Back

Leveraging Large Language Models to Identify In-Hospital Cardiac Arrest

Vo, J.; Weissenbacher, D.; Reinier, K.; Chugh, H.; Uy-Evanado, A.; O'COnnor, K.; Gonzalez-Hernandez, G.; Chugh, S. S.

2025-10-17 cardiovascular medicine
10.1101/2025.10.16.25338028 medRxiv
Show abstract

Manual chart abstraction is the gold standard for identifying in-hospital cardiac arrest (IHCA) but is resource-intensive. Diagnosis codes are a widely used alternative given their accessibility and automated nature, but this method has poor sensitivity and positive predictive value. We present a novel large language model (LLM) approach to identify IHCA and location, highlighting the potential of LLMs for rapid, accurate, and automated IHCA identification from clinical notes.

Published in JACC: Clinical Electrophysiology · training set

Matching journals

The top 4 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.