J.J. Borckardt, D.H. Henninger, K. Bath, D.A. Annan-Coultas, L. Langdale, K. Kascak, C. Pelic, M. DeArellano, K. Brady
Medical University of South Carolina and Figment Learning Labs,
United States
Keywords: AI, healthcare, education, simulation, controlled trial, STTR, NIH
Summary:
Purpose: Artificial intelligence (AI)–powered virtual patient simulations offer a promising new approach to clinical education by combining adaptive interactivity with standardized case delivery. This NIH STTR-sponsored randomized controlled trial evaluated whether an AI-driven simulation platform (Figment Learning Labs Virtual Clinic) could improve learner knowledge, performance, and engagement compared with traditional written case studies. Methods: Health professional students from medicine, nursing, dentistry, pharmacy, and allied health programs (N = 58) were randomized (1:1) to complete either three written case studies or three AI-powered virtual patient encounters after viewing a standardized 15-minute instructional video on pain management. Knowledge (8-item test), case discussion accuracy, and engagement were assessed pre- and post-intervention using mixed ANOVA and independent-samples t tests. Results: Knowledge improved across conditions, F(1,56)=36.32, p<.001, with a significant time × group interaction, F(1,56)=4.39, p=.041, indicating greater gains for the AI simulation group. Virtual patient participants demonstrated higher case discussion accuracy (M = 0.78 vs. M = 0.65; p<.001) and greater engagement on 13 of 17 items (p<.05), including perceived realism, enjoyment, and clinical relevance. Conclusions: AI-enabled virtual patients outperformed static case studies in promoting learning and engagement, suggesting that AI simulation can extend the benefits of traditional simulation-based training with enhanced scalability, interactivity, and standardization. Integrating AI-driven virtual patients into curricula may represent a next-generation model for competency-based health professions education.