This study was conducted as a prospective cluster-randomized controlled trial. From January 25 to June 25, 2024, clinicians in the coronary heartdisease ward of Fuwai Hospital, Chinese Academy of Medical Sciences were selected as the research object. Using the method of cluster-randomized allocation, the four wards were randomly assigned 1∶1, with physicians and their medical records assigned to the corresponding group based on the ward. The experimental group utilized the intelligent medical record writing assistant, with 46 physicians included and 4105 medical records collected. The control group used traditional writing methods, with 41 physicians included and 4680 medical records collected. Primary outcome measures included quantitative analysis of medical record writing efficiency and medical record writing quality. Secondary outcomes assessed physicians' satisfaction with the use of the intelligent medical record writing assistant.

The average writing time for discharge records in the experimental group was significantly shorter than that in the control group(5.73 min vs. 8.69 min, P < 0.001). Stratified analysis revealed writing times medical records for patients with ≤5 days and > 5 days of hospitalization(experimental group vs. control group) were 5.59 min vs. 8.34 min and 8.50 min vs. 12.27 min, respectively(all P < 0.001). The rate of quality medical records(evaluated immediately after system submission) was significantly higher in the experimental group than in the control group(94.06% vs. 81.82%, P < 0.001). The questionnaire survey indicated that the vast majority of physicians(80%) used the assistant frequently, and were highly satisfied with the accuracy, completeness, and user experience of the generated information.

The intelligent medical record writing assistant can significantly enhance the writing efficiency and optimize medical record quality concurrently, and physicians are highly satisfied with it. This study validates the effectiveness of the new model of intelligent medical record writing applied to clinical practice, and provides a paradigm for the in-depth application and promotion of this model in the future.