Evaluating the accuracy of a state-of-the-art large language model for prediction of admissions from the emergency room
Journal of the American Medical Informatics Association · 2024
DOI: 10.1093/jamia/ocae103 ↗Kurz gefasst
Evaluationsstudie in der Zeitschrift Journal of the American Medical Informatics Association, 2024. Verfasst von Benjamin S Glicksberg und 11 weiteren Autor:innen. Beteiligte Einrichtung: Division of Data-Driven and Digital Medicine, Department of Medicine, Icahn School of Medicine at Mount Sinai, New York, NY 10029, United States. Hauptthemen laut MEDLINE: Emergency Service, Hospital, Electronic Health Records, Patient Admission. Die Arbeit wurde bisher 84-mal zitiert (Crossref) und nennt 15 Literaturangaben. Der Volltext ist frei in PubMed Central verfügbar. Gefördert von: NHLBI NIH HHS, National Heart Lung and Blood Institute, NIH HHS.
Abstract (Originalsprache)
Background
Artificial intelligence (AI) and large language models (LLMs) can play a critical role in emergency room operations by augmenting decision-making about patient admission. However, there are no studies for LLMs using real-world data and scenarios, in comparison to and being informed by traditional supervised machine learning (ML) models. We evaluated the performance of GPT-4 for predicting patient admissions from emergency department (ED) visits. We compared performance to traditional ML models both naively and when informed by few-shot examples and/or numerical probabilities.
Methods
We conducted a retrospective study using electronic health records across 7 NYC hospitals. We trained Bio-Clinical-BERT and XGBoost (XGB) models on unstructured and structured data, respectively, and created an ensemble model reflecting ML performance. We then assessed GPT-4 capabilities in many scenarios: through Zero-shot, Few-shot with and without retrieval-augmented generation (RAG), and with and without ML numerical probabilities.
Results
The Ensemble ML model achieved an area under the receiver operating characteristic curve (AUC) of 0.88, an area under the precision-recall curve (AUPRC) of 0.72 and an accuracy of 82.9%. The naïve GPT-4's performance (0.79 AUC, 0.48 AUPRC, and 77.5% accuracy) showed substantial improvement when given limited, relevant data to learn from (ie, RAG) and underlying ML probabilities (0.87 AUC, 0.71 AUPRC, and 83.1% accuracy). Interestingly, RAG alone boosted performance to near peak levels (0.82 AUC, 0.56 AUPRC, and 81.3% accuracy).
Conclusions
The naïve LLM had limited performance but showed significant improvement in predicting ED admissions when supplemented with real-world examples to learn from, particularly through RAG, and/or numerical probabilities from traditional ML models. Its peak performance, although slightly lower than the pure ML model, is noteworthy given its potential for providing reasoning behind predictions. Further refinement of LLMs with real-world data is necessary for successful integration as decision-support tools in care settings.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Division of Data-Driven and Digital Medicine, Department of Medicine, Icahn School of Medicine at Mount Sinai, New York, NY 10029, United States.
- Institute for Healthcare Delivery Science, Icahn School of Medicine at Mount Sinai, New York, NY 10029, United States.
- The Charles Bronfman Department of Personalized Medicine, Icahn School of Medicine at Mount Sinai, New York, NY 10029, United States.
- Department of Emergency Medicine, Icahn School of Medicine at Mount Sinai, New York, NY 10029, United States.
- Department of Population Health Science and Policy, Icahn School of Medicine at Mount Sinai, New York, NY 10029, United States.
ORCID-Profile: Benjamin S Glicksberg, Ashwin Sawant, Donald Apakama, Robert Freeman
Förderung
- NHLBI NIH HHS
- National Heart Lung and Blood Institute
- NIH HHS
- NIH
Zitieren
APA
Glicksberg, B. S., Timsina, P., Patel, D., Sawant, A., Vaid, A., Raut, G., Charney, A. W., Apakama, D., Carr, B. G., Freeman, R., Nadkarni, G. N., & Klang, E. (2024). Evaluating the accuracy of a state-of-the-art large language model for prediction of admissions from the emergency room. Journal of the American Medical Informatics Association. https://doi.org/10.1093/jamia/ocae103
Vancouver
Glicksberg BS, Timsina P, Patel D, Sawant A, Vaid A, Raut G, et al. Evaluating the accuracy of a state-of-the-art large language model for prediction of admissions from the emergency room. Journal of the American Medical Informatics Association. 2024. doi:10.1093/jamia/ocae103
BibTeX
@article{Glicksberg2024,
author = {Glicksberg, Benjamin S and Timsina, Prem and Patel, Dhaval and Sawant, Ashwin and Vaid, Akhil and Raut, Ganesh and Charney, Alexander W and Apakama, Donald and Carr, Brendan G and Freeman, Robert and Nadkarni, Girish N and Klang, Eyal},
title = {Evaluating the accuracy of a state-of-the-art large language model for prediction of admissions from the emergency room},
journal = {Journal of the American Medical Informatics Association},
year = {2024},
doi = {10.1093/jamia/ocae103}
}
EUVIMED