We evaluated the association of an EHR-integrated LLM note summarization tool with emergency physician medical record review time and user experience.
We conducted a prospective, within-participant crossover time-motion quality improvement study in a single urban quaternary-care academic ED from December 2024 through July 2025.
Ten board-certified emergency physicians were each observed during 3 artificial intelligence (AI)-enabled shifts and 3 control shifts.
During AI-enabled shifts, physicians used an EHR-integrated LLM-based note summarization tool that generated concise summaries from unstructured clinical notes within the EHR.
In this quality improvement study, we did not detect a statistically significant difference in total or per-patient medical record review time with LLM-based note summarization.
Together, these findings highlight the complexity of evaluating workflow tools in dynamic clinical environments such as the ED, where perceived improvements may reflect factors not captured by time-based metrics such as improved information salience and reduced cognitive burden.
Greater objective efficiency gains may emerge as clinicians gain familiarity with the tool or as more advanced implementations improve integration and performance.
【MORE】