Expert benchmark for the GI Mentor II^TM

Abstract

Background

There is increasing interest in the use of virtual-reality simulators in general surgery residency training. Many simulators lack a benchmark against which trainees can measure competence and skill.

Methods

Surgeons who had performed over 1,000 colonoscopies were evaluated on module 1, case 5 of the GI Mentor I or II™ virtual-reality endoscopy simulator (Simbionix, Cleveland). Participants were given 5 min to familiarize themselves with the simulator, and then performed the study case with standardized instructions. Metrics were recorded by using the previously calibrated simulator.

Results

Twenty-three surgeons (21 male, 2 female) participated. Mean height was 69.6 ± 2.6 inches, mean age 51 ± 9 years, median surgical glove size 7.5, and surgeons had 18.8 ± 10.1 years of practice, and did 8 ± 6 colonoscopies weekly. Ten participants had advanced training in endoscopy, laparoscopy or colorectal surgery; eight had used the simulator before, of whom six had used it once. Mean time to complete the study case was 13.6 ± 5.3 min and time to reach the cecum was 6.5 ± 4.3 min. Participants examined 92.3 ± 3.6% of the simulated colonic mucosa with a clear view of the lumen 89.5 ± 4.2% of the time. Total time the colon was looped was 22 ± 35 s (range 0–133 s). The overall efficiency of screening was 70.33 ± 23.45% (range 20–94%). Participants tended to mistake normal simulated colonic structures as pathology.

Conclusion

Performance on a virtual-reality endoscopic simulator has a wide amount of variability even among a group of experienced endoscopists. Expert benchmark tests should be performed on simulators that will be used for resident assessment prior to any attempts at certification of competence.