Natural Language Processing Technologies for Public Health in Africa: Scoping Review.

Area of Science:

• Public Health Informatics
• Computational Linguistics
• Health Systems Research

Background:

• Natural Language Processing (NLP) offers significant potential for advancing public health initiatives.
• However, the application of NLP in African health systems is hindered by limited digital infrastructure and computational resources, particularly concerning the continent's linguistic diversity.

Purpose of the Study:

• To conduct a scoping review mapping the evidence on NLP technologies for public health in Africa.
• To identify public health needs addressed, influencing factors, deployment stages, integration levels, and measurable impacts of NLP technologies.
• To synthesize recommendations for improving the quality, cost, and accessibility of health-related NLP in Africa.

Main Methods:

• Systematic literature search of academic and gray literature from January 2013 to October 2024 across major databases (MEDLINE, ACL Anthology, Scopus, IEEE Xplore, ACM Digital Library).
• Data extraction and mapping of NLP functions to WHO essential public health functions and UN SDGs.
• Analysis following PRISMA-ScR guidelines, with a publicly available protocol.

Main Results:

• 54 studies were included, revealing uneven language coverage with limited support for major African languages and minimal support for most of the continent's >2000 languages.
• Most NLP technologies are in the prototyping phase, with only one fully deployed chatbot; evidence of measurable public health impact is scarce (4% of studies).
• Recommendations emphasize expanding language support, addressing local needs, building trust, integrating solutions into health systems, and using participatory design.

Conclusions:

• Significant gaps persist in the deployment, linguistic inclusivity, and health outcome evaluation of NLP for public health in Africa.
• Future research must adopt cross-sectoral, needs-based approaches, engage local communities, align with existing health systems, and include rigorous evaluations to improve public health outcomes.