Upskilling, Deskilling, or Never Skilling: Who Benefits From Artificial Intelligence in Colonoscopy?
In this issue of Gastroenterology, Dominitz and colleagues1 describe the effect of a large computer-aided detection (CADe) program in the US Veterans Administration (VA). The VA randomized 42 sites to CADe implementation with 97 control facilities. The investigators applied a difference-in-difference design to compare adenoma detection rates (ADRs). The ADR is a primary metric to assess the quality of colonoscopy performance. Presumably it reflects the thoroughness with which precancerous lesions are found and removed.
