Key points

Introduction

Colonoscopy continues to be the dominant method used for colorectal cancer prevention, and optimizing training and teaching methods is critical to ensure consistent high-quality procedures in practice. Gastroenterologic and surgical societies now acknowledge the need for all trainees to master a series of motor and cognitive skills before graduating from their respective programs. High-quality supervision is important for high-quality training ; however, there is now evidence that the use of technology in the form of simulators and position sensors, and retraining experienced endoscopists, could further improve competency and quality in colonoscopy, a practice once seen as requiring apprenticeship-directed training. This article reviews the key studies and themes in training and teaching innovations in colonoscopy.

Introduction

Colonoscopy continues to be the dominant method used for colorectal cancer prevention, and optimizing training and teaching methods is critical to ensure consistent high-quality procedures in practice. Gastroenterologic and surgical societies now acknowledge the need for all trainees to master a series of motor and cognitive skills before graduating from their respective programs. High-quality supervision is important for high-quality training ; however, there is now evidence that the use of technology in the form of simulators and position sensors, and retraining experienced endoscopists, could further improve competency and quality in colonoscopy, a practice once seen as requiring apprenticeship-directed training. This article reviews the key studies and themes in training and teaching innovations in colonoscopy.

Learning curve of colonoscopy

The American Society for Gastrointestinal Endoscopy Training Committee most recently published a position statement on the colonoscopy core curriculum, identifying key features in motor and cognitive skills to enable endoscopists to perform colonoscopies safely and competently. The core technical and cognitive skills are shown in Box 1 . These core skills can be classified within the competency outline described by the Accreditation Council for Graduate Medical Education (ACGME).

Although there is no reliable number of colonoscopies needed to complete before ensuring competency, recent studies have shown that there is a learning curve for this procedure. One of the most heavily reviewed studies is by Sedlack, in which 41 gastroenterology fellows at a single center performed 6635 colonoscopies from July 2007 through June 2010. The fellows’ core cognitive and motor colonoscopy skills were assessed using the Mayo Colonoscopy Skills Assessment Tool, a validation tool created to provide a valid means to objectively assess individual cognitive and motor skills throughout colonoscopy training. The study concluded that a Mayo Colonoscopy Skills Assessment Tool score of 3.5, cecal intubation rates of 85%, and intubation times of less than 16 minutes were recommended as minimal competency criteria. On average, 275 procedures were needed to achieve even this conservatively defined competence in colonoscopy ( Fig. 1 ). This critical number differs from the current recommendation that fellows perform a minimum of 140 colonoscopies before assessing competency. This difference is even greater when considering the threshold of 50 procedures recommended by the family practice and general surgery residency guidelines. With such a wide range of learning curves, programs must ensure that adequate volumes of procedures be available to trainees in the outpatient and inpatient setting, and that ongoing competency assessment be performed throughout the stages of training.

Overall skills learning curves. The learning curve of average Mayo Colonoscopy Skills Assessment Tool scores for overall motor ( solid line ) and overall cognitive ( dashed line ) skills are shown. Average scores are labeled with error bars showing the 95% confidence interval. By 250 procedures, cognitive skills have achieved the 3.5 minimal competency criteria goals, with motor scores reaching this by 275 procedures. By 300 procedures, all scores and the 95% confidence interval are above this threshold.

( Adapted from Sedlack RE. Training to competency in colonoscopy: assessing and defining competency standards. Gastrointest Endosc 2011;74(2):360; with permission.)

Impact of type of training on quality in colonoscopy

Assessment of training and competency in performing colonoscopy in general surgery has also received scrutiny. Surgical societies, the ACGME, and the Residency Review Committee require surgical residents to perform 50 colonoscopies before assessment of proficiency. In a study by Spier and colleagues, 21 surgical residents performed a mean of 80 ± 35 total colonoscopies during the 2-month outpatient endoscopy rotation. Average cecal intubation rate was 47%, and resident comfort level for independently performing a total colonoscopy was scored a mean 3.6 on a scale of 1 to 5 (5 = most comfortable). Thus, although surgical residents can fulfill the prescribed number of colonoscopy procedures (N = 50) during an allotted time, achieving technical competence in colonoscopy as defined by a 90% cecal intubation rate was not successful. Similarly, Selvasekar and colleagues demonstrated that colorectal surgery fellows were found to achieve screening colonoscopy competency, defined by a significant reduction in total procedure and 80% cecal intubation rate within 35 minutes, after completing 94 to 114 procedures.

There is also evidence that having separate endoscopy training programs among gastroenterologists and surgeons leads to variable quality of colonoscopy performance. In a retrospective review by Leyden and colleagues, all colonoscopy procedures performed in a single endoscopy unit by both gastroenterology trainees and surgical trainees with more than 2 years of endoscopy experience were audited. Among 3079 single-endoscopist colonoscopies, there was an observed disparity in endoscopic performance between surgical and gastroenterology trainees, with higher completion rates, polyp detection rates, adenoma detection rates (ADR), and greater withdrawal time correlating closely with polyp detection rate for gastroenterology trainees compared with surgery trainees.

One possible option to reduce such differences in colonoscopy competency could be to no longer mandate that all surgical trainees be trained in colonoscopy if their subspecialty interest does not involve application of endoscopy. Thus, only those whose career track requires routine use of endoscopy may need to take part in a more intense colonoscopy training program, which could be integrated into the gastroenterology trainee curriculum.

Colonoscopy simulators

Traditionally, colonoscopy is learned by hands-on training under the supervision of an experienced endoscopist. However, with increasing demands to perform more colonoscopies in a streamlined fashion with minimal interruptions and delays, significant cost considerations have been demonstrated from lost productivity at institutions where novices are being trained to perform colonoscopy. In a study analyzing data from the Clinical Outcomes Research Initiative Project, involvement of fellows prolonged the procedure time by 10% to 37%, with an estimated loss of reimbursement to the academic institution of $500,000 to $1,000,000 per year. Simulators have therefore been developed to potentially close the gap between increased procedure duration and financial burden.

Computer simulators

Computer simulators, first developed in the 1980s, enhanced traditional endoscopy teaching. They have been the subject of much attention for possibly better preparing trainees to perform colonoscopies on real patients after completing a finite number of simulations. Medical training has many similarities to aviation training, which catalyzed computer-based training simulators. Simulators are safe, and ethical concerns are minimized when training with a mannequin. Second, simulators are multipurpose, allowing training in different environments and clinical scenarios. Third, simulators are cost-effective, providing a great deal of training for the monetary investment initially placed on the equipment.

The typical electronic simulator contains a functioning colonoscope, real dials, and buttons that allow for steering and torquing of the instrument, with the exception that the hydraulic and pneumatic functions do not actually cause the scope to eject water or air. A mannequin with an anus is provided, in addition to a screen for visualization. Currently, the two commercially available simulators include the Simbionix GI Mentor (Simbionix USA Corporation, Cleveland, OH) and the CAE Healthcare AccuTouch Endoscopy Simulator (CAE Healthcare [formerly Immersion Medical], Gaithersburg, MD) ( Fig. 2 ). The endoscope has a closed tip that contains forced feedback sensors, thus allowing the trainee to experience the “haptic” resistance as the endoscope is advanced through the mannequin. Each case creates a different scenario with different indications for colonoscopy, vital signs, patient history, and evaluations pertaining to the chief complaint. During each simulation session, the software tracks a broad range of metrics and presents the user with a report at the end of each colonoscopy. Some of these metrics include information about looping, overinsufflation, and loss of mucosal visualization. Data are collected automatically, and at the completion of the module, trainees can review their overall performance, and instructors can follow their progress longitudinally.

CAE Healthcare (formerly Immersion Medical) Accu Touch endoscopy simulator.

( Courtesy of CAE Healthcare. © 2012 CAE Healthcare; with permission.)

Several clinical trials have shown benefit with using a computer simulator early on in the fellowship training. Sedlack and Kolars developed a computer-based colonoscopy simulator course for first-year gastroenterology fellows using the Immersion Medical simulator (as of 2010, Immersion Medical is CAE Healthcare). The study demonstrated that the most benefit for trainees was during the early stages of training. One of the largest trials conducted by Cohen and colleagues, a multicenter trial, randomized 45 first-year gastroenterology fellows from 16 hospitals over 2 years to 10 hours of simulation training during the first 8 weeks of fellowship. Following this, 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. Results showed that the simulator-trained group demonstrated higher overall competency during the first 80 clinical cases; however, the median number of cases required to reach 90% objective competency was identical in the simulator and no-simulator groups (mean, 160). Most recently, in a multinational, randomized, blinded, controlled trial by Haycock and colleagues, 36 novice colonoscopists were randomized to either 16 hours of simulator training or patient-based training and then completed three simulator cases before and after training; three live cases were assessed after training by masked experts. Results demonstrated that novices trained on the computer simulator demonstrated superior technical skills on simulated cases but no difference was observed between the groups in the live patient cases.

Several validation studies have also shown that computer colonoscopy simulators could distinguish between expert and novice endoscopists, but with varying types of measurable parameters. In the studies by Kim and colleagues and Sedlack and Kolars, computer-based simulators were able to discriminate levels of expertise only for parameters related to procedure time and not for actual endoscopic skills. McConnell and colleagues demonstrated that the only significant differences between novice and expert endoscopists were seen in performance metrics related to procedural time, air insufflation, sedation management, endoscope force, and patient comfort. In contrast, the studies by Grantcharov and colleagues, Mahmood and Darzi, and Eversbusch and Grantcharov demonstrated that significant differences in performance existed among endoscopists of varying levels of experience for procedure-related times and endoscopic skills. Furthermore, one validation study demonstrated that trainees failed to improve without feedback from a mentor.

Although current simulators have demonstrated a benefit in skill acquisition for the first 20 to 80 cases performed by novices with an acceleration of the learning curve, no reduction in the median number of cases required to achieve technical and cognitive competency has been observed. It is in the expert opinion of the Preservation and Incorporation of Valuable Endoscopic Innovations committee that to justify the expense and effort involved in purchasing simulators and incorporating them into the training program, a reduction in training times or procedure numbers of at least 25% is required. Most recently, new ACGME training guidelines implemented in July 2012 require all gastroenterology fellows to participate in some type of simulation-based training during their fellowship.

Computer-based simulators can accelerate the learning curve for trainees in colonoscopy and provide trainees with low-risk, low-stress hands-on time with endoscopic equipment before performing on real patients; however, the duration of these benefits has been shown to be limited to the very early stages of novice training.