The FINe-Health Foundry research project will develop a nationwide foundation health model. Supported by Business Finland, the initiative aims to address fragmented health data and build a versatile and scalable AI system for clinical and population-level decision-making.
An aging population and growing demands are straining healthcare systems globally. At the same time, medical data remains fragmented across different databases and clinical systems, preventing its effective use. To solve this, a major research initiative led by the ELLIS Institute Finland in collaboration with Aalto University, the University of Helsinki, and the University of Turku is building a nationwide foundation health model.
Dubbed the FINe-Health Foundry, the project has been awarded nearly five million euros in funding from Business Finland for an initial three-year period. The project is one of six bold initiatives selected under Business Finland's Rise to Challenge (Näytönpaikka) funding call.
Instead of building narrow, task-specific AI models, the FINe-Health Foundry is building a single, versatile foundation health model capable of multiple applications:
- Clinical Support: Assisting doctors and medical staff during active patient care with intelligent, context-aware decision support.
- Patient-Targeted Prevention: Predicting individual disease risks for over 200 different conditions for every living person in Finland, allowing for early, personalized preventative interventions.
- Population-Level Simulation: Enabling healthcare leaders and policymakers to run what-if scenarios to simulate the impact of changes in care guidelines, resource allocations, or policy shifts.
Current AI systems primarily identify patterns and correlations. The FINe-Health Foundry aims to push AI capabilities forward by introducing cause-and-effect (causal) reasoning. By reasoning about the outcomes of different clinical paths, the model can help doctors find the most effective treatments while avoiding unnecessary interventions.
To ensure the AI remains safe and trustworthy, the project is creating virtual laboratories where healthcare experts can audit, interact with, and integrate their knowledge into the model's decision-making loops. Our group contributes directly to these efforts through three main research activities: pretraining foundation health model, medical reasoning and grounding, and agentic method for virtual laboratory.
Read the official news post at the ELLIS Institute Finland website.