ChatGPT as a Virtual Patient: Written Empathic Expressions During Medical History Taking

Area of Science:

• Medical Education Technology
• Artificial Intelligence in Healthcare
• Clinical Skills Training

Background:

• Virtual patients are established tools for teaching medical history taking.
• ChatGPT has been increasingly adopted in various medical education contexts.
• The potential of AI chatbots to foster specific clinical skills like empathy is an emerging area of research.

Purpose of the Study:

• To evaluate ChatGPT's efficacy in training empathic history-taking skills among medical students.
• To assess whether using ChatGPT for history-taking exercises impacts students' subjective autonomy.
• To explore the potential of ChatGPT as a novel tool in medical pedagogy.

Main Methods:

• Third-year medical students conducted history-taking sessions with ChatGPT 3.5 using a cardiology-focused prompt.
• Student autonomy was measured via a post-interaction questionnaire.
• Chat transcripts were analyzed using the Empathic Communication Coding System to quantify empathic opportunities and student responses.

Main Results:

• Of 659 interactions, 93 presented empathic opportunities, primarily initiated by ChatGPT reporting emotional content.
• Students demonstrated implicit recognition of patient perspectives, though occasional inadequate responses were noted.
• Students reported a high degree of experienced subjective autonomy during the interactions.

Conclusions:

• Preliminary findings suggest ChatGPT can serve as a virtual patient for empathic history-taking training.
• ChatGPT shows potential as a supplementary tool alongside traditional simulated patient programs.
• Medical institutions may consider integrating ChatGPT into curricula, potentially using methods like the flipped classroom model.