Trusted research environments are core components of the European Health Data Space (EHDS) for secondary use of data. Yet they also play a larger decisive role in the primary use of data, with the increased use of artificial intelligence (AI) in supporting clinical decision processes. It is the bridge between primary and secondary uses of data which is now of growing importance!

Trusted Research Environments (also known as TREs) underpin federated data approaches because they enable secure, governed access to sensitive data where it resides. Their scope now encompasses all types of data, even though genomics was often the first focus of initial initiatives. Federated data occurs when decentralised, independent services or components collaborate to achieve a common goal. The environments preserve local data stewardship and at the same time provide the technical and governance frameworks needed for privacy-preserving, interoperable, and collaborative research across multiple institutions. A variety of major initiatives co-exist today, but all have all been developed independently without relying on a common framework.

***

Earlier in June 2026, the UK-based Public Policy Projects company – which holds events and publishes report on life sciences-related policy  – published the outcomes of an expert round table organised in March 2026. The company runs a global genomics programme, and the topic of genomics was central to the round table’s exploration of the secondary use of health data.

The round table was sponsored by Syntropy, which is a joint venture between Merck – the life sciences, healthcare, and electronics company  – an EHTEL member, and Palantir, the data integration and analytics software company. Syntropy is a data platform that empowers researchers/research centres to work across organisational boundaries.

EHTEL digital health facilitator, Luc Nicolas, attended the round table alongside personnel from Merck offshoot, EMD Digital. As attendees, they were among senior participants representing the healthcare, industry, and research sectors. Attendees came from several Canadian, Commonwealth, US, and UK locations.

During the round table, Luc spoke of work done by EHTEL and COMFORTage, a project that works at large-scale in Europe to foster personalised preventions for the prevention of dementia and frailty through pilots, hubs, and fora.

What was the round table about?

The round table participants delved into questions around: Are trusted research environments enabling collaboration at scale in practice? Or, instead, are highly secure environments being built that are difficult to connect to, navigate, and use effectively?

The round table focused on the practical realities that can make a trusted research environment work and offered participants’ concrete answers to the questions posed. Of chief interest is the solution described by Syntropy on pancreatic cancer.

What are trusted research environments?

The round table introduced a definition of trusted research environments (TREs) in the following way:

Source: Round table report, p5

What are the environments’ challenges?

Among the persistent difficulties facing larger scale-up of trusted research environments are barriers related to interoperability, governance, usability, accessibility, and trust.

Hence, among the many ongoing challenges with trusted research environments are:

• A lack of a shared focus on the practicalities of making the environments work.
• A fragmented landscape with plenty of parallel, local developments.
• Ongoing problems with interoperability and duplication.
• Operational difficulties with governance, systems, and platforms.
• Limited clarity on the measurement of success.

How to scale up a trusted research environment on pancreatic cancer?

In the round table report, a two-page information box describes the outcomes of a Syntropy-based investigation of data related to 800+ patients with pancreatic cancer, a highly aggressive form of cancer. The illustrative case study tells the story of how the initiative completed the extraction, integration, and analysis of multimodal clinical and molecular data from multiple sources and systems. The large cohort of patient data spanned information from over two decades: it came from electronic health records, next generation sequencing, pathology and radiology records, and tumour registries. The same approach has later been applied to datasets at a variety of institutions.

Source: Syntropy presentation by Gabrielle Wong (EMD Digital), 19 January 2026,
EHTEL match-making event at Radical health Festival 2026

Practically speaking, for implementers, the initiative showed that “routinely collected, multi-platform clinical and genomic data can be harmonised and made analysis-ready at scale, without manual re-extraction for each study” (Round table report, p11). The Syntropy trusted research environment is designed to scale, and can be seen as “a meaningful shift in what becomes tractable” (Round table report, p11).

Four key insights of interest

The Public Policy Projects round table report cites 10 key insights in total.

Four of these insights have distinct relevance for EHTEL members. They fit well with EHTEL’s own work on the EHDS and beyond, the bridge between primary and secondary uses of data, and the importance of AI.

• Interoperability remains one of the biggest unresolved challenges

Much effort still needs to be invested today in data harmonisation pipelines. In the European Union (EU), the 2025 EHDS regulation establishes the first EU-wide references which will reduce this burden. AI will, of course, also play a decisive role in this respect!

• Governance is necessary, but is often experienced as friction

Regulatory approvals, arrangements, and negotiations can be slow and inter-institutional relationships complex. This slowness and complexity may delay and potentially restrict (research) collaboration. Governance is thus a must. But governance must be as agile as possible and should aim at maximal administrative simplification.

• Usability remains a major barrier to wider adoption

EHTEL’s working paper 3 on electronic health record systems, AI, and algorithm-based tools makes similar comments on usability. For example: “Several [barriers] – especially data quality gaps, vendor lock-in, and insufficient clinical usability – remain the primary reasons why AI-[Clinical Decision Support Systems] are difficult to deploy at scale.” (Working Paper 3, p12).

• AI could improve accessibility and functionality, but only if the foundations are in place

Among the foundations to be in place are the need to include the simplification of workflows, reductions in technical barriers, and help for users to be able to interact with complex datasets more easily. Essentially, “[AI’s] value depends on strong data quality, clear standards, sufficient computational infrastructure and robust validation processes.” (Round table report, p7.)

Data quality requires that data providers have created the necessary data governance board which will validate automation, and exercise continuous monitoring.

What to conclude?

Trusted research environments are an essential infrastructure, and their potential is widely recognised. Solutions need to be systemic and interlocking. Today’s task is to “make them more connected, usable, equitable and effective in supporting research and patient benefits” while, of course, ensuring that they are secure (Round table report, p19).

Holding an ongoing dialogue on the scaling-up and bridging of primary and secondary uses of data is invaluable for the future.

For more information

For more detail, including on the work of Syntropy, read Fulfilling the promise of trusted research environments (Round table report).

See EHTEL’s working paper 3.