Evidence Hub

Efficacy of large language models in detecting postoperative delirium from unstructured clinical notes: A retrospective cohort study

Da-In Eun, Hyung-Chul Lee, Gang Heo, J. Jeong, Jipyeong Lee, Hyeonhoon Lee, Sang-Youn Park, K. Nam et al.

npj Digital Medicine, 2025

Abstract

Early identification of postoperative delirium (POD) remains challenging. This retrospective observational study compared the performance of large language models (LLMs), Llama-3-70B and GPT-4o, and physicians in predicting clinically significant POD, defined as either requiring antipsychotics or diagnosis of delirium by neurologists following consultation for delirium-related symptoms. The c-statistics of Llama-3-70B and GPT-4o were 0.74 and 0.76, respectively. LLMs showed higher sensitivity (Llama-3-70B, 0.900; GPT-4o, 0.868; physicians, 0.723) and lower specificity (0.463, 0.547, and 0.814, respectively) than physicians. Inter-rater agreement was almost perfect for both Llama-3-70B and GPT-4o (Fleiss’ kappa = 0.852 and 0.854, respectively) but fair for physicians (0.219). Both LLMs detected clinically significant POD approximately one day earlier than physicians (Kaplan-Meier analysis, median time to diagnosis: Llama-3-70B, 34.5 h; GPT-4o, 37.5 h; physicians, 62.9 h; log-rank P < 0.001). The integration of LLMs as a complementary screening tool under physician supervision may improve the early, reproducible diagnosis of clinically significant POD.

How strong is the evidence?

Evidence Hub reads this as a cohort study, done in people. Its evidence level is 3 of 5 (Moderate · trial / cohort / secondary analysis). Trials without a large randomized design, cohorts and secondary analyses give moderate evidence.

How the five levels of evidence work. The level is assigned automatically from the abstract and can be wrong, so check the paper.

Read the paper

Related studies

Go deeper in Evidence Hub

Evidence Hub adds a study data sheet (design, sample size, evidence level, key numbers), an AI summary and a chat grounded in the paper's text.

Open this study Sign up free