Large Language Models in Medicine: Opportunities, Limitations, and Future Directions – A Scoping Review

Abstract

Introduction. Large language models (LLMs) are increasingly used in healthcare and health-related fields, including clinical decision support, personalized medicine, digital health, and sports performance monitoring. However, concerns regarding reliability, transparency, and safety continue to limit their broader implementation.

Materials and Methods. A scoping review of 24 PubMed-indexed studies published between 2023 and 2026 was conducted to assess current applications, benefits, limitations, and future directions of LLMs in healthcare.

Results. LLMs are being applied in clinical decision support, diagnostics, biomedical research, personalized healthcare, and digital health solutions. They can enhance information retrieval, clinical reasoning, literature synthesis, health monitoring, and sports-related applications such as training support and performance optimization. Emerging trends include retrieval-augmented generation, multimodal systems, and specialized AI agents. Key challenges include hallucinations, limited explainability, privacy concerns, and insufficient clinical validation.

Conclusions. LLMs have significant potential to improve healthcare, biomedical research, and health-oriented sports applications. Currently, they should serve as supportive tools rather than autonomous decision-makers. Further progress requires robust validation, regulatory oversight, and effective human–AI collaboration.