Building HDRS through learning

By Alsdair Cowie-Fraser, Chief Technology and Product Officer, Health Data Research Service (HDRS) 

Building national health data infrastructure is not simply a technology challenge. It is an integration challenge. 

Bringing together different systems, standards, organisations and ways of working demands listening, learning what works (and what doesn’t), and a commitment to building on what already exists. 

Across the UK, there is already considerable expertise and investment in making health data available for research. Our job at HDRS is not to start again, but to connect that expertise, build on proven approaches and create the shared capabilities that will enable researchers to work more effectively across the UK. 

We cannot wait until HDRS is fully established to start learning how best to do that. That is why the 43 technology transformation projects being developed across the UK’s health data ecosystem are so important. They allow us to test solutions to real-world challenges, understand what can be scaled and deliver value while we build. 

Spanning five areas, these projects will provide practical learning to inform the technology, products and services we develop. 

1. Making health data research-ready 

Data being available does not necessarily mean it is ready to use. 

Different standards, terminology and levels of completeness can make preparing health data for research complex and time-consuming. 

Projects ranging from making hospital prescribing data research-ready to developing common datasets for diabetes, dementia and cardiovascular, renal and metabolic conditions are tackling these barriers. Others are testing how unstructured clinical information can be transformed into usable research data. 

For HDRS, the opportunity is to identify which standards and approaches can be reused, improving data quality and reducing the time researchers spend preparing information before they can use it. 

2. Connecting data across systems and borders 

The most important research questions rarely fit neatly within a single dataset, organisation or geography. 

Projects exploring federated linkage, cross-environment cohort discovery and the integration of primary care, genomic and specialist datasets are testing how information can be connected and analysed securely across different environments. 

Their learning will help HDRS develop the interoperability and shared capabilities needed to support research at UK scale. 

A national service does not mean making every system identical. It means enabling different systems to work together effectively. 

3. Simplifying secure data access 

Making data easier to access must go hand in hand with maintaining public trust. 

Researchers currently navigate different processes for discovering data, securing approvals and accessing analytical environments. While safeguards are essential, unnecessary variation and duplication can create avoidable delays. 

Projects testing shared enquiry routes, standardised approvals, improved metadata and reusable processes for checking research outputs are exploring how to make access simpler and more consistent. 

For HDRS, they offer valuable insight into what can be standardised or automated, and where human judgement and local accountability must remain central. 

The ambition is faster, more straightforward access without compromising security, transparency or public confidence. 

4. Enabling the next generation of research and AI 

We must build infrastructure for the research of tomorrow, not just today. 

AI, multimodal analysis and advances in computing are creating new opportunities for research, but also new demands on infrastructure. 

Projects exploring secure access to AI computing resources, federated learning, medical imaging and large language models are helping us understand those requirements. Others, including SENTINEL-AI, are testing approaches to evaluating clinical AI and monitoring its performance once in use. 

These projects will inform how HDRS evolves its technology and services to support increasingly sophisticated research, data-driven clinical trials, regulatory evidence and post-market surveillance. 

5. Standardising and scaling what works 

Innovation should not mean solving the same problems repeatedly. 

Across the UK, technical teams have developed valuable tools and capabilities, often independently, to address similar challenges. 

Projects focused on shared architectures, interoperable platforms, reusable software and common approaches to analytical environments are exploring how these capabilities can be adopted more widely. 

For HDRS, the challenge is understanding what should be standardised nationally, what should remain flexible locally and how successful approaches can be scaled without imposing a single technology solution. 

Our success will depend not just on what we build, but on how effectively we enable the wider ecosystem to work as one. 

Learning through delivery, not designing in isolation 

Together, these 43 projects represent more than individual technical developments. They are an opportunity to shape national infrastructure through practical experience, rather than assumptions. 

From a technology and product perspective, three principles are particularly important. 

First, build around real problems, not assumed requirements. Working alongside delivery teams gives us direct insight into the challenges facing researchers, data custodians and technical specialists. That evidence must shape our priorities, product design and investment decisions. 

Second, design for reuse and interoperability from the outset. A solution that works in one environment is valuable. One that can work across many has the potential to transform research nationally. 

Third, test, learn and adapt. Not every prototype will become a national service, and not every approach will prove scalable. Discovering what doesn’t work is as important as identifying what does. It allows us to make better decisions, earlier. 

This is what learning while we build means in practice. 

HDRS has an opportunity to bring together the UK’s existing strengths and create something greater than the sum of its parts. These projects will help us understand how to do that – while delivering improvements for researchers along the way. 

Because building a truly national health data infrastructure is not about starting from scratch. It is about making what already exists work better, together.