Improving postsurgical fall detection for older Americans using LLM-driven analysis of clinical narratives.

Autor:	Pillai M; Veterans Affairs Palo Alto Health Care System, Palo Alto, California, USA.; Department of Medicine (Biomedical Informatics), Stanford University, Stanford, CA, USA., Blumke TL; National Center for Collaborative Healthcare Innovation, Palo Alto VA Healthcare System, Palo Alto, CA, USA., Studnia J; Institute for Computational and Mathematical Engineering, Stanford University, Stanford, CA, USA., Wang Y; Department of Medicine (Biomedical Informatics), Stanford University, Stanford, CA, USA., Veigulis ZP; Claimable Inc., Sacramento, CA, USA., Ware AD; National Center for Collaborative Healthcare Innovation, Palo Alto VA Healthcare System, Palo Alto, CA, USA., Hoover PJ; National Center for Collaborative Healthcare Innovation, Palo Alto VA Healthcare System, Palo Alto, CA, USA., Carroll IR; Department of Anesthesiology, Stanford University School of Medicine, Stanford, CA, USA., Humphreys K; Department of Psychiatry and the Behavioral Sciences, Stanford University, Stanford, CA USA., Osborne TF; National Center for Collaborative Healthcare Innovation, Palo Alto VA Healthcare System, Palo Alto, CA, USA.; Department of Radiology, Stanford University School of Medicine, Stanford, CA, USA., Asch SM; Veterans Affairs Palo Alto Health Care System, Palo Alto, California, USA.; Division of Primary Care and Population Health, Stanford University School of Medicine, Stanford, CA, USA., Hernandez-Boussard T; Department of Medicine (Biomedical Informatics), Stanford University, Stanford, CA, USA., Curtin CM; Veterans Affairs Palo Alto Health Care System, Palo Alto, California, USA.; Division of Plastic and Reconstructive Surgery, Department of Surgery, Stanford University School of Medicine, Stanford, CA, USA.
Jazyk:	angličtina
Zdroj:	MedRxiv : the preprint server for health sciences [medRxiv] 2024 Jun 26. Date of Electronic Publication: 2024 Jun 26.
DOI:	10.1101/2024.06.25.24309480
Abstrakt:	Postsurgical falls have significant patient and societal implications but remain challenging to identify and track. Detecting postsurgical falls is crucial to improve patient care for older adults and reduce healthcare costs. Large language models (LLMs) offer a promising solution for reliable and automated fall detection using unstructured data in clinical notes. We tested several LLM prompting approaches to postsurgical fall detection in two different healthcare systems with three open-source LLMs. The Mixtral-8×7B zero-shot had the best performance at Stanford Health Care (PPV = 0.81, recall = 0.67) and the Veterans Health Administration (PPV = 0.93, recall = 0.94). These results demonstrate that LLMs can detect falls with little to no guidance and lay groundwork for applications of LLMs in fall prediction and prevention across many different settings. Competing Interests: Competing Interests None declared.
Databáze:	MEDLINE
Externí odkaz:	Zobrazit plný text záznamu