Development and initial validation of an instrument for video-based assessment of technical skill in ERCP

• Authors: Joseph B Elmunzer, Medical University of South Carolina; Catharine M Walsh, Hospital for Sick Children; Gretchen Guiton, University of Colorado; Jose Serrano, National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Bethesda; Amitabh Chak, Case Western Reserve University; Steven Edmundowicz, University of Colorado; Richard S Kwon, University of Michigan; Daniel Mullady, Washington University St. Louis; Georgios Papachristou, Ohio State University; Grace Elta, University of Michigan; Todd H Baron, University of North Carolina; Patrick Yachimski, Vanderbilt University; Evan L Fogel, Indiana University; Peter Draganov, University of Florida; Jason R Taylor, St Louis University; James Scheiman, University of Virginia; Vikesh K Singh, Johns Hopkins Med Inst; Shyam Varadarajulu, Advent Health; Field Willingham, Emory University; Gregory A Cote, Medical University of South Carolina; Peter B Cotton, Medical University of South Carolina; Violette Simon, University of Colorado; Rebecca Spitzer, Medical University of South Carolina; Rajesh Keswani, Northwestern University; Sachin Wani, University of Colorado
• Journal Title: GASTROINTESTINAL ENDOSCOPY
• Volume: Volume 93, Number 4
• Publisher: MOSBY-ELSEVIER | 2021-03-16, Pages 914-923
• Type of Work: Article | Post-print: After Peer Review
• Publisher DOI: http://doi.org/10.1016/j.gie.2020.07.055
• Subjects: Health Sciences, Nutrition
• Keywords: Science & Technology; Life Sciences & Biomedicine; Gastroenterology & Hepatology

Abstract

Background and Aims: The accurate measurement of technical skill in ERCP is essential for endoscopic training, quality assurance, and coaching of this procedure. Hypothesizing that technical skill can be measured by analysis of ERCP videos, we aimed to develop and validate a video-based ERCP skill assessment tool. Methods: Based on review of procedural videos, the task of ERCP was deconstructed into its basic components by an expert panel that developed an initial version of the Bethesda ERCP Skill Assessment Tool (BESAT). Subsequently, 2 modified Delphi panels and 3 validation exercises were conducted with the goal of iteratively refining the tool. Fully crossed generalizability studies investigated the contributions of assessors, ERCP performance, and technical elements to reliability. Results: Twenty-nine technical elements were initially generated from task deconstruction. Ultimately, after iterative refinement, the tool comprised 6 technical elements and 11 subelements. The developmental process achieved consistent improvements in the performance characteristics of the tool with every iteration. For the most recent version of the tool, BESAT-v4, the generalizability coefficient (a reliability index) was.67. Most variance in BESAT scores (43.55%) was attributed to differences in endoscopists’ skill, indicating that the tool can reliably differentiate between endoscopists based on video analysis. Conclusions: Video-based assessment of ERCP skill appears to be feasible with a novel instrument that demonstrates favorable validity evidence. Future steps include determining whether the tool can discriminate between endoscopists of varying experience levels and predict important outcomes in clinical practice.