SeRP in collaboration with its partner tenants DPUK and SAIL Databank is invested in developing synthetic data to provide greater access to data.
Synthetic data is artificially created data, typically created by generative models trained on real world datasets. Synthesiser models aim to learn the statistical properties, distributions and correlations of the real data, to be able to create realistic artificial samples which are statistically similar to the data it was trained on.
Synthetic data has many different use cases in sensitive health data research. For example, we can use synthetic data to enable researcher/student training in how to analyse health data without the need for lengthy governance approval processes. It can also be used by researchers to develop their algorithms and workflows while waiting for data access. Additionally, we can use synthetic data in federated data scenarios where the real data cannot be seen.
This collaborative programme of work has already resulted in the creation of an open-source Python package within the DPUK Data Portal, SynthOpt, to generate varying levels of synthetic data for these different purposes.
Report: Perspectives & Recommendations on the Development of Synthetic Datasets in Trusted Research Environments
The UK Synthetic Data Community Group report, Perspectives & Recommendations on the Development of Synthetic Datasets in Trusted Research Environments, provides a practical and timely guide to the responsible development and use of synthetic data across the UK research ecosystem. Funded by DARE UK, the report brings together insights and recommendations from experts working in synthetic data, Trusted Research Environments (TREs), and administrative data research over the past year.
At the core of the report is the VSTAR Framework, which offers a structured approach to implementing synthetic data in secure research settings. The framework addresses key themes including governance, public trust, disclosure control, and real-world implementation, helping organisations balance data utility with privacy and security requirements.
The report highlights the growing role of synthetic data in enabling safer access to sensitive datasets, particularly in health and administrative data research. It outlines how synthetic data can support researcher training, accelerate methodological development, and enable collaboration in scenarios where real data cannot be directly accessed or shared.
Strong engagement from across the research community demonstrates a shared commitment to responsible innovation and the need for clear standards and best practices. Overall, the report positions synthetic data as an increasingly important tool within TREs, while emphasising that its successful adoption depends on robust governance, transparency, and maintaining public trust alongside technical innovation.
Read the full report here.