Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Using large language models as decision support tools in emergency ophthalmology.

Ante Kreso1, Zvonimir Boban2, Sime Kabic1

  • 1University Hospital Split, Department for Ophthalmology, Croatia.

International Journal of Medical Informatics
|March 27, 2025
PubMed
Summary

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Diet, Gut Microbiome, and Microbial Metabolites in Inflammatory Bowel Disease: From Functional Dysbiosis to Precision Nutrition.

International journal of molecular sciences·2026
Same author

Temporal Dynamics of Catestatin Across Gestation: Links to Metabolic and Hemodynamic Adaptation.

Life (Basel, Switzerland)·2026
Same author

Lean Mass and Musculoskeletal Preservation in GLP-1-Based Obesity Treatment: Nutrition, Exercise, Supplementation, and Monitoring Strategies.

Metabolites·2026
Same author

Molecular determinants of thromboinflammatory activation in inflammatory bowel disease.

Journal of molecular medicine (Berlin, Germany)·2026
Same author

Extra-Virgin Olive Oil Phenolics in IBD-Associated Vascular Risk.

Molecules (Basel, Switzerland)·2026
Same author

Natural Bioactive Compounds Targeting Gut Barrier Integrity and Metabolic Endotoxemia in Cardiometabolic Disease: Mechanistic Insights and Translational Perspectives.

Molecules (Basel, Switzerland)·2026
This summary is machine-generated.

Large language models (LLMs) show promise in emergency ophthalmology decision support. GPT-4 and Llama-3-70b performed comparably to human experts in a real-world case study.

Area of Science:

  • Ophthalmology
  • Artificial Intelligence
  • Medical Decision Support

Background:

  • Large language models (LLMs) show potential in medicine.
  • Their utility in emergency ophthalmology decision support is unproven with real-world data.

Purpose of the Study:

  • To evaluate state-of-the-art LLMs (GPT-4, GPT-4o, Llama-3-70b) as decision support tools in emergency ophthalmology.
  • To compare LLM performance against human expert ophthalmologists.

Main Methods:

  • Prospective comparative study involving 73 anonymized emergency ophthalmology cases.
  • LLM-generated diagnoses and treatment plans were assessed against those by certified ophthalmologists.
  • Two independent experts graded LLM and human reports using a 4-point Likert scale.
Keywords:
Artificial intelligenceDecision support systems, clinicalNatural language processingOphthalmology

Related Experiment Videos

Main Results:

  • Human experts achieved a mean score of 3.72.
  • GPT-4 (3.52) and Llama-3-70b (3.48) demonstrated performance comparable to human experts.
  • GPT-4o scored lower (3.20), with statistically significant differences noted between human scores and GPT-4o (P < 0.001).

Conclusions:

  • LLMs show accuracy as decision support tools in emergency ophthalmology.
  • GPT-4 and Llama-3-70b performance was comparable to human ophthalmologists.
  • LLMs hold potential for integration into clinical emergency ophthalmology practice.