Grounding AI in reality: Advancing African Healthcare

And establishing trust requires a clear framework. Dr. Kyula outlined a readiness checklist that includes clinical risk assessments, patient data governance, and rigorous vendor audits. The message was clear: we cannot simply rely on intuition.

We need verifiable checks and balances, including "human-in-the-loop" processes for diagnosis and emergency protocols for when AI gets it wrong. Governance isn't optional – it's mandatory for building a system that clinicians and patients can rely on.

The vision of an African Health Data Space

Dr. Franz Leisch of the Austrian Society for Telemedicine and E-health shared details of the creation of the European Health Data Space (EHDS). Dr. Leisch outlined how a continental data framework could empower citizens, enable cross-border research, and provide a secure foundation for AI.

However, he cautioned that this vision must be grounded in African realities. It requires inclusive governance, continental alignment, and a focus on innovation. But while Europe faces the challenge of overcoming legacy data fragmentation, Africa may have the advantage of bypassing these when creating their own African Health Data Space.

Olivier Twagirayezu of the Smart Africa Alliance also presented the goal of a "single digital market" where data, capital, and services move seamlessly across borders.

He described the current reality: a patient navigating a siloed system, carrying paper records from one facility to the next, losing valuable time and data at every step. A unified data space could transform this experience, creating a continuum of care that is more efficient and effective.

Real stories of adoption: Lessons from the front lines

Dr. Robert Korom, Chief Medical Officer of Penda Health, then shared results from their AI Consult study in a primary care setting. He noted that in Nairobi County, diagnostic accuracy is just 43%, and adherence to treatment guidelines is only 41%. These are not just statistics; they represent millions of lives impacted by gaps in care.

Penda Health introduced an AI-powered decision support tool that uses a traffic light system to alert clinicians to potential issues. The results showed significant quality improvements, but the key takeaway was not about the technology itself. As Dr. Korom summarised,

Success required careful change management, building trust in the system through peer champions, and providing individualised feedback. And this focus on the human element was a recurring theme.

A key challenge identified was balancing the eagerness of AI proponents with the practical needs of a hospital.

• Dr. Toseef Din, CEO of M.P. Shah Hospital, spoke of the need to build trust with clinicians first before introducing advanced tools, assuring them that technology is there to assist, not replace. Her hospital is already drafting its own AI policy, recognising that governance must keep pace with technology.
• And Geoffrey Gitagia, IT Manager at Avenue Health Care, stressed the importance of speaking the clinicians' language and walking a mile in their shoes.

Dr. Din also noted the importance of conducting a digital maturity audit to identify where AI can provide the most value, focusing on assistive intelligence before aiming for autonomous systems. Above all, she reminded us not to forget compassion and our commitment to the Hippocratic oath: "Do no harm."

Looking ahead: Responsible, data-driven healthcare

The event concluded with a workshop that brought these themes to life, demonstrating how clinical decision support (CDS) systems like UpToDate® can drive clinical excellence.

We explored how analysing usage data can:

• Reveal gaps in care.
• Guide resource allocation.
• Inform training needs.

For example, when data showed Diabetic Ketoacidosis (DKA) was a top-searched topic in Kenya, it prompted the question: how many hospitalisations are there, and how can we better support clinicians in managing this condition?

Data raises questions; it doesn't always provide the answers. The true power lies in combining data-driven insights with clinician expertise. As we look to the future, it’s clear that AI’s role extends beyond clinical diagnosis. It can optimise administrative workflows, predict supply chain needs, and help leaders with financial planning.

The discussions in Nairobi affirmed my belief that the future of healthcare is a blend of technology and humanity. As Dr. Korom said,

By grounding AI in the reality of trust, readiness, and relevance, African health leaders are forging a new path toward a more equitable and effective healthcare future for the continent.

Learn more about UpToDate Expert AI, the next evolution of UpToDate providing health systems and clinicians with the fast, reliable GenAI clinical decision support they need.