EUVIMEDEUVIMED entdecken
Crossref

Feasibility and Acceptability of a Large Language Model-Based Electronic Health Record Summarization Tool in Inpatient Care

Kevin D. Smith, Rebecca S. Silverman, Heather L. Himelhoch, Cheng-Kai Kao

Applied Clinical Informatics · 2026-10 · 17 (05) · 1033-1041

DOI: 10.1055/a-3001-9600 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Applied Clinical Informatics, online erschienen am 20.11.2026. Verfasst von Kevin D. Smith und 3 weiteren Autor:innen. Die Arbeit nennt 14 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Current health care practice requires providers to manually navigate through different electronic health record (EHR) sections to perform chart review and construct a complete picture of the patient’s condition. This process is both time-consuming and inefficient. This initiative evaluated the feasibility, acceptability, and perceived usefulness of a large language model (LLM)-based EHR summarization tool in streamlining inpatient care delivery by aggregating and synthesizing patient data into concise summaries. A three-phase evaluation was conducted involving 32 providers across seven specialties, using a tool called “Inpatient Insights Summary” in the Epic Systems Hyperdrive EHR product. The tool summarizes data in three summary types: “recent events” (discrete data), “recent notes” (clinical notes), and “patient story” (longitudinal narrative). Feedback was collected through embedded and external forms, including a modified Physician Documentation Quality Instrument (PDQI-9). The tool generated summaries that clinicians rated as accurate and comprehensive; clinicians detected hallucinations in 6.2% of summaries during routine use. No significant differences in user feedback were found across summary types. Statistically significant differences in feedback were noted between the proof-of-concept and pilot phases for thoroughness, usefulness, organization, and clarity, indicating improved summaries through prompt engineering refinement. These findings suggest that LLM-based EHR summarization tools can produce summaries that clinicians rate as accurate and comprehensive, with hallucinations infrequently detected, and may support clinical workflow efficiency. Further research is needed to validate these findings in larger settings and assess the effect on clinical practice and patient care.

Zitieren

APA

Smith, K. D., Silverman, R. S., Himelhoch, H. L., & Kao, C. (2026). Feasibility and Acceptability of a Large Language Model-Based Electronic Health Record Summarization Tool in Inpatient Care. Applied Clinical Informatics, 17(05), 1033-1041. https://doi.org/10.1055/a-3001-9600

Vancouver

Smith KD, Silverman RS, Himelhoch HL, Kao C. Feasibility and Acceptability of a Large Language Model-Based Electronic Health Record Summarization Tool in Inpatient Care. Applied Clinical Informatics. 2026;17(05):1033-1041. doi:10.1055/a-3001-9600

BibTeX
@article{Smith2026,
  author = {Smith, Kevin D. and Silverman, Rebecca S. and Himelhoch, Heather L. and Kao, Cheng-Kai},
  title = {Feasibility and Acceptability of a Large Language Model-Based Electronic Health Record Summarization Tool in Inpatient Care},
  journal = {Applied Clinical Informatics},
  year = {2026},
  volume = {17},
  number = {05},
  pages = {1033--1041},
  doi = {10.1055/a-3001-9600}
}
Kostenlos starten →