Multi-Stage Patient Role-Playing Framework for Realistic Clinical Interactions

2026-01-16

AI Summary

This study addresses the lack of realism and diversity in existing Chinese clinical dialogue simulation data, which hinders effective evaluation of large language models (LLMs) in patient role-playing scenarios. To bridge this gap, the authors construct Ch-PatientSim, the first Chinese patient simulation dataset grounded in the Big Five personality dimensions, and propose a novel multi-stage patient role-playing framework that requires no model fine-tuning. By integrating few-shot generation, human verification, and staged dialogue mechanisms, the framework produces personalized and authentic doctor–patient conversations. Experimental results demonstrate that the proposed approach significantly outperforms existing methods across multiple patient simulation dimensions, effectively mitigating the common issues of overly formal responses and insufficient personality expression, thereby enhancing both the realism and linguistic diversity of simulated patient behaviors.

Technology Category

• Cognitive Modeling & Cognitive Systems: Simulating Human Behavior
• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: Conversational AI/Dialog Systems

Application Category

• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Large language models for search

Abstract

The simulation of realistic clinical interactions plays a pivotal role in advancing clinical Large Language Models (LLMs) and supporting medical diagnostic education. Existing approaches and benchmarks rely on generic or LLM-generated dialogue data, which limits the authenticity and diversity of doctor-patient interactions. In this work, we propose the first Chinese patient simulation dataset (Ch-PatientSim), constructed from realistic clinical interaction scenarios to comprehensively evaluate the performance of models in emulating patient behavior. Patients are simulated based on a five-dimensional persona structure. To address issues of the persona class imbalance, a portion of the dataset is augmented using few-shot generation, followed by manual verification. We evaluate various state-of-the-art LLMs and find that most produce overly formal responses that lack individual personality. To address this limitation, we propose a training-free Multi-Stage Patient Role-Playing (MSPRP) framework, which decomposes interactions into three stages to ensure both personalization and realism in model responses. Experimental results demonstrate that our approach significantly improves model performance across multiple dimensions of patient simulation.

Problem

• clinical interaction simulation
• patient role-playing
• LLM evaluation
• dialogue authenticity
• medical education

Innovation

• Multi-Stage Patient Role-Playing
• Ch-PatientSim
• five-dimensional persona
• few-shot generation
• clinical LLMs