Embracing the Future of Medical Education With Large Language Model-Based Virtual Patients: Scoping Review.

Area of Science:

• Medical Education Technology
• Artificial Intelligence in Healthcare
• Virtual Patient Simulation

Background:

• Rapid advancements in large language models (LLMs) are driving innovation in medical education.
• LLM-based virtual patients are emerging as a novel tool for medical simulations.
• This field is experiencing significant growth, with most research published recently.

Purpose of the Study:

• To systematically review current applications of LLM-based virtual patients in medical education.
• To identify emerging research trends and challenges in this domain.
• To explore future development directions for LLM-based virtual patient technology.

Main Methods:

• Adherence to PRISMA-ScR guidelines for a scoping review.
• Systematic search of five major databases (Web of Science, PubMed, IEEE Xplore, Embase, Scopus) from 2018 to 2025.
• Comprehensive analysis of 28 selected studies on LLM-based virtual patients, covering design, application, and evaluation.

Main Results:

• The research on LLM-based virtual patients is nascent, with 92.9% of studies published in the last two years.
• Applications primarily focus on medical training across various disciplines, often integrating with VR/AR technologies.
• Evaluations predominantly focus on user experience, lacking standardized methods and objective outcome measures; privacy/security concerns are noted.

Conclusions:

• LLM-based virtual patients offer significant potential for medical training, particularly in communication skills.
• Current limitations include lack of standardization, absence of nonverbal cues, and privacy/security issues.
• Future research should prioritize enhancing reliability, realism, safety, and scientific efficacy.