Mastery learning guided by artificial intelligence is superior in flexible bronchoscopy training: a randomized controlled trial.

<bold>Introduction:</bold> Simulation-based training has proven effective for learning flexible bronchoscopy. However, no studies have tested the efficacy of training towards established proficiency criteria i.e., mastery learning (ML). <bold>Study Aim:</bold> To test the effectiveness of ML compared to self-regulated learning SRL on novice bronchoscopists’ end-of-training performance. <bold>Methods:</bold> In a standardized simulated setting, novices without prior bronchoscopy experience trained using an artificial intelligence (AI) guidance system that automatically recognizes the bronchial segments. They were randomized into two groups: the ML group and the SRL group. The ML group trained until they completed two procedures meeting the proficiency targets: 18 inspected segments, 18 structured progressions, <120 seconds procedure time. The SRL group trained until they no longer perceived any additional benefits from training. Both groups then did a finalizing test, without the AI-guidance enabled. <bold>Results:</bold> 24 participants completed the study, with 12 in each group. The ML group had a high number of inspected segments (mean±sd, 17.2±1) and structured progressions (15.5±2). This was not significantly better than the SRL group, -.1 segments (P=.85) and +.7 correct progressions (P=.58), but the ML group performed the test procedure significantly faster (−73 seconds, P<.001). The ML did not spend significantly longer time training (+5 minutes, P=.84). ML is a very efficient training form in bronchoscopy allowing novice trainees to learn how to perform a thorough, systematic, and quick flexible bronchoscopy. ML does not require longer time spent training compared to SRL.