Artificial Intelligence-Enhanced Colposcopy Training Improves Diagnostic Reasoning and Communication Competencies: A Randomized Controlled Study.

Published in Journal of gynecology obstetrics and human reproduction. Abstract: Cervical cancer remains a major global health issue, and colposcopy training struggles with standardization and scalability. AI has potential to improve medical education, but its effectiveness in colposcopy training remains underexplored. To evaluate the efficacy of an artificial intelligence (AI)-driven virtual patient platform for colposcopy training among resident physicians. In this randomized, assessor-blinded, wait-list-controlled trial, 60 obstetrics and gynecology residents (PGY-2/3) from two affiliated teaching hospitals were enrolled (May 2023-September 2025). Participants were randomly assigned to either the Immediate AI Intervention group (traditional teaching plus an AI platform throughout) or the Wait-List Control group (traditional teaching only during Month 1, followed by the AI platform). The primary outcome was the between-group difference in the Case Analysis score at 1 month. Secondary outcomes included theoretical knowledge, communication performance, diagnostic justification, and operative competence assessed via OSCE. At 1 month, the Immediate AI group showed superior performance in Case Analysis (78.57 ± 3.72 vs 71.03 ± 5.82, mean difference = 7.54, p<0.001, Cohen's d=1.54), Diagnostic Justification (81.63 ± 6.21 vs 74.13 ± 6.77, p<0.001, Cohen's d=1.15), Communication Performance (80.97 ± 3.90 vs 75.77 ± 6.08, p<0.001, Cohen's d=1.02), and Theoretical Knowledge (84.67 ± 6.27 vs 80.43 ± 5.07, p=0.006, Cohen's d=0.74). Operative competence showed no si...