Background
Skills in gastrointestinal endoscopy mainly depend on experience and exercise. Patients, in whom a trainee does the first endoscopies, are likely to suffer more pain and prolonged procedures. Training on endoscopy simulators, their latest generation being electronic virtual reality devices, is said to decrease time to reach competency in endoscopy. The purpose of the study was to determine whether the GI-Mentor® simulator (Simbionix, Israel) (Figure 1) can distinguish between beginners and experts in endoscopy, and to evaluate, if training for a limited time period can improve the performance of beginners.


Figure 1: Outside view of the GI-Mentor including torso,
endoscope and computer unit

Methods
Testing and training took place on the GI-Mentor® (Simbionix, Israel), a virtual endoscopy simulation. Thirteen beginners and 11 experts (1000+ procedures, both EGD and colonoscopy) in GI-endoscopy were included. The assessment of basic abilities included both two virtual gastroscopies and colonoscopies as well as two virtual skill tests. (Figure 2-4) The beginners were then randomised to a training (n=7) and a non training (n=6) group. The training group was allowed to work with the simulator two hours a day. After three weeks all participants were re-examined with two new endoscopy cases and the same virtual skill test. Insertion time, correctly identified pathologies, adverse events (inappropriate or unsuccessful retroflection, excessive wall pressure, impaired lumen view), and number of bubble hits in the virtual skill test were recorded. Chi-square test and Mann-Whitney tests were used for comparison.



Figure 2: Study design as described in the methods section

Figure 3+4: Endoscopic view of the virtual skill tests. Task in
figure 3 is to move a ball placed in the biopsy forceps into the
basket, task in figure 4 is to stitch the bubble with the needle

Results I
Basic assessment revealed significant differences in favour of experts in the virtual skill tests for the time used (p<0,05) (Figure 5) and for the number of bubble hits (p<0,01). Differences in the endoscopic tests differences could be seen for the number of adverse events in colonoscopy (p=0,02), for successful retroflection during EGD (p<0,005), for insertion time of gastroscopy and colonoscopy (p<0,001), and for the percentage of correctly identified pathologies (p<0,02) At the basic assessment, training and non training group did not differ in any tested parameter.

Results II
Final evaluation after three weeks showed significant differences between the two beginner groups in favour of the exercising group for number of adverse events during gastroscopy (p=0,02) and colonoscopy (p=0,04), for insertion time (p<0,03) and time of unrestrained lumen view during colonoscopy (p<0,02) and for the number of bubble hits (p<0,01). Significant differences were no longer seen between experts and exercising group at the final evaluation for any tested parameter
(Figure 6).

Figure5: Virtual skill test results demonstrating differences between beginners and experts before training.

Figure 6: Virtual skill test revealing similar performance between experts and exercising beginners after training

Discussion
Testing with the GI-Mentor® virtual endoscopy simulator is capable to disclose differences between beginners and experts of GI endoscopy. A training period of three weeks and two hours per day improves the performance of beginners, verified by significant differences in several tested parameters. The value of virtual reality simulator training for accelerating the development of hand-eye skills in endoscopy is obvious. Experts and beginners appreciated the realistic three-dimensional movements. Complex manipulative techniques- e.g. loop formations- are not simulated, which would make the simulator more attractive for the experienced endoscopist.


Figure 7-10: Virtual image of reflux esophagitis IV, angio-dysplasia (corpus), pseudo-polyposis (caecum), colon cancer

Up