Should simulator use become mandatory in endoscopy training?

Flight simulators were first introduced to the US army in 1934 after a series of accidents following the advent of flights using navigational instruments. They are now an essential component to both individual pilot and flight crew training and are thought to save money, time, and most importantly, lives. While complications from endoscopy are unlikely to be as graphic, nor usually as life threatening, there have been very significant recent advances in endoscopy simulator development. Simulators are available in the form of mechanical, animal, and computer models.1 Mechanical models represent some of the earliest endoscopy simulators and help to provide familiarity with endoscope use, but fail to reproduce ‘real-life’ situations. Animal models provide the most realistic tactile response and tissue characteristics similar to that in human endoscopy, but have substantial obstacles to their frequent use, including their expense, space requirements, and labor required to prepare and maintain them. Computer simulator technology, such as that seen with the AccuTouch endoscopy simulator (Immersion Medical, Gaithersburg, MD, USA) or the GI mentor II (Simbionix, Cleveland, OH, USA), provide certain advantages over mechanical and animal models; these include ease of use and maintenance, and relative compact storage. Using a realistic endoscope, the user is able to see on a monitor the effects of inserting and withdrawing the instrument while manipulating the controls. These technologies can be applied to gastroscopy, colonoscopy, endoscopic retrograde cholangiopancreatography, and endoscopic ultrasound. On the GI mentor system, an accessory channel is able to simulate biopsy, polypectomy, and other interventions, and may vocalize to the endoscopist discomfort from over-insufflation or looping. In this edition of the Journal of Gastroenterology and Hepatology, Shirai and colleagues assess, for novice endoscopists, whether prior training with a computer simulator facilitates manipulation and examination skills when the endoscopist embarks on ‘real-life’ gastroscopy.2 The aims of these simulators are to give the trainee familiarity with endoscope manipulation and facilitate the recognition of anatomy and pathology. This familiarity will hopefully translate into improved patient comfort, time efficiency, better visualization of mucosa, and most importantly, prevent complications. These benefits may accelerate the training time of novice endoscopists, which ultimately could help trainees to work independently earlier, freeing up more time for both teaching and training doctors. When computer simulators are used to train endoscopists, they should be considered as only one facet of an integrated approach that should involve a significant period of one-on-one training with experienced endoscopists who are competent trainers. Further, it needs to be noted that computer simulator training focuses predominantly on the ‘endoscopic’ component of the procedure; it does not include other important components, such as sedation, airway management, or cardiovascular resuscitation. While being fairly realistic in terms of manipulation and passage through the upper gastrointestinal tract, there is poor reproduction of cricopharyngeal and deep duodenal intubation, or of mucosal pathology. However, computer simulators may not need to be high fidelity to achieve significant training benefit. To date, the uptake of computer simulator use has been variable between countries. Further, it is generally confined to independent medical training institutions, usually those aligned with universities or gastroenterological societies. The main disincentive to the introduction of computer simulators is cost, which for the GI mentor system is approximately $AUD96 000. There is an increasing body of evidence that simulator use early in gastroscopy training facilitates learning and scope manipulation. An Italian group evaluated the differences in manipulation skills and complications of novice endoscopists randomized to receive training for 2 h daily for 3 weeks, compared with those who did not.3 Those participants who underwent training were subsequently better at endoscopy performed on the GI mentor system. Another group then demonstrated that 10 h of computer simulator use prior to gastroscopy in humans improved procedure duration and completeness, and diminished requests for assistance.4 Shirai and colleagues randomized endoscopy trainees, all of whom received bedside training, to either 5 h of simulator training or no simulator training. They then assessed their ability to perform their first two upper gastrointestinal endoscopies and examine important, relevant areas.2 Unique to this study, the two assessors were blinded to the previous amount of training received. Overall, the group that received simulator training was superior in manipulating the endoscope through the upper gastrointestinal tract, but there was no measured difference in observational skillsbetween the two cohorts. Even more benefit may have been observed had the time spent on computer simulators been supervised. Similar benefits in early training have been shown for colonoscopy.5,6 We believe that much of the initial focus of a trainee is on endoscope manipulation, to the detriment of appreciating mucosal pathology. It should therefore follow that accelerated learning of manipulative skills will subsequently enhance the acquisition of observational and interpretive skills. There is an increasing body of evidence in endoscopy and various surgical fields that simulator training improves dexterity which translates into improved learning in real-life settings.7–9 There are also increasing expectations from society, our patients and for professional governance reasons, that simulators be used before learning on humans. In the future, we feel that novice endoscopists will be required to spend a specified period of time on simulators before progressing to performing endoscopy on patients. Perhaps, with the accumulated evidence available, that time should be now!