Patient Data: Research

My Lords, after listening to the last debate, I am tempted to speak only about how health research data might help economic growth to alleviate all the debt that we were talking about, but I will not. I thank the noble Baroness, Lady Merron, for taking this debate and ,all noble Lords who signed up to take part, despite the brief time for which they are allowed to speak. I also thank the Library for its excellent brief and POST for the publication of its report on the subject of health data research. I fully support the Government’s initiative to strengthen the collection and use of health data for research but I hope that the commitment from the Government will be longer term and that they will be prepared to do whatever is necessary to make the UK a world leader in health data research. The use of health data to improve healthcare is not new. In the 14th century, plague killed one-third of the population because they did not have the ability to use data and track the epidemic. By 1854, however, John Snow tracked the cholera outbreak in Soho and controlled the spread of the disease. There are many examples of the use of data and epidemiological studies leading to improvements in healthcare, such as the association made in 1954 between smoking and lung cancer and chronic lung disease and, later, the association between HPV and cervical cancer. Most recently and quite importantly, the UK Cystic Fibrosis Registry, kept by the Cystic Fibrosis Trust, identified genetic mutations that occur in some children and adults who suffer from cystic fibrosis, and identified an effective drug that improves their ventilatory function. The advent of AI and machine learning, genomics and pharmacogenomics, and the linking of such data through machine learning and imaging technology will transform the research using health data for better healthcare and innovations. Better data means fewer deaths. A strong, efficient health data platform that can be accessed for research is crucial for the success of the life sciences strategy and for economic growth. Currently, the life sciences sector has a turnover of £146.9 billion and employs over 360,000 people. Apart from delivering better healthcare, health data research is the fuel that will accelerate that by adding more than £10 billion annually in value added growth. The UK’s strands for health data curation have the ability to track longitudinal data from a diverse population of 63 million, but the current system is fragmented and scattered across thousands of GP practices, hospital trusts, community health clinics and others that deliver healthcare. However, the UK health data environment is changing from a siloed system to a national framework. One hopes that the active phasing-out of data sharing, which often led to data breaches, will now stop that from happening. Data sharing will be replaced with the secure data environment, SDE, and the data stay-put model, where the data will stay on one platform and not be shared but rather acted upon. Researchers will be allowed to access only that data, and it will not be exported. The Five Safes model for data access and outputs will make the data secure. The launch of the HDRS, the Health Data Research Service, which is backed by £600 million of Government investment, keeps the promise going. That is all very good, but the verdict is that, while it is a great blueprint, delivery is the test. We have good plans, but let us hope that delivery will follow. What are the challenges? Currently, regional secure data environments operate in silos in over 180 trusts. Often, they are written and uncoded, which means that using them is difficult. They are diverse, handwritten and uncoded, and they occur in 180 trusts and thousands of GP practices, and are therefore not connected. Data is currently written and collected in different hospitals and general practices but is not unified in any data models. The governance to access data for research is multi-layered, causing bottlenecks. One has to answer to so many different regulators just to be allowed to use the data. If the UK is to become a leader in clinical trials, health data needs to be coded and easily available in every hospital. To maintain public trust, there need to be clear, transparent, and legally binding arrangements as to how data can be accessed and used for the pharma and tech industries, and how the NHS will benefit from the benefits that accrue, both in healthcare improvement and money. I personally approve of the opt-out model rather than the opt-in model, because the latter will not work for health data research. What are the consequences of not addressing these challenges? If we do not address them, it will affect the NHS; it will affect patients, because they will not receive modern care; and it will affect the UK economy. I therefore have four key questions for the Minister. First, what specific legal mandate will be needed for longitudinal data streams that link GP and hospital data? Secondly, what effective plans will be put in place to end the fragmented secure data environment? Thirdly, on public trust, will the Government legislate that any benefits from the use of and access to NHS data by commercial companies flow back to the NHS? Fourthly, are there plans for trusts to have a workforce, such as data engineers, to clean up the data and produce the appropriate codes that will be used by AI and machine learning to national standards and, if so, how will they be funded? I look forward to the Minister’s answers.

My Lords, it is a great pleasure to follow the noble Lord, Lord Patel. I congratulate him on securing this very important debate and on his insightful and excellent introductory speech. The National Health Service is one of Britain’s greatest institutions. After 30 years as a front-line NHS consultant in women’s health, I believe it can be more than a healthcare provider. The NHS holds the richest longitudinal health dataset in the world. It can help to improve outcomes, accelerate innovation, create intellectual property, patents and high-value jobs, and attract global investment. The new Health Data Research Service presents a significant opportunity. Our ambition should go further, to create the high-quality representative datasets needed to develop and validate AI safely. AI can analyse data at a large scale in a way that humans cannot, predicting risk, promoting early detection, accelerating drug discovery and clinical trials, and enabling the much-needed shift from treatment to prevention. However, AI is only as good and equitable as the data it is trained on. We need richer data points—not simply age and sex, but ethnicity, disability, socio-economic deprivation, postcode and wider determinants of health—to identify who is missing out and why, why outcomes differ and how to address them. Diseases do not care about borders, and nor should any medical research. The UK should ensure regulatory interoperability with international standards and lead global research with our data. This is essential in particular for rare diseases and cancers, and for ethnic minority communities who are underrepresented in research. Public trust must be the foundation. Patient consent, privacy, security, transparency and appropriate safeguards must be embedded from the outset. If the challenges my noble friend Lord Patel described are addressed, the NHS patient database can become a powerhouse for health research and AI-driven innovation, and economic growth. Can the Minister say whether the Government will ensure that the Health Data Research Service develops representative datasets for AI research and also pursue international data interoperability, so that Britain can lead in equitable and rare diseases innovation, while driving economic growth? Good patient data can drive good medical research and help deliver good health and good growth together.

My Lords, I thank the noble Lord, Lord Patel, for securing this debate. I declare something of an interest, in that I am just back from the ninth World One Health Congress in Lisbon, partially supported by DSAC. I go to the UKRI to start with the basic fact that the social determinants of health determine up to 80% of health outcomes, with clinical care accounting for only 20%. We must not lose sight of this reality, particularly given that healthy life expectancy is down two years in the past decade. We do not need data to know that poverty, inequality and austerity have enormous health costs. Data collection does not really add to our understanding of a reality that includes the impacts of financial insecurity, poor housing, filthy air and discrimination. The noble Lord, Lord Patel, spoke about economic growth having an impact on health. I would argue that we need a major redistribution of the resources in our society to tackle those things. Waving the magic wand of AI and saying we will fix it with that just does not satisfy. However, when we come to health data, and there is no doubt it is extremely important and useful, we have a huge problem of trust: patients trusting the system and trusting the Government to govern if their data is handed over. If there are no transparent controls, patients are not going to allow it. I note that the Government themselves, in a letter to the chair of the Commons health committee, acknowledged that, in recent times, tens of thousands of patients have withdrawn their data from use in research projects, with very many of them expressing concern about the involvement of the US defence and health tech company Palantir. There really is no wonder about that. We need to think about this. Peter Thiel and Alex Karp have said some very disturbing things, but I also go to Palantir’s UK head of operations, Louis Mosley, at last year’s Alliance for Responsible Citizenship. He said Western civilisation was in an existential battle against the “armies of fact-checkers and experts”, lawyers, academics and journalists. Palantir is supposed to be supplying unbiased, fact-based data analysis for the NHS: this is the company that we are talking about. We know that it had very little track record in health data before it enlisted the aid of the company of the former Member of this House, Lord Mandelson—Global Counsel—and secured this contract. The Government have a really important decision to make on health data and whether they continue with this contract. We must not allow ourselves to be trapped into using these companies. Also, as the noble Lord, Lord Bridges of Headley, referred to in the last debate, we have to ask questions about their financial stability; if we rely on them and they are not financially stable, that is another issue as well.

My Lords, I am very grateful to my noble friend Lord Patel for securing this timely debate, in which I only have time to speak about UK Biobank. I declare an interest, as I have been not only a participant since 2007 but also a co-investigator in a research study analysing UK Biobank data to develop AI algorithms for the early detection of atrial fibrillation. Ahead of this debate, there have been some very biased, negative briefings about UK Biobank—a not-for-profit charity—with multiple claims backed by no evidence whatever. UK Biobank is a visionary project in modern science, admired the world over. It has assembled a unique dataset from 500,000 UK volunteers. Over 22,000 scientists have been approved to analyse the data, contributing to 18,000 peer-reviewed papers. Now, we all know that access to UK Biobank data was offered for a very short period in April on a Chinese e-commerce website. There was swift action to deal with the data breach and the advice from China was that there had been no sales of data. Since mid-2024, Biobank has been making a transition to its data being available only within a trusted research environment, a TRE, with an airlock. The data from UK Biobank does not contain any personally identifying information. It is de-identified. No participant has ever been re-identified without their active co-operation. All participants were informed of the data breach in China and given the opportunity to ask for their data to be removed. The number of those who did so is staggering: 344 out of 500,000 participants, less than 0.1% of the total number of participants. This voter confidence, I believe, stems not only from the altruism of the participants but from the awareness that UK Biobank has already enabled new diagnostics and treatments that are saving thousands of lives: for example, through polygenic risk scores being added to heart disease risk measurement, through non-invasive diagnosis of fatty liver disease and through near real-time automated cardiac MRI analysis. The current UK Biobank model for patient data research could be scaled up as a result of the introduction of the single patient record. I hope that the Minister will be able to confirm that the NHS Modernisation Bill will make this possible.

My Lords, I very much thank the noble Lord, Lord Patel, for introducing this debate. I very much agree with the points he made and I hope the answers will be forthcoming. Clearly, I support the use of patient data for research; the benefits from medical innovation, prevention and treatment are substantial. But I want to focus on a narrower issue: the gap between what we know about public opinion and how policy responds to it. The evidence consistently shows strong public support for the use of health data in research, yet the same evidence also shows continued concern about how data is handled and, in particular, the role of private companies in accessing and using that data. These concerns are acknowledged in many reports but are too often treated as secondary. This is especially important for mental health data. Research indicates that willingness to share mental health information remains high and differs only marginally from willingness to share physical health data. However, trust is critical. People’s confidence depends not simply on whether data is used but on who is using it and for what purpose. For those with stigmatised conditions, the consequences of any misuse or breach can be profound. Recent events have demonstrated that public confidence cannot be taken for granted. I was not going to mention UK Biobank, but since it has been mentioned, I hear the defence clearly. I am a fan of it and a participant in it, but that case illustrates the problems we face, because once trust is lost participation may decline, undermining the very research that depends on public co-operation. I therefore ask the Minister two questions. First, will the Government consider establishing a standing oversight body with substantial patient representation to review access to particularly sensitive categories of data, not least on mental health, and to publish the reasons for its decisions? Secondly, where commercial organisations benefit from access to NHS data, what steps will the Government take to ensure that the knowledge gained is returned to the NHS and to patients?

My Lords, this House’s Science and Technology Committee, which I chair, has been investigating innovation in the NHS, including the use of patient data for research. We expect to produce our report in the next few months. There is no question that the UK has unique and exceptionally rich health data. Better use of it would save lives, benefit the NHS and grow the economy. However, enabling access to this data for research has been a long-term policy problem. Professor Cathie Sudlow’s review warned that access is in danger of going backwards since the pandemic. The £600 million Health Data Research Service, the HDRS, is intended to help address this. It has the potential to be transformative, but there are two key areas of concern: funding and public trust. First, on funding, after its initial five-year budget period the Government have set a target for the HDRS to be self-funding through commercial partnerships. But since there is not yet clarity around which services it will offer, what its expenditures will be and what prices it will charge, we cannot assess whether this is realistic. High access fees on existing health data research projects already lock out life sciences SMEs and could push researchers overseas, where they could access health data more easily. The self-funding model risks the HDRS ending up dependent on large pharmaceutical companies just to maintain its operations. Could the Minister explain how and why the department decided that the HDRS should be self-funding, and will she publish the analysis behind that? Secondly, there is the question of public trust. Its importance has already been referred to by my noble friend Lord Patel in his excellent opening speech and by other noble Lords. There is a long list of well-meaning policy initiatives to enable access to patient data, including care.data and the GPDPR, which were undermined by a loss of public trust. Public support for using health data for research is real but conditional—it is undoubtedly weaker when it comes to commercial uses. Trust cannot be assumed. It has to be earned through transparency, clarity about how data is used and visible, tangible benefits to patients. Can the Minister tell us how the HDRS will become a trusted data custodian? What safeguards will it use for patient data and how is it going to engage with the public? I hope the Minister will tell us how the HDRS will address these important issues about funding and public trust. It is vital to ensure that we can realise the immense opportunity offered by full use of patient health data.

My Lords, I join others in thanking the noble Lord, Lord Patel, for securing this debate. Across a range of medical conditions, we are hopefully on the brink of medical advances, underpinned by robust research, which have the opportunity to save lives or transform them for the better. With the structure of the National Health Service, we have a unique asset which puts us in a much better position than most other countries across the globe. Data can be of major benefit when it is collated correctly by government, its quality is uniform in nature and it is used effectively. I will give an example from outside the health service. During the pandemic, in Northern Ireland we had a single education authority which held all data for every person in Northern Ireland on free school meals and the uniform grant. That meant that, when support needed to be given to families in need, we could identify them straightaway. About 98% of families could be helped from the very start, which contrasted with most areas of Great Britain, where the data was quite often of a different nature and fragmented between different local authorities. I think that the Government’s response to the Sudlow review, in terms of the HDRS, is the right way forward, but there are four conditions which are critical to implementation. First, it is about not simply the quantity but the quality of data, particularly as most health data tends to be qualitative in nature. As has been highlighted, it is critical that data is codified, but must also be coded in a consistent way, and that we have linkages between datasets. Secondly, it is critical that we provide clarity and certainty for medical companies and researchers on what can be available. We need to ensure that data is research ready and that companies know what data they are getting, what the cost will be to them and how long they can have that data. That will enable them to plan research for the future. Thirdly, as has been highlighted, public trust is critical. One of the main barriers whenever the Government looked at some form of national identity was public concern over how their data would be used and whether it would be secure. Particularly now with malevolent foreign states and criminal gangs, we know that there is a risk of data breaches and data capture, so we need to ensure that what is put in place is secure from the start. Finally, it is important that we are comprehensive in our approach to geography and subject matter. We need to ensure that, for example, if we are going to integrate social care, that is factored into data provision, and that we have systems which take account not simply of NHS England but of services across the UK, involving all the devolved institutions working together to provide that data.

My noble friend Lord Patel, in introducing this important debate, spoke about health data secure systems. I will speak about the Secure Anonymised Information Linkage system—the SAIL database—based in Swansea and funded by the Welsh Government, which runs on the mantra of “public data for public good”. It has been going for almost 20 years and has data reaching back a quarter of a century. The SAIL Databank was highlighted in the Health Data Research Service (HDRS) Digital Ecosystems Analysis report, because it has expertise in data pipelines, advanced data linkages and secure research infrastructure, powered by the Secure eResearch Platform, or SeRP. These capabilities are essential for the future of health data research.