SeRP delivers final report on federated data project TELEPORT

In January 2023, as part of the DARE UK programme, UK Research and Innovation (UKRI) commissioned five driver projects to inform the design of a UK-wide network of trusted research environments (TREs) – highly secure digital environments that provide access to sensitive data for analysis by accredited researchers.

These driver projects ran for nine months, from February to October 2023, and their outcomes have now been captured in a series of final reports.

In April this year we reported on SeRP’s successful bid for a share of £2m UKRI investment to lead one of the five projects and partner in the delivery of another; both projects focused on the development of federated data analysis across the UK’s population data landscape.

Federation of data allows access to data held in physically separated environments without the need to move the data physically, through the creation of networks to provide access pathways to data through a single virtual environment, yet data residing in its original hosted location. A second method of federated analytics is through the movement of secure research objects between TREs (but not the data itself) while still supporting the TRE Five Safes principles that govern and protect sensitive data. Research objects are a standardised way of describing and packaging the digital information needed to pose a research question and report the answer.

SeRP Federation enables these options, both in the form of a ‘pop-up’ virtual TRE which provides a secure and governed environment in which federated analytics can take place, and in the form of an algorithm or database query code which is containerised to be deployed on standardised data structures. Federation navigates the diverse governance protocols of different TREs to return the desired insights to the researcher without ever compromising data security.

Led by SeRP, the TELEPORT driver project’s mission was to create a software-defined solution resulting in a novel model of ‘pop-up’ TREs by federating health data from Wales and Scotland in collaboration with the University of Edinburgh and Public Health Scotland. This included the build of the ‘pop-up’ TRE itself, a connected interface with pre-loaded software accessible from within project partner TREs, as well as the networking needed to provide access through to underlying data sat in databases or file stores for those contributing data to the ‘pop-up’.

A key part of this work was also to explore how similar but different approvals processes can be brought together through a technical solution so that the data controllers – the people who are ultimately responsible for the data – can have the confidence it is being used appropriately and securely by researchers and in line with the uses they have approved, without ultimately diverging too far away from standard and trusted current processes.

In this final report, ‘Connecting big data to users at light speed’, the authors successfully demonstrate:

  1. Proof-of-concept for a 100% software-defined ‘pop-up’ TRE for efficient, governance-preserving, multi-TRE federated data analysis.
  2. Practical application between Dementias Platform UK and SAIL Databank to enable access to routine NHS data and the corresponding longitudinal cohort data within the governance confines of each respective TRE.
  3. How collaborative problem-solving between two national data safe havens can overcome inherent data sharing barriers, promote efficiency in data access for the research community, and provide a framework and solution for less mature TREs to work towards federation.
  4. How the expertise of public panels (SAIL Databank Consumer Panel and Public Benefit and Privacy Panel for Health and Social Care in Scotland) can provide valuable insights into a project of this nature. The involvement of a Lay Lead in project meetings provided an important public perspective.

Chief Technical Officer at SeRP, Professor Simon Thompson, said, “Data federation across whole UK datasets and environments has the power to transform the efficiency, speed and scale at which researchers can unlock population-level discoveries. It will reduce the duplication of efforts of analysing data in disparate environments, boost sample sizes to enable breakthroughs into rare diseases and produce better scientific outputs for all conditions through the inclusion of entire UK population data. The TELEPORT project has proven that data federation and federated analytics can work. We’ll now explore the application of data federation across wider UK TREs and international TREs, initially between the UK and Australia. Its potential for new solutions to globally significant health challenges is enormous.”

Read the full ‘DARE TELEPORT: Connecting big data to users at light speed’ final report here – https://zenodo.org/records/10055358

Concurrent federated analytics project TRE-FX, on which SeRP collaborated with the universities of Nottingham and Manchester, addresses the challenges faced by researchers to perform analyses across TREs in a slightly different context through research object deployment, and has demonstrated ‘that it is possible to streamline how analysis tools access multiple TREs while enabling the TREs to ensure that the access is safe’.

Read the full TRE-FX final report here – https://zenodo.org/records/10055354

A DARE UK review of sensitive research data infrastructure formed the foundations and context of these projects. Between November 2022 and June 2023, the DARE UK Delivery Team conducted a programme of surveys, interviews and ad hoc consultations with organisations providing infrastructure and other services in support of research using sensitive data with a particular focus on trusted research environments (TREs).

These results fed directly into the first version of the DARE UK ‘Federated Architecture Blueprint’ to address the principal governance and technical operation challenges of creating a federated network of secure research resources for sensitive data research across the UK, paving the way for TELEPORT and TRE-FX.

Read the full ‘UK Sensitive Data Research Infrastructure: A Landscape Review’ here – https://zenodo.org/records/10082545