Evaluation of a Medical Interview-Assistance System Using Artificial Intelligence for Resident Physicians...

Area of Science:

• Medical Informatics
• Clinical Decision Support Systems
• Artificial Intelligence in Healthcare

Background:

• Artificial intelligence (AI) is poised to revolutionize medical interviews, yet its practical application and benefits in Japan remain largely unassessed.
• Current AI-based systems for medical interview support are not widely adopted in Japan, necessitating research into their efficacy.

Purpose of the Study:

• To evaluate the usefulness of a commercial AI-powered medical interview support system.
• To compare the diagnostic accuracy, interview duration, and questioning efficiency of resident physicians with and without AI assistance.

Main Methods:

• A randomized controlled trial involving 20 resident physicians was conducted across two trials.
• Participants were divided into two groups: one receiving AI support via a flowchart application and one without.
• Key metrics included correct diagnosis rates, time to complete interviews, and the number of questions asked.

Main Results:

• The AI-assisted group demonstrated a significantly higher rate of correct diagnoses compared to the control group (0.561 vs. 0.393; p = 0.02).
• A significant reduction in overall consultation time was observed in the AI-assisted group (370s vs. 390s; p = 0.04).
• Data from 192 differential diagnoses were analyzed.

Conclusions:

• AI-assisted medical interviews enhance diagnostic accuracy and efficiency for resident physicians.
• The integration of AI systems into clinical practice has the potential to elevate the standard of medical care.
• Further adoption of AI tools could address current limitations in medical interview support systems.