NHS Digital’s director of research and clinical trials, Michael Chapman, looks at the organisation’s development of a trusted research environment (TRE) in the NHS. He is joined by Health Data Research UK’s Susheel Varma who explains the benefits of using TREs for data sharing and Professor Angela Wood and Dr Will Whiteley of consortium CVD-COVID-UK to discuss how the new environment is working for them on the front line of clinical research.

There has been much debate in the past twelve months about the future of health data sharing. The announcement last spring of a proposed change to the way primary care data will be collected in England has energised public debates with conversations across the country about the need for sharing health data in a transparent, ethical and secure way and how to best achieve that.

Partly as a result of that debate, we are changing how we share data, by moving from a data release system to data access system.

In a data release system, we produce a minimised cut of data that has been requested and send it to the requestor who then analyses it in their own system. This approach has been in place for many years, with contractual safeguards around the data, oversight of releases by independent bodies to ensure appropriate decision-making, and a robust audit process to make sure that data is handled by recipients ethically and securely.

In a data access system, we flip this on its head by keeping the data within the full control of the data custodian (such as ourselves) and only providing secure access to data to approved researchers within the security perimeters of a Trusted Research Environment (TRE).

The evolution of technology solutions and innovative governance approaches has provided us with an opportunity to strengthen the security, integrity and confidentiality of the data and help reassure the public about their privacy, whilst balancing the utility of the data for research, both scientific and operational.

Professor Ben Goldacre’s recent review of the use of health data for research and analysis set out in detail the challenges of the current data release system as well as the advantages of a data access system.  Organisations such as the Office for National Statistics, SAIL Databank, Public Health Scotland and the Honest Broker Service Northern Ireland as well as projects such as OpenSafely, the British Heart Foundation Data Science Centre and others have been working towards providing secure access to data within a Trusted Research Environment for several years.  We, too, are working towards the same goal.

Trusted Research Environments provide approved researchers from authorised organisations with timely and secure access to health and care data, then enabling them to collaborate, link data and share code and results, all within the same research project.

NHS Digital’s service, funded by the NHS England Data for Research and Development Programme, is a secure data platform with the analytical and statistical tools to support researchers to conduct productive research on the deidentified data they have been granted access to.

Their findings, but not any raw data, can then be exported safely, subject to disclosure checks and ensuring the formats and analyses are approved and sent only to authorised users. The service can then monitor and track who has accessed the data and it can never be shared with unauthorised people or linked to other datasets without our knowledge.

So, by its very design, the service aims to protect the privacy of those whose data is held in them, so patients can rest assured that their health records are handled safely and privacy and confidentiality risks are minimised.

We demonstrated this approach during the pandemic, working with the partnership between Health Data Research UK (HDR UK) and the British Heart Foundation Data Science Centre who have been using our service to answer some of the most high-profile and important questions surrounding the cardiovascular impacts of COVID-19 and the vaccines developed to fight it.

Unlocking the power of secure data access

The TRE approach is one that has been long advocated by HDR UK, the national institute for health data science.

Susheel Varma, their interim chief technology officer, says that the organisation is committed to the approach of data access through privacy-enhancing solutions like TREs, that have robust and independent accreditation, monitoring and auditing, and promotes good data governance.

“Trusted Research Environments have so many benefits, to the public, to the health and care sector and to clinical researchers,” he says.

“TREs provide authorised researchers who have an approved project with access to the data that they require, which is de-identified and provided within a secure environment to perform their analysis and generate research outputs without the risk of disclosing a person’s identity and privacy.

“This provides greater assurance that sensitive data is handled securely, as its use can be tracked, with technical safeguards to ensure that no potentially disclosive data leaves the secure environment.

“Therefore, they make research safer, so that people can be confident that their personal health data is accessed securely and their privacy is protected.

“Researchers are provided with a single location to access the valuable datasets they are approved for, so the data and analytical tools are all in one place, like a secure reference library.

“Research is therefore made more efficient, collaborative and cost-effective, providing rich data that enables the deep insights necessary to drive innovation in healthcare.

“By using these more trustworthy approaches to linking UK health data, we are enabling discoveries to help save and improve people’s lives.

“We are working with data custodians and TRE providers across the UK to establish this approach as best practice and have published, in partnership with the UK Health Data Research Alliance, a principles paper for trustworthy data research through TREs.”

Population-level data at subset access speeds

TREs stand to unlock huge benefits for patients and the public.

Angela Wood, a Professor of Health Data Science at University of Cambridge has been collaborating with the BHF Data Science Centre (hosted by HDR UK) through the CVD-COVID-UK consortium to investigate the relationship between COVID-19 and cardiovascular disease.

“There are many benefits to accessing and performing research on population-wide healthcare data in the NHS Digital TRE during the pandemic,” she says.

“First, it provides a novel scale and depth of information, leading to enhanced precision of findings so we can be more confident in our conclusions, and enabling a wider spectrum of research studies, including being able to investigate extremely rare outcomes over short periods of time.

“Second, it produces results which have unquestionable generalisability across important subgroups, including age, ethnicities, geographical locations and socioeconomic characteristics. Examining, modelling, and reporting Covid-19 disease trends across a whole population reliably informs public health policies.

“The application process to use the data is rapid – the BHF Data Science Centre team are great at getting the secure and trusted data access approvals in place very quickly, and once access is approved, the data and analytical tools are accessed within the same secure environment.

“There is also a growing team of health data scientists, who are dedicated to guiding analysts with the data curation process so they are able to get going with the research more quickly.

“Access to linked population-wide structured healthcare data within the NHS Digital TRE,  including data from hospitalisations, primary care, death registrations, prescriptions and Covid testing and vaccination, has been key for Covid-related research.

“Using information from 46 million adults, we have been able to study the rare but serious clotting events after Covid vaccination; however we’ve shown these side effects are outweighed by the vaccination benefits of reducing Covid morbidity and deaths. We are also studying the long-term effects of COVID on people’s vascular health to pinpoint when they get back to their pre-COVID risk level.

“If we did not have access to this data resource for research, we wouldn’t be able to tackle these important policy related questions.

“I would certainly advise other researchers to consider how this population-wide data resource would help them address future health-related research priorities.”

Democratisation and collaboration

Dr Will Whiteley, consultant neurologist and Reader at Edinburgh University and senior research fellow at the MRC Population Health Research Unit has also been part of CVD-COVID UK team.

“The TRE has definitely democratised the process of data access,” he says.

“Firstly, it allows a whole project to be carried out in a junior researcher’s two-three year contract; before, it took years just to access the data.  It’s really important for developing researchers to experience designing a project, curating the data for it and delivering it in the course of their postdoctoral or PhD fellowship – otherwise they don’t experience creating and finishing a project.

“Many of my peers across academia and clinical research don’t use NHS Digital data because very few want to start on a years’ long process of applying for the data.

“Having a TRE to access healthcare data is much easier than starting from scratch.  Using the NHS Digital data in the TRE, you can study a million people and have it done within a year or even less.  That will transform our knowledge about the delivery of healthcare.

“Finally, it’s great that we can now get teams working in one place on the same data.  The team that I work with are based in Swansea, Cambridge, Bristol and Edinburgh, so it’s a multi-institution project working with the same data.

“The TRE is currently not without its challenges and the future is very uncertain for non-COVID-19 data sharing.  However, we have a duty to the patients and people who use the NHS who expect us to use this data effectively for improving their care.”

The future for TREs

At NHS Digital, we will deliver a flexible and scalable solution that supports users’ needs in a safe and secure way.  The change to a data access model will take time but, over the next few years, this will become the default route to access the data we hold.  The benefits to patient care and how we learn about disease cannot be underestimated.

This has been recognised centrally with a joint funding package of up to £200m recently announced by the Government,  a significant proportion of which NHS England has committed to investing in Trusted Research Environments.

Dr Claire Bloomfield, deputy director and senior responsible officer for the Data for Research and Development programme in NHS England’s Transformation Directorate said, “We have been delighted to work with the NHS Digital Trusted Research Environment team as funders during 21/21, which has led to fantastic work such as that outlined by Dr Whiteley and Prof Wood. This foundational work has opened the door for further funding to expand the scope, capacity and capability of national health data infrastructure underpinned by patient and public engagement and involvement.

“We look forward to continuing to work with NHS Digital to expand the range of research possible within the TRE and to ensure the quality of experience is sustained across key use cases beyond COVID, including in cancer and cardiovascular disease.”

Data sharing has protected vulnerable people through shielding during the pandemic, and it identified dexamethasone as a viable treatment, saving over a million lives. It could do the same for other diseases, like cancer, obesity, diabetes, strokes or kidney/liver disease.

These are the benefits that underscore the need for ethical, secure health data sharing that everyone can be confident in, and we are determined to deliver them.