Skip to main content

European Health Data Space (EHDS)

Jump straight to Avatar compliance regarding the EHDS

The European Health Data Space (EHDS) is established by Regulation (EU) 2025/327 of 11 February 2025, published in the Official Journal on 5 March 2025. It covers two uses of electronic health data:

  • Primary use: access by natural persons to their own health data and exchange of these data between healthcare providers across the EU.
  • Secondary use: reuse of health data for research, innovation, policymaking, statistics, education and the improvement of care.

This page focuses on secondary use (Chapter IV of the regulation), where anonymised and synthetic data play a central role.

Timeline​

The regulation is in force but applies in stages (Article 105):

  • 26 March 2027: general date of application. By then, Member States must inform the Commission of the health data access bodies they have designated (Article 55(6)).
  • 26 March 2029: the rules on secondary use (Chapter IV) apply to most categories of health data.
  • 26 March 2031: they apply to the remaining categories listed in Article 51(1), points (b), (f), (g), (m) and (p): determinants of health, genetic, genomic and other omic data, clinical trial data and data from research cohorts.

Implementing acts are being adopted in the meantime. For example, Implementing Regulation (EU) 2026/2098 sets the minimum metadata that health data holders must provide to describe their datasets for secondary use, based on the HealthDCAT-AP profile. It applies from 26 March 2029.

Purposes of Secondary Use​

Health data access bodies may only grant access to health data for the purposes listed in Article 53(1), including:

  • public interest in public or occupational health,
  • policymaking and regulatory activities,
  • official statistics,
  • education and teaching in the health or care sectors,
  • scientific research related to the health or care sectors, including development and innovation activities for products or services and the training, testing and evaluation of algorithms, including AI systems,
  • improvement of the delivery of care.

Access is granted by a health data access body through a data permit (Article 68) or, for anonymised statistical answers, a health data request (Article 69).

Anonymised Data First​

The EHDS makes anonymised data the default format for secondary use. According to Article 66(2):

“Health data access bodies shall provide electronic health data in an anonymised format, where the purpose of processing by the health data user can be achieved with such data, taking into account the information provided by the health data user.”

Pseudonymised data are the exception. They are only provided where the health data user “has sufficiently demonstrated that the purpose of processing cannot be achieved with anonymised data” (Article 66(3)). The access application must include this justification (Article 67(2), point (e)).

The regulation also asks for state-of-the-art techniques. When providing access to a pseudonymised or anonymised dataset, a health data access body “should use state-of-the-art pseudonymisation or anonymisation technology and standards, ensuring to the maximum extent possible that natural persons cannot be re-identified by health data users” (Recital 72). Health data users are prohibited from re-identifying or attempting to re-identify the persons concerned (Article 61(3)).

Recital 92 adds that some categories of health data remain sensitive even once anonymised, for instance for rare diseases, because a residual re-identification risk can remain “even where state-of-the-art anonymisation techniques are used”.

Secure Processing Environments​

Health data are only made available through a secure processing environment (Article 73(1)). Only non-personal data can leave it. According to Article 73(2):

“Health data access bodies shall review the electronic health data included in a download request to ensure that health data users are only able to download non-personal electronic health data, including electronic health data in an anonymised statistical format, from the secure processing environment.”

Right to Opt Out​

Natural persons have the right to opt out, at any time and without giving reasons, from the processing of their personal electronic health data for secondary use (Article 71(1)). Member States may provide for limited exceptions for public interest purposes (Article 71(4)).

Each Member State sets up its own health data access bodies and the national rules needed to apply the regulation.

The EHDS does not define anonymisation itself: whether a dataset is anonymous is assessed under the GDPR, taking into account all the means reasonably likely to be used (see the GDPR page). The regulation does not mention synthetic data either.

Within that framework, the Avatar method addresses several needs created by the EHDS:

  • Anonymised data first (Article 66): health data holders and health data access bodies need anonymised datasets that are useful enough for the requested purpose. The Avatar method generates synthetic data that preserve the statistical properties of the original data, so that more requests can be served without pseudonymised data.
  • Outputs of secure processing environments (Article 73): only non-personal data can be downloaded. Avatar data whose anonymity has been demonstrated are non-personal data and can therefore leave the environment, unlike pseudonymised extracts.
  • State-of-the-art anonymisation and residual risk (Recitals 72 and 92): each avatarization comes with a privacy evaluation that measures the residual re-identification risk, including for rare records, to document that the generated data are anonymous in their context of use.

Generating avatars processes the original personal data. This step therefore remains subject to the GDPR and to the EHDS rules that apply to personal data, including the right to opt out: data of persons who have opted out should be excluded upstream, before generation.