A Methodological Framework for Designing Clinically Coherent Generative AI Simulated Patients: A Physiotherapy Education Case Study

Generative artificial intelligence (AI) offers novel opportunities for scalable health professions education. However, several AI-driven simulated patient initiatives rely on ad hoc approaches. This study describes the development and evaluation of a methodological framework for designing generative AI mock patients. Using a physiotherapy education chatbot as an exemplar, a framework is presented integrating clinical reasoning and communication, layered prompt architecture, scenario constraints, and guardrails. Results suggest that the framework supported consistent patient-persona behaviour and scenario progression, while also revealing select interaction challenges. Mitigation strategies are described. This work contributes to the health informatics education community by offering a practical methodology for the evidence-based design of generative AI simulations in health professions education.