Scoping Review Examines LLMs in Health Communication

A scoping review published in the Journal of Medical Internet Research (JMIR) maps the current landscape of large language models (LLMs) applied to health care communication. The review identifies use cases where LLMs are being deployed or studied in clinical and patient-facing communication contexts, documents the challenges practitioners and researchers encounter, and outlines directions for future work.

Health care communication spans a wide range of interactions -- clinician-to-patient explanation, documentation, patient education, care coordination, and clinical training. LLMs have been applied or proposed in multiple of these contexts, though the evidence base remains heterogeneous. A scoping review methodology maps the breadth of published work without requiring quantitative synthesis, making it suited to a rapidly expanding field where studies differ widely in populations, model types, and outcomes measured.

Clinical deployment of LLMs in communication settings raises recurring concerns across the published literature: accuracy and hallucination risk in medical contexts, bias in outputs across demographic groups, regulatory and liability considerations, privacy requirements for patient data, and the lack of standardized evaluation frameworks. The review surfaces these as persistent gaps rather than solved problems.

For teams building or evaluating LLM applications in health care settings, the review provides a structured overview of where work is being done and where evaluation methodology and deployment standards remain underdeveloped. It is most useful as a literature orientation and gap analysis rather than as guidance on specific model choices or deployment architectures.