Anthropic and the medical knowledge platform OpenEvidence have announced a collaboration to provide a specialized clinical decision support tool free of charge to healthcare professionals in approximately 100 low- and middle-income countries. OpenEvidence answers doctors' questions using peer-reviewed literature and clinical guidelines, and the platform is already free for clinicians in the U.S. and Europe.

The rollout includes countries such as Uganda, Angola, Sudan, Haiti, and Mongolia. Anthropic provides backend model capabilities, while OpenEvidence tailors the tool to local healthcare infrastructure, disease profiles, diagnostic conditions, and available medications. Founder Daniel Nadler stated that access to medical knowledge should not depend on geographical location. Anthropic President Daniela Amodei mentioned that market incentives alone cannot deliver products to low-resource areas.

Nadler noted that U.S. clinicians completed approximately 42 million consultations through the platform in August and expects that by 2026, hundreds of millions of Americans will have received treatment from doctors using this tool. The platform has previously been largely unavailable in most of Africa due to infrastructure and market structure issues, but has partnered with the Rwanda Biomedical Center and Resolve to Save Lives, with about 45 Rwandan physicians and nurses testing it for customization, and is connecting with Botswana's health institutions and the University of Pennsylvania's Botswana project.

The World Health Organization estimates a global health workforce shortage of about 11.1 million by 2030. OpenEvidence raised $250 million in January this year at a valuation of $12 billion, with investors including NVIDIA and Google Ventures. The company has also deployed the same evidence retrieval system to over 10,000 clinicians in the University of Pennsylvania medical system, extending to the Botswana collaboration network.

The buyers are frontline doctors lacking access to literature, specialist consultations, and continuing education, while the sellers are model companies donating foundational model capabilities into public health channels. Funding does not operate on a pay-per-use basis but rather through free coverage supported by valuation. The beneficiaries are verifiable grassroots diagnosis and treatment narratives needing global case distribution for model training, while the pressured parties are publishers selling knowledge via subscription and generic chatbots unable to localize drug and testing menus.

The specialized version emphasizes that each answer can be traced back to the literature. The two companies have not disclosed the timeline for launching in each country or the minimum conditions for connectivity.

Source: Public Information