35
Examination of witnesses
Baroness Blackwood of North Oxford and Dr Melanie Ivarsson.
Q139 The Chair: Welcome to the Select Committee on Science and Technology, where we are investigating personalised medicine, AI, and the impact on and potential for the NHS. We are very pleased to welcome our two witnesses: Baroness Blackwood of North Oxford, chair of the Health Data Research Service, and her colleague Dr Melanie Ivarsson, who is the CEO of the Health Data Research Service.
My first question is for Baroness Blackwood, and then the next part is for Dr Ivarsson. Lady Blackwood, when you are introducing yourself, with the Health Data Research Service still in the process of being set up—it was set up fairly recently—could you set out for us your vision for it? What will its core functions be and what are your priorities over the next year? When the HDRS is fully operational, what will researchers or companies be able to do with UK health data that they cannot do today? With that, would you like to introduce yourself and to start?
Baroness Blackwood of North Oxford: Thank you, Lord Mair. I would like to also declare my interests as set out in the register. In addition to being chair of HDRS, I chair Genomics England, which I know you have already heard from in this inquiry, and I am a member of the board at BioNTech and the RTW Biotech Opportunities fund, which are a little bit more distant but relevant.
Thank you for the question. This is a hugely exciting moment for innovation in healthcare. It is currently too slow in the UK to set up clinical trials and drive innovation and new therapies and technologies for patient benefit and the improvement of the NHS and wider societal benefit. HDRS has been set up with a clear mission, which is to unite the UK’s fragmented health data, which we are all familiar with. We intend to act as a single front door for health data for research, to streamline and accelerate data access, to unlock data-enabled clinical trials, to support regulators and accelerate risk-proportionate regulatory approvals, and to support post-market surveillance.
With these key priorities, we believe that the UK will become a better place to do clinical trials but it will also improve the health and life expectancy that we see in the public. I will let my colleague Dr Ivarsson take over on how we will operationalise this, to the extent that we understand it at this stage. It is still very early in our set-up and incorporation, but we have done a lot of thinking about this and I know that she would like to give a statement on that.
The Chair: Dr Ivarsson, when you are talking about the operational side, perhaps you could tell us your thoughts about the timeline for bringing the service online and provide some more information on that, please.
Dr Melanie Ivarsson: Of course. I am four months into my tenure as the CEO, and I have spent most of my time talking a great deal to academics, people who do clinical research, industry, charities, those who use the data, as well as those who have trod this path in other ways before, to understand how our health data is currently located and how it is used as we build out the Health Data Research Service. I have also talked a lot to patients and the public about what is important to them around setting up a service that has public trust and the safety and security of health data built in from the very beginning.
We have laid out our plans in our first-year business case, which is working its way through the Government and the Treasury at the moment. That is broken into five key activities within the first year.
We have a number of existing research infrastructure elements that we will look after and fund in a run-and-maintain setting this year. That includes the secure data environments that are set up regionally: OpenSAFELY and DigiTrials. We will also be mobilising HDRS into a functional company; that is, hiring the talent and starting to set up activities such as HR, finance, procurement, technology, patient and public trust experts.
We then are spending a lot of time in parallel setting the blueprint for our future-looking strategy as we look to submit the next business case, our two- to four-year business case, towards the end of the year. That has multiple elements to it, including the technical architecture and the commercial model that we will be putting forward as part of that blueprint.
We are also embarking on the long-term transformation of the existing ways that approved researchers access data for research. That means putting together a number of what I would call “no regret” moves, so trying to understand how we might be able to improve the existing infrastructure by making some appropriate investments to strengthen the foundations.
Finally, we will be launching a number of driver projects. We will be launching the call for those driver projects in the coming weeks, and we will be investing in a number of driver projects in this first year. Those will, I hope, have many elements that are UK-wide, as HDRS is being set up to include England, Scotland, Wales and Northern Ireland. Those driver projects will help us learn about the services that we wish to offer and ensure that we are setting things up appropriately for the long-term sustainable service we are trying to create.
The Chair: Thank you. That is very helpful. Before we move on to more questions, are you able to say anything about the key metrics that you envisage as being a measure of your success?
Dr Melanie Ivarsson: We are trying to work out at the moment exactly what we want our metrics to be, so we will have a number of those that will be reported to the board as we develop the service. We are very clear that HDRS needs to make access quicker, so we will undoubtedly have a metric that looks at what it takes to access data for research today versus in the future.
As Baroness Blackwood has already stated, we think the impact of HDRS will also be about stimulating more clinical trials coming to the UK, as well as, I hope, looking at risk-proportionate access to innovation, as well as creating a post-marketing surveillance system for health interventions, such as devices and drugs. Therefore, the impact of those on patients, researchers and the economy will also be what we are trying to evaluate as we build the service.
The Chair: Do you envisage setting targets for those three aspects: access to data, economic benefit, and benefit to the NHS itself?
Dr Melanie Ivarsson: We are still in discussions with the team and, as the board is created, we will have those discussions, absolutely. I think that the first one we will be committing to is the access to data, because that is one that we can set up as part of the service and measure. I would like to think we would be able to measure the other elements as well as we progress with developing the service.
Baroness Willis of Summertown: Can I ask a quick follow-up question? You mentioned driver projects. Are these scientific projects? Are they looking at process? Could you give an example of what you mean by a driver project?
Dr Melanie Ivarsson: Yes, that is a very good question because I did not know what a driver project was myself until I came into this role and needed to learn. I am not entirely sure where the term comes from, but it is what the Government, as they started to draft the first-year business case, called the investment in potentially pilot projects.
For example, when health data has been used in a safe environment, at the moment the information that is then released to the researcher, output checking is a very manual process. We would like to run a pilot looking at whether you could automate that process, for example. We have lots of others that are very focused on taking the safety and security of health data first and improving the way things are working today to ensure that, as we build the service, we are not retrofitting new technology but building the service with the most advanced technology available to us, so that we can build public trust along the way as we are able to transparently show how we are building the service.
The Chair: Lady Blackwood, please come in.
Baroness Blackwood of North Oxford: It is just a short follow-up. What has been notable in the short time that we have been establishing the principles that we will operate on is the tremendous appetite among partners from industry, the NHS, charitable partners and so on to support these driver projects. The main criterion from a board perspective is a strong awareness that we need to build social consent and maintain that social consent for engaging with this data. Any project that comes through needs to demonstrate early public benefit while having long-term transformational potential. It is that balance that we will be looking at.
The Chair: Lord Willis is coming in online.
Q140 Lord Willis of Knaresborough: Good morning, Lady Blackwood. I start by saying that “health data” is a very broad term. Could you set out for the committee which datasets the Health Data Research Service will prioritise for access?
Multiple witnesses have stressed the importance of primary care and GP data and proposed that the HDRS should become the data controller for primary care data, taking liability away from 6,000 individual GP practices. Do you regard this as your highest priority? If so, will you pursue this? If not, who will fix the primary care data problem? That remains up in the air.
Baroness Blackwood of North Oxford: I will let Dr Ivarsson speak to some of the technical aspects of this. Obviously, safe and secure access to GP data is a high priority for HDRS. It is a significant gap at the moment in the research landscape, and in setting us up the Government set a mandate that the backbone of HDRS would be GP data with national secondary care data prescribing and mortality data. We will, of course, also look to engage with the multimodal and the consented datasets, which are very important for certain use cases.
The question is how we go about that and what the process is to achieve it. Obviously, GPs themselves are not the obstacle; they want their patients to receive the best possible care and are broadly content that this should be used for research. It is just how we structure that in a transparent model with the right governance, with the right security and with the right technology in order to enable this, which has advanced significantly even during the debate that you have been having in this inquiry and enables us to safely access data in a much more appropriate way that would reassure the public. I will let Dr Ivarsson speak to some of that.
Dr Melanie Ivarsson: This is one of the most important points of our consideration right now, which is: what is the data that is most valuable to researchers to be able to access? As Baroness Blackwood has stated, we have heard from academic and commercial researchers that the great amount of data that is collected within the GP record is of enormous use but, when combined with secondary, HES and other data, that allows a wealth of potential insights into disease progression and how different treatments might be able to benefit patients. We consider that to be the longitudinal spine or backbone of the data that we would like to make available to approved researchers to do approved research when the service is up and running.
With regard to the technology around that, we are working across the potential for different solutions to how to ensure that access is available. We are likely to start building the service around understanding what data is where and how access can be provided to researchers, and start with an appropriate data spine that we can carefully monitor and control.
As researchers identify more data types that may be useful to them, part of the role of the service will be to go and provide that data. For example, there is certain research within oncology that has very deep multimodal needs. It is the same with dementia and neurodegenerative diseases, where we are aware that researchers may want specific data available and we would try to enable their safe and secure access to that data.
With regard to the GP and controllership, we also have a piece of work looking at information governance, data access, controllership, which will need to collect all voices and opinions. We are going through that at the moment and that will form part of our strategy, which we will be happy to share more of over the summer.
Lord Willis of Knaresborough: What we cannot understand here is how the Health Data Research Service fits in alongside other bodies. How does it fit in alongside the other bodies in the area that also talk about the issue of health data, such as HDR UK, Genomics England, Our Future Health and UK Biobank, all of which seem to be in control of it? How do they all fit together and how do you make sure they all fit together?
Dr Melanie Ivarsson: This is a very important aspect for us, and I meet the leaders of those organisations regularly. I know that Professor Morris from HDR UK has been in to speak to you. We are actually based in the same building—the Wellcome building—so we meet regularly and talk about this very topic, as I do with other leaders of the other organisations.
They also have a different purpose and have been set up for a different reason; for example, if you look at HDR UK and the work that it does with predominantly academic research. Biobank and Our Future Health again have been set up to have a very different purpose. The data that they hold within is also different—consented cohorts or the Biobank population, again, consented.
The HDRS is accessing all data across England, Northern Ireland, Scotland and Wales. The entirety of that data is the potential data within that system from birth to death. That is an extraordinary amount of data that has the potential for researchers to access.
The way we are thinking about it, and we have been talking very openly and collaborating very closely about this, is that as we set up the HDRS front door there will be times when a researcher comes to us and we know that the best thing we should do is offer a concierge service and say, “Please go and talk to UK Biobank. We know what they do and that is the right place for you to go”.
There will be other occasions where we could work together with one of those other organisations, whether it might be setting up a national-level clinical trial or accessing different data types that different organisations are able to access more easily. I think that is one of the incredible opportunities for the UK right now: to have all these groups and organisations working very collaboratively, understanding each other and connecting closely, and creating a service for researchers that enables them to do the best possible research in the UK.
The Chair: Lord Willis mentioned Genomics England. Lady Blackwood, you are in a unique position: you are chair of Genomics England as well as chair of HDRS. How do you see the relationship between those two organisations?
Baroness Blackwood of North Oxford: It is important to state that I will be standing down as Genomics England chair at the end of the year. I told my board yesterday, so you are the second to know. The reason for that is partly so that I can concentrate on building what is a very important national asset, and partly to avoid any potential future conflict. We are just about to start recruiting my replacement.
A good way to think about this is through use cases. At the moment you can have some use cases where the data that sits within the existing datasets meets the requirement for those use cases, and some of those have already been set out by Dr Ivarsson. However, in cases where you need to join together maybe performance data with health economic data, with genomic data and so on, in order to establish early on in your development of the pipeline whether you are on the right track or you should shift, that is very difficult to do in the health system at the moment, as I know you will have had evidence for. You have to have multiple contractual agreements across the health system, which delays things and makes the UK less attractive to do that research in, despite the fact that the quality of the data, once you get it, is excellent.
The value that we can add is by working out when you can increase the value of the genomic data for research by adding it to other things, by enabling that access and accelerating it. That is how we envisage it. When we have spoken to the different bodies, they also see the value of this and the importance of it and how we should engage with it. We have not set out the exact nature of those relationships yet because we will have to go through that piece by piece.
Q141 Lord Patel: First, thank you for coming. I know that the two of you are the most competent people to lead such an exercise, so I have confidence but that does not mean I do not have questions.
Baroness Blackwood of North Oxford: I would expect nothing less, Lord Patel.
Lord Patel: You have covered already a lot of the ground in the questions that I had set out to ask. But, around that, you also created some more questions. If I was challenging you, I would say that collaboration with other datasets—UK Biobank, HDR UK, Genomics England, Our Future Health, and so on—is important. Being a concierge is just like being a concierge carrying the bags; why do you not just control the whole lot?
Baroness Blackwood of North Oxford: There are some technical reasons for that, which I will allow Dr Ivarsson to answer. It is primarily because we do not need to any more. There is appropriate discomfort from the public at the idea of having one single data lake that is controlled by one single controller. It is appropriate to understand that there is confidence in more local management of data, and we can think about how we engage with that.
Technology advances more recently mean that it is possible to accelerate access while maintaining the existing datasets—I will get there—but one of the things that we can do from the position that we have been set in with the political will is not only reach into change processes in the health system, such as procurement expectations and contract expectations, but also reach into the Government in terms of thinking about what regulatory or policy reforms need to happen in order to transform the landscape. We expect that there would be some of that as well. I do not know what you would like to add.
Dr Melanie Ivarsson: The only thing to add is that we recognise that we are being set up as a service to enable researchers to access the incredible wealth and richness of UK health data. At the moment, part of the frustration for many researchers is that they need to go to different places and it takes a huge amount of time. I hear stories of people going off and having one if not two children and full maternity leave and coming back and the data is still not available to them.
Lord Patel: They could be grandparents.
Dr Melanie Ivarsson: Where we can have a huge impact is understanding the landscape and where the data is and making sure that researchers can get the access they need. The first thing that we need to do is to have a good understanding of what is where and how we can enable a much more fit-for-purpose information governance that enables researchers to get the access they need.
Lord Patel: That is important about the governance and access going together. That leads me on to the other question, which is: say, if a researcher was working in preventive aspects of health—and you might answer me—which dataset will be the core dataset that they need to access? As you said, your ambition is to be a single door for access to health data for research purposes. How would that then be accessed and how would that be joined up? I can access UK Biobank data, but to allow that to be linked up to your core NHS data, how would that happen?
Dr Melanie Ivarsson: It will depend entirely on the research project that that researcher wants to do, and obviously the different datasets that they want to access. We have not yet got into the technical details of working with data that sits within these different organisations. Our starting point is the NHS data that currently resides within the regional secure data environment. That is largely the 11 SDEs in England, and then working with our colleagues in Scotland, Wales and Northern Ireland. That would be starting off by identifying primary care, secondary HES, mortality and pharmacy.
That will answer the vast majority of health-related research questions, we believe, from what we have heard so far, and will be a huge benefit if we can reduce that access time from years or months to weeks. As researchers request more complex datasets, we will work very closely with them to identify those. Those may take longer to pull together, and we are bringing in a group of technical advisers to help support how we can do that.
Baroness Blackwood of North Oxford: That is some of the purpose of the driver projects: to work through those technical questions and make sure that we select them in a way that answers these questions for different use cases.
Lord Patel: To be quite clear about where the access for data is for Genomics England, for UK Biobank, for HDR UK, and so on, what is your access point for data?
Baroness Blackwood of North Oxford: We have not built it yet. We would imagine it would be not that dissimilar in terms of having the governance structure that would have a single front door that you could access. We could maybe write to you with more detail on that as it emerges through the development process.
Lord Patel: Just now NHS core data is fragmented and having to pull it together is difficult. If I have a research project that is likely to involve your core data and maybe data from Biobank or Genomics England, as you mentioned, would you be the service that focuses on my research project, collects the data from other sources and then gives it to me as a service?
Dr Melanie Ivarsson: We would work with the different groups to create a trusted research environment for that data to reside in, yes.
The Chair: Lord Patel, you were going to ask about the—
Lord Patel: They answered the question about Scotland and other—
The Chair: —clinical trials.
Lord Patel: The clinical trials was answered because they said they will build an environment that will help clinical trials. It depends on the environment but, to give you an example, if I wanted to test a drug that was likely only to benefit avoiding cardiac failure, for instance, in a certain group of patients who have had a heart attack and therefore are at risk of cardiac failure, your database will enable the selective testing of that.
Dr Melanie Ivarsson: One of the things we want to build—and there has been a call from the industry for HDRS to do this. The ABPI has already issued a report with its recommendations that the Health Data Research Service creates a data-enabled clinical trial capability.
We recognise that currently when clinical researchers want to do clinical research, they have to send feasibility out to lots of different sites, which will then try to identify whether they have patients within their database that they are able to invite to take part in research. DigiTrials, for example, is another initiative that I think has done a good job in trying to help catalyse clinical research feasibility.
The service that we would like to build would primarily do three things for clinical trials. The first would be to enable researchers to understand the patient journey within the UK; for example, the standard of care, the incidence of heart attack, how the patients are travelling through that particular experience. The second is where they may be located in the country, so the geomapping, and then, thirdly, working closely with our NIHR colleagues and others to identify where the sites might be that they could then perform that study.
Where I would love us to be able to get to is to eventually build a service that would allow patients to be invited to partake in research via their doctor or healthcare professional. I think that would help enormously to stimulate the clinical trial ecosystem in the UK. We have such a large amount of health data that we can access about people to understand our patients, where they are and what their experience is. I would love to see us stimulate clinical research as a form of care, that people are offered trials as a form of care in the future much more than today.
Lord Patel: I get it. We ought to become the centre for cancer research, because with Genomics England’s genomic hubs and ability to have a universal database of DNA analysis or molecular analysis of each and every tumour in the country and then to link it with clinical trials would be important, but that is a big ask.
Baroness Blackwood of North Oxford: It is a big ask, and a good way to think about it maybe is that would support more rapid feasibility assessments for the trials; site selection—going where you know that you will have the right uptake and the right data, as Dr Ivarsson said; and enabling pre-screening analyses to reduce the screen failure rates.
This is all about lowering the risk of trials and accelerating the process, which also lowers the cost, of course, for the end result. We will not be a regulator, we will not be in the position to move then to more flexible trials, but working with the HRA and the MHRA, which are very keen to do this, we can provide a lot of evidence for that.
I do think that HDRS will help to make sure that when clinical trials start, they are starting in the right place, with the right patients and focused on the right question, which will lower the risk and improve the environment in the UK for trials.
Q142 Baroness Willis of Summertown: I wanted some clarity on the data that will be available through this front door, in particular the environmental data. You have talked about longitudinal data. You have talked about the hospital data, the National Health Service data, but things such as pollution data, climate change data, access to green space, where increasingly those datasets are understanding people’s risk of and vulnerability to different environmental aspects of disease, I am interested to know how you plan on linking that in or will this front door be only for clinical data?
Dr Melanie Ivarsson: We are starting with health and care data, but we are already in conversations with other parts of the Government around other data types that might be available: education, environmental, and so on. We recognise that we have a way to go in building out the initial purpose of HDRS, such as health and care.
To include any other data types would have to be done very carefully. We would obviously want to maintain public trust throughout this process. Our primary purpose is health and care data. We are having conversations with other parts of the Government. As we build out the use cases and understand what might be helpful, we will continuously revisit this, but it needs to be done very carefully and thoughtfully, with public trust.
Baroness Willis of Summertown: I would urge you to, not least because I had a long discussion with a very senior surgeon from Great Ormond Street Hospital who was deeply frustrated that he had to go digging around to try to find the pollution data for where the children were who had a very high susceptibility to asthma. This is not just a nice to have. Some of this data is critical when we are thinking about preventive medicine approaches as well. Researchers cannot join together right now.
Dr Melanie Ivarsson: No, we recognise that.
Baroness Blackwood of North Oxford: It is a very well-made point and understood among the clinical and research community. I think that Lord Patel’s point is important, though. This is a significant undertaking. Not every undertaking like this previously has succeeded tremendously. One of the priorities of the board will be to make sure there are realistic and deliverable priorities at the outset, which can demonstrate from the beginning benefit to patients and the public and clinicians, so that we maintain that public trust. That does not mean that the long-term vision is not more significant as we go through, but it is important to have achievable aims.
Q143 Baroness Nicholson of Winterbourne: Your real success, surely, in the long term of this enormously exciting work must be from the confidence you give to the electorate, to the patients, to parents and teachers and so on, and, of course, to the GPs, who are at the moment a little bit of a blot on the landscape. Are you thinking about running alongside assessments all the time with special groups as to how you are effecting this, because you will be putting forward very big and exciting changes and they will continue all the way through, certainly with your pilot schemes, and I think for a long time afterwards? Would you think of having a constant flow of discussions with the patients, with the GPs and so on, or is that too artificial?
Dr Melanie Ivarsson: No, this is an important point for us as we are building out the Health Data Research Service. We have already undertaken extensive public and patient engagement work, and the one thing we have heard loud and clear is: “We do not want this to be transactional. We do not want to be a box to be ticked. We want to sit alongside you and build the research service”. There are many patient groups that are strongly supportive of what we are trying to do and are thrilled when they hear that researchers may be able to get access to that data much more easily.
They want us to take a long-term approach with them so that they understand our governance, understand the safety and security provisions we are putting in place around data, understand the transparency around how decisions will be made, and sit alongside us through the structure of advisory committees supporting me and Baroness Blackwood and the board.
With regards to GPs, industry and other healthcare professionals, that is absolutely all the same, and I have spoken to probably 250 different individuals and groups in the last four months. Everybody is very willing to come with us on the journey, and people have strong opinions and they can be different. It is important that we listen to those different experiences and we value the needs of the different researchers, because, as you say, the outcome of what HDRS might be able to offer the research community has the potential to be quite substantial and significant. We are creating an infrastructure of stakeholders and groups that will work with us throughout the build and maintenance of the Health Data Research Service.
Baroness Blackwood of North Oxford: It is important to state that, first, the public expect their health data to be used for improving direct care but also for research and innovation. That is the message we receive all the time, but there are conditions on that. One is that we operate with complete transparency and are clear about what we are doing and why we are doing it and the beneficial public benefit reasons for which the data is being used.
Secondly, it is the pillars on which that is built, so having the right technical architecture that operates and is future-proof and does not clog things up, and enables very high levels of cyber security, and governance structures that make sure that there is ethical access for the right reasons so that the data flows safely.
Finally, it is communicating the benefit of the data being used. You have strong consent but you have to make sure that you continuously maintain that public confidence. You do not win the public confidence and then move on. It is a daily effort. At Genomics England, for example, we fully integrated participant views all the way up to board level and had communications engagement at C-suite level, and we envisage having similar structures at HDRS.
The Chair: Lord Winston, this is the topic you want to pursue, I think.
Q144 Lord Winston: Thank you very much for you coming; it is good to see you both. To some extent I have been pre-empted because the question that I was set to ask you has just been asked, but I would like to go into that in a bit more detail. I am not convinced that we are not a little bit too sanguine about the issue of public trust. I think that there is a great deal of work that needs to be done, and I wonder whether you agree with that and how you would start that.
Specifically, clearly one of the issues is: if you are going to do this with health data research, where do you start? The general suggestion has been general practice, for example. Would you want to take control away from GPs to some extent of the data that they have?
I might just make the question a bit longer. Over 20 years ago, the Five Safes framework was set up, but it seems that there is a real issue here with data, because data seems at one level to be quite a dry issue. Data is about us, our personal data, at the most difficult time when we are often going to be most vulnerable. I do not think that that has been properly assessed. It would be helpful if you could give us some strategies that you feel that government could do, because I think that there are many things that are not being done with regard to that area of public engagement and public trust. It is very obvious in recent history, the last five years, for example, as you have seen. Lady Blackwood, might I start with you?
Baroness Blackwood of North Oxford: These are obviously core questions for HDRS, but also for the UK in how we manage health data. It will not be entirely delivered through HDRS, but we have an important role to play. The first is having, as Dr Ivarsson already set out, structured and long-term continuous engagement with the different groups, because obviously concerns that come from GPs about trust in the use of data, their relationship with patients, and liabilities are completely different from those of a patient who may be going through for different reasons. Making sure that we have that 360 perspective and maintain an understanding of the sentiment as it changes due to cultural changes, but also events that happen, will be core to what we do.
I also think that there needs to be a conversation about the technical progress that has been made, which ensures that you can understand who is looking at the data, what they are using it for, and have very tight permissions on who takes it out and why they take it out. We are in a different landscape from five years ago, so that provides some reassurance. It is also important to note here that we would be focusing on the non-consented research and de-identified data in the first instance, so some of those other questions are reserved for a little bit later on.
I would just note, with my Genomics England hat on for a moment, that the Genomics Medicine Service engages with rare disease and cancer patients who have often been searching for an answer for many years. The evidence is that they consent for their data to be used for research for the benefit of others at an extremely high rate—over 90%—and that they are a motivated group, and it is understandable that they would want their data to be used in that way.
What has been interesting is that with the more recent generation study sequencing of newborns, a similarly high consent rate has been found among healthy all-comers. When there is a proper discussion on consent and engagement about data about the benefits, the safety structures, the governance structures and so on, the consent rate is high, but that conversation has to be maintained and engaged on a long-term basis.
Dr Melanie Ivarsson: Following on from Baroness Blackwood’s point, I have spent a fair bit of my career developing new treatments for rare disease patients. You want to create a service that enables the important research to be done for everybody, from those chronic conditions that are more population impactful to those incredibly rare conditions. We know that the understanding of how data is used for research differs depending on your own personal experience through the system.
We certainly want to ensure that we are able to have an informed dialogue about research as a form of healthcare, and in some countries, for example—I have spent some time talking to the Nordic countries—the social contract already is one where the population largely expects their data to be used to help researchers better understand cancer, rare disease and how new treatments can be developed.
I think you asked whether there is more we can do around this. Yes, absolutely. I think that it will take more than just HDRS on its own trying to lead that conversation, so we are very happy to be part of that dialogue. I would love to get to a point where everybody understands how data is used for research, the value of research, and the value of the options that it may provide them or a loved one when they need access to a clinical trial or an innovation as a form of care. But there is still much to do in that debate.
Lord Winston: You mentioned dialogue, which is quite interesting because you will probably remember 15 years ago the Government looked at the process of dialogue in a different way where they had groups discussing. It was the research councils that started this. I remember this when I was on EPSRC. We started to have dialogue with people who were completely disconnected from the research. That was hugely successful in understanding where the real issues were emotionally for many people. We looked at particular issues within that. Do you think that approach, perhaps stimulated by government, might be helpful?
Dr Melanie Ivarsson: Yes. There has been a series of focus groups, national deliberations, and PPIE work that has been done. One of the things that I have also been learning about in my new role is the work that is done at the regional level by the regional secure data environment groups: south-west, north-west, London, Thames Valley, east of England. There are 11 regions within the country. They do enormously impactful regional work and have regional conversations. As we build the Health Data Research Service it is important that we build that local understanding into how we build the service.
We understand also, talking to colleagues in Scotland, Wales and Northern Ireland, that this is very important and something they do well. That continual public dialogue occurring locally will be important as we build what will be a national service.
Lord Winston: Is primary care the way to start?
Dr Melanie Ivarsson: When we listen to the vast majority of users, I think that will be important. Last week I was at Great Ormond Street talking to researchers, and how they want to build data for children with rare diseases looks a little bit different. I think that is where we have to build a service that meets a number of those important use cases for research.
I think that primary care will enable the majority of researchers to do the majority of things they want to do with data, but we will rapidly follow that with those very bespoke services that researchers such as those at Great Ormond Street would like to see.
Baroness Blackwood of North Oxford: There is also something in the transparency point about providing GPs and other health professionals with the information, the evidence and the data they need to have those conversations about use of data with patients, whether it is through a clinical trial or other means. Because the landscape changes quite a lot and it is fragmented, that can feel a bit overwhelming in the midst of a day-to-day job. I think that there is something there that would be very helpful for the Government to do.
Lord Winston: It is quite a huge ask in many ways.
Baroness Blackwood of North Oxford: It is, although developing consent models and talking about how data will be used is normal practice in a clinical trial. This is just extending it on to a broader question.
Lord Winston: I will ask more questions on commercial issues later on, but one thing is that sometimes seeing a massive health success made a very big difference to public perception. You will remember many years ago how there was huge resistance in both Houses to in vitro fertilisation, for example, but when we started to get successes it was extraordinary how people suddenly realised this was useful medically to a whole range of people who had never considered it before. Could we do that with data?
Baroness Blackwood of North Oxford: As we said, starting out with use cases that demonstrate clear public, patient and clinician benefit so it is evident early on how using the data improves care and improves lives, and communicating that very effectively and clearly, and just in general providing the information for those within the system through case studies and examples rather than simply through high-level generic information, is exactly the right approach.
Baroness Willis of Summertown: I have a slightly different question as well but along the same lines. I was sitting here reflecting on this discussion and taking it down to the GP level. I think, Lady Blackwood, that you made the comment about non-consented de-identified data. I was just thinking of the workload. Just trying to get consent from the individual in a GP’s surgery must be huge, but there will also be cultural norms. I have seen it with data we have looked at, where there will be particular groups that will not consent, so then your data is skewed. Is there any thought moving to more of a non-consented process of de-identified data to move this on more quickly?
Dr Melanie Ivarsson: This is an important topic and we are working closely with the Government and policy teams on this. We know that through the opt-out aspect it is a smaller number of people who opt out: around 3% to 5% of the population. You are right, that can vary according to where you might be in the country and your background. So we are working with groups such as Understanding Patient Data and others to try to look at this a lot more carefully.
I would like there to be more of a concept of informed opt-out so that people understand why they are opting out of something. I am not sure that is where we are at the moment. That is something we would be very keen to look at as well as part of this.
Lord Patel: Can I follow that up? The modernisation health Bill that is currently going through Parliament has for data collection an opt-out proposal without any initial feedback. Consent is obviously not there because it is an automatic opt-out, which suggests that most patients will be automatically opted in for data collection without ever knowing that they had an option to opt out.
Baroness Blackwood of North Oxford: That is the existing situation with the secondary uses for data. The most important thing is that where there is a concern that a patient wants to understand more, it is possible for them to have an informed and sensible discussion with their GP, or whoever it may be, but also where we have gained the confidence of that sector in understanding the security, the transparency, the structures that we will use so they can feel confident in recommending opting in. That is where we want to be.
Lord Patel: I am tempted but I will not be tempted to ask you the obvious question whether the Bill needs amending. I will not ask that question.
Baroness Blackwood of North Oxford: It is not a question for HDRS, Lord Patel.
The Chair: I think that is coming later.
Q145 Baroness Jones of Whitchurch: You are dealing here with a very precious commodity, potentially with huge economic value. I know that a lot of commentators feel that up until now we have given away the data too easily and we have not extracted the economic value from it that perhaps we could have done.
On the one hand, we are balancing the Treasury that is expecting a return in due course from HDRS. It is expecting you in some ways to use it as a vehicle for more inward investment, to attract more companies to come and invest here in the UK, which is an admirable ambition. On the other hand, we have the NHS that is desperate, for all the reasons that we have been exploring, to say that it wants more access to that data.
I think, Lady Blackwood, you said that your priority was the early public benefits. We understand all that, that the NHS can provide that public benefit as well, but those are quite difficult things to balance. Although you made a fleeting reference at the beginning in your key priorities to the economic advantages of all this, I wonder whether you could explain a little bit more about how you will balance all that so that we get the economic benefit of that very unique data that we have and the NHS benefits, the industry benefits, the Government benefit, and so on. How will you do all that?
Dr Melanie Ivarsson: This is a very important part of our strategy. The intent for HDRS is to be self-sustaining. We will need to charge for our services in order to become a self-sustaining entity. We have been, first, ensuring that that is something that would be very acceptable to the public. The PPIE work that has been conducted to date suggests that the patients and public are largely comfortable with their data being used in a commercial arrangement as long as it serves the public good and as long as it generates value back into the health system in some way. That is the premise on which we are building out our set of services.
Data access services today come with a commercial aspect, so that will continue. We are hopeful that we will be able to enter into a series of partnerships and relationships that enable researchers to come to access the data but stay to do the clinical trials because they are able, and that would increase the investment back into the UK.
We also have such potential to stimulate innovation within the UK, risk-proportionate approval of innovative therapies, and be the place where companies see high-quality data access and data research is occurring. We think that that offering, together with very thoughtful commercial terms, will provide for HDRS to become a self-sustaining organisation and that the investments made and the result of that would be returned back into technological advancement and improving HDRS services, but also that the outcomes would go back into the healthcare system.
One of the things that we are being very thoughtful about is part of the health economic model that is required as you look at the impact of HDRS. For example, if researchers were able to show that you could delay the onset of dementia because of X intervention, for example, that would be hugely valuable to the NHS in the UK. If you were able to change the clinical pathway, save on the amount of treatment that somebody receives or how they are managed through the system, those impacts could be significant. That will take time. Basic R&D has a very long timeline when conducted by researchers, and we are hopeful that in addition to the R&D and the clinical trials, the research itself will have a real impact on healthcare.
We are bringing in some analysts and health economists to help us look at the value of that, so that will be part of how we evolve our strategy this summer.
Baroness Jones of Whitchurch: When you are looking at the commercial contracts, which you might have to have with commercial research companies, is there a potential that you get more favourable terms if you are based in the UK or that the research is carried out in the UK? How much of that are you able to do because otherwise it seems like we are just handing over the research data to other global entities and we do not get the UK benefit from it?
Dr Melanie Ivarsson: We have had some fabulous conversations with researchers, small, local UK-based biotechs as well as some of the larger industry companies. There is real focus, interest and passion in seeing the UK life science ecosystem thrive. Everybody knows that our health data is unique in the world. The investment that companies, biotechs or researchers want to make in accessing the data and the results of that is all part for them of also stimulating the UK, because it means bringing more clinical trials and more research to the UK.
With regard to how we would set out commercial terms, we will be following all the same rules as all other government departments, so we will be following those closely to ensure that we set up appropriate relationships.
The Chair: Lord Patel, you have some more questions on this.
Q146 Lord Patel: I am acting as Lord Drayson if he was here, because he would have asked some tough commercial questions.
If you take the model of UK Biobank, which is low-fee and easy access, to, for instance, how the future health will work, which is full cost recovery model—and your model is different from the Genomics England model—there is an issue also about whether this data is so rich in information, if it goes right, that everybody in the world, particularly the United States companies, will want to access it. It is a question of whether it will be available only if you work in the United Kingdom or whether the data can be accessed offshore, which is the model that the UK Biobank uses.
Baroness Blackwood of North Oxford: Yes, it is important to clarify that UK Biobank is an independent charity and we will be a government-owned company. As such, we are bound by Managing Public Money, by the subsidy Act, the Procurement Act, by all these structures and regulations.
As a result of that, we are looking at how we can have flexibility and make sure that commercial benefit returns to the UK environment while still operating within those rules, and perhaps thinking about what could be improved in some of those rules. We are considering at the moment the full range of examples as they exist—and we recognise that there is a range of consequences of those decisions—but also looking at some examples that do not regularly appear, such as IP rights and things like that.
Lord Patel: I was going to get to that.
Baroness Blackwood of North Oxford: We have not made any decisions. At the moment we are working out what is feasible within the landscape and how we can navigate it, but the points that have been made about the need to benefit UK life sciences particularly are well understood and we understand that.
Lord Patel: Okay. If we get on to charging regimes, is the charging regime likely to be different for commercial companies, big companies as opposed to SMEs, and individuals, particularly individual academics working in NHS who have a brilliant idea about using some data? Just now, if I give you an example, you can access SDR UK data, any NHS researchers can do so, but the cost of accessing the data is huge because it is held in an environment that charges money. Will there be a differential in charging and will there be a preferential treatment for our own NHS academics?
Dr Melanie Ivarsson: At HDRS, we will operate under the rules that our charges will cover the full cost of the delivery of the service and, as Baroness Blackwood has noted, we are bound by the same rules as government departments are in how we charge, so Managing Public Money, the subsidy Act, the Procurement Act, and so on.
You are asking whether there is differential pricing according to the sector that the customer belongs to, and there are some examples, such as Genomics England, that have been successful in working through this and looking at the services offered across the sector. We will be learning from that experience as we build out the commercial model for the service.
Lord Patel: You said you were thinking about IP. Who will decide?
Baroness Blackwood of North Oxford: It will be part of the commercial strategy as we develop it and obviously we will get technical advice to make sure that it would work in practice and be long-term beneficial but also sit within the rules within which we operate. However, ultimately it will be the board that decides, with agreement from the Department of Health and Social Care and the Treasury.
Q147 Lord Patel: There has always been this problem about public trust, and the public do trust us in using their data for research, but they are a bit sceptical, as we hear, if that data is then available to big pharma or big companies that make profit out of that data. How will you manage that?
Dr Melanie Ivarsson: The patients absolutely have to be the north star of HDRS. We are creating a service actually for patients because they are the ultimate beneficiaries. This is an important topic that I get asked about a lot. That is in part because we know that developing a drug is very expensive and very lengthy, and therefore it does fall to commercial industry to do the majority of drug development. I think that over 90% of drugs are actually developed by the pharmaceutical industry.
We need to have an important balance and an important open dialogue with the public around the importance of all types of research—academic research, industry research conducted by biotechs that have a particular deep focus on one area, as well as large industry that will do research across multiple different therapeutic areas to develop drugs—and the value that that brings to them as patients who might one day need to take those medications. It is important that we bring the public along with us in not only understanding what HDRS is about but how that important research will end up developing innovative medicines.
Lord Patel: It will be part of your job to persuade the public that although it is a commercial enterprise that might make profit, it is for the benefit of universal good?
Baroness Blackwood of North Oxford: We have evidence that that can be done. It is done at UK Biobank, it is done at Genomics England, it is done at Our Future Health. That engagement and discussion are possible. It requires evidence and to show your working—to be transparent, to explain why the data is safe, how it will be used appropriately and what measures you have in place to ensure that—and to be very good at communicating the benefits and outcomes as they appear, because then it is understood, as we were discussing with Lord Winston. Having use cases and case studies to understand why it is worth it must be part of the discussion.
The Chair: Following on from Lord Patel’s question about large pharmaceutical companies, in the process of your negotiations over IP and subscription charges, how are SMEs put at some advantage over the very large pharma companies?
Dr Melanie Ivarsson: This will be an important topic and we have run some round tables with SMEs and the biotech sector to really understand their needs. Accessing data means very different things to different groups and they have very specific needs to help them get the access to the rich multimodal data that they need to do their research.
In many cases, SMEs have not even tried to do that in the UK. Because it has taken so long for them to get access to the data that they want, they have to go elsewhere, often purchasing access in other countries. It would be very positive if we were able to offer a service to bring SMEs, particularly those that are based here, to the UK to be able to use the access that will sit within the Health Data Research Service. I am very hopeful that we will be able to stimulate some growth in investment in the SMEs in the UK through HDRS.
The Chair: Do you see a way of even giving some preference to UK SMEs compared with the international pharmaceutical companies?
Dr Melanie Ivarsson: Obviously, there are various rules around how you engage that the Government have set out, which we will follow. For us, the important thing is that it is an approved researcher with the experience to do approved research and that is how we will be starting out the service.
Q148 Lord Duncan of Springbank: I want to understand how you use the health data to train in the AI world. Could you give us a bit of an overview of that first of all, and I will pick up perhaps some questions thereafter?
Dr Melanie Ivarsson: Yes. We are still very early in building the service, but it is very important that we have an AI strategy as we build out the technological architecture for the Health Data Research Service. We see the UK health data as a major opportunity and we know that it is very important that we do this very responsibly, that we maintain the safety and security of data and that we have a very specific approach to this. We are being very thoughtful.
Training on health data will be permitted but only within air-locked, trusted research environments—that is our position at the moment. Data cannot leave the secure environment, all activity will need to be audited, and any outputs will need to be reviewed before being released. That is something that we are working through with AI experts as we set up the service, because we have a unique opportunity, an advantage, for UK patients for AI to be trained on UK data rather than somebody else’s data, and applied to the UK patient journey. The outcome of being able to do this could be very beneficial, but to get there we need to be responsible and thoughtful about our approach.
Lord Duncan of Springbank: If you were to analyse the data and the output of what would be used as part of a wider research programme, do you treat the security of that data differently from how you would treat the sandbox element of the training methods used to educate the tool?
Dr Melanie Ivarsson: No. When you are handling sensitive health data, it must all be treated as sensitive health data itself, regardless. It needs to be processed in a very safe and secure way. Some of our earliest investments, actually, will be looking at this because we know it is such an important topic. Ensuring that we have the most safe and secure provision for data handled in this way will be really important.
Lord Duncan of Springbank: In my head I can see a difference between those who would interrogate the data, so people clinically qualified to do so, and those who would train the data engine. Who are these people? They are presumably different groups. Do you have bespoke trainers whose purpose is then to use clinical knowledge, or are they simply AI technicians? Who is this cohort?
Dr Melanie Ivarsson: Well, not yet, because there are only nine people in the Health Data Research Service. But that is actually the great opportunity, because we can go and talk to a wealth of experts and really understand the specific skills that we need to bring in.
We have a technology lead and we are building out a technology advisory committee, and they will be supporting us as we do this. When I talk to some clinicians, there is great excitement about the potential use for AI in both the foundation models and the application. HDRS’s responsibility will not be in clinical delivery, but we do want to create an environment where researchers can come with a proposal around doing something with data in an AI algorithm and that we can create a safe and secure place for them to do that.
We are trying to understand what the use cases are for this and then build the technologists within HDRS to support that work in a safe and secure way.
Lord Duncan of Springbank: As my final question, and I know there are others to follow, I am then conscious as to how you create the audit trail for the training so that people have confidence that the training itself is sound. I have used AI. You can ask the same question twice and get quite different answers—for no obvious reason, to me.
Dr Melanie Ivarsson: Again, that is likely to be one of our driver projects around AI. AI for data that is federated and sits in lots of different places has some unique technical challenges that we are working through at the moment. I would be happy to follow up with a more technical answer, but it is something we are still working through at the moment.
Lord Duncan of Springbank: That is very helpful. Thank you.
Baroness Blackwood of North Oxford: I am so sorry, but I just think it would be helpful to have some examples of where this has already happened safely, to patient benefit. You may have already received this but Serena Nik-Zainal, who is one of the leading cancer researchers in the world, worked with Genomics England to develop an algorithm called PRRDetect, which identifies cancer tumours but also, most importantly, genetic mutations so that they can be more effectively targeted for treatment. That is already operational.
InstaDeep also worked with Genomics England to develop the nucleotide transformer, which works like ChatGPT for genetic information, non-coding areas, which is important in the early development of a product. This means that you can go from that to a treatment in the clinic faster and cheaper. This is the sort of output that we are seeing, managed in a safe and effective way.
The Chair: Lord Patel, you want to pursue—
Q149 Lord Patel: Yes. Listening to you there, which is fantastic, that says we should aim to develop sovereign AI for health data itself, including cloud?
Dr Melanie Ivarsson: We are well placed in the UK, I think, to be world-leading on AI in health. With the dataset that we have available we are having some interesting conversations with the Government as part of the compute power that might be available to us and the AI capabilities that might be available to us. We are building HDRS through that lens.
Lord Patel: You would build your own compute power or would you have access to, for instance, Isambard Bristol University computing power?
Dr Melanie Ivarsson: We are in conversations with the Government about what access we may be able to achieve.
Lord Winston: What compute power do you think you would need and how much power resource would you need to drive it?
Dr Melanie Ivarsson: I would have to follow up with the specifics of how much we think we would need from a technical perspective. At the moment what we have been looking at is what research we would want to do or what research environments we might need to be able to create that would need access to compute power, particularly for some of the AI training that uses more compute. That is the thing we are thinking about, but it is early stages for that element of our strategy.
Lord Patel: Back to the question about building, which is the ambition, I understand that you are at the beginning of the work and it will take time, but so that we understand and so that we are able to help you and the powers that be, this is a good idea to have sovereign AI health capacity with compute power and develop a health cloud. The most expensive things are, first, compute power and, secondly, access to the cloud, because otherwise those that run the cloud now will charge a fortune and benefit from it. Is that a good proposal, that you should be thinking about going that way and that the Government should help you do it?
Baroness Blackwood of North Oxford: I think HDRS—and Dr Ivarsson will want to answer on this—needs to make sure that it focuses on what it is good at and can deliver effectively. I think that HDRS will be a sovereign capability. It is a government-owned company and will enable the development and validation of AI, clinical safety, regulatory approvals, and the operational deployment of tools. If you think that through, it is things such as access to population-level training data, data curation and quality assurance, federated learning infrastructure, and working with industry and partners on algorithm development. That is where the capabilities of the organisation will lie and I think that is the appropriate route to go. I do not know whether there is anything you would add, Mel.
Dr Melanie Ivarsson: No. It is an important question and we are developing our technical strategy in this area this summer. One of the things we are very keen to do, in addition to the operational capabilities that Baroness Blackwood has highlighted there, is to enable a service that can embrace technology as it advances. We all know that AI is where we are looking at the moment to understand those advances.
That is something that we are very mindful of as we build the service, that we are not ever retrofitting technical capability into the service but that we are building it with a very future-looking approach. That strategy will be something we will be sharing very transparently towards the end of the summer.
Lord Patel: As a parallel question, and from the answers you have been giving, you clearly have a lot of thinking you have done already and I am impressed by it, but there are issues about getting personnel with expertise; for instance, data managers who have expertise in being able to manage a huge set of databases, unclean, then it becomes usable. The number of people you require might not be available in the UK. Like sovereign AI companies that may bid for AI development, now the Government have agreed they get rapid 24-hour visa grants and seven years of free visa, and so on, are you likely to benefit from a similar arrangement, or should you?
Dr Melanie Ivarsson: At the moment, we have just started looking at the talent in the UK around the technology aspect. There is actually quite a lot, I have to say. I think that the UK has a lot of good technical talent. We also want to be very thoughtful as we grow the organisation and not overbuild. We need to be very thoughtful and build a lean organisation that is able to deliver on the services. At the moment, I am feeling comfortable with the early indication of the talent marketplace, that we can bring in the talent in the UK.
Baroness Blackwood of North Oxford: My experience from Genomics England, where we need a lot of technical expertise, tells me that while it would look like it would be challenging to recruit into a public body with the relevant pay scales, in fact the mission orientation and the obvious benefit of those organisations is really attractive as part of recruitment efforts. It matters a lot, actually.
Q150 Baroness Willis of Summertown: Moving on, particularly to the secure data environments, we have all been aware of what has happened with UK Biobank quite recently and the lack of trust, and the fact that that environment has been accessed by others. In Professor Sudlow’s review, Uniting the UK’s Health Data, she argued that there are too many secure data environments across the UK with inconsistent standards. Now, I can see this is a huge issue. I know you are talking about UK Biobank and you made the point that that is a charity, it is a charitable organisation, as opposed to the Government. So you are starting with different standards before you even set out.
Do you think it is possible that you are going be able to set these accreditation standards that will then supersede all these other standards that are out there to make it a much more fluid and faster-moving process?
Dr Melanie Ivarsson: I think that you are speaking to the very purpose of why Health Data Research Services was created, and it is an enormous privilege to be able to work with our regional SDEs to understand how they are currently set up and to have the opportunity to create the right service for the future.
I alluded earlier to the incredibly important regional public trust work that the SDEs do, and I think that should not be underestimated. It is not simply a case of merging a fragmented system together and going, “There you go, there is your national system”. It is important that we bring public trust along because I think that people take comfort from and have an increased understanding of their own local and regional services, so that is a very important element of this.
We are conducting a piece of work at the moment that looks at the regional secure data environments. That is to help us understand how they are operating today and to ensure that we can very responsibly find ways to federate or connect them together to bring the power of more health data for research use. The goal is to have consistent standards and reliable access but, again, we have to be thoughtful about this because we do not want uniformity just for the sake of uniformity, so where the technology is in some cases we can actually work with things as they are.
One thing that will be important is that for the different use cases the technological solution might be quite different. It is not about just creating a federated system and going, “That is it”. For some of the things where we have been talking to other government bodies, such as the MHRA, the way we will be creating the data that they want to access to look at post-market surveillance may look very different from a researcher who wants to come in and identify where chronic kidney disease patients are in the country, for example, and their journey. That is one of the things that will be really important.
Baroness Willis of Summertown: Can I come back slightly? This is great that it is getting the patients on board and everyone agreeing to this, but those layers of bureaucracy that are already within the system are huge, I would say. I have seen some of them. I think that if we are not careful, in 10 years’ time we will be in exactly the same place, because unless there is an appetite at the top level to really try to get these things to work in a streamlined way and have a federated system, I do not see what the problem is with that if that will bring about the solution of actually having fast access to this data to solve some of these issues. I just feel that we will carry on talking and talking and we will be no further along in three years’ time.
Dr Melanie Ivarsson: You are talking about the important aspect of the information governance, the data access governance. Yes, it is so many layers and one of the challenges is that each of the regions have their own access. There are just so many different access points, so one of the first things we will be doing—and I think that it is more than a driver project, it is actually how we have to build the service—is starting to look at all that and come up with a much more streamlined, less bureaucratic, less duplicative access service.
We have already spoken to many of the bodies that are involved, the HRA and others, around how we will approach that. We will also be working with the trusts and the data controllers as we lay out our strategy to eliminate layers. There is no point in putting a front door on a bunch of different rooms and everybody then having to just literally walk inside the door and then go and do the same thing. We have to make this a much more open-plan house where people come in safely, securely, to do the right research but we have eliminated a lot of the existing bureaucracy.
Baroness Willis of Summertown: This is one area where AI might really help.
Dr Melanie Ivarsson: Yes, it may.
Baroness Willis of Summertown: We will be able to understand the processes. I think that it is a process. A lot of it seems to be a process. I have some experience of trying to access some of these databases and I certainly have some sympathy with people saying that they had two children and still cannot get the data.
I am based in a university, so the next question there is: is your plan to accredit universities or companies to take that data and hold it off the secure platform, put it on to another secure platform to be able to do the analysis, or will it all have to go back to the central resource?
Dr Melanie Ivarsson: The starting position is that we create modernised, trusted research environments and that it is more of a reading library than a lending library approach. That is the starting position.
Baroness Willis of Summertown: Do you think eventually it will become a lending library or not?
Dr Melanie Ivarsson: I think that would depend a great deal on technology, whether we feel we could do that in a safe and secure way, but that is not our initial approach.
Baroness Willis of Summertown: Do you have a model of other countries or other areas where they have set up equivalents and what have they done? What does good look like there?
Dr Melanie Ivarsson: I have had some conversations with some other countries. Nobody has quite the same situation as the NHS. No one has so much data in one system as we do. There are other good examples of where people have been able to look at certain data types, but the UK is unique and that is part of what it offers researchers.
Baroness Willis of Summertown: My final question on these things is about interoperability. I cannot even say the word. We have had a lot of witnesses saying that is a really big problem. Is that the job of this new agency to try to sort that out as well, or who will do that? Who will make the systems align and work?
Dr Melanie Ivarsson: Yes, we are definitely looking at this as part of our technical strategy that we are developing over the summer. We recognise that we have an important near-term role in standardising and curating data for research use, but we do not want to create an inefficient manual situation. So what we really are trying to do is build scalable automated pipelines that improve the data quality systematically rather than case by case and looking at the technical solutions that enable this. This is very likely to be one of our technology projects.
Baroness Willis of Summertown: This is my question. There presumably are some things here that are bespoke projects that could be actually set to—
Dr Melanie Ivarsson: Yes.
The Chair: Are you confident that you will be able to achieve a standardisation of how data is curated throughout the NHS? Because that is another big ask. We have used “big ask” a few times in this session.
Dr Melanie Ivarsson: Yes, there has been a lot of progress on standardisation. It is not overnight; it takes time. I do not know that HDRS’s purpose is to entirely dictate the standardisation, but one of the things that we have definitely heard from many of the users of the data is that if the quality is not high enough, it cannot be used to do the research that they want to be able to do. Having that conversation will be important as we set up HDRS.
The Chair: Setting standards means what?
Dr Melanie Ivarsson: If you are doing real world evidence, for example, if there are gaps in the data you can pull data together, but obviously if it does not have the information that you need you cannot get the insights you are looking for. There has been some success in creating standards that ensure that the data is completed to a high level of quality and accuracy. That, for some use cases, would be very important, particularly in post-marketing surveillance where you are looking at the safety or efficacy of a new treatment. One of the conversations we have been having with the MHRA is ensuring that the data that is collected meets a minimum standard, for example.
Lord Patel: If you think that you can train Lord Winston and me to a standard way of writing things down so that the data is collected, you have a huge ambition.
Baroness Blackwood of North Oxford: You have made an important point, Lord Patel, which is we have to fix the technical infrastructure and find a good strategy for doing that that is realistic and achievable, but also the human infrastructure. We have to think through how we do that, and some of that will be the responsibility of HDRS but some of it will be sitting with NHS England and regulators and other bodies, so we work very closely.
Lord Patel: I am assuming the NHS app will be your major data access.
Dr Melanie Ivarsson: Not for researchers.
Lord Patel: No, for you, for data collection.
Dr Melanie Ivarsson: Actually, we will not go through the app, no.
Baroness Jones of Whitchurch: I just have a very simple question, listening to this discussion, which is: do you think that you have sufficient powers within the way that you have been set up to tackle these inconsistent standards and the fragmentation and so on? There is a danger that this will just be another layer of bureaucracy when the Government are actually saying that we want to streamline everything and reduce bureaucracy. Is there something more that government need to do in order to help that process that you have identified or do you have the powers already?
Dr Melanie Ivarsson: This is a critical element of setting up HDRS. I feel strongly that HDRS has the mandate from the Government to go and look at all aspects of the health data ecosystem and to come forward and make recommendations. What happens with those will be really important.
Where we see the duplicative access, for example, when we want to eliminate access, that might mean having a conversation with the body that currently governs the access and trying to make that change happen. I have a very strong feeling of support from everyone I speak to that they want to see this be successful. No one is saying, “Well, as long as you do not touch my bit, we are fine”. Genuinely everyone is saying, “We have to do more. We have to do this better”.
As we develop our strategy this summer and identify those bottlenecks, the bureaucracy, and try to evoke change, that is when the reality occurs, when we actually get to have those important conversations. I do feel the full support of the Government and, in particular, the Department of Health and my colleagues at NHS England in particular for the England aspect of this, enabling that.
Q151 Lord Verjee: I have a three-part question, one on the funding and commercial arrangements, one on the legal powers and then the other on talent and recruitment. I think that you have answered on the powers that you hope HDRS will have and I think that you have answered that you see sufficient talent to be recruited for the business plan you have. Maybe you can answer that and then I will get to the commercial arrangements.
Dr Melanie Ivarsson: If I take the talent, yes, I believe that HDRS will be able to attract and recruit the skills, as Baroness Blackwood has noted. There is a very compelling opportunity to be part of building HDRS and I have had, wherever I go, a large number of people talk about wanting to join us on that journey. At the moment, I am feeling really optimistic about the great talent we have in the UK.
You asked about the legal powers around controllership. Again, it is important that we are having these conversations with the right people in the Government and the policymakers and trying to understand the legislative framework. We are working closely with the Government to understand the use of data for research and the legislative framework that will be required as we set up HDRS to succeed. Those conversations are happening.
Lord Verjee: Okay, great. Maybe you could explain a little bit more about the £600 million funding. Will that be enough for your endeavours? What will the majority of this funding be used for? Are you a cost centre, are you a profit centre, or is your mission just to be sustainable? I would like to understand the commercial aspects of HDRS.
The Chair: Could you also clarify that £600 million funding? Is that over a five-year period?
Dr Melanie Ivarsson: Correct. We have received £600 million in funding: £500 million from the Government and £100 million from Wellcome. The intent, or the design, is that HDRS would become self-funding or a self-sustained entity by approximately 2030. That is how we are designing our business plan at the moment to achieve that.
It is a substantial foundation for the building of HDRS and we are hoping that as we demonstrate value through faster research, approvals, bringing more clinical trials into the UK, starting to see the patient and economic outcomes, this will help us attract additional funding streams as necessary. We are working out our strategy at the moment on how we spend that money.
As I mentioned earlier, in our first-year business case, about 50% of our funding is to maintain existing health data infrastructure, so that is the regional secure data environments, the national secure data environment, and so on. As we look to put together our two- to four-year road map, we will have a portion of our funding that will be to maintain the existing infrastructure as we evolve that to be the end state. There is a portion that will be spent on the technical transformation. There is a portion that will be spent on mobilising the company and recruiting the great talent that we will be needing to build that organisation. Then, as we build the services around that, we will be able to start entering into commercial relationships.
We have literally just started modelling out some of the commercial models, but that work is nascent and we expect to have worked that through more in the coming months. We thought it was important to spend a lot of time understanding the value of data and the use cases and actually the services that we are trying to build first. That is where we are now. I think that we have a good sense of what HDRS needs to deliver and then we are pulling that commercial model together this summer to be submitted as part of our two- to four-year business case.
Baroness Blackwood of North Oxford: It is worth noting just in the context of your questions about powers and capabilities that the level of political will—cross-party political will—that we have experienced is incredibly high, and this is replicated through the parts of government that know that we need to do something differently this time and want to make sure that we are enabled with the funding, the support and the powers in order to do that. Obviously, the ambition is that we will become self-sustaining and so on, but it is important that that political and Whitehall will is maintained in order for us to achieve our outcomes.
Q152 The Chair: You have both answered many questions. My final question is: as you will know, our report will be coming to conclusions and we will be making recommendations to the Government and to the NHS more widely. What do you feel you need from the Government to achieve your mission that you currently do not have? The remaining part of the question is: what are your top priorities that you think we should recommend in order to get this right? Lady Blackwood, would you like to start?
Baroness Blackwood of North Oxford: I started off with Lord Verjee on the question of continued political will. We have set out with a mandate to make change across the ecosystem, to remove duplication, to enhance what works, and I think that it is important to maintain and continue that mandate whatever may come.
Subsequent to that, a very strong signal from the centre, answering some of Baroness Willis’s questions around the support, to cut through layers of governance, to make sure that we work with others and have a good engagement and we work out the incentive structures and the levers that we can pull in order to make this a reality, is important, and we have touched on a lot of that today. From a board and company perspective, we also need the permission and the freedom to operate and the agility that we need in order to navigate what is a very fast moving and changeable environment, technically and globally.
In order for us to be competitive, we need to be able to move fast and we need to be able to work out how we can do that within the government structures which we have discussed today. I know that there are some other points that Mel might like to add.
Dr Melanie Ivarsson: Thank you. We are obviously very appreciative of the funding, but there can be a little uncertainty when you are starting a new organisation and company from scratch around how much you will spend in each year, so some flexibility around how we might be able to roll funding over from one year to the next would be good.
The Chair: Do you feel you do not have that flexibility?
Dr Melanie Ivarsson: I think that it would be good to know I could have it, so that would be very helpful. That is not the standard Treasury approach, so being able to have flexibility one year to the next while we are so early in building the company would be incredibly helpful.
Many of you have brought up the point of the social contract and the conversation with the public and championship of the value of data for research. That would be really important. We cannot do that alone. We need lots of voices to support that and the benefit to the public of research as a form of care and how data is used. That continued dialogue openly and transparently will be critical for us.
The Chair: Thank you both very much. You have answered a lot of questions from us. We are very grateful. We have learned a lot about the aspirations and ideas behind HDRS. We wish you both every success with this very important enterprise. Thank you very much for coming. Now we will close the session. Thank you.
