DARWIN EU: EMA's federated real-world evidence network on OMOP CDM

The Data Analysis and Real World Interrogation Network of the European Medicines Agency: federated architecture, partners, completed studies, role of Erasmus MC as Coordination Centre, relationship with EHDEN and the European Health Data Space.

Digital HealthR&D DARWIN EUEMAOMOPReal-World EvidenceEHDENPharmacovigilanceDigital Health

An infrastructure for regulatory assessment

The European Medicines Agency (EMA) and the Heads of Medicines Agencies (HMA) network — which brings together national medicines agencies of EU and EEA countries — have a recurring need: assessing drug effectiveness and safety in real clinical practice, not only in the controlled conditions of registration clinical trials. Post-authorisation safety studies (PASS) and efficacy studies (PAES), updated benefit-risk assessments, analysis of prescribing patterns across European countries require access to real-world healthcare data, structured homogeneously, usable at timelines compatible with regulatory cadence.

To answer this need, EMA launched on 9 February 2022 the Data Analysis and Real World Interrogation Network — DARWIN EU: a federated network of European healthcare data sources designed to produce real-world evidence in support of regulatory assessments. Over two years after launch, the network is fully operational.

Federated architecture on OMOP CDM

DARWIN EU’s technical model is federated and based on OMOP Common Data Model. Every data partner keeps its healthcare data locally, converts them to OMOP CDM following the typical OHDSI Extract-Transform-Load, and exposes them to a set of pre-defined aggregate queries. Individual data never leave the data partner.

The operational flow of a DARWIN EU study:

  1. EMA or a regulatory committee (PRAC for pharmacovigilance, CHMP for authorisations) requests a study to answer a specific regulatory question
  2. The Coordination Centre of DARWIN EU — hosted by the Medical Informatics Department of Erasmus MC (Rotterdam) under the lead of Peter Rijnbeek — designs the protocol using OHDSI methodologies (CohortMethod, SelfControlledCaseSeries, PatientLevelPrediction, IncidenceRate)
  3. Analysis scripts in R are prepared using HADES and go through review
  4. Scripts are distributed to the data partners relevant to the question
  5. Each partner runs them locally on its CDM; aggregate results are generated
  6. Aggregate results flow back to the Coordination Centre, which produces the combined analysis
  7. The report is made available to EMA and, once published, accessible publicly through the DARWIN EU Results Repository

The structure ensures privacy-by-design (only aggregates travel) and analytical standardisation (every partner executes the same script on its harmonised data).

The Coordination Centre

The Coordination Centre is the operational heart of the network. It was awarded to Erasmus MC through an EMA public tender completed in 2021, with a five-year contract. Responsibilities:

  • Study protocol design
  • Analysis software development and maintenance
  • Data partner onboarding (together with EMA)
  • Coordinated execution of studies across the network
  • Methodological support to regulatory requesters
  • Publication of results

The link with EHDEN (European Health Data Evidence Network) is direct and historical. EHDEN — IMI project launched in November 2018 and approaching its conclusion in 2024 — has built the technical and community base on which DARWIN EU could grow: OHDSI tools at European scale, the network of certified SMEs for OMOP conversion, data partners already mapped. In many cases the same centres that became EHDEN data partners are now DARWIN EU data partners.

Data partners

The DARWIN EU network has grown progressively. The first onboarding waves involved centres consolidated in the European OMOP world:

  • CPRD (Clinical Practice Research Datalink, UK) — one of the largest European primary care databases
  • THIN (The Health Improvement Network, Spain) — primary care database
  • IPCI (Integrated Primary Care Information, Netherlands, Erasmus MC)
  • SIDIAP (Sistema d’Informació per al Desenvolupament de la Investigació en Atenció Primària, Catalonia)
  • National Danish Registries (Denmark)
  • Finnish National Registries (Finland)
  • National Health Insurance Research Database (various countries)
  • Italian regional databases (among them the ARIA / Lombardy Region database)

As of mid-2024 the network has about twenty operational data partners, with more waves under way. The stated goal is to reach forty or more data partners by 2025, with geographically and clinically representative coverage of the European Union.

Studies conducted

In the first twenty-eight months of operation DARWIN EU has completed or initiated dozens of studies, including:

  • Background rates of adverse events in European populations, useful as reference for post-vaccination COVID-19 and new drug pharmacovigilance assessments
  • Prescribing patterns of drugs with use restrictions or safety alerts (valproate in women of childbearing age, SGLT2 inhibitors, etc.)
  • Drug utilization studies comparing countries on newer therapeutic classes
  • Post-Authorisation Safety Studies (PASS) requested by PRAC for specific drugs
  • Rapid assessments of emerging safety signals
  • Studies in priority areas: oncology, cardiology, rare diseases, paediatrics

Protocols and results are published on the DARWIN EU Results Catalogue, with varying detail level — some studies stay for EMA internal use, others are published in full and cited in peer-reviewed articles.

Relationship with EHDEN

DARWIN EU and EHDEN have distinct and complementary roles:

  • EHDEN (2018-2024) — IMI research project that has built capacity: training (Academy), certification (SME), conversion (data partners mapped to OMOP). Enduring legacy: the community and the tools
  • DARWIN EU (2022 →) — EMA regulatory service that uses that capacity to produce evidence supporting regulatory decisions. Long-term operational model

With EHDEN’s conclusion in 2024, part of the community naturally migrates towards DARWIN EU — and other European healthcare data programmes that will be structured under EHDS.

Relationship with EHDS

The European Health Data Space (EHDS) — EU regulation politically adopted in spring 2024, awaiting formal publication — builds a broader European frame for healthcare data flow. DARWIN EU sits in the secondary use area of EHDS, as a federated network already operational dedicated to the specific topic of regulatory drug assessment.

Once the EHDS regime is operational, convergence is expected: the national Health Data Access Bodies (HDABs) that EHDS mandates on Member States will be able to draw on the same OMOP data partners that feed DARWIN EU today, through data permit mechanisms. The network will not be replaced; it will be integrated in a broader architecture that will include non-regulatory clinical research, health policy, and technological innovation.

Italian participation

Italy participates in DARWIN EU through regional data partners and research centres. Participation has grown in conjunction with the PNRR work on healthcare information systems: the construction of regional Health Data Ecosystems (EDSs) envisaged by FSE 2.0 is the natural infrastructural base on which new Italian data partners can plug into the network.

Constraints remain the known ones: SNOMED CT access without Italian membership of SNOMED International, governance of consents for secondary use, alignment of Italian terminologies (Italian ICD-9-CM, AIC, ATC) to OMOP standards. These are themes that will find framing in the EHDS regulation once operational.

What this means for healthcare data producers

For an Italian healthcare agency that wants to make its database available for regulatory evidence, participation in a network like DARWIN EU implies:

  • Converting the database to OMOP CDM (via EHDEN path or certified SME)
  • Signing a framework agreement with the Coordination Centre
  • Technical capability to run R scripts on request
  • Local governance process to authorise each study
  • Audit and traceability of data access

The benefit — beyond scientific contribution — is access to consolidated infrastructure for observational studies, to an international methodological community, and to reusable protocols. It is the model that will likely also structure Italian participation in EHDS secondary use in the coming years.


References: DARWIN EU — Data Analysis and Real World Interrogation Network, launched by EMA on 9 February 2022. Coordination Centre at Erasmus MC (Rotterdam). DARWIN EU Results Catalogue. OHDSI — OMOP CDM. IMI EHDEN (2018-2024). EMA Framework for real-world evidence in regulatory decision-making.

Need support? Under attack? Service Status
Need support? Under attack? Service Status