The increasing documentation workload in medical practice, particularly for clinical notes, has driven the development of AI-driven solutions. This study introduces an AI Doctor Assistant (DA) that generates drafts of outpatient progress notes. The DA focuses on two main tasks: generating the operation summary section and the clinical findings section of progress notes. Using datasets from four medical specialties, this study compares the performance of individually fine-tuned models for each professor with a general model trained on data from all professors. Experimental results comparing individual and general models revealed that the general model's performance varied depending on the task type. These findings underline the AI DA's potential to reduce documentation burdens and improve generalized ability across diverse clinical scenarios.
Yoo et al. (Thu,) studied this question.