(Poster 46) AI Simulation Tool for Urogynecology and Pelvic Surgery Board Preparation

Objectives: To describe the development of an artificial intelligence (AI)-driven educational platform designed to simulate certifying examinations (CE) in Urogynecology and Reconstructive Pelvic Surgery (URPS), structured around medical knowledge and clinical reasoning based on the AUGS Guide to Learning in URPS, landmark clinical trials, and selected primary sources.

Methods: A generative AI engine was developed to simulate CE-style questioning aligned with the 2024 URPS Fellowship Guide to Learning and the updated ABOG Qualifying Exam Blueprint. Using either user-uploaded or simulated case lists, the AI dynamically generates case-based questions across a range of core topics including pelvic organ prolapse, urinary and fecal incontinence, clinical counseling, perioperative decision-making, management of surgical complications, and complex pelvic floor disorders.

The platform integrates data from high-impact publications as well as recent evidence from AUGS and AUA guidelines. Following each user response, the system provides structured feedback, including sample responses, cited trial statistics, and guideline references. The AI challenges vague or incomplete answers and adapts follow-up questioning based on user reasoning, mimicking real examiner behavior.

Results: The AI generated hundreds of board-style case prompts across the URPS competency spectrum. These incorporated trial-based nuances such as the long-term outcomes from OPTIMAL, considerations from OPUS on prophylactic slings, and ESTEEM’s implications for mixed incontinence management. The engine also leveraged data from newer studies on mesh safety and surgical outcomes to refine recommendations.

The system dynamically adjusted case complexity and prompted users to reevaluate management in the context of changing patient variables (e.g., age, comorbidities, or prior surgical history). Learners were repeatedly challenged to apply evidence-based medicine and demonstrate clinical adaptability. The modular structure supported both self-directed study and structured mock exam sessions.

Conclusions: This AI-based simulation tool provides a realistic, scalable, and guideline-aligned method for CE preparation in URPS. Additional applications include case-based didactic learning at the resident and fellow level with minimal faculty time investment. By integrating landmark trials, simulating examiner behavior, and delivering adaptive feedback, the platform promotes deep learning and clinical reasoning consistent with fellowship training goals. Importantly, this tool offers a more accessible and cost-effective alternative for trainees, addressing the current gap in board preparation resources, which are often limited and financially burdensome.