This week I had the opportunity to participate in the Health Data Forum ↗. I really enjoyed it, not only for the level of the sessions, but especially for the conversations with other colleagues sharing experiences and discussing ideas.
I also had the pleasure of participating in the panel Systems, AI & EHDS: From Data to Health Systems Impact. This aligns very closely with my work at OHDS and the 1HealthAI Factory, where my focus is on turning data, AI and infrastructure into practical solutions that can deliver real impact. One of the ideas I tried to bring to the discussion is that even if EHDS is a major step establishing the legal framework for exchanging health data, this alone is not enough. If we want to turn data into solutions we have to design EHDS taking also into account the infrastructure that will support it. This means that we have to do the design with both parts in mind, and we have to address the full data lifecycle from data sharing to analysis, processing and, ultimately, deployment and use of the final solutions by the citizens.
If we do not do it this way, we can then have serious problems later on to integrate all the heterogeneous infrastructures of the different health systems to implement what is defined by the EHDS.
The good part is that we already have many of the pieces: the EHDS governance framework, national and regional health systems, data spaces, AI Factories, HPC and cloud infrastructures, and a broad ecosystem of research institutes and organizations around it willing to create applications that will improve the health system.
So the challenge is how to join these pieces and make them work together in a common EHDS.
This requires both top-down and bottom-up approaches. Europe can provide common frameworks, infrastructures and standards, but health systems are heterogeneous and operate at national, regional and local levels. At the same time, concrete use cases and real needs from researchers, clinicians and citizens should guide us in the way, so we finally create something useful.
Data sharing works best when people can clearly see the value they receive from it. Millions of people already generate and share continuous health-related data through wearables such as smartwatches. And increasingly, people are willing to bring their own medical information to AI assistants to better understand it and ask questions about their health — perhaps without much thought about the implications.
There is an important lesson here for EHDS: trust requires good governance and privacy safeguards, but it also depends on providing useful services that deliver value to citizens.
This is also one of the ideas behind our work at the OneHealth DataSpace ↗ at CESGA: going beyond data sharing towards an infrastructure supporting the complete lifecycle, so that researchers and other users can move from discovering and accessing data to actually exploiting them with Big Data, HPC and AI resources.
Governance and infrastructure need to evolve together, and if we co-design them around real use cases, listening to citizens, researchers and clinicians, we have a much better opportunity to turn our joint effort into solutions with a real impact in the health system.
Many thanks to the organizers, fellow panelists and everyone I had the opportunity to talk with during the conference. The exchange of ideas was, for me, one of the most valuable parts of these days.