EHDS: framework in place, infrastructure still missing

Adigens Health helps pharmaceutical and rare disease companies design real-world evidence strategies built around the target trial framework: from early development planning through HTA submission. Get in touch at info@adigenshealth.com.

New guidance from TEHDAS2 clarifies how the European Health Data Space will govern pharmaceutical access to real-world data. The problems it leaves unresolved matter more for observational studies built in a target trial framework than the problems it solves.

The European Health Data Space entered into force in March 2025. In the months since, the TEHDAS2 joint action—the European Commission’s expert body for EHDS implementation—has published a series of guidelines for the national Health Data Access Bodies that will govern pharmaceutical companies’ access to real-world data. Three documents, released between February and March 2026, address fees and penalties, permitted uses, and the obligations of health data holders. Together, they give the clearest picture yet of what EHDS will mean in practice for companies building real-world evidence strategies around the target trial framework.

The picture is instructive in ways the headline implementation timeline does not convey. Several of the most commercially significant questions remain unresolved. The marketing and research boundary—which will determine what pharmaceutical companies can lawfully do with EHDS data—is still subject to case-by-case HDAB judgement. Mutual recognition between national HDABs does not exist: a single cross-border study can receive different decisions in different member states. Nearly half of EU member states had no formalised pricing practices for health data provision as of late 2024, and fee unpredictability has been explicitly flagged as a problem for research planning. The technical maturity of health data holders varies so widely that the variable-level extraction a pre-specified target trial requires cannot be assumed to exist in every market a company needs. And the guidance is silent on perhaps the most fundamental operational question: how a company is supposed to establish, before committing to a pre-specified protocol, whether the data required to execute that protocol exists.

The EHDS framework is advancing. Companies planning multi-country observational studies for EU JCA or other HTA submissions need to engage with what has not yet been built, not only with what has.

What the new guidance establishes

The guidance confirms that pharmaceutical-sponsored real-world evidence studies fall within the EHDS secondary use framework, provided they can demonstrate scientific rigour. HDABs are advised to require a study protocol, pre-registration, a data management plan and, where national law demands it, ethics approval. They will assess methodology, aims and expected outputs against the stated research purpose. Belonging to a scientific institution, or labelling a project as research, is not sufficient.

On the supply side, Article 60 of the EHDS Regulation legally obliges health data holders—hospitals, registries, insurers, electronic health record operators—to make their data available when a data permit or request is issued by an HDAB. The cross-border data infrastructure that observational research at scale has always depended on is, in principle, being put on a statutory footing. Once a permit is issued, data holders have three months to provide data, extendable by a further three months. HDABs have two additional months to make the data available in the secure processing environment.

On fees, the guidance proposes a centralised single-invoice model, with costs staged as applications advance. Non-EU companies and commercial users will pay more than academic or public-body applicants; pharmaceutical companies are not among those eligible for fee reductions. Applicants can withdraw at any point but are charged for costs incurred to that point.

What it does not resolve

The marketing and research boundary is the most commercially exposed unresolved question. Article 54(c) of the EHDS Regulation prohibits the use of health data for advertising or marketing activities. The guidance draws the line carefully—using prescription data to identify low-prescribing physicians for promotional outreach is prohibited; health economics and outcomes research conducted under HTA frameworks is permitted—but the decisive test, whether data are used to promote goods or services or to support scientific understanding, will be applied case-by-case by individual HDABs. The same dataset can be used either way. Companies that lack internal governance separating legitimate RWE programmes from commercial data activities face a compliance exposure the guidance quantifies clearly: fines of up to four percent of global annual turnover for serious infringements.

Cross-border studies face legal and technical fragmentation that the guidance acknowledges but does not fix. Each national HDAB assesses applications under its own legal framework. Ethical review requirements are not harmonised: some member states require formal ethics approval for secondary use of health data; others do not. Enforcement decisions by one HDAB do not automatically apply in others. Mutual recognition mechanisms are flagged as necessary but are not in place. A company planning a target trial emulation study across multiple markets cannot rely on a single protocol or a single ethics approval to cover all jurisdictions.

Technical maturity among health data holders is the supply-side problem the guidance is most candid about. Two data holders may both list electronic health record data in an EHDS catalogue while differing fundamentally in what they can extract. Less mature holders may only be able to provide entire datasets, leaving variable selection to the HDAB. For a target trial protocol that specifies disease stage at initiation, prior therapy lines and dosing history, the difference between a mature and an immature data holder is the difference between an answerable and an unanswerable study.

Fee unpredictability remains real. The guidance explicitly flags the problem for researchers who need cost estimates twelve to twenty-four months before data access but stops short of resolving it. For pharmaceutical companies building EHDS data costs into development budgets—often two or three years before a study begins—early estimates under the staged fee model may diverge substantially from actual charges.

The timeline that matters

The EHDS Regulation has not yet started to apply. HDABs must be connected and ready to receive applications by March 2029. Full coverage of all Article 51 data categories—including the electronic health record, claims and registry data that most observational studies depend on—does not apply until March 2031. The implementing acts from the European Commission that will make these obligations binding are still pending.

For pharmaceutical companies planning real-world evidence studies, this has a direct and practical implication. The data categories a study requires may not be available through EHDS on the timeline the development programme demands. Companies need to know not only what the framework will eventually provide, but what it will provide when they need it.

The target trial framework’s unresolved problem

The TEHDAS2 guidance resolves one question clearly in favour of companies that have already invested in target trial methodology: the documentation an HDAB wants to see maps almost exactly onto what a pre-specified target trial protocol provides. Defined eligibility criteria, treatment strategies, outcomes, time-zero and causal contrast are precisely what HDAB assessment criteria are designed around. Well-specified applications are also processed more efficiently by data holders, reducing the risk of receiving incomplete or misaligned data—a risk that compounds across multiple jurisdictions.

The guidance also confirms that the purpose declared at application is binding throughout the study lifecycle. Data accessed for scientific research cannot subsequently be reused for marketing activities. This reinforces why companies need to define their comparative effectiveness question—in our view, through a target trial protocol—before approaching an HDAB, not after.

But there is a tension the guidance does not address, and the target trial framework cannot resolve on its own. Pre-specification is only meaningful if the study it describes can actually be executed. A protocol that specifies variables a data holder cannot extract, or a population too small to exist in the available data sources, is not a defensible study design: it is an unanswerable question committed to paper. Establishing that a study is feasible before pre-specifying it requires access to patient-level data—the very data that EHDS makes conditional on submitting a pre-specified protocol. The circularity is real.

Neither TEHDAS2 nor the EHDS Regulation describes how companies are supposed to navigate this. Preliminary data discovery—catalogue-level information about what datasets exist—is available under EHDS without a formal permit. But catalogue entries confirm that a dataset exists, not that it contains the specific variables, at the required granularity, for the patient population and follow-up period a target trial demands. The gap between what a catalogue describes and what a data holder can actually extract is precisely what the D6.1 maturity framework documents. Closing that gap requires the ability to interrogate patient-level data early enough to inform protocol design—before the protocol is locked, not after it has been submitted.

EHDS is the right long-term answer to the data access problem that target trial emulation has always faced. The new guidance is progress. But the pathway from evidence need to feasibility-confirmed, pre-specified, HDAB-approvable study protocol remains undefined. Until it is defined—whether through formal EHDS mechanisms, national HDAB guidance, or collaborative frameworks between companies and data holders—pre-specification and data access will remain two requirements that each assume the other has already been satisfied.