Impact of fellowship training level on colonoscopy quality and efficiency metrics.

BACKGROUND AND AIMS: Previous studies have described variable effects of fellow involvement on the adenoma detection rate (ADR), but few have stratified this effect by level of training. We aimed to evaluate the "fellow effect" on multiple procedural metrics including a newly defined adenoma management efficiency index, which may have a role in documenting colonoscopy proficiency for trainees. We also describe the impact of level of training on moderate sedation use.

METHODS: We performed a retrospective review of 2024 patients (mean age, 60.9 ± 10 years; 94% men) who underwent outpatient colonoscopy between June 2012 and December 2014 at our Veterans Affairs Medical Center. Colonoscopies were divided into 5 groups. The first 2 groups were first-year fellows in the first 6 months and last 6 months of the training year. Second- and third-year fellows and attending-only procedures accounted for 1 group each. We collected data on doses of sedatives used, frequency of adjunctive agent use, procedural times, and location, size, and histology of polyps. We defined the adenoma management efficiency index as average time required per adenoma resected during withdrawal.

RESULTS: Of the colonoscopies performed, 1675 involved a fellow and 349 were performed by the attending alone. There was no difference in ADR between fellows according to level of training (P = .8) or between fellows compared with attending-only procedures (P = .67). Procedural times decreased consistently during training and declined further for attending-only procedures. This translated into improvement in the adenoma management efficiency index (fellow groups by ascending level of training: 23.5 minutes vs 18.3 minutes vs 13.7 minutes vs 13.4 minutes vs attending group 11.7 minutes; P < .001). There was no difference in the average doses of midazolam and fentanyl used among fellow groups (P = .16 and P = .1, respectively). Compared with attending-only procedures, fellow involvement was associated with higher doses of fentanyl and midazolam and more frequent use of diphenhydramine and glucagon (P < .0001, P = .0002, P < .0001, and P = .01, respectively).

CONCLUSIONS: ADR was similar at different stages of fellowship training and comparable with the attending group. Efficiency of detecting and resecting polyps improved throughout training without reaching the attending level. Fellow involvement led to a greater use of moderate sedation, which may relate to a longer procedure duration and an evolving experience in endoscopic technique.