Current Landscape and Future Directions Regarding Generative Large Language Models in Stroke Care: Scoping Review

Area of Science:

• Artificial Intelligence in Medicine
• Neurology
• Health Informatics

Background:

• Stroke significantly impacts global health, causing disability and straining healthcare resources.
• Generative large language models (gLLMs) offer potential solutions for stroke care challenges.
• A comprehensive review of gLLM applications and performance in stroke care is needed.

Purpose of the Study:

• To consolidate evidence on gLLM-based interventions in stroke care.
• To examine the current landscape, limitations, and future directions of gLLM applications in stroke.
• To identify shortcomings in the design, reporting, and evaluation of these interventions.

Main Methods:

• Scoping review adhering to PRISMA-ScR and PCC framework.
• Searched 6 major scientific databases in December 2024 for gLLM interventions in stroke care.
• Mapped key characteristics and outcomes of included studies.

Main Results:

• 25 studies were included, with retrospective designs predominating (64%).
• Key gLLM applications: clinical decision support (40%), administrative assistance (36%), patient interaction (20%).
• Identified challenges: factual alignment, robustness, interpretability, efficiency, and clinical adoption.

Conclusions:

• gLLM applications in stroke care are nascent, with most being early-stage implementations.
• Critical gaps exist in research and clinical translation.
• An actionable framework is proposed for developing impactful and trustworthy AI tools in stroke care.