Utility of Generative Artificial Intelligence for Japanese Medical Interview Training: Randomized Crossover Pilot...

Area of Science:

• Medical Education
• Artificial Intelligence
• Clinical Skills Training

Background:

• Generative artificial intelligence (AI) is increasingly explored for medical education.
• Its application in culturally specific contexts like Japanese medical interview training is underexplored.
• This study investigates generative AI's utility for training physicians in Japan.

Purpose of the Study:

• To evaluate generative AI as a medical interview training tool.
• To compare AI-based training with traditional face-to-face methods.
• To assess AI's effectiveness in a Japanese medical context.

Main Methods:

• A randomized crossover pilot study with 20 physicians.
• Comparison of AI-based (GPT platform) and traditional (simulated patient) interview stations.
• Evaluations in Japanese using 6 metrics, including clinical reasoning and communication.

Main Results:

• AI-based stations scored lower in patient care and communication (P=.009).
• AI-based stations showed comparable performance in clinical reasoning (P=.10).
• Traditional methods outperformed AI in interpersonal aspects of the interview.

Conclusions:

• Generative AI demonstrates potential as a supplementary tool for clinical reasoning practice.
• AI enables self-directed learning, offering an advantage over traditional methods.
• Hybrid training models integrating AI and traditional approaches are recommended for comprehensive medical interview training in Japan.