2025 年 45 巻 3 号 p. 143-153
Background : Physician shortages and limited medical access contribute to healthcare disparities. Artificial intelligence (AI) offers a potential solution, but existing AI consultation systems lack the clinical accuracy needed for safe implementation. This study evaluates an AI Medical Consultation System using ChatGPT's GPTs and Python.
Methods : The AI Medical Consultation system was designed with structured medical questions covering 390 clinical symptoms. Four chief complaints were selected for each of four major medical domains : respiratory, circulatory, digestive, and nervous systems. The AI conducted consultations and generated structured medical histories, which were compared to actual transcripts. Performance was evaluated based on completeness, accuracy, redundancy, classification appropriateness, and terminological consistency.
Results : The AI Medical Consultation system provided comprehensive consultations with high accuracy in structured history generation. However, it frequently omitted negative findings and had minor inaccuracies in symptom duration and frequency. Despite these limitations, it effectively categorized information, minimized redundancy, and maintained consistent terminology across domains, indicating its potential to reduce healthcare disparities.
Conclusions : The AI Medical Consultation System enhances efficiency, reduces physician workload, and expands access to care. This case study highlights the potential of AI-driven consultations as a step toward equitable healthcare.
