As part of our commitment to build capacity and skills for the longitudinal research community, we’re supporting the development of a new training Trusted Research Environment (t‑TRE) containing synthetic data.
We’ve appointed UK Longitudinal Linkage Collaboration (UK LLC) to lead this work as part of an £850K project. The new t-TRE will provide a safe environment for researchers to understand and work with data that has been linked to increasingly diverse data sources.
It will also help demystify TRE tools and practices by providing access to a realistic TRE with proportionately lowered governance requirements compared to those that protect sensitive data.
Familiarising users with t-TRE working practices – such as requesting and retrieving results from the environment – will allow risk experimentation and enable rapid access so users can begin developing syntax ahead of working with real data.
We spoke with Richard Thomas, Senior Data Manager at UK LLC and co-investigator for this project to understand what’s coming next.
Richard, this project aims to create a learning environment populated with synthetic data for researchers. What’s the vision for this new t-TRE?
The vision is to enable interdisciplinary longitudinal researchers to confidently transition to using newly available combinations of study data, linked datasets, and the contemporary data environments and tools that support them.
This is essential because unfamiliarity with these settings and tools – along with limited methodological understanding of issues such as sample representation and linkage bias – can discourage researchers from engaging with these resources.
The t‑TRE will be designed to empower users by offering TRE‑like access in a convenient, low‑burden environment, supported by high‑quality training and linked to exciting new research opportunities.
The training will use statistically generated synthetic data that resembles real data but does not relate to actual individuals. As a result, it carries lower sensitivity and risk, allowing access through a proportionate and appropriately designed governance framework.
Why is a t‑TRE important given current trends in data access and security?
This project responds to a gap in the current Longitudinal Population Studies (LPS) landscape.
The design and testing of this t‑TRE are essential because data provision for longitudinal research is rapidly shifting toward models in which data are stored, processed, and analysed within TREs, also referred to as Secure Data Environments (SDEs).
This shift is driven by increasing availability of participants linked routine health and socio-economic records, as well as environmental data.
These newly accessible data – via UK LLC – provide exciting new scientific opportunities. But we also know that participants expect the confidentiality and security of these data to be maintained, and that it’s very clear as to who is using which data for what purpose at all times. Hence TREs becoming the default access route.
This transition isn’t expected to be temporary; rather, it represents a new norm for longitudinal research.
What are the main challenges this pilot t-TRE will help mitigate?
This pilot t‑TRE will address challenges in at least three key areas:
- Skills and methodological gaps for linked data research: Users need training to work effectively with the growing volume and diversity of linked longitudinal data. Many are unfamiliar with analysing pooled LPS datasets or addressing issues of representation, requiring targeted methodological development and hands‑on practice.
- Lack of familiarity with TRE systems, tools, and working practices: Many analysts have limited experience with TRE software, processes, and expectations. This creates barriers and burdens for both users and operators. A low‑risk training environment can demystify TRE tools, build confidence, promote good analytical and Findable, Accessible, Interoperable, and Reusable (FAIR) practices, and reduce errors.
- Reducing barriers to early project development: Accessing real TRE data can be relatively slow due to reviews and contracting, UK LLC has accelerated access to linked data substantially, with a typical time of application to data provision of 10-12 weeks. However, some researchers will want to move faster than that. A t-TRE with rapid access and a synthetic “mirror image” of the real data will allow users to develop and test syntax on synthetic data while they wait for access to the real data.
How will the t-TRE be built, governed, and made accessible to the wider research community?
In consultation with data owners, we’ll commission the development of synthetic data generators within the UK LLC data pipeline. This will enable the consistent and sustainable production of synthetic data at scale.
This approach will give trainees a realistic understanding of the types of data and structures they will encounter – cross sectional data representing a survey wave, longitudinal event-based data such as those found in health records, and the complex database structures required to integrate these data.
The training content will be designed specifically to interact with the synthetic data, ensuring a complete and practical learning experience. The availability of synthetic data will also allow users to develop and refine syntax and, where appropriate, test the feasibility of their project plans before accessing real data.
Access to t-TRE will be governed by UK LLC following the co-development principles it has already established with its network of partners.
What partnerships will support the development and sustainability of the t-TRE?
UK LLC will host and manage the training environment in partnership with Secure eResearch Platform (SeRP), who will provide the technical infrastructure.
The methodological research and the development of synthetic data‑generation tools will be commissioned through a process managed by Population Research UK (PRUK).
Although the pilot will be developed jointly by UK LLC and SeRP, what is learnt from this UK first – in fact, we are not aware of any other initiative worldwide – will support wider initiatives and community benefit.
How do you see the community contributing to the t-TRE?
We’ll work closely with users and professional groups to identify both opportunities and risks, helping us shape the generation and protection of synthetic data at varying levels of statistical granularity.
As synthetic data becomes more realistic, its utility increases – for example, enabling more accurate feasibility assessments. However, higher realism may raise disclosure risks, which could, in turn, require stronger governance measures. The level of any potential risk will need careful consideration.
We’ll therefore balance risk, benefit, and utility to deliver a suite of functions that meets stakeholder needs while maintaining appropriate safeguards.
Closing thoughts
We are progressing toward PRUK’s delivery plan objectives – building capacity and skills across the UK’s longitudinal research community.
This initiative is also informed by what we heard during our stakeholder consultation, particularly the need for a learning environment supported by synthetic data where researchers can train and practise.
It is a direct response to those requests, ensuring the LPS community’s priorities shape the path ahead.
“This project will form a crucial part to the PRUK vision for a seamless end-to-end infrastructure landscape which supports users from data discovery, through application, and to data use. We are excited to be leading this from UK LLC and working with the research community to understand training needs and the opportunities t-TRE will bring.”
Andy Boyd, UK LLC Director and PRUK Co-Infrastructure chair
This work is due to commence in April 2026. We look forward to sharing progress and welcoming the community to engage with the new resources as they emerge.
Stay connected with us
To stay informed about upcoming events, funding opportunities and news, we encourage you to sign up to our newsletter. If you have ideas, insights or feedback to share about our work, please reach out to us via email – info-pruk@bristol.ac.uk
About Richard Thomas
Richard is a Senior Data Manager at UK LLC and is responsible for the management of the Trusted Research Environment. He leads the data architecture, engineering, and curation work streams. From spending four years working as part of the data linkage team at Children of the 90s study (ALSPAC) within the University of Bristol, Richard has vast experience in managing data pipelines of linked records from a variety of data domains. Prior to working with the University, Richard spent time in Australia and Saudi Arabia managing geological exploration data; ensuring data integrity, security and availability within large organisations.
About UK LLC
UK Longitudinal Linkage Collaboration (UK LLC) is the national Trusted Research Environment (TRE) for data linkage in longitudinal research, led by the Universities of Bristol and Edinburgh. It simplifies access to linked data for research in the public good, enabling discoveries through linked associations between health, social, behavioural, and economic research.
About our funding models
At Population Research UK, we’re committed to supporting a more interconnected data landscape, where users of LPS benefit from a seamless journey.
To achieve this, our delivery plan uses three complementary funding models: PRUK Hub to centrally deliver community engagement and advocacy; open calls to invite competitive applications where specialist delivery is needed; and partner-led funding applied where our UKRI partners – CLOSER and UK Longitudinal Linkage Collaboration – are best placed to deliver activity, building on existing UKRI supported infrastructure and avoiding duplication.
The t-TRE for the longitudinal research community project falls under the partner-led funding model, leveraging UK LLC’s expertise and established networks to deliver the work and that community consultation identified a high demand for training on both TRE use and linked data use.
