Syndromic Analysis of Sepsis Cohorts Using Large Language Models

Can large language models (LLMs) accurately extract presenting signs and symptoms from clinical notes to identify associations between symptoms, multidrug-resistant infections, and in-hospital mortality in large cohorts of patients with possible sepsis?In this cohort study of 104 248 patients with possible infection, LLMs extracted signs and symptoms from admission notes with accuracy comparable to that of physicians performing manual medical records review. Hierarchical clustering identified 7 symptom-based syndromes that correlated with infection sources, risk for methicillin-resistant and multidrug-resistant gram-negative organisms, and in-hospital death.Findings of this study suggest that LLMs can enable the efficient, large-scale extraction of signs and symptoms from clinical notes and the differential correlation of syndromes with infection sources, multidrug-resistant infections, and mortality; the value of large-scale sign-and-symptom data in models of antibiotic choice, effectiveness, and outcomes in patients with sepsis warrants further study.

Corresponding Author: Theodore R. Pak, MD, PhD, Department of Population Medicine, Harvard Medical School and Harvard Pilgrim Healthcare Institute, 401 Park Dr, Ste 401 E, Boston, MA 02215 (tpak@mgh.harvard.edu).

Link to the article in your story

We encourage you to link out to this article in your story using the link below. It includes an access token that will give free access to the article for your readers up to one year after publication. (The link will be live after the article publishes and embargo is lifted.)

Please see the article for additional information, including full author list, author contributions and affiliations, conflict of interest and financial disclosures, and funding and support.

Need more information? Contact us.