Large Language Models in Healthcare and Medical Applications: A Review

Area of Science:

• Medical Informatics
• Artificial Intelligence in Medicine
• Natural Language Processing

Background:

• Large language models (LLMs) offer advanced natural language processing (NLP) capabilities with significant potential to revolutionize healthcare.
• Current LLM applications span clinical decision support, medical education, diagnostics, and patient care.

Purpose of the Study:

• To systematically examine the potential and challenges of LLMs in the healthcare domain.
• To provide a comprehensive understanding of LLM background, evolution, architecture, and multimodal capabilities in medicine.

Main Methods:

• Exploration of domain-specific data acquisition, large-scale pre-training, and supervised fine-tuning.
• Analysis of prompt engineering and in-context learning for healthcare use cases.
• Categorization of prominent application areas and examination of methodological aspects.

Main Results:

• LLMs demonstrate promising applications across various medical fields, enhancing capabilities in decision support, education, diagnostics, and patient care.
• Identified key trends and prominent application areas for LLMs in medicine.
• Highlighted critical challenges including privacy, ethical deployment, factual accuracy, model bias, interpretability, and regulatory compliance.

Conclusions:

• Responsible integration of LLMs in healthcare requires addressing significant technical and social challenges.
• Future research should focus on emerging directions and strategic recommendations for developing and deploying healthcare LLMs ethically and effectively.