¹University Hospital Bonn, Institute for Hygiene and Public Health, Bonn, Germany
²University Hospital Bonn, Institute of Medical Microbiology, Immunology and Parasitology, Bonn, Germany
³University Hospital Bonn, Department of Pediatrics, Bonn, Germany
⁴University Bonn, Medical Faculty, Dean of Studies Office, Bonn, Germany
⁵University Hospital Bonn, Institute of Medical Education, Bonn, Germany

Text

Background: Clinical reasoning (CR) is a critical skill for medical students to develop in order to provide high-quality patient care and reduce error rates. CR is a complex cognitive process involving data collection, cue interpretation, data analysis, planning and implementing interventions, patient communication, and reflection on decision-making. The advent of large language models (LLMs) has opened new possibilities for education. At the Institute for Hygiene and Public Health at the University Hospital Bonn, we are developing a course on CR, specifically focused on antibiotic treatment and antimicrobial resistance, using an inverted classroom format and LLMs to enhance students’ reasoning skills.

Methods: In this e-learning course, students are introduced to various reasoning strategies and common biases. Virtual Patients (VPs), which are effective tools for teaching CR, have been integrated into the course. Human-curated VPs are hosted on the eCampus platform, allowing students to practice their skills. We use LLMs to generate virtual patient vignettes for student interaction. These AI-generated vignettes are reviewed by an interdisciplinary expert panel. The chat capabilities of LLMs are leveraged to simulate Patient-Doctor and Doctor-Doctor interactions, which are crucial for clinical reasoning. A feasibility study is being conducted to evaluate whether LLMs can function as tutors, testing students’ reasoning abilities, providing feedback, and assessing their performance.

Results: The development of the course began in 2024. Initial results and materials will be presented at the GMA annual conference in 2025.

Discussion: LLMs have the potential to transform clinical reasoning education by simulating realistic medical interactions and providing personalized feedback, thereby enhancing decision-making skills in a scalable and adaptive manner. This approach offers a pathway for medical education to better meet the needs of individual learners, especially in complex decision-making contexts like clinical reasoning