From the newsletter
The biggest barrier to rolling out AI tools in African healthcare is a lack of credibility. Physicians and hospitals are unsure which new tech is safe and valuable. A new $60 million initiative, Evidence for AI in Health, will now fund locally led assessments of AI tools to generate the evidence required for adoption in disease prediction, diagnostics and health records.
The project, backed by Wellcome, the Gates Foundation and the Novo Nordisk Foundation, is implemented by the African Population Health Research Centre. It evaluates tools that already exist or are close to deployment in health systems.
Hospitals, donors and other health partners will be able to procure AI tools based on local data on effectiveness, which is mostly missing now, limiting proliferation.
More details
The initiative directly targets the AI evidence gap. It will support locally led studies to test how AI tools perform in real clinical settings, especially in primary and community care. That includes decision-support tools that help frontline health workers with triage, diagnosis, referrals and clinical documentation. These are operational tools.
The Evidence for AI in Health (EVAH) will assess a range of technologies already being used or proposed for use in health systems. These include prediction models that flag high-risk patients, computer vision systems analysing X-rays or ultrasound scans, large language model tools that assist with documentation and decision-making and multimodal systems that combine clinical, visual and health record data. Priority will be given to tools designed for resource-limited settings.
Studies will include implementation research to see how tools work in practice, randomised controlled trials to measure impact, economic evaluations to assess value for money, and acceptability studies involving patients and clinicians. For buyers, this type of evidence is far more useful than pilot reports or technical demonstrations.
There is also a structural implication for African innovators. Locally developed AI tools often struggle to secure independent evaluation funding, which weakens their credibility compared to tools validated in high-income markets. By financing locally led evaluations, EVAH may help shift evidence generation closer to the markets where these tools are actually deployed.
The EVAH initiative is now inviting proposals for its first call. Researchers can apply through two funding pathways. Pathway A offers up to $1 million for early-stage deployment evaluations, looking at usability, workflow integration, adoption, and safety. Pathway B provides up to $3 million for tools ready for scale, measuring health outcomes and system performance. Proposals are due 1 April 2026.
The African Population Health Research Centre, which is the African implementing partner of the project, supports EVAH by scoping tools for evaluation, developing evaluation frameworks, conducting or guiding real-world and impact studies, analysing results and sharing findings with investors. They ensure that evaluations are locally led and methodologically sound for partners considering adoption or scale of AI-enabled clinical tools.
Our take
African innovators need structured support to generate evidence in their own markets and build credibility with investors.
Local developers often cannot afford independent evaluations, which limits adoption. The new support for regionally led evaluations could level the playing field and make African health AI solutions competitive globally.