Hospital data from four countries, analysed together - and never moved

Institutions and health organizations in Portugal, Italy, Poland, and Hungary hold decades of clinical records, biobank samples, genomic sequences, and medical imaging. Each institution holds a piece of a research puzzle that could accelerate drug discovery, improve cancer diagnostics, and transform patient outcomes but combining those pieces is near-impossible. GDPR, national health laws, and institutional liability rules mean patient data cannot cross a hospital's boundary, let alone a national one. The result: researchers working on the same conditions at different institutions cannot pool their datasets. AI models trained on a single hospital's records produce limited, institution-specific results. Drug efficacy analyses, predictive risk models, and cancer research programmes that could benefit millions of patients are blocked by the gap between what the data could enable and what the rules allow.

HeliX (Health Ecosystem for Learning and Innovation Across Sovereign Data Spaces) is a 36-month, €9.2M EU-funded project under the Interregional Innovation Investments (I3) framework, launched in September 2025 in Porto, with E-Group as the key technology developer. The HeliX platform is built on E-Group's FedX federated learning engine: instead of pulling patient records to a central server, it sends AI models to each institution, trains them locally on local data, and returns only anonymised statistical insights nothing. The project's the kick-off meeting was held in Porto in October 2025. identifies an individual, nothing that leaves the hospital's control. The platform connects real clinical data holders: ULS Coimbra (nine hospitals, 26 primary care centres, serving 1.8 million patients), SYNLAB SDN (100,000 imaging examinations and one million in-vitro tests per year, with an accredited biobank), the National Cancer Institute Pascale Foundation, and Pineta Grande Hospital, alongside other institutes across four countries. Prologica standardises all data to the OMOP Common Data Model so institutions using different systems in different countries can be analysed in the same analytical framework. Three cascade funding schemes will recruit at least 45 additional SMEs and 15 further data holders into the ecosystem expanding coverage to more hospitals, more conditions, and more European regions.

HeliX is live and progressing from TRL 6 toward TRL 9, the point at which the platform moves from a validated prototype to a fully operational commercial product available on the market. The project simultaneously addresses all three barriers that have kept European health data locked: regulatory compliance (privacy-preserving by design, aligned with GDPR, the EU AI Act, and the forthcoming European Health Data Space); technical interoperability (OMOP-standardised data enabling cross-institutional analysis); and commercial sustainability (a governance, licensing, and IPR framework built to operate beyond EU funding). The competitive advantage over major cloud platforms is structural: where AWS, Azure, and Google Cloud require data to move to their infrastructure, HeliX delivers equivalent AI capabilities while the data stays precisely where regulation demands it stays. For E-Group, HeliX is the vehicle through which FedX advances from an established research tool to a market-ready commercial platform, the pivotal step in E-Group's trajectory from federated learning innovator to the sovereign health AI platform of record in Europe.