Unveiling the Potential of Large Language Models in Transforming Chronic Disease Management: Mixed Methods Systematic...

Area of Science:

• Artificial Intelligence in Healthcare
• Digital Health
• Chronic Disease Management

Background:

• Chronic diseases represent a significant global health challenge, causing a majority of worldwide deaths.
• Large language models (LLMs) offer potential for optimizing chronic disease management, but evidence is limited.

Purpose of the Study:

• To review the feasibility, opportunities, and challenges of using LLMs in chronic disease management.
• To synthesize evidence across the spectrum from prevention to long-term care.

Main Methods:

• Systematic review following PRISMA guidelines, searching 11 databases.
• Included intervention and simulation studies on LLMs for chronic diseases.
• Evaluated methodological quality and conducted meta-analyses for feasibility.

Main Results:

• 20 studies examined LLMs for cancer, cardiovascular, and metabolic disorders.
• LLMs demonstrated feasibility, generating accurate health recommendations (71% accuracy).
• Retrieval-augmented LLMs showed higher accuracy than general-purpose LLMs.

Conclusions:

• LLMs show potential to transform chronic disease management at multiple levels.
• Clinical application is nascent; requires focus on data security, fine-tuning, and integration with wearables.
• LLMs can become valuable adjuncts for healthcare professionals with further development.