Large language models (LLMs) have emerged as significant tools in healthcare, demonstrating their potential for clinical reasoning and patient care. This survey examines recent advancements in medical LLMs, focusing on reasoning applications and requirements. We adopt a dual-view approach that connects clinical practice with computational methods.
On the clinical side, we establish a five-level competency scheme following Miller's Pyramid, progressing from knowledge recall to dynamic case management. On the computational side, we link deductive, inductive, and abductive reasoning patterns to common medical goals and tasks.
Additionally, we introduce a benchmark dataset spanning five levels of medical reasoning capability and report results on 18 state-of-the-art models. The findings reveal that medical specialist models excel in diagnosis-centric tasks, while general models lead in decision support and dialogue.
In conclusion, we discuss current progress and open challenges, including data limitations, hallucination, and grounding issues, and outline directions toward safer, more reliable, and workflow-ready systems.
Blogger's Review: This survey provides a deep dive into the application of LLMs in the medical field, emphasizing the critical need to align clinical demands with computational capabilities. While progress has been made, key issues such as data integrity and model reliability must be addressed to advance the practical use of AI in healthcare.