NLP in health education: Review of 64 studies finds promise and gaps

A scoping review published in arXiv and accepted at IEEE ICHI 2026 systematically maps how natural language processing (NLP) and AI are transforming health professions education. The authors—Javad Mohammad Alizadeh, Pegah Saebi, Mukesh Kumar Patel, and Huanmei Wu from Indiana University—followed the Arksey and O'Malley framework and PRISMA-ScR guidelines, analyzing 64 studies published between 2015 and 2026. They searched databases including PubMed, ERIC, IEEE Xplore, and Google Scholar.

Seven thematic domains emerged: automated assessment (e.g., grading essays), large language models (LLMs) as student-facing tutors, virtual patients and clinical simulation, curriculum analysis and program evaluation, personalized adaptive learning, public health and health promotion education, and educator/institutional integration. The review found significant technical promise—particularly in automated assessment, clinical simulation, and curriculum analysis—but also persistent challenges: hallucination and accuracy concerns, algorithmic bias, data privacy risks, the digital divide, overreliance on AI, and a lack of standardized outcome measures.

Strikingly, only four of the 64 studies explicitly addressed public health education, representing a major evidence gap. The authors argue this is a priority area for future research and investment. The review serves as a roadmap for evidence-informed adoption of NLP and AI across medical, nursing, pharmacy, and allied health training programs. It highlights the need for rigorous benchmarking, ethical safeguards, and cross-disciplinary collaboration to avoid repeating mistakes seen in other AI deployments.