Until recently, little prospective data existed regarding training and the rates of skill acquisition in either paediatric or adult endoscopy training. As a consequence, recommendations in paediatrics of the minimum numbers of supervised procedures required for competence in training have often been based on anecdote and personal opinion. Some prospectively collected data exists in adult endoscopy (1,2), and although this is often inconsistent, it forms the basis for training requirements that have been used historically. During the last 10 years, changes in paediatric endoscopy training guidelines have occurred in the absence of a prospectively gathered evidence base to support them. In 1997, the North American Society of Pediatric Gastroenterology, Hepatology and Nutrition recommended a minimum of 50 colonoscopies. This was revised to 100 in 1999 and also included recommendations for therapeutic procedures such as snare polypectomies (20), balloon dilatation of strictures (15) and management of bleeding GI lesions (20) (3–5). However, a requirement for an absolute number of procedures does not take into consideration potential differences between individuals both in terms of technical scope–handling skill acquisition rate and lesion recognition. Anecdotally, it has been suggested that there may be significant differences between trainees in their progression in endoscopic skill acquisition. Although this would seem likely and has been demonstrated in other surgical specialities (6), no attempt to prospectively investigate this exists within paediatric endoscopy.
It has been previously suggested that endoscopy simulators have the ability to differentiate between beginners and experts in endoscopy, supporting the validity of the model (7). Other studies have also suggested that virtual training may reduce the time needed to reach competence in endoscopy. Fertlitsch et al. randomly allocated beginners in adult ileocolonoscopy to either receive training on the Symbionix-GI Mentor (SGM) endoscopy simulator or no training for 3 weeks after baseline assessment (8). They were subsequently reassessed on the simulator and a statistically significant difference between the 2 groups was seen. Our department has had the use of an SGM to which trainees had access. The aim of the present study was to assess the differences in the rates of skill acquisition in trainees, looking for evidence of intertrainee variability and the impact of exposure on virtual endoscopy.
The Centre for Paediatric Gastroenterology at the Royal Free Hospital is one of the largest tertiary referral centres in the United Kingdom, performing approximately 600 endoscopies per year, of which approximately 400 are ileocolonoscopies. It is a nationally accredited unit for the training of paediatric gastroenterologists. During the 7-year period 1997 to 2004, the Consultant (R.H., S.M., or M.T.) supervising each endoscopy list contemporaneously assessed each ileocolonoscopy performed by the trainee using an assessment form. Prospective skill acquisition was assessed by rates of unassisted attainment of caecum, unassisted intubation of terminal ileum, and rates of accurate lesion recognition. At the end of each 10 completed ileocolonoscopies the percentage of success in each of these categories was calculated.
The department acquired an SGM in the fourth year of the 7-year study period. All trainees arriving after this point (n = 7) received at least 4 h/wk during a 6-week period of supervised training with the virtual ileocolonoscopy simulator before commencing ileocolonoscopy training on children. This allowed a comparison of rates of skill acquisition between trainees who had supervised training on the SGM (virtual-trained group [VT]) and those who arrived before we had this resource (standard-trained group [ST]). None had prior exposure to training on ileocolonoscopy before the period of assessment and training. Both groups carried out colonoscopies using the same equipment, the Olympus PCF240L, until 1999, and then the Fujinon EC410 subsequently; neither group used variable-stiffness colonoscopies, which have been suggested to enhance endoscopic performance (9). All of the trainees had obtained some experience of performing diagnostic upper gastrointestinal endoscopy. Some of this was not at our centre, and hence their relative abilities/rates of skill acquisition at performing upper gastrointestinal endoscopy were not prospectively assessed.
The outcomes considered were the proportion of successful attempts in terms of attainment of caecum and terminal ileum and lesion recognition rate. To assess the group effect on these outcomes, we used mixed linear regressions with random intercepts allowing for the trainees' cluster effects. Number of attempts was treated as a within-subject factor.
Completed assessment forms were available on 14 trainees who had rotated through our department during the 7-year study period. One of the trainees had completed less than 30 ileocolonoscopies by the end of the study period, so this person was excluded from analysis. The reminder of the trainees (13 individuals) had each completed a median of 80 colonoscopies (range, 50–140) by the end of the study period. Of these 13 trainees, 6 (1 woman, 5 men) completed their training before the arrival of the endoscopic simulator (ST), and 7 (2 women, 5 men) received exposure to this equipment before commencing ileocolonoscopy on children (VT).
The individuals in the ST group completed a median of 95 colonoscopies (range, 90–140) and those in the VT group completed a median of 60 colonoscopies (range, 50–100) during the study period. The number of attempts was seen to have a statistically significant effect on the 2 outcomes. On average, each additional attempt made by the trainee brings up the proportion of successes and the lesion recognition rate and the subjective score by 0.81% and 0.74%, respectively. Consequently, the multiple regressions were adjusted for the effect of the number of attempts. After adjusting for the effects of the number of attempts and site, a highly significant difference between the 2 groups was detected. On average, independent of site and number of attempts, the proportion of successes was 36% more for the VT group (95% CI, 25%–47%; P = 0.0000). Similarly, the VT group did significantly better in terms of the lesion recognition rate (mean difference, 34%; 95% CI, 23% –45%; P = 0.0000; see Fig. 1), comparing mean lesion recognition rates between the 2 groups.
Previous studies of adult patient–based trainees have suggested that there are significant differences in the rates of skill acquisition in ileocolonoscopy among trainees. Tassios et al. noted that depending on individual skill acquisition rates, between 100 and 180 supervised colonoscopies need to be performed before competence was achieved, which was defined by attainment of the caecum, with no reference made to intubation of the terminal ileum (2). Our study demonstrates the same wide interindividual variability in all of the assessed criteria, and it is particularly interesting that there was an accelerated rate of lesion recognition in the VT group. It could be hypothesised that the more competent a trainee becomes technically, the less he or she is concentrating on the skill required to negotiate the colonoscope, and consequently more attention is focused on looking at the mucosa. Although the requirement for a minimum number of procedures performed during training is likely to remain, our study suggests that this may be less helpful than ongoing assessment of trainees' skill acquisition rates, with competency-based training as the determining route to accepted completion of ileocolonoscopic training. It would therefore seem appropriate that standardised logbooks of trainees' procedures should be mandatory throughout recognised training centres, with accepted levels of competence to be achieved clearly stated. Consequently, it would be necessary to ensure that those involved in training were themselves competent to teach and instructed in standardised assessment of trainees. This suggests the need for periodic accreditation of those acting as trainers, as exists in other specialities (10). A case can also be made for routine use of virtual endoscopic training models before the exposure of trainee ileocolonoscopists to patients. No comment can be made on the argument that training in adult patient–based ileocolonoscopy would be useful for the paediatric trainee before training in children on the basis of this study. However, it is well recognised that the clinical environment, including use of sedation rather than general anaesthetic, the anatomy of the paediatric colon, the imperative to biopsy, and the different, albeit overlapping, pathology spectrum between adult and paediatric ileocolonoscopy practice, may indicate that a virtual paediatric model and a specific paediatric-based training environment may be more appropriate. The possibility of complications occurring may be diminished if a trainee is technically more competent when he or she begins to train with the human subject, although no difference was found between the 2 groups in our study because no complications occurred. Subsequently, it could be proposed that assessment of competence in the VT model could influence when a trainee is allowed to commence clinical training.
Along with previous publications in adult practice (11,12), but uniquely in paediatric endoscopic training, our study suggests that learning curves may be hastened by prior exposure to virtual endoscopy. This supports the view that the SGM ileocolonoscopy construct accurately resembles the experience of performing the procedure on actual subjects. Whether it accurately represents such clinical situations as sigmoid loops remains conjectural. Felsher et al. have previously objectively evaluated the skills of experienced and nonexperienced endoscopists with the SGM and noted significant performance differences between the 2 groups (7).
It seems clear, therefore, from the experience of both adult and paediatric endoscopy trainees that exposure to virtual endoscopy before using these procedures on human subjects accelerates skill acquisition. We would consequently suggest that, where possible, exposure to simulators before performing endoscopy on patients should be encouraged. Furthermore, we would suggest that objective assessment using a standardised form is carried out with all paediatric endoscopy trainees, with the possibility of additional exposure to virtual endoscopy in those who appear slower to acquire skills. We would also welcome a prospective randomised multicenter trial among paediatric gastroenterology trainees to further assess the role of not only the colonoscopy construct but also the other endoscopic models, which are available.
1. Cass OW, Frekkeman ML, Peine CJ, et al
. Objective evaluation of endoscopy
skills during training. Ann Intern Med 1993; 118:40–44.
2. Tassios PS, Ladas SD, Grammenos I, et al
. Acquisition of competence in colonoscopy
: the learning curve of trainees. Endoscopy
3. Fox V. Clinical competency in pediatric endoscopy
. J Pediatr Gastroenterol Nutr 1998; 26:200–204.
4. Hassall E. Requirements for training to ensure competence of endoscopist performing invasive procedures in children. J Pediatr Gastroenterol Nutr 1997; 24:345–347.
5. Rudolph C, Winter H, NASPGN Executive Council et al. NASPGN guidelines for training in pediatric gastroenterology. J Pediatr Gastroenterol Nutr
6. Vossen C, Van Ballaer P, Shaw RW, et al
. Effect of training on endoscopic intracoporeal knot tying. Hum Reprod 1997; 12:2658–2663.
7. Felsher JJ, Olesevich M, Farres H, et al
. Validation of a flexible endoscopy
simulator. Am J Surg 2005; 189:497–500.
8. Fertlitsch A, Glauninger P, Gupper A, et al
. Evaluation of a virtual endoscopy
simulator for training in gastrointestinal endoscopy
9. Shumaker DA, Zaman A, Katon RM. Use of a variable stiffness colonoscope allows completion of colonoscopy
after failure with standard adult colonoscope. Endoscopy
10. Salisbury C. The Australian Quality Assurance and Continuing Education
Programme as a model for the re-accreditation of general practitioners in the United Kingdom. Br J Gen Pract 1997; 47:319–322.
11. Hochberger J, Matthes K, Maiss J, et al
. Training with compactEasie biologic endoscopy
simulator significantly improves hemostatic technical skill of gastroenterology fellows: a randomised controlled comparison with clinical endoscopy
training alone. Gastrointest Endoscop 2005; 61:216–218.
12. Clark JA, Volchok JA, Hazey JW, et al
. Initial experience using an endoscopic simulator to train surgical residents in flexible endoscopy
in a community medical center residency program. Curr Surg 2005; 62:59–63.