Dataspace4Health, housed in Luxembourg and operating as a Gaia-X Lighthouse initiative, is transitioning from theoretical design toward functioning, open-source building blocks that enable secure health data sharing across organisations. The system creates technical infrastructure for governed, sovereign and auditable data movement, aligned with the European Health Data Space (EHDS) framework for secondary use while respecting existing health governance structures and avoiding data centralisation.
The initiative brings together healthcare providers and research bodies through a distributed architecture where data holders retain ownership, data seekers locate and request information via standardised pathways, and transactions are managed through identity verification, metadata standards, policy frameworks, contracts and audit trails. Two priority use cases—diabetes digital twin and precision oncology—demonstrate the model's value in settings where data fragmentation across hospitals and research centres currently hampers progress.
Technical progress and connector deployment
A significant achievement has been demonstrating a complete Eclipse Dataspace Connector workflow linking HRS (Hôpitaux Robert Schuman) as a data provider with LIH (Luxembourg Institute of Health) as a data consumer. The flow incorporates Federator-based search capabilities, policy creation, contract handling, negotiation and data movement via a shared dashboard. The pilot now includes active connector instances for both organisations, Federator-managed registration, and HealthDCAT-AP metadata documentation.
Trust and credential mechanisms
Establishing trust between participants has advanced through a Gaia-X Danube-based credential system. The Federator creates and signs LegalPerson Verifiable Credentials, with registration details validated via the Gaia-X Notary. These credentials function within the consortium's trust framework and have been extended with Dataspace4Health-specific claims covering participant identifier, data-sharing protocol endpoint, protocol support, legal jurisdiction and participant classification.
A metadata capture tool built on HealthDCAT-AP standards enables dataset descriptions, while a synchronisation mechanism automatically refreshes connector assets from published metadata. This connection between rich health metadata and the exchange layer reduces manual effort for organisations sharing data.
Privacy and data preparation
MedAnon, a rule-based privacy system in development, complements the dataspace by handling de-identification, pseudonymisation, safeguarding of clinical text, preservation of FHIR data structures, privacy assessment and documented clearance procedures. The tool bridges privacy obligations and technical safeguards needed before data reuse.
A Secure Processing Environment pilot offers trusted computation capabilities. Direct integration between the connector and the processing environment remains under development; the project is not yet declaring a fully integrated chain but is working toward controlled deployment.
Broader applications and future direction
Beyond its health focus, Dataspace4Health is creating reusable components for Luxembourg's broader dataspace ecosystem. The same technical elements can be adapted for other sectors as part of the Dataspace4Luxembourg initiative and support the Data Spaces Hub Luxembourg community. The project also lays groundwork for cross-border health data sharing, including planned cooperation with Japan through the Artificial Intelligence for neurodegenerative diseases Care (AID-Care) project.
Upcoming priorities include strengthening the operational workflow, completing connector-to-processing-environment integration, improving catalogue usability, implementing transaction logging and audit functions, establishing data quality and readiness metrics, and closer alignment with EHDS secondary-use processes. The initiative demonstrates how Gaia-X principles for health dataspaces translate into practical, trustworthy and interoperable systems.
