LLM Agents for Healthcare: Concepts, Methods, and Applications

The rapid advancement of Large Language Models has paved the way for autonomous AI agents capable of complex reasoning, planning, and tool use in specialized domains. In healthcare, these agents offer transformative potential by integrating heterogeneous data sources, including electronic health records, clinical notes, medical images, and genomic profiles and so on, to support clinical decisionmaking and patient care. This paper provides a comprehensive survey of LLM-based agents in healthcare, characterizing their core architectural components: input processing, memory management, reasoning engines, and action modules. We categorize existing methodologies into distinct reasoning paradigms, ranging from simple chain-of-thought processes to complex, tool-augmented, and self-correcting frameworks. Furthermore, we review diverse applications across clinical documentation, diagnostic assistance, and biomedical research. Finally, we discuss the critical challenges of safety, interpretability, and privacy that must be addressed to ensure the responsible deployment of AI agents in clinical workflows.