Combining real-time AI and in-person expert instruction in simulated surgical skills training - Randomized crossover trial

Abstract

Abstract Traditional surgical training has significant limitations, lacking objectivity and standardization. Deploying AI tools with conventional expert-mediated teaching may uncover areas where AI could complement experts and enhance surgical training through real-time performance assessment and feedback alongside risk mitigation. This randomized crossover trial assessed learning outcomes in two training sessions involving in-person expert instruction and real-time AI feedback using previously validated tumor resection simulations. Receiving expert feedback before real-time AI instruction led to greater performance improvement in trainee performance scores compared to the opposite order, with a mean difference of 0.67 95%CI [0.43–0.91], p < 0.001. Diminishing returns were observed with human expert feedback, which were not seen with AI feedback, such as increased injury and bleeding risk. In surgical procedural training, AI feedback may efficiently maintain peak performance after an initial learning phase led by human experts. AI-integrated surgical curricula should consider the relative benefits of both AI and expert feedback.

Ähnliche Arbeiten