G-3098

Development and Evaluation of an Al-Based Adaptive Training Platform for Enhancing Clinical Communication Skills among Medical Trainees.

 Seyedeh Bahar Shahidi Marnani ¹ ℗, Farahnaz Shahdadian ¹, Dr Arezoo Vasili ² ©   

 
Abstract

Introduction: Effective clinical communication is a key part of medical practice, strongly influencing both patient outcomes and therapeutic relationships. Despite its importance, many medical education systems lack structured ways to teach these skills. This study aims to address that gap by developing and applying an AI-powered training platform to improve communication training for medical trainees. Methods and Materials: We conducted a quasi-experimental study at Isfahan University of Medical Sciences with 32 medical trainees and 150 patient encounters. The training platform had three main parts: (1) a scenario bank of 150 validated, culturally adapted communication cases, (2) an adaptive learning engine using natural language processing (NLP), sentiment analysis, and both supervised and reinforcement learning to give real-time feedback, and (3) a multidimensional assessment tool to evaluate five key areas of communication. For outcome measurement, we used three standardized tools: patient satisfaction surveys (Cronbach's α=0.82), objective structured clinical examinations (OSCEs), and structured faculty assessments. Statistical analysis was done with SPSS version 26, using paired t-tests, multivariate regression to control for confounders, and inter-rater reliability checks. Results: After using the platform, we observed clear improvements. Patient satisfaction scores increased by 13.7% (87.5±4.10 vs. 81.7±8.76; p=0.02). OSCE communication scores improved by 15.2%. In specific skill areas, clinical encounter structuring scored 89.2±3.85, treatment explanation 88.7±3.90, and empathic response 82.3±6.15. The platform covered 92% of common clinical scenarios and cut down skill acquisition time by 40% compared to traditional methods. User satisfaction averaged 4.6 out of Conclusion and Discussion: This study highlights the value of AI-based platforms in communication training. Personalized learning paths, instant feedback, and culturally relevant scenarios helped learners improve more effectively. The assessment framework also brought more objectivity to evaluating skills. This model can be scaled for other clinical training needs and shows promise for improving how communication is taught in medicine.

Keywords: Clinical Communication; Artificial Intelligence; Medical Education; Natural Language Processing; Computer-Assisted