In March 2025, the European Union published and brought into force the regulation establishing the European Health Data Space, whose operational deployment is progressive. The EHDS is the first regulation establishing a common European space for data specifically dedicated to health at this scale, but it extends pre-existing cross-border infrastructures, notably MyHealth@EU. The EHDS is presented by the Commission as potentially generating approximately 11 billion euros in savings over ten years; the precise effect of a lack of trust or massive opt-outs is not quantified in this estimate.

The Essentials

  • The European Health Data Space (EHDS) gives healthcare professionals in the 27 Member States direct access to patient records, with mandatory pseudonymization and a single contact point per country.
  • The European Commission projects 11 billion euros in savings over a decade; 68% of healthcare professionals believe that unified access would improve continuity of care (EU Commission).
  • The system is based on dual use: cross-border primary care and large-scale public health research, with distinct access rules for each component.
  • Pseudonymization protects patients but creates technical friction that each state will need to manage with its own health information systems.
  • The EHDS is the first full-scale test of the European digital sovereignty model: its results will determine the extension of the model to other sensitive data sectors.

Hundreds of Thousands of Cross-Border Healthcare Reimbursement Requests Each Year

The problem that the EHDS seeks to solve is mundane and costly. A retired German hospitalized in Spain cannot transmit his record to the local medical team. A Belgian cross-border worker followed up in France must reconstruct his medical history at each consultation. A Portuguese student treated in the Netherlands leaves without his primary care physician knowing what was prescribed to him. Data available under the directive reveal hundreds of thousands of annual requests, not tens of millions of people.

Until now, European coordination relied mainly on a voluntary framework and on MyHealth@EU, whose interoperability and coverage remained partial and unevenly deployed across Member States. Parliament approved the text in 2024; the Council adopted the regulation on 21 January 2025; the final text is dated and signed 11 February 2025, then published 5 March 2025. It creates a legal obligation for interoperability for all 27 states, coupled with a common technical architecture. Each state must designate authorities and contact points for several functions; the regulation does not necessarily create a single national contact point. Pseudonymization is a guarantee of secondary use when anonymization is insufficient; it is not mandatory before any cross-border exchange, particularly for care.

Patients can restrict access for care purposes; secondary use provides for a simple and reversible right of opt-out, subject to strictly defined exceptions for certain uses in the public interest, while opt-out from primary cross-border exchange depends on the choice of each Member State.

The figure of 68% of healthcare professionals who believe that unified access would improve continuity of care indicates that demand exists on the practitioners’ side. The question remains whether national infrastructures will keep pace with political ambition.

Limits and Contributions of Pseudonymization

Pseudonymization is the central mechanism of the system. It replaces direct identifiers with codes that do not, alone, allow the patient to be identified, while allowing the reconstruction of a coherent record for the doctor receiving the data. On paper, this satisfies both GDPR and specific data protection requirements for health data, classified among sensitive data under Article 9 of the European regulation.

In practice, pseudonymization raises a problem of system correspondence. Each country has built its health infrastructure with its own identifiers, its own formats, its own software. France uses a national identification number linked to social security; Germany still operates largely with decentralized systems by Länder; Nordic countries have unique civil digital identifiers that facilitate interconnection but raise other sovereignty questions.

According to Regulation (EU) 2025/327, secondary use covers in particular research, innovation, public policy, regulatory activities and patient safety; the Haas Avocats article of 24 March 2025 mentions research and innovation more briefly. The regulation sets pseudonymization rules; the Haas article summarizes or details them little, which does not allow for concluding a difference in rules. For primary care, the healthcare professional can, under conditions, access identifiable data if the clinical situation requires it. For research, data is anonymized when possible, or pseudonymized if necessary; researchers must not re-identify individuals, but the re-identification key may be retained by the competent body or a trusted third party. This distinction is relevant, but it creates real implementation complexity that hospital information systems will need to absorb, and whose compliance cost is not yet consolidated at the European scale.

The 11 Billion Projected Rests on Unequal Digital Maturity Assumptions

The projection of 11 billion euros in savings over a decade, put forward by the European Commission, deserves to be read with precision. It rests on several assumptions: reduction of redundant examinations, better coordination of chronic treatments, gains in clinical research through access to larger data cohorts. These effects are plausible and documented in sectoral studies, but the expected benefits depend on effective deployment, interoperability and adoption of the system; the quantified effect of a high opt-out rate on the 11 billion is not explicitly established in the official projection.

Yet levels of digital maturity in health vary widely from one state to another. The Nordic countries, Finland, Sweden, Denmark, have for several years had national electronic health records accessible in real time. Estonia, a recognized pioneer of the digital state, has dematerialized virtually all of its prescriptions and hospital reports. Conversely, several Central and Eastern European states remain largely on paper systems or fragmented IT infrastructures, without national interoperability prior to European interoperability.

This asymmetry is structural. It means that the 11 billion euros in savings will be captured very unevenly across states, more where digital infrastructures are already mature, less where they remain to be built. The potentially largest savings are, paradoxically, in the countries where adoption will be most difficult, because they are the ones accumulating the most redundancies and care pathway gaps. The EHDS regulation provides technical support mechanisms, but implementation timetables leave open the question of real catch-up or consolidation of health digital inequalities within the Union.

This dynamic joins that observed in other areas of European digital regulation, where regulating without building the common infrastructure produces well-designed rules but unevenly applied.

The Luxembourg Experience Under European Observation

Luxembourg is one of the first states to have made public its national implementation mechanism, via its health and social security platform. The Luxembourg singularity is instructive: a country of 670,000 inhabitants with a massive border population, nearly 50% of employees working in Luxembourg reside in France, Belgium or Germany; the Grand Duchy is exposed more than any other state to care pathway discontinuities. It has an immediate functional incentive to deploy EHDS quickly and seriously.

Its national architecture rests on several digital health tools and secure sharing between professionals. The regulation requires an easily accessible and easily understandable opt-out mechanism; its modalities, including online access, fall within national implementation. The regulation sets no number of clicks to exercise opt-out. The European Commission is monitoring this deployment as a reference case for states of comparable size that have not yet finalized their national system.

The Luxembourg experience illustrates a broader phenomenon: the states that have the most to gain functionally from the system are also those with the most political reasons to invest in its effective implementation. The regulatory framework and EHDS architecture are sound on the technical level. The question is how many states will have, by 2027, the infrastructures and political will necessary to go beyond formal deployment.

The European Digital Sovereignty Model at Stake

The EHDS constitutes, in its design, an original response to a tension that Europe has not yet resolved in other digital domains: creating a common infrastructure without creating centralized dependency. Current geopolitical fragmentation weighs on European digital sovereignty ambitions, and health is not exempt from this pressure.

The model chosen is that of distributed governance: states designate competent authorities and contact points; data remains in principle with its holders, public or private, under applicable national and European rules. Cross-border transfers pass through standardized protocols without centralization in a single European database. It is deliberately different from a centralized federal cloud model, and deliberately different also from the American model where health data is aggregated by private actors. The architectural identity of the system is clear: shared sovereignty, interoperability through standards, public control.

This model has coherence, but it exposes the Union to two opposite risks depending on the trajectories Member States will take in the years to come. The first scenario is that of progressive consolidation: digitally advanced states pull the others along, European standards become the de facto norm for national health systems, and EHDS effectively becomes the first pillar of a trustworthy digital space replicable to other sectors, finance, education, mobility. In this scenario, 2027-2028 will be readable in the data: increasing adoption rate among healthcare professionals, measurable reduction in examination redundancy in border zones, first pseudonymized data cohorts mobilized for research.

The second scenario is that of institutional freezing. Not through outright opposition, no state has an interest in publicly blocking a care continuity system, but through under-investment, compliance delays, and tensions over sharing secondary data for research. This secondary component is potentially the most contentious: it allows, with authorization and in a secure environment, access to anonymized data or, if necessary, pseudonymized data for research, public policy, innovation and regulatory activities. The value is immense; a cohort of 450 million individuals with integrated health histories would have no equivalent worldwide. But political consent to this sharing, state by state, is less certain than in-principle adherence to the primary care component.

The issue of political trust takes on concrete dimension here. Governments whose national health data will contribute to studies led from Brussels or other capitals will need to answer their national parliaments on access conditions, confidentiality guarantees, and result sharing. This debate has not yet taken place in most Member States. Its outcome depends as much on the quality of the technical system as on the capital of trust accumulated by European institutions during 2025-2028, a period marked by tensions on other digital sovereignty issues, such as the question of who ultimately owns the governance of algorithms processing sensitive data.

The signals to monitor in the next two years are precise: the number of states that will have effectively designated and operationalized their national responsible body, the opt-out rate observed among patients in pioneering countries, the capacity of hospital information systems to interface with European protocols without excessive local investment, and the first scientific publication using EHDS data cohorts crossed between multiple Member States. These four indicators will draw, before any financial projection, whether the shared sovereignty model holds under real load.

The Patient Remains the Link the Regulation Has Not Entirely Resolved

The entire technical architecture of EHDS rests on a presupposition that deserves to be stated clearly: patients must trust the system enough not to mass opt-out of it. If a significant portion of the population, through institutional distrust, cultural sensitivity to health data, or simply through failure to understand the mechanism, chooses to restrict access to their records, part of the projected gains can diminish, as can the value of research cohorts.

National experiences are instructive. In England, the care.data programme, announced in 2013 and developed notably through experiments scheduled for 2014, was officially closed in July 2016 without data extraction from general practitioners, following the recommendations of the National Data Guardian Review, controversies over patient information, consent and opt-out. In France, the Health Data Hub has evolved its governance and communication on the trust of professionals and patient associations. These precedents show that technical mechanics are insufficient: trust is built through transparency on actual uses, clarity on third-party access rights, and transparent demonstration that data use provides public benefit and equitable benefit to patients and the NHS, including when external partners are involved.

The EHDS benefits from a more solid legal framework than care.data or the first version of the Health Data Hub. Communication to patients remains largely in construction in most Member States. Part of the political work of 2025-2027 will take place in doctors’ offices and citizen drop-ins at town halls, as much as in server rooms.

Europe is setting up a public framework of governance and interoperability that also frames the participation of private actors, rather than a model that excludes them. It remains to be demonstrated that its institutions are capable of accompanying this model with enough pedagogy and consistency for citizens of the 27 to accord it the trust it requires. The next two years will tell whether the ambition of the regulation was commensurate with the patience necessary to embody it.

Sources

  1. Launch of the European Health Data Space, Luxembourg Health Social Security
  2. European Commission, EHDS Regulation 2024 (official text available on EUR-Lex)
  3. Haas Avocats, Analysis of the EHDS Regulation, December 2025 (available on haas-avocats.com)
  4. Esanté.gouv.fr, Documentation on the European health data space