Data Safe Havens are an essential way of managing risk of unauthorised re-identification of individuals from de-identified data. This risk heightens with increased data linkage across a wider range of data sets. To date, most of the focus of Data Safe Havens has been on the people, processes and technology required to deliver effective Information Security. 

Our aim is for Alliance Trusted Research Environments (TRE) to be more than Data Safe Havens by meeting researchers’ technical and functionality requirements both now and in the future, whilst providing confidence to data custodians and the public through implementing a rigorous system of data access based on an evolution of the Five Safes model. This evolution is required to meet research needs for Machine Learning capability, High Performance Computing or High Throughput Computing e.g., for analysis of genetic and imaging data, and also to support the convergence of research and care to enable a future of personalised medicine. 

In March 2020, we hosted a workshop that brought together a community of experts interested in improving the use of data for research in a safe and ethical way through the development of TREs.   The outputs from this workshop and a consultation run between 30 Apr and 26 May 2020 have been used to update our Green Paper.   The paper outlines our position on the characteristics required for TREs to support safe and ethical research using health data assets from data custodians within the UK Health Data research Alliance. 

We are very grateful for the responses received which covered a broad range of stakeholder views including patient and public representatives, NHS and other data providers, academia, system-level stakeholders, TRE service providers and consultancies. Whilst it has not been possible to incorporate all of the comments received, the input will also be used in the development of the implementation plan. We continue to welcome comments and contributions to help shape the development and next steps: 

  1. Consistent and proportionate accreditation of safe people.
  2. Consistent accreditation of safe settings, with a focus on use of cloud computing.
  3. Involvement of public and patient representatives in the data access decision making process.
  4. Improved lay explanations of the design and functioning of TREs.
  5. Enhancing the researcher experience whilst minimising risks to privacy.  
  6. Addressing the technical, governance and process challenges of creating a network of TREs through federation. 

Read the latest version of the Green Paper

Outputs from the workshop

If you are interested in this work and would like to be involved, please email us.