BACKGROUND: Endoscopic retrograde cholangiopancreatography (ERCP) is a technically demanding procedure that can cause substantial complications. Competence in performing ERCP and the learning curve for achieving competence are poorly understood. OBJECTIVE: To evaluate the number of supervised ERCPs that physicians must do to achieve procedural competence. Competence was defined as a 0.8 probability of successfully completing specific technical components of ERCP and an overall grading of competence as judged by the attending physician. DESIGN: Prospective study. SETTING: University training program for gastroenterologists. PARTICIPANTS: 17 gastroenterology fellows at various stages of training. MEASURES: Experienced therapeutic endoscopists prospectively graded gastroenterology fellows during 1796 consecutive ERCPs. Fellows were graded on their overall level of competence for the procedure and on specific technical components of ERCP. RESULTS: Grading data were available for 1450 ERCPs (81%). The number of ERCPs done before adequate skill was achieved was 160 for cholangiography, 140 for pancreatography, 160 for deep cannulation of the pancreatic duct, 120 for stone extraction, and 60 for stent insertion. Fellows achieved overall competence after completing 180 to 200 ERCPs. The predicted probability of overall competence was 0.8 after 137 ERCPs and 0.9 after 185 ERCPs. CONCLUSIONS: At least 180 ERCPs were required before these gastroenterology fellows could be considered competent in ERCP. This number is much greater than that previously recommended, and these findings have substantial implications for training guidelines and issues of competence and certification in ERCP. The methods used to define and evaluate competence in ERCP could also be used to assess competence in other medical procedures.
BACKGROUND: Endoscopic retrograde cholangiopancreatography (ERCP) is a technically demanding procedure that can cause substantial complications. Competence in performing ERCP and the learning curve for achieving competence are poorly understood. OBJECTIVE: To evaluate the number of supervised ERCPs that physicians must do to achieve procedural competence. Competence was defined as a 0.8 probability of successfully completing specific technical components of ERCP and an overall grading of competence as judged by the attending physician. DESIGN: Prospective study. SETTING: University training program for gastroenterologists. PARTICIPANTS: 17 gastroenterology fellows at various stages of training. MEASURES: Experienced therapeutic endoscopists prospectively graded gastroenterology fellows during 1796 consecutive ERCPs. Fellows were graded on their overall level of competence for the procedure and on specific technical components of ERCP. RESULTS: Grading data were available for 1450 ERCPs (81%). The number of ERCPs done before adequate skill was achieved was 160 for cholangiography, 140 for pancreatography, 160 for deep cannulation of the pancreatic duct, 120 for stone extraction, and 60 for stent insertion. Fellows achieved overall competence after completing 180 to 200 ERCPs. The predicted probability of overall competence was 0.8 after 137 ERCPs and 0.9 after 185 ERCPs. CONCLUSIONS: At least 180 ERCPs were required before these gastroenterology fellows could be considered competent in ERCP. This number is much greater than that previously recommended, and these findings have substantial implications for training guidelines and issues of competence and certification in ERCP. The methods used to define and evaluate competence in ERCP could also be used to assess competence in other medical procedures.