JMIR Publications

ABSTRACT

Background:

Hypertension is the leading risk factor for many disease, and health education is essential for its effective management. The emergence of artificial intelligence has created new opportunities for implementing digital health education interventions. However, the application of AI in hypertension health education has yet to be integrated.

Objective:

This scoping review aims to map the existing evidence on the use of AI in hypertension health education. It seeks to identify the types of AI technologies employed, the research methodologies used, the outcome measures assessed, and the potential advantages and current challenges.

Methods:

This scoping review was conducted following the Arksey and O'Malley framework and is reported in accordance with the PRISMA-ScR guidelines. We systematically searched PubMed, Cochrane Library, CINAHL, Embase, and Web of Science for peer-reviewed original research published in English between January 2015 and January 29, 2026. We used a Population-Concept-Context framework to guide screening and data extraction, focusing on studies that applied AI to hypertension health education.

Results:

Our initial search yielded 384 articles. We then performed snowball sampling by searching the reference lists of the included studies. After screening, 14 studies from 7 countries met the inclusion criteria and were included in the final analysis. The included studies employed diverse methodologies: mixed-methods studies (n=3, 21.42%), qualitative studies (n=1, 7.14%), quantitative randomized controlled trials (n=3, 21.42%), and quantitative non-randomized controlled trials (n=7, 50.00%). The AI technologies identified included machine learning, generative AI, natural language processing, and knowledge bases. The outcome measures assessed were system availability, accuracy, readability, blood pressure, adherence, robustness, engagement, and so on.

Conclusions:

AI demonstrates significant potential to enhance hypertension health education. However, current applications face critical limitations, including concerns about content reliability, readability, and privacy. Future research should focus on developing hybrid AI architectures (such as integrating large language models with knowledge graphs), and conducting rigorous clinical validation through randomized controlled trials. Clinical Trial: OSF Registries 4wv3f.