EUVIMEDEUVIMED entdecken
BeobachtungsstudieKohortenstudieOpen Access · CC BY 4.0

Evaluation of two large language models for intensive care unit discharge decisions: a prospective observational cohort study

Engin İhsan Turan, Abdurrahman Engin Baydemir, Ebru Kaya, Zehra Polat Turan, Ayça Sultan Şahin

Brazilian Journal of Anesthesiology (English Edition) · 2026-07 · 76 (4) · 844768

DOI: 10.1016/j.bjane.2026.844768 ↗

Kurz gefasst

Beobachtungsstudie in der Zeitschrift Brazilian Journal of Anesthesiology (English Edition), 2026-07. Verfasst von Engin İhsan Turan und 4 weiteren Autor:innen. Beteiligte Einrichtung: Istanbul Health Science University Kanuni Sultan Süleyman Education and Training Hospital, Department of Anesthesiology, Istanbul, Turkey. Electronic…. Hauptthemen laut MEDLINE: Large Language Models, Intensive Care Units, Patient Discharge, Critical Care. Die Arbeit nennt 17 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Background

The aim of this study was to evaluate the effectiveness of two general-purpose Large Language Models (LLMs), ChatGPT and Gemini, in predicting Intensive Care Unit (ICU) discharge decisions (discharge vs. non-discharge). By comparing their outputs with decisions made by ICU physicians, we sought to determine the alignment of AI-generated recommendations with expert clinical judgment and assess their potential as decision-support tools in critical care.

Methods

This prospective observational cohort study was conducted in a tertiary ICU between September 2024 and May 2025. Adult patients (≥ 18 years) requiring ICU discharge decisions were included. Standardized clinical prompts were generated from electronic health records and input into ChatGPT and Gemini. The models’ binary discharge decisions were compared to those of ICU physicians. Model performance was assessed using accuracy, sensitivity, specificity, F1 score, Cohen’s kappa, and McNemar’s test. Discharge was defined as the positive class for all diagnostic performance analyses.

Results

A total of 398 patients were analyzed. ChatGPT demonstrated higher accuracy than Gemini (87.2% vs. 66.3%), with higher sensitivity (85.9% vs. 46.9%) and F1 score (0.890 vs. 0.628), whereas Gemini showed higher specificity (96.2% vs. 89.2%). Agreement with clinician decisions was substantial for ChatGPT (κ = 0.737, p = 0.024) and fair for Gemini (κ = 0.379, p < 0.001). Laboratory markers such as lactate, hemoglobin, and procalcitonin significantly differed between discharged and non-discharged patients.

Conclusion

Large language models may support ICU discharge decisions when guided by structured, guideline-informed prompting. ChatGPT achieved higher overall accuracy, sensitivity, and F1 score, whereas Gemini demonstrated higher specificity.

Trial registration

Externation (Discharge) of ICU, NCT06584890, registered 03 September 2024, prospectively registered, https://register.clinicaltrials.gov/prs/beta/studies/S000EVXZ00000029/recordSummary.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Large Language ModelsIntensive Care UnitsPatient DischargeCritical CareHumansProspective StudiesFemaleMaleCohort StudiesMiddle AgedAgedGenerative Artificial IntelligenceAdultSensitivity and Specificity

Stichwörter der Autor:innen

Artificial intelligenceDecision makingDecision support systems, ClinicalIntensive care unitsNatural language processingPatient discharge

Autor:innen und Institutionen

  • Istanbul Health Science University Kanuni Sultan Süleyman Education and Training Hospital, Department of Anesthesiology, Istanbul, Turkey. Electronic address: enginihsan@hotmail.com.
  • Basaksehir Cam ve Sakura City Hospital, Department of Anesthesiology, Istanbul, Turkey.
  • Istanbul Health Science University Kanuni Sultan Süleyman Education and Training Hospital, Department of Anesthesiology, Istanbul, Turkey.
  • Istanbul Health Science University Sisli Hamidiye Etfal Education and Training Hospital, Department of Anesthesiology, Istanbul, Turkey.

Zitieren

APA

Turan, E. İ., Baydemir, A. E., Kaya, E., Turan, Z. P., & Şahin, A. S. (2026). Evaluation of two large language models for intensive care unit discharge decisions: a prospective observational cohort study. Brazilian Journal of Anesthesiology (English Edition), 76(4), 844768. https://doi.org/10.1016/j.bjane.2026.844768

Vancouver

Turan Eİ, Baydemir AE, Kaya E, Turan ZP, Şahin AS. Evaluation of two large language models for intensive care unit discharge decisions: a prospective observational cohort study. Brazilian Journal of Anesthesiology (English Edition). 2026;76(4):844768. doi:10.1016/j.bjane.2026.844768

BibTeX
@article{Turan2026,
  author = {Turan, Engin İhsan and Baydemir, Abdurrahman Engin and Kaya, Ebru and Turan, Zehra Polat and Şahin, Ayça Sultan},
  title = {Evaluation of two large language models for intensive care unit discharge decisions: a prospective observational cohort study},
  journal = {Brazilian Journal of Anesthesiology (English Edition)},
  year = {2026},
  volume = {76},
  number = {4},
  pages = {844768},
  doi = {10.1016/j.bjane.2026.844768}
}
Kostenlos starten →