervised training on the GI Mentor or no simulator experience during the first 8 weeks of fellowship. After this period, both groups began performing real colonoscopies. The first 200 colonoscopies performed by each fellow were graded by proctors to measure technical and cognitive success, and patient comfort level during the procedure. Main Outcome Measurements: A mixed-effects model comparison between the 2 groups of objective and subjective competency scores and patient discomfort in the performance of real colonoscopies over time. Results: Forty-five fellows were randomized from 16 hospitals over 2 years. Fellows in the simulator group had significantly higher objective competency rates during the first 100 cases. A mixed-effects model demonstrated a higher objective competence overall in the simulator group (P < .0001), with the difference between groups being significantly greater during the first 80 cases performed. The median number of cases needed to reach 90% competency was 160 in both groups. The patient comfort level was similar. Conclusions: Fellows who underwent GI Mentor training performed significantly better during the early phase of real colonoscopy training.