Welcome to this one-off session of the Health and Social Care Committee where we will be looking at the Federated Data Platform. I will ask the witnesses on the first panel to start by introducing themselves and what they do.

I am currently deputy director of research at the Nuffield Trust, which is a health policy think-tank, and we aim to improve health and care through bringing evidence to debate. By training, I am a quantitative researcher, and I previously worked for many years in the NHS as an analyst and with informatics expertise.

I am Nigel Shadbolt. I am a professor of computer science at the University of Oxford. I chair the AI at Oxford research initiative, but in this context I am executive chair and co-founder of the Open Data Institute, which I founded in 2012 with Sir Tim Berners-Lee. Its mission is to build a trusted and open data ecosystem.

Hello. I am Matt Hennessy. I am the chief data and analytics officer for NHS Greater Manchester, and I am here as a representative of the professional Chief Data and Analytical Officer Network.

Thank you for being here.

I can give you one style of answer, and I think Matt can give you lived experience as well. Federation is exactly that. It is the bringing together of a set of components, and the components in this case are different types of data held by the NHS. That data might have to do with procedures that have been undergone by a patient, or with particular prescriptions or care that is available. There are many different islands of data as a patient goes through their NHS journey. Very often data lives in silos, it can be hard to pull it together and it can be expensive. The Federated Data Platform is an overlay that allows you to pull on the different data so that when you make a query that involves different types of data, it is served up by a system. You don’t see the silos; you see a joined-up response.

I think that is a very good description.

Yes, indeed, and some specialist forms of care, such as cancer patient management, can look to integrate across these things as well. That is exactly the way to think about it, the major areas where a patient touches the NHS system.

NHS England’s website reports impressive improvements. In the debate we had in Westminster Hall on 16 April, the then Parliamentary Under-Secretary of State told us that the Federated Data Platform could deliver up to £2.4 billion of benefit over time. What is not to like about it?

There are definitely some benefits, as Nigel described, from bringing data together. It is a good direction for the NHS to go, and the approach builds on previous problems with NHS IT, which has tried to sort everything out with one single system whereas this is a system that lives locally in NHS trusts. I think there are a couple of challenges in understanding the benefits. Some organisations are already doing a lot of this, and they may not see enough benefit from a central national system that is available to everyone. For other trusts that are further behind on their digital journey, this could really help to move them along. The second challenge with the benefits is that, to some extent, we don’t know what would have happened if we did not have the platform. NHS England has provided data on activity, how many trusts are part of the system and so on, but it is very difficult to be sure of the benefits without doing a proper evaluation that looks at what might have happened anyway. You will get benefits from other digital areas as well. At the moment we don’t have that evaluation. One has only just been commissioned and will not report for a number of years. There is quite a lot of unknowns still.

You will have seen the FT article today that goes through some of those issues, looking for the cause-and-effect relationship of the Federated Data Platform and the outcomes. According to this, it seems that the Chelsea and Westminster trust has achieved an 84% decline in out-patient waiting lists, yet about a third have seen a reverse in progress. Surely it must be possible in those circumstances, particularly where you have people on the ground being able to see the effect of the Federated Data Platform, to draw conclusions about the benefits of it.

To some extent, the benefits depend on the implementation and the extent to which NHS organisations use it as part of transforming and improving patient care, but initially the Federated Data Platform is likely to throw up a lot of data quality issues. To address the silos of data, you will have to make sure that you can match up the patients and the information about them in different systems. There will need to be a lot of work initially, and with all of these systems it can take time to see benefits. I think there is likely to be quite a mixed picture across different organisations, but I am sure others on the panel—

Can I simply draw the conclusion that you are sceptical and not yet convinced?

I don’t think we know yet. From our work evaluating other initiatives of this kind, it is often difficult to draw conclusions quickly about the impact of implementing some of these changes and to work out the cause and effect.

Matt, you are very close to the ground, so presumably you can see the benefits.

There is a mixed picture. We have surveyed the members of the network, and it seems quite clear that those trusts or ICBs that were less digitally mature are deriving benefit from it, because if you move from—

Less digital prior to the rollout?

Less digitally mature, yes. If you are reliant on spreadsheets and written records, any form of digitisation will feel better. As the article talked about, one of the challenges is trying to get the right metric to assess progress. A lot of the metrics that have been established are just about uptake or waiting list size, and you can see the corollary of an advancement in digitisation that happens at a trust or an ICB. Therefore, they are able to identify people who should not be on waiting lists because they are deceased, have already been treated or don’t need treatment any more. You can clean waiting lists quite quickly, but that does not necessarily mean the people who remain on the waiting lists will be seen any quicker, because in fact these are sort of ghost patients. It will feel very positive on the ground, but trying to account for the genuine outcomes of implementing something like this requires a lot more rigour and analysis.

Presumably you don’t need a federated data platform to do the job of cleaning—as you put it—making true the waiting lists, removing people who no longer require the procedure. Surely there must be systems available to do that. When local authorities go through their housing waiting lists, I know they have their own mechanical or data methods in their systems. They don’t need to have a federated data platform to do that, do they?

No, they don’t and I think that is part of the uptake story where some trusts and ICBs, including my own, have said that they already have some of the capability locally.

You said in your letter in February that many ICBs “already have similar tools in use that presently exceed the capability and application”. You are saying that it is only the ICBs that were less digitally enabled that might have seen benefit. What systems were in place, and do you not think that those simply could have been used rather than the FDP brought in?

Waiting list validation tends to happen at the trust level, so some solutions exist as part of electronic patient records or different software solutions that sit at a trust level.

I should not have confused the two. I am sorry.

On the work that happens at ICB level, it is worth remembering that some have come together very recently. Greater Manchester has been a thing for a considerable period and so has been able to advance its thoughts on what the digital infrastructure should look like over a long period, but some ICBs are relatively newly formed and, therefore, the opportunity presented by the Federated Data Platform is an off-the-shelf package that would allow them to move forward quickly.

Do any of you know what the Government’s rationale was in 2023 for signing the contract with Palantir to deliver the FDP in the first place, rather than to expand the existing systems or make them more widely taken up?

I think you would have to look back and ask the Government. I don’t think we can reasonably be against the notion of a federated platform at all. You do want to bring together the different perspectives and components of how that harness works. How you build a “best of breed” environment for that is partly where the contest is. We will perhaps come on to talk about the issues of interoperability and the use of open standards. Is there actually a competition, an open environment, for other solutions to be integrated?

On understanding the rationale, do you think interoperability was the justification to take a federated view, rather than the piecemeal view that might have been taken by each of the ICBs having their own systems? Is that the way of thinking about it?

There are a couple of things to bear in mind here. One is that after the national programme for IT, which did not end well, the NHS pulled back from trying to do things nationally. It then invested in particular trusts, so you had what were called global digital exemplars and fast followers. There was a lot of investment in some organisations. It ended up with quite a mixed picture, with some organisations doing much better and others really struggling with basic systems. You can see the appeal of a federated data platform being something that everybody could contribute to, and also the benefit, realised during the covid pandemic, of being able to get a better national picture of data than was previously possible. It would be for the Department to answer that question fully. However, those would be some of the benefits of a federated platform, with more consistency nationally as well.

I have a few questions that lead on perfectly. In part, we are here because of a key moment in the move to digital as part of the two shifts, but also because of disquiet about Palantir as a company and in particular lock-in. I think there is a lot of scepticism. On the face of it, there is a discontinuation clause in the contract, which is probably a better way of putting it—and we will come to that with colleagues a bit later. If we can go backwards to 2023, to the best of your knowledge has the NHS sufficiently managed the issue of supplier lock-in? I will direct that to Matt because you have various examples of people coming in and out of this. Is it possible to disentangle from Palantir if we wanted to?

I think the honest answer is that the Palantir product is an off-the-shelf, all-in-one product, and that is what has been commissioned. The places that have sought to do something different, or have experience and have a mature platform, might be based on a selection of technologies that reduce the risk of lock-in because you can swap out a component of the technology stack without ruining the whole thing.

Am I to understand that the real concern is that the people who had very little before and are building from scratch are now most at risk of staying locked into this system? Have I understood that correctly?

Yes, that is correct.

That is very helpful. Is there any data that is not currently in the public domain—we know there is a study coming—that we need as scrutineers of this move to make a decision on whether the right thing to do is to continue, or not, with Palantir? Do you feel there is enough information in the public domain? What else should we be looking at?

We want to assure ourselves that the mechanisms being appealed to, to allow swapping in or out of providers or to support a genuine marketplace for solutions, are in fact the ones we have. You may be aware that within the Palantir framework there is the concept of a solution exchange. One of the questions is whether that is actually an environment where another entrant could come along for its application that might run, for example, the cancer patient experience. Is there an opportunity even to contest or show how that product is working? That is not the situation we have at the moment.

You mentioned earlier that open data standards might be a way around this. Some have described what we have bought into as, in effect, a Netflix model, in that we are paying for an off-the-shelf product. We are paying for a subscription that takes the data, puts it together in a way that is understandable and, theoretically, you can bolt things on. We will get to the app store in a second, but if we stop paying that subscription, the intellectual property of what it has been pulling together and the behind the scenes of all that just disappears. Is that your understanding?

Not all of it, necessarily. One of the things that was committed to was called—it sounds technical—a canonical data model. That is just a way of describing a consistent way of representing very tedious things, such as the way we would represent dates, all the different ways and formats you can lay out a calendar date, and just extend that to a whole range of medical-related terms. To some extent, the achievement of the canonical data model is to say that we can come to a consistent view on all the data assets that we have around the NHS, as to how we unify this concept. That might sound very appealing, but it can be quite tricky to do. One of the things the Palantir product offers is that it has this thing called an ontology, again a very elaborate name for a set of terms and relationships between terms—patients, beds, the relationship between a patient and a bed, occupancy, availability—but that is not a unique piece of IP. There are other alternatives. There are agreed standards in the medical profession. One of the questions is: how easy is it to disentangle the extensive amount of characterisation, not just of patients and procedures but of workflow, for example? It is important to understand what is open and what might be proprietary.

You may know this or you may not. Who actually developed it? Who developed the data model? Was it the NHS who told Palantir what to do, or did Palantir say, “Look, we have this solution for you”?

There is a lot of domain knowledge that will have come from NHS employees and experts.

Who owns it?

There will also be software engineers from the Palantir side who will be looking to map it into their own format. There is quite a mixed environment, but the question is: if we wish to make a different decision about a supplier, in a certain sense, do we own enough of a sufficient characterisation—in open data terms and open model formats—that is unencumbered by any of the IP another company has generated?

I guess that is a question we should ask directly to our second panel.

Absolutely. I think it would be very wise for this Committee—whatever happens in any retendering—to be confident that that model is clean and does not contain IP elements from a provider, and to get warranties from the provider that that is the case.

On the app store, very briefly, is it your understanding of the bolt-on environment that will come late—if this is intended, and if it does not go with Palantir in the next phase—that we could build something and still crack on? What is your view of whether that would still be possible?

That would be possible, yes. As Sir Nigel said, a lot of this is trying to understand what is proprietary and what is not. I like to think of the FDP as a kind of bank vault with safety deposit boxes in it. The safety deposit boxes are instances that are provided to trusts or ICBs and you go into the bank vault. There are some tools in the vault for you to use to build models, and you can bring some of your own tools into the vault. If you were to change bank or move to another vault, you would be leaving some of the tools that helped you build what you have built. That would be a portability problem because you would know how you built it, but you would be using some of the tools that perhaps you did not necessarily have a full understanding of, and you were just able to use them.

We will not necessarily know the impact until we do it. Is that what you are saying?

Yes. The key is that the NHS’s canonical data model standardises the language and the format for data across all the trusts. That is key. It has been a critical step, and I applaud that. Ultimately, it is the data that is the foundation of the federated data model, and ensuring that those standards are owned and controlled by the NHS rather than being proprietary. It is the data that is sovereign and not the application on top of it.

That is very helpful.

I want to talk a bit about what is being called the break clause. Does it exist? Is there really a break clause? My understanding—I just wanted to get your understanding to clarify this, and we will also clarify it with the Minister later—is that the contract ends next year. There is the potential to extend it by two years, then by a further one year and a further one year, so it is a standard two-plus-one-plus-one contract. Unless the Government exercise their choice to extend the contract next year, it will end and so we should not be talking about a break clause, should we? We should be talking about choosing whether to extend. Matt, is that your understanding?

I think that is a good description of it, yes. I am not over the absolute detail, so it is something to put to the Department and the Government. I think the idea that it is being portrayed as a break clause is partly around the amount of notice needed to be given. It is not that you would not continue the rest of the contract. You have to act to say that it is finishing.

I have some other questions, but does anyone want to add anything on that just to clarify about the break clause versus choice to extend? No. Thank you. I think we are talking about March next year, but the Minister stated that recently, so we will check it with the Government. We are talking about 90 days’ notice needing to be given. If the Government were to choose not to extend the contract, they would have to do so at the end of December, near to Christmas, or presumably mid-December. Perhaps picking up a bit on what you were discussing with the Chair, if they decide not to carry on with Palantir, they need to make sure that something similar can be in place because the basic concept is a good one. First, do you think that the capability exists in the UK tech sector to make a sovereign offering, a UK offering to replace Palantir? Secondly, could one exist, and could it exist in the time available?

I think Manchester is proof of what you can do. It is not an exclusively sovereign stack, but it has a number of components.

If the Government were to take a decision in December not to continue with the current contract, on a scale of one to 10, how confident are you that we would have the capability in British companies, as part of a sovereign AI approach, to deliver what is needed instead of what is there now?

Fully sovereign is a stretch. We would want to look to the market and understand. Understanding what has been abstracted and built into the existing harness is complex. We would need to be thoughtful about the time we gave ourselves to implement other solutions. For example, there are alternatives to be had for everything from data presentation to the underlying data storage.

I agree. It is important to make a distinction between what FDP offers ICBs and what it offers trusts. The infrastructure in the stack that has been built for Greater Manchester was built specifically to meet the needs of a strategic commissioner and the role that the ICB has.

Does it include GP data as well?

It does include GP data, yes, but that is distinctly different from an operational workflow that might be in a trust. I know that a lot of trust electronic patient records are able to do some of the things that are offered within the Palantir stack. I think it is treating them as two different entities.

Palantir may say—I don’t want to put words in Palantir’s mouth, although it is not backward in coming forward with comments generally—that, “Without us, it will not work. You cannot do it, or you cannot do it in the time available”.

I think the distinction to make—and this is some of the feedback we have had from the network members—is that what has been built or has been provisioned is an off-the-shelf, all-in-one “jack of all trades, master of none” that absolutely benefits some of its users.

According to today’s FT, most of its benefit is to Chelsea and Westminster hospital, which I must declare is in my constituency. It is getting 84% of the benefit.

If you break up the components, the data storage components, what they call the ETL or the visualisation or the analysis parts of that single offering, it is very easy to put together a sovereign offer that has all those components but not necessarily in one company.

That is interesting. There is time. If I can explore the issue of time, I don’t know if you are aware that, in February 2026, NHS England awarded a contract to Imperial College Projects—that is not Imperial NHS trust; it is Imperial College London—to independently evaluate the FDP, including the economic impact. However, the results of the study are not expected to be published until after a decision has been taken about whether to continue with the contract. Do you think the fact that the NHS let an evaluation contract that is due to report after the time that they will have to take a decision on whether to extend the contract suggests that they have already made up their mind and they are going to extend it willy-nilly?

To be completely honest with you, I suspect it just took them longer to award the evaluation contract, because I think the evaluation was discussed quite a long time ago. I suppose one of the challenges is that the timescale is a real challenge. I am not a software expert, but I think you just need to observe the history of some of these projects to know that a lot of these things take an awful long time to procure, let alone to implement and transition from one system to another.

Yes, but if you are going to do an evaluation, if you are going to procure a big thing—

Oh, sorry, I am not talking about the evaluation now. I am talking about changing systems. If you were to change from the Palantir system to another supplier, you would probably have to go through a lot of steps to do that. The other point that has not been mentioned yet is that there is also a national instance of the FDP, as well as the ones that are in trusts and ICBs, so that is bringing together the data for NHS England and DHSC as well.

That sounds very helpful. I have a couple more questions related to this. Do you consider that if we had all the time in the world, there are British companies that could provide the range of services that you have just described?

As I say, I am not an expert on the detail, but the fact that there are organisations using different approaches suggests there would be alternatives. However, I would have thought that the timescale would be challenging.

I think the timescale sounds challenging. If the Government are minded not to continue with the contract after March, would it make sense for them to make that much clearer now in order to give British companies time? Obviously, at the same time as announcing their decision not to extend, they would announce their intention to re-procure. Would it make sense to do that now rather than leave it to December? That is the final date they can do it, which would give British companies and others time to prepare and be in the best place to tender—and I would not say move seamlessly from the existing service, but move as easily and properly as possible from the existing service to the new contractor. Therefore, they should not leave it until December before making their decision. Matt, I can see you nodding your head.

I think so, yes. Because each organisation—and these are sort of sovereign organisations, trusts and ICBs—is having to make decisions now about its workforce, and about what it commissions above and beyond what might be provisioned nationally. Therefore, the earlier there is a Government directive on the direction of travel, the easier it will be for NHS organisations to plan their own staff.

Realistically speaking, is there any way on God’s earth that we would be able to get other companies into a position to tender and to take over from next March, if we do not make it very clear that that is what is going to happen in the next month or so, and that leaving it to December would be much too late?

Those procurement challenges are clear. There are a variety of providers for the underpinning databases, and there are a variety of providers for orchestration between them. So there would need to be some kind of coming together, some consortium model, and that all takes time. There is also the fact that the current provider will have extracted a huge amount of insight and valuable knowledge about what the overall workflows, patterns, and the essential sorts of queries across datasets are. It is genuinely something to reflect on: how do we, in the future or perhaps even now, imagine capturing some of that value that the British taxpayers have paid for, for a continuance of a service, whether we do or do not use the break clause? I think there needs to be a much stronger sense of capturing the value that is created out of the workflows and activities that the trusts and integrated boards are doing.

Just to nail this point, and I think Ben has covered it very well, are you saying that it is already too late to begin the process of alternative procurement—I think this follows the point that you were making, Nigel—and would it be too disruptive even to consider it? Is it too late, and would it be too disruptive?

That largely depends on what you are seeking to commission and procure.

A complete change in sovereign provider.

The Minister said, “No more this company”. Is it too late to do that in time for March next year?

I would say no. The reason I say that is because I think it is about describing what it is you are seeking to deliver. One of the challenges that the network has put back to NHS England is to say, “What exactly are you seeking to achieve with the Federated Data Platform?” What national teams and the national programme team are uniquely positioned to do is to set standards and policy, and to convene communities of interest. However, in the case of this programme, they have also procured solutions, and I think the question for the Federated Data Platform is: what are we actually seeking to do? If it is about interoperability and open standards, that is slightly different from buying something that has everything in one go.

There is nothing to suggest that the UK lacks the talent or capability to produce an equivalent to what Palantir offers. Indeed, the same applies for many technologies produced in the US or elsewhere. The challenge is whether we can deliver these fast and as efficiently as some of the dominant US companies. I think that is some of where the conversation is going, which is a procurement decision for the NHS, rather than technical. With open standards, effective governance and some proper investment, the public sector and UK-based firms should be able to provide a solution for analysing NHS data, but do we have realistic timescales in which we can do that, and can we imagine engineering a procurement process where we can do that?

So it might be slightly more expensive and slightly slower, but it needs a policy decision to decide if that is what we are going to do?

Yes. Because behind all of this, there will be issues that amount to trust in platforms.

It is a pleasure to guest on your Select Committee, Chair. Matt, how much do you already have in place in Greater Manchester of what the Federated Data Platform is trying to achieve?

Separating the trust’s operational workflow from the integrated care board’s, first the integrated care board has what it needs, and the technology that we have can match anything that the Palantir FDP can do. Where I think our functionality is greater is that we have spent a lot of time working to build public trust, working to get the social licence and public legitimacy to connect GP data. The same task that perhaps is done in Palantir software can be done in Greater Manchester, but it has a greater utility because it includes all the GP data, adult social care data and local flows that support the neighbourhood health agendas.

Therefore, you have everything you need, you have everything the FDP can do, and more, because you have the single patient record already existing across Greater Manchester?

It does. We honour opt-outs, so those people who choose not to—

Of course. Is Manchester unique?

I don’t think so. I am aware of other systems. Perhaps not at an ICB level, but Leeds has a very good integrated dataset. Dorset, Kent, OneLondon—

Thames Valley.

Yes, Thames Valley.

I understand that about 30%, at least, of the population is already covered by the sort of functionality you have in Greater Manchester.

I would not say to what percentage, but yes, definitely, there are other areas across the country.

There are other suppliers, other than the ones you are using, that are doing similar roles in other trusts, in other places.

Yes, and part of that, especially for ICBs again, is that when they were first formed, they were required to develop population health management solutions, and some were able to move faster than others, so there is quite a lot of provision out there.

These are using a standardised thing called Cypher, which ran from 2020 to 2023. Is that right?

Yes, that solution was built for all those customers of Graphnet as a provider. Graphnet was providing the shared care record and population health management platforms for a number of areas. The Cypher programme united those so that people could see each other’s and template the dashboards or the solutions right across that network.

We have sophisticated technology doing all of the things that we want to do in terms of federated data, data analysis and the single patient record already provided by UK-based companies already running, already providing benefits arguably greater than Palantir has done. And that already covers, from what I can see, more trusts and more people than Palantir does today.

Again, I cannot validate the numbers you talk about, but yes, there is definitely an existing infrastructure.

NHS England’s medium-term planning framework, from October last year, suggests “all providers in acute, community, and mental health sectors are onboarded to the NHS Federated Data Platform (FDP) and using its core products” to the exclusion of all others. Is that something you think can be achieved by 2028-29, which is what it asks for?

I do not think so. I think each trust and each ICB would want to make its own decisions about the added value. While the total local cost of ownership for adopting FDP is not well established, that presents a risk to organisations to adopt something without knowing the extent to which they might need to carry on and pick up future bills.

This touches on something that I think is important to consider. We are talking about the present, but as you look to the future, one of the things we understand is that there will be—and we have this demonstrated already—extraordinary value potentially in linked patient records, where we can see patterns of disease, areas of socio-demographic association with disease patterns; a range of things. We know this is highly sensitive data. It is data where we need to have a high level of trust between patients and providers. We have a model for this that is tricky enough on its own. It is actually Ben Goldacre’s OpenSAFELY project, which is linking, at scale, English patient records and revealing all sorts of fascinating issues around patterns of disease and comorbidities. The question we should look forward to is how we combine trusted research environments, which have very sensitive data, with the wider health experience. That is an architecture for the future, if you will, but it is one we need to be able to imagine getting to.

Thank you for that. That is quite useful, because it moves me on to data security, trust and rolling up into the national level. It came to light again in the Financial Times recently that NHS England has removed the project-by-project permissions for engineers moving data into the national instance from the local trusts through the so-called National Data Integration Tenant, and has given those engineers admin access to it. Do you think it is reasonable that the National Data Guardian did not know about it and was not consulted? Do you think that data is safe? Should Palantir engineers be having unlimited access to the admin data and other people working on that system?

One important point for me is that NHS England is a data controller and Palantir is the data processor. Under data protection law, there will be a data protection impact assessment that sets out how the data are going to be used. If the data are used in a way that is not covered by that data protection impact assessment, that is something that needs to be understood. That would be a question, I would have thought, to follow up as to how that has happened, if people are using the data in ways that are not covered by the information governance assessments that would set out those requirements.

I think what the story of Manchester is telling us is that one of the things we need to do is to bring the stakeholders with us, bring the patients with us and bring the trust with us—not the NHS trust but the trust of patients and GPs. Do you think this is sounding like a trustworthy system at the moment, despite the legal protections that may be there?

I think that the perception of risk and the risks are really important—both of them—and I think that the Federated Data Platform has done more public engagement and having a public oversight arrangement than perhaps earlier national programmes. So I think there has been progress, but I think it is an ongoing requirement to maintain that trust, and I think there are going to be particular issues. There is also a broader question, because how the public perceive trust in data and health data is also how they perceive trust in other data. We know from some of our work in other countries, in Denmark, that there is a very strong culture of there being a digital record, and everybody accepts that. They expect to do things digitally and have one login, and we do not have that environment in the UK. Therefore, it is particularly important to think about the trust issues, and the connotations of Palantir’s military outputs heighten the importance of that issue, so it needs to be taken even more seriously.

Operating with public trust is absolutely crucial. It must be absolutely crucial, not just because it is a health-related issue, but because of the broader awareness of how data may be repurposed and may be linked. At the Open Data Institute, we believe fundamentally that you have to design data architectures that prevent future misuse or abuse. It is not enough to rely on intent or goodwill. In our 2021 report we have 18 mechanisms to control who can access data, what can be accessed, how the data can be accessed, what it is used for, and those design decisions—

We are running a little short on time. Can I just put one further question to you? You mentioned research, is it your understanding that the FDP is designed as a place for getting research data from? You were talking about future-proofing and getting research data out of the system.

I would want to apply much stronger stipulations around data held in platforms, if we were to regard them as trusted research environments. That is a very high bar indeed, and you want to make sure that the data and the uses it is put to are held in very strong sovereign control.

That is a no, then.

You would want to assure that.

Okay. Thank you very much.

In closing, I am going to ask you a few questions about our current digital capability in the NHS. At a more senior level, are we able to manage these digital contracts?

At a senior level, the network I represent shows that we have the skills and knowledge to offer services to policymakers and programme leads in managing that.

Are there enough of you?

Probably not, and there is a diminishing number.

Why is there a diminishing number, given the importance of this area?

With some of the NHS reforms, we have lost a number of people with some of the skills to voluntary redundancy.

We have lost some digital experts to redundancy because of the reforms?

Yes. As a result of the changing roles, for example, of ICBs as strategic commissioners. It is not just a shrinking; there are new roles and responsibilities. We want to do more in the market management space—

Just to be clear, are you saying that we have lost capacity and capability, or that it just looks a bit different in these new worlds?

I think we have lost, just in the sheer number of people who have gone from the system.

Sorry, just to be really clear, did we need them?

I think we did.

All right. You do not think that these redundancies were actually the right thing for digital capability at that senior level?

I think that each area has taken a decision that it feels it needed to take in order to live within the means that the trust or the ICB had adopted.

Understood, but that has left gaps?

It has left gaps, yes. That is probably the way to look at it, because some have support. My own ICB has been very supportive, for example, of data and digital.

You have talked about the cuts in the ICBs. There are also cuts at the centre, in NHS England. Do you have any understanding of those cuts in that digital capability world and what they have done there?

Not specifically the cuts, but I think it is worth bearing in mind that there has been a long-running challenge of the NHS not having sufficient capability in analytics, informatics and digital.

Why do you think that is?

It is a combination of challenges, such as competing with other sectors, with the private sector, and how roles in the NHS are perceived, with non-clinical roles perhaps often not getting the profile that they merit. The Government’s review, commissioned from Ben Goldacre a few years ago, covered these issues very well. It came up with a number of recommendations about how to reinforce and develop capability and analytics in the NHS. There is a wealth of information to refer to, but I am aware that, although I do not work in the NHS now, colleagues I worked with when I was in the NHS are leaving because of the changes in the roles and organisations and uncertainty about commissioning support units that provide a lot of expertise in a lot of areas.

I think I am right in saying that is concerning, given the scale of the challenge and the ambition that we are talking about today. My colleagues have been talking about the current coverage, perhaps a sort of re-contracting. Nevertheless, I think we all agree that FDP and this digital capability are essential. I think what I am hearing, however, is that the resource available currently is not sufficient. Therefore, to scale this, to hit this ambition, am I right in saying that the level of human resource needs to be seriously looked at? Is that a fair reflection of what I am hearing?

I would agree with that.

Yes. I also think that perhaps the failure to develop the workforce in the right way leads to some of the challenges, which we have already discussed, around being overly reliant on the skills that exist within the suppliers.

I hear you, Matt. Obviously, there is a lot of workforce planning going on in the health service at the moment; lots of talk about the workforce plan. Are you assured that, within that workforce planning, the gaps that you have identified for us—and the future gaps that will presumably arise given the scale of ambition in this digital specialty—will be realised and met in the workforce planning that is under way? Are you assured in that sense?

I am not close enough with the workforce plan to know. That might be a question for the Department.

I hear that there is a lot of disquiet. I would again commend that important pieces of work get done, which contain helpful suggestions on what the shape and nature of that workforce might be. I reference the Goldacre report, which has some fantastically powerful recommendations.

It has been around for a while.

Yes, it has.

Even I have read it.

Yes, it has, and it bears re-examining.

Excellent. I think I will leave it there. Thank you.

Thank you very much. The Minister is with us now. I ask you each to pick one thing that you would like to leave uppermost in our minds, or perhaps to plant in the Minister’s mind while she is in the room with you. What would that be?

It would be to develop a programme and approach that makes best use of the unique position that trusts, integrated care boards or the national team and the regional teams are in. Focusing the national team on establishing standards and policy and bringing communities of interest together, allowing solution development to happen at the frontline. That would be my one thing.

Whatever happens, ensure the fundamental interoperability of technology, the ability to have a data model for all the writers of data that we have. The canonical data model should be genuinely clean, open and maintained.

I would add a focus on transparency about what the FDP is, engaging the public and thinking about how to develop further public trust in data sharing, which can bring huge, huge benefits and is really necessary for the NHS, but public trust cannot be taken for granted.

That leaves me just to thank all of you. Witnesses: Preet Kaur Gill MP, Rob Thompson and Ayub Bhayat.

Thank you very much to our second panel, and especially to the Minister, for taking time out of their busy diaries this afternoon to be in front of this Committee. It is the Minister’s first time, so welcome, and also welcome to your officials. Very briefly, for the sake of those who may not know what everyone does, please introduce yourselves and what you do. We will start with the Minister.

Thank you so much, Chair. It is great to be here. I am Preet Kaur Gill. I am the Minister for Health Innovation and Safety. I am also the Member of Parliament for Birmingham, Edgbaston.

I am Ayub Bhayat, the director for data analytics at NHS England and the senior responsible officer for the FDP.

I am Rob Thompson, the relatively new chief digital, data and technology officer for DHSC and NHS England.

Thank you so much. I appreciate your question. Let me start by saying that I know there has been a lot of discussion about what the Federated Data Platform is. In this role, I very quickly wanted not just to read about it but to go to see it in practice. Therefore, I visited the Chelsea and Westminster hospital and the Royal Surrey, both of which are deploying the Federated Data Platform. What is the Federated Data Platform? It is an operation system. The NHS designs it. It is for the NHS, and it is helping the NHS to make better use of the information that it already has. All of this information already exists in the NHS. We know that staff are making decisions about who needs treatment, who has been waiting the longest, where bed capacity is and whether the cancer pathways are quick enough. I know that patients care about this because, far too often, many of us hear from patients who say, “My GP has referred me to the hospital, and I just don’t know what has happened. It is an urgent referral.” The problem is that the NHS has spent so much time accessing information from different systems, so clinicians are having to go into different systems or they are having to reprioritise patients based on a spreadsheet. Just imagine the hours it takes for somebody to go through a spreadsheet. So what is it? The Federated Data Platform helps solve that problem.

The reason I was explaining what it is, is because that is really the first premise. I speak to so many people who do not know what it is and what it seeks to do, because it does not replace patient records. The platform is live across 139 trusts, and 170 trusts and 35 integrated care boards have signed up to it. I met the chief executive of my care board. I asked him how he uses the FDP. What he said and was able to show me is that he could see how many ambulances are sitting outside of the hospital right then, so making decisions about resource allocation. It has helped 100,000 patients undergo procedures.

The approach of how it is being used in hospitals?

Of course I am, because I went to physically see it. I spoke to clinicians, and they were able to show me how it works on the ward. You will know from your previous occupation that in a ward some hospitals have replaced the whiteboard with a digital system that tells them how many patients have had their ops, and who is ready for discharge. You can see that in real snapshots.

You asked whether I am happy with the FDP and how it is being deployed?

I am happy with the FDP and how it is being utilised.

Thank you all for coming, and congratulations on your new position, Preet. It is great to see you all here. I would like to ask a few questions because, in looking to the future and what happens next, we need to look at the past. Could I look a bit at the procurement of the Palantir contract? My concern is that we have a contract with an American company, with all the data and other issues that are increasingly of concern to us. We had internal tenders from the NHS, and I wonder whether we gave the NHS the chance to build its own system. If we did that, why did we then decide to go out to tender? That is where I will start. First, what assessment was made of the capacity of the NHS to build its own solution before deciding to procure externally? Mr Bhayat, maybe you could tell us about that, because you were there at the time.

Let me start with that. I am assured, and have looked at it, that before the contract was put out for tender through the Government’s normal processes, there was a horizon scan about who had the capability. The NHS is not procuring a supplier; it is procuring capability.

What assessment was made of the capability?

We looked at British companies to see which had that capability. There was an exploration, and those companies simply did not exist. The Palantir contract includes companies that it was working with when it pitched for the contract. Three of the four are British companies. The other company is based in Northern Ireland. This is about capability. Let me just be clear—I am happy to bring in my official—but the contract is in its initial term, and like any major Government contract, it will be subject to rigorous review before any extension decisions are taken, which is happening.

There is a review taking place, which Imperial College Projects is undertaking. You will have to take a decision. If I am correct, the contract ends unless the Government decide to extend it. We are talking about break clauses, but my understanding is that there is no such thing as a break clause here. We are talking about the contract coming to an end—we have had legal opinion on this—in March next year, and the Government have to decide, 90 days or more beforehand, whether they wish to extend it for a further two years and another one and another one; it is a two-plus-one-plus-one contract. That is correct, isn’t?

Yes. Imperial College is doing the independent evaluation of the effectiveness of FDP, and it is also making sure that it is meeting its terms and conditions and contractual—

The contract will come up for review.

The contract will come up for review this year, in line with the standard contract management processes.

Is the Imperial College evaluation going to report before or after the Department completes its review of Palantir, to help inform the decision it takes on whether to extend the contract?

Let me come back to that. Imperial College is looking at the effectiveness of the FDP.

I know what it is looking at. You must assume that we have done a lot of homework. We know what people are looking at.

Minister, can I please urge you to directly answer the questions as they are presented? We have carefully thought out the order.

I am answering the question, but the question was: what is Imperial College doing?

It was not, so I will ask it again. That wasn’t the question. Forgive me, Minister—and I appreciate that you are doing a terrific job—but that was not the question. The question is: why are you going to wait? Why have you let this independent evaluation run later than the time you have to take the decision on whether to extend the contract? You have to take a decision by December on whether to extend the contract. You are not getting the independent evaluation to deliver its result until later. Does this rather suggest that whoever gave the contract to Imperial assumed that we were just going to roll over the contract? I am going to ask Mr Bhayat because he was involved early on, and he might have something technically interesting and informative to say on this.

Thank you for the question. If I can just expand on it. We ran a process to procure an independent evaluation partner. That process was run late last year, and we got into the process of award. We awarded Imperial College Health Partners as the preferred organisation to provide the independent evaluation services. There was no assumption made about whether the contract will be extended or terminated on that basis. This was an independent process. It was planned as part of the initial procurement and award of the FDP, and we now have Imperial College Health Partners in place. I had a call with them last night to make sure their evaluation will continue.

I understand. Thank you for explaining that. You are going to be deciding by December whether to extend Palantir’s contract. Is that correct?

We have to provide notice by December if we are not extending it. That is the 90-day notice period.

You are undertaking an analysis at the moment that will inform that decision. Is that correct?

The independent analysis that we are undertaking, not the Imperial College evaluation, is happening right now. There is a review process of the contract by our commercial department, which is in line with the Cabinet Office playbook for contract review.

That will be concluded when?

We are hoping for that to be completed early autumn. That is the current timeline.

If you did wish to continue with the existing contract, fine, but if you decided that you did not wish to continue the existing contract, we have just been hearing that it would require quite a lot of time to get a new tender out, to get other people to procure. Leaving it until December would be a bit late. Do you think that you should reach your decision sooner than that, ideally by the summer?

Through the process that we have undertaken over the last three years of implementing the Federated Data Platform to make sure our patients are getting the most—

Is that a yes or a no?

There is context to this. I am offering the context first to make sure that the context is understood. As you said, and rightly so, we have a two-plus-one-plus-one contract—

The question I am asking—forgive me, if I am not being clear, I do apologise—if you take the decision not to extend Palantir’s contract from next March, you have to take that decision by December, but if you leave it till then, it will be quite late to run a procurement. If you are minded to take the decision not to extend the contract, why not do it by the summer?

The process that needs to operate right now is a thorough commercial process for reviewing the current agreement. That will report back in early autumn. The latest we need to inform the supplier is December.

How long do you think it would take to run a new procurement?

In a typical scenario, it depends. There are loads of government frameworks out there. Commercial frameworks exist through which you can choose, but just being—

One month, two months, three months, six months?

For this process, from a full business case for the FDP to procurement, we ran an open procurement process to make sure we got the right solution for the NHS. It took us over 15 months.

Over 15 months? Are we saying from that, given what you have learned, and you have obviously learned a lot working with the current provider—do you think it would take over 15 months to run a new procurement if you chose to do that?

Government legislation has changed since we ran this procurement. There have been improvements through the Procurement Act that allow a more defined process to be created. If the decision is to move away from Palantir as a supplier for the FDP, we could choose to run a different type of process.

Are you confident that, if you took the decision in December, you would be able to have a new provider in place by April?

The analysis is happening right now. It would be wrong of me to prejudge the analysis that is going on right now.

Are you confident that if you took a decision in December not to continue with the Palantir contract in March, that you could have a procurement run and a new supplier in place by March 2027? Are you confident of that?

Just to be clear, though—

No, I want to be clear. Forgive me, I should not interrupt you, but I want to be clear. I keep asking the same question. I do not want to sound like Jeremy Paxman talking to Michael Howard all those years ago, but forgive me for repeating the question. Are you confident that, if the Government take a decision in December not to extend Palantir’s contract, that will allow sufficient time for a procurement to be run and a new contractor to be in place by March 2027? Are you confident of that?

Our intention is to inform by early autumn, like I said earlier, so although December is the latest period we can inform the supplier, our intent is sooner.

Thank you, that is very clear and I appreciate it. If you do it by early autumn, are you confident that you could have a new contract in place by the following March?

We will wait for commercial advice to come back on that. Commercial teams are working that through at this stage.

You do not know yet. Is that the answer?

I do not have an answer to that.

You do not know yet.

Can I ask a question about British sovereign AI? The Chancellor recently—the Financial Times has run it; so have all the other papers—said that we need to buy British in critical sectors. She said that the critical sectors she has identified are artificial intelligence, alongside energy infrastructure, shipbuilding and steel. The Government at the moment—she and Minister Chris Ward—are preparing guidance to prioritise contracts for British businesses. Is that something you are going to bear in mind when deciding whether to continue or to extend the existing contract, or will you be taking the invitation the Chancellor has given you, and in anticipation of the guidance she will be producing, to think that maybe what we ought to do is we ought to take a sovereign AI approach to continue with the Federated Data Platform? Will we effectively be seeking to buy prioritised domestic procurement?

It will always be my priority to buy British where that capability exists. However, as I said, the NHS is not procuring a supplier. It is procuring capability, and we have to horizon scan and look at who has that capability here in the United Kingdom.

Is there an argument to be made that, if you give enough notice to British companies and potential British suppliers, they will have the opportunity and be prioritised in a tender process? If they have enough notice, they can get their teams together, they can get their thoughts together. We need to do a little bit more to get them competition-ready, but then the benefits will be so great to this country if we have a sovereign AI. Is that something that you will consider?

Absolutely. We will always horizon scan, but let me just be clear—

No, that was not a question about horizon scanning.

It is about talking to businesses where we already know that they have some level of capability. Some of those businesses just do not, and this contract was about partnering with British companies. Let’s not forget that three of the four are British companies. They have British employees. One of them is a Northern Ireland company, but these are British employees who are working on behalf of the contractor, the supplier, so there is a bit of that within this contract already, but like I say, it is about capability in the round.

Talking about British companies, I just want to come on to something else, which is about where we go now. I must say, I tip my hat to the Financial Times, which is producing a lot of information, which I know Committee members have been following closely, and I certainly have. The Financial Times says at the time of the procurement, Samantha Jones—who is now the permanent secretary of the Department of Health and Social Care—was an adviser to Carnall Farrar, which is one of the companies that you are referring to and is part of the consortium that won the contract. It also said that she helped Global Counsel—Peter Mandelson’s company, which had Palantir as a client—organise a roundtable this year, once she was in place as the new permanent secretary of the Department of Health and Social Care, on a 10-year health plan. Therefore, we have someone who is quite close to Palantir at the head of the NHS. Equally, we have a situation where Penny Dash, who is now the chair of NHS England, was Chair of the North West London ICB, which has pushed Palantir extremely hard. Indeed, she was chair when Matthew Swindells, who resigned recently, was joint chair of the former North West London Hospitals trust. He advocated extending Palantir to GP data, despite Chelsea and Westminster hospital board papers stating he was to be excluded from any decision-making in relation to Palantir. The idea he had in the email in the Financial Times on 5 March says that Penny Dash also backed his idea, and she said that the NHS chief—one of the NHS subsidiary officers, a digital officer—was offering money for Palantir. That may explain why Chelsea and Westminster hospital alone accounts for the vast majority, as today’s Financial Times makes clear, of the outcomes that Palantir has achieved, ostensibly across the country, but mostly in my constituency of Chelsea and Fulham. Should we be concerned, as we look forward, about potential conflicts of interest at the highest level of the Department of Health and Social Care and NHS England, in deciding whether to continue with Palantir?

We should always be mindful of conflicts of interest, and you would expect any civil servant, as you would expect any politician, to make sure that they declare any conflicts of interest that may arise. My understanding is that when the contract was issued, none of those people cited in the FT were part of that conversation about who got the contract. As I said, there was a rigorous process. This is about capability. This is not about who owns the company; it is about who controls the data, fundamentally. As the Minister for Health Innovation and Safety, I think that is what is key here.

Decisions taken about what is done with the data are decisions taken by the people who let the contract. They are not taken by Palantir; they are taken by people in the Department who write a contract and say, “This is what we want, this is what we will allow, this is what we will not allow”. You write the contract, people pitch for it and they follow the rules. There was a case where the chair of four NHS trusts in north-west London—where I was on the integrated care board but did not know any of this at the time—was advocating to extend Palantir’s contract to GP data, although that was not in the contract. That was something that he should not have been doing because, according to his own board, he was not meant to be involved in any decision-making because he had been close to Palantir. Again, the chair of the ICB at the time backed this and said that there would be money available from the NHS. As we look forward with Palantir, do you think that there are some people who should be recusing themselves from the decision-making process because they are too close to what has happened up to now?

I started off by saying that this was an operation system designed by the NHS for the NHS. Who controls the data? The NHS. Who decides how information is used? The NHS. Who decides how access to that information happens? That is the NHS. In the end, this is about patient outcomes, and that is what we have to look at.

The NHS is not a thing. The NHS is made up of people who, from the very top down, take decisions based on the best information available and based on other things. Palantir won the contract. It lowballed and got the contract. This is happening right across the place. It is lowballing, it is giving lots of free advice, it comes in and wins the contract—with the Ministry of Defence and with the police here in London—and then it is locked in. I am reassured a little about what has been said, but I am a bit concerned that there is not serious work—and I am looking for reassurance that there is—taking place to look at whether we should extend the contract, and that the decision may already have been taken and it is a matter of managing the fallout, if you want. However, the contract will be extended willy-nilly come next March and a British company, despite what the Chancellor is saying about supporting British sovereign AI, will not get a chance to provide the service that the NHS needs.

I hear everything that you are saying, and I do not dispute that, but the question we have to be asking ourselves is: is this delivering for patients? Is it delivering for the taxpayer?

Surely we have to ask ourselves, could not a British company do it as well?

Is it operating safely and securely, because in the end, we want transformation.

I do not think that that is what the Chancellor was talking about when she talked about designating procurement in AI as critical for national security.

I support the Chancellor.

I am not disputing that it may be doing those things. Other people may, but I am not disputing that. I am simply asking whether we should not—when it comes to taking forward from next March what happens in the NHS—be looking instead at building Britain’s own sovereign capability in AI in this area.

Yes. I agree with you. I am not disputing that.

I am a little concerned that some of the people who may be involved in taking the decision about that may be a little too close to what has happened to date.

I can assure you that I absolutely agree with you. Governments should never become dependent on a single supplier. That is the premise of this Government. That principle applies not just here in the NHS but across Government. The NHS cannot afford to become locked into any individual technology provider. That is why the contract already contains provisions relating to intellectual property, exit planning and transition arrangements. That is all. The detail is in the contract.

Palantir should not assume that it will just get the contract rolled over?

Absolutely not.

Minister Gill, this is very interesting. I have been reading the contract for quite a long time. It is the first contract I have ever seen that talks about know-how, which belongs entirely to Palantir. You are right that what the NHS takes into the contract, it takes out again at the end, but most of the rest of it ends up being shared. I am surprised that you talk of buying a capability rather than a system, because the original contract was to procure a Federated Data Platform for the NHS. “NHS England requires a supplier for the provision of Federated Data Platform and associated services. The data platform will be owned and controlled by the NHS to unlock the power of NHS data to understand patterns, solve problems, plan services”, and so on, all of which is very good stuff. I do not think any of us is suggesting that we do not need this level of analysis at the top of the NHS, that we do not need to help trusts and ICBs with the ability to do data analysis. However, there are concerns with the way this is being operated and is working. I am delighted to see the numbers going up. They have changed quite considerably since April. I suspect they are making a big effort to get those numbers up. However, there have been considerable worries about some of the practices that are going on and the data that is flowing through. We had a query from the National Data Guardian after it became clear that in the system, where it transfers data from the independent trust tenant up to the NHS national area, it is going through in cleartext—identifiable data is being passed through—which is then used on the national system. Rather than going through the usual processes of requiring permissions to access it, it transpires that Palantir engineers and others have now been given admin access to it whenever they want, and they have to provide no reason to do that. Do you think it is right that the National Data Guardian is concerned?

First, I am very sorry that happened, but nobody has access to the system without having a legitimate purpose, and all access is time-limited. It is audited. You know who would have access and what they need it for. There was an error by NHS England that meant that, in the data protection impact assessment that was referred to, NHS staff were mentioned, rather than clarifying that actually the access was for authorised users and support staff. I know that NHS England has now reviewed that information and is strengthening the impact assessment to make sure it aligns with the best data protection practice. I hope that is clearer. We have also responded to the National Data Guardian’s concerns, and we are working with her to provide more information and to make sure that we implement her recommendations, as she has shared.

It does not give me confidence when you say that the NHS is controlling all the data and it knows what it is doing with it when it is clearly making errors like this, which is one of the obvious points.

It did not change who has access to the data, it just improved the description of who has access. There will be people who need to work on the backend, as you know, for a specific reason. They would have to have a legitimate purpose. It would have to be time-limited. The transparency in the system means that you can audit who that person was and what they did, because any breach of any contract, in accordance with UK law and in accordance with NHS governance, will be taken very, very seriously. If there was a breach, we would be aware of that. As I say, there is a proper audit process around access, and it is not just that anybody can have access, they have to have a legitimate purpose to do so.

Who is running that audit?

The audit is accessible.

Just to be clear. First, the environment that you referenced is a highly governed environment where everything is logged, it is audited and it is audited by a separate team, and it happens within NHS England. The access, as the Minister just said—

NHS England are auditing NHS England?

Correct. It is a separate team within NHS England, which is normal process. It is a clear data access process that is referred to. The clear data access process is renewed every three months for external contractors. It is a highly governed state. In the impact assessment, as the Minister was referring to, there was the wording that was incorrect. The FAQs on our website have been clear since January ’24.

Hopefully they will do better, but how will we get confidence back for patients across the country that they can trust the system?

Because as I said, in the end, public trust matters, as you know, with all of this stuff. The NHS remains in control of NHS data. It is data that has already existed, but in different systems. What the FDP has done is it has just brought that information together so that clinicians in hospitals can make day-to-day decisions about cancer pathways, how many beds are available, theatre utilisation, discharges. This is important stuff, because in the end patients want to know that their outcomes and their value for taxpayers’ money are why we are doing this.

Yes, we fully understand that it is all about patients and that it is all about value to the taxpayer and the rest. There are questions about how it is being delivered. I am delighted that you went to see the Westminster and Chelsea, which has improved its performance since the immediate post-pandemic days, which it is measuring against, according to the British Medical Journal article that questions the scale of results that are being delivered there. However, we are seeing very similar, if not better, results and better capabilities, we heard earlier, in the Greater Manchester area, Merseyside, Thames Valley and a whole range of others as well. At least 30% of the population is covered by these UK-developed systems, so why are you asking NHS England to have those removed by 2028-29?

We are not. I think it is great that there are local systems and that they are being utilised. Let me be clear, in Greater Manchester the hospitals are using the FDP; the ICB is not. We want to make sure that there are different tools within the FDP. Some hospitals use different tools because some of this is about adoption, it is about capacity and they know what they need. This is not about us mandating to the many trusts that already have their own systems. I was able to see at Chelsea and Westminster that there is an electronic patient record for GPs. If they want to make a referral, it goes straight on to the consultant’s list. The consultant can give feedback to the GP straightaway if they do not think they need to be on a waiting list or that they may need non-urgent care. That list becomes—

Forgive me, Minister, but your time is tight, so we do not need to go through the benefits to the patient. We understand that. Manchester can do that without using the FDP. It can do that already in its systems, as can all these other trusts. They can see from the ambulance exactly what medication a GP has provided, or the patient has had in the hospital. It is already there. They have gone one stage beyond the FDP. There may be, but they have not discovered them yet, areas where the FDP can offer them new services, but those do not exist in today’s offering, so they are not interested. Yet the NHS England medium-term planning framework, released in October last year, says that they have to remove all of their systems and replace them with the FDP. Are you going to change that?

My understanding in Greater Manchester is that all of the hospitals are using the FDP and the ICB is not, so there is a difference. The hospitals are using the FDP. We are saying that where there are good systems, we need to look at the different tools within the FDP, because there are a range of tools and we want to make sure—for example, waiting lists, whether it is theatre utilisation, whether it is discharge—because most of us want to see patients get the best possible care. That is an area that we need to be speaking to trusts about, because different trusts are using different tools. They are not all using the same ones. There is some flexibility within that.

That is surely good—as we said before, we do not want a single supplier—having different trusts using different tools. So long as they comply with the overall data model, so long as they comply with the NHS standards, surely that should be encouraged and indeed built upon. Clearly there are areas—my own hospitals in Devon have only just installed new systems for electronic patient records. They are somewhat behind the times, and they may need assistance from areas like perhaps Manchester to grow their systems. How do we mix these two desires to get something that is centrally driven, like the FDP, yet maintains the ability for Manchester to build upon their existing systems that are highly performing today.

Yes, I think that we need that flexibility. Where there are really good systems and they are working, we need to build on them. The electronic patient record is quite different from the FDP, but it does interact with the FDP and for primary care that is important. Therefore, I do not see why we should not have the level of flexibility and different systems. In the end, though, what I do want, in sharing best practice, is that some of the tools within the FDP should be rolled out much more, but this is about clinicians sharing best practice, it is about people going to see what works, because adoption will be different in different places. Some hospitals will not have the IT personnel, for example, to be able to deliver some of this. Some of the hospitals may not need this because they have a good system, as you say, but they may need some other bits of the FDP tools, and that is important. We all want patients to be treated safely; we want their information to be kept securely. This is all about value for the taxpayer, but in the end it is about patient outcomes and making sure that we can give the best possible care to patients.

Would you be happy to look again at the proposed rollout as suggested in the medium-term plan?

Yes, I would.

In the medium-term plan, the commitment for 2028-29 is that organisations should move towards using FDP and should start to use it where those tools are relevant. We have since clarified that statement a number of times to explain to NHS organisations, because that is a query that we received and we have clarified it to say that it is not a mandate. We said, “If you have better tools locally, you may still carry on using those.” We clarified that after publication.

The status quo is that most hospitals do not have a system that is as good as the FDP at the moment. Manchester is quite different, from what I can see.

I am not sure that is accurate, from the numbers I am hearing from people. I hear that 30% of the population have Manchester-style systems, and other providers provide similar things, perhaps not with all the bells and whistles, that provide up to the level of and including the single patient record that exists in Manchester today.

Would you dispute that?

I look at the number of 139 acute trusts having signed up to use it as an indicator of whether they want or need it, and that for me is the benchmark.

They are obliged to. They have no option.

We do not mandate.

You do not mandate. This is a hypothetical, but it is an incredibly important one, Minister. There is scepticism on a number of fronts that we could get out of this if we wanted to, and about what the consequences and opportunities are in that. You have been very clear that you would rather buy British, that if the capability were there you would seek to try to use it. However, I want an assurance—and this Committee wants an assurance on behalf of the public—that it is a very genuine choice, and that when it comes to the decision that you will have to make in December, there is a plan for migrating to new systems, what that might look like, has this been tested and all the rest of it. Can I just start with a simple question? What have you done to ensure that there is a meaningful ability to leave this contract when you want to, if that is indeed what you want to do?

We are not locked into Palantir, as you have heard.

No, but imagine you make the decision to say no, then what?

The Government have already exited contracts with Palantir recently. MHCLG has, the Homes for Ukraine scheme has also done so. As you say, there will be disruption. Any end of this FDP, as I say, of course would cause disruption.

What disruption? Do you have a programme that is spelling this out? I am assuming you do not want to share it with us yet, but are you preparing for what that might look like? That could potentially be very important for whether it happens or not. You need to be planning now if that is a potential.

Yes, this is a national digital programme, and we always talk about transformation and long-term planning. Sometimes what we will also be doing is constantly saying that we have this, that people have raised legitimate concerns, and suddenly we need to end the contract. We have to have legitimate, clear reasons as to why we would want to change supplier or change our approach. The provisions are that we and the NHS own the intellectual property on the tools; therefore, you can transfer and keep those with a new supplier.

I will come back to that point in a moment, but go ahead.

The conversation will especially be with the current suppliers that have partnered with Palantir in this contract, to talk to some of them and to big providers as to what their capability is, because we have to be assured that the end of any contract means continuity for patient care, because this is about patients.

Can you give me a yes or no? Do you have a programme of work within the NHS to make sure that if you decide in December, for whatever reason—and we have heard that trust could be reason enough. To realise its full potential down the line, you may well decide that all the other activities that Palantir takes, not to do with the NHS, is sufficient reason and that you want to buy that trust back from the public and that it is worth going slower, and so on. Do you have a programme of work that it is developing for you so that you can decide whether it is worth the disruption and what exactly that disruption will be?

Yes.

Excellent. Will you at any point aim to share that with us so that we can scrutinise it?

I will answer for the Minister there. The review process that we are undertaking is the programme of work that we have established to make sure that we can—

That is the contract, not the Imperial review. That is a different thing.

Correct, yes. The contract review process is that process, and the programme of work, through which we will ask how we exit and how can we exit. The contract itself has exit provisions within it. As the Minister has already said, the intellectual property is owned by the NHS, as is the canonical data model that underpins the NHS Federated Data Platform. All of those are transferable. When we set out and let this contract, we created those provisions at the outset.

You did not quite answer the question. Would you at that point be sharing with us what those risks might be, that piece of work?

That work will happen internally, it will go through ministerial approval, we will work that through and at that point, if the decision is to exit, that will be a public decision. I am sure we will work to be open about that, as we have been transparent all the way through. We publish pretty much every single thing on the NHS FDP on the website.

I just need to make a point that the Government routinely do this with technologies and providers. Technologies are phased out when better technologies turn up. It is usual and highly regular for us to have contractual provisions in place for the termination of contract, termination assistance as you transition from one to another, usually at the same time as you run a procurement. Those processes can take anywhere between one and three years, depending upon the complexity, size and scale of what you are doing. My view is that if we were to exit, we would be spending quite a bit of time and effort arresting the benefits from the programme as it currently stands in favour of a transition.

Say that sentence again, “arresting the benefits”?

The benefits will stop as we exit.

You would have to go slower to potentially go faster if that is the decision that the Minister wanted to make.

That is the point. You would have to stop developing on the platform to give the new platform time to catch up.

A downside, potentially, of moving away from the Palantir contract is that the digital transformation, that third arm, might have to slow down. That is already one. Are there others that you are able to tell us about that you are concerned for as a potential downside?

From a technical point of view, we will be moving from a supplier that has a high-performance platform—it works, and it works incredibly well—and we will be taking on trust that the alternative will work as well. Usually you find out that there are always upsides and downsides to these things. We would have to work those risks and those mitigations through.

Are all of those already part of this process, so that the Minister can make a decision about the process?

You would only be able to do that once you have made a decision to exit the contract, you have run a procurement and you know who the winning bidder is, because at that point you then know what you are buying, what the technical implementation is likely to be and what the transition path will be.

Fundamentally, what I understand from this, Minister, is that you will have to make, to an extent, a decision of principle whether longer-term trust in the system, British capability and sovereignty trumps the potential downside. We do not know for sure, because until we have a new bidder and a new company taking over, we will not fully understand what that capability will be, but you will have to take it on trust. Is that your understanding of the decision that you will have to make?

Public procurement has to be conducted lawfully and there are terms and conditions as part of any contract. I have to be assured that that is being met, and balance that with patient outcomes, because of the disruption for patients and NHS staff in making decisions about waiting lists, co-ordinating discharges, supporting vaccination programmes and improving theatre utilisation. There is public trust here. The public do understand what the FDP is trying to do, because this is simply about bringing systems together. I have to look in the round at whether this contract is meeting its requirements as per the contractual requirements that it was provided on. Is it doing the things that we said it would do? Are clinicians using it, and are they behind this? Is it making a difference?

Many are not, as you know.

No, they are. They are, as you know. I shared the figure with you.

Some are, but some are actively not choosing to use the platform because it is being run by Palantir. That is factual.

No, that is not factual, because 139 trusts are live, 170 have signed up. That does not tell you that trust—

Minister, both of those things can be true. It can be true that all those trusts have signed up and also that there is huge disquiet among some of the Royal Colleges and also individuals who have contacted this Committee, and indeed journalists, that they do not want to use it. In fact, it extends to patients, who have said quite categorically that while Palantir is the main contractor for the FDP—which I do not think anyone is disputing here. It is worth being clear about that. The problem here is the supplier. Do you accept that there is disquiet about the supplier itself, and that that itself needs to be weighed against all the factors you have just spoken about?

I would say that the NHS cannot operate a procurement system based on political preference or opinion. At the end of the day, any supplier has to operate on the basis of evidence, capability, value for money and compliance with UK and international law. While there may be allegations and while there may be some concerns, I have to focus on and balance patient outcomes and transformation for the NHS. That has to be the guiding principle. Most importantly for the public is whether their data is secure. Will they get the best possible care? Will they be sat on a waiting list for years and years, with their health being impacted? All of these things I have to balance alongside the Cabinet Office procurement rules and contractual rules, and I have to make decisions based on that. I cannot pass opinion or judgment on every global company that exists or delivers here. These are British employees who have partnered with Palantir—three British companies and one from Northern Ireland. They are British employees who are working on our NHS. This is a system that has been designed by the NHS for the NHS. It controls the data, it has access to the data, and the data is not being sold.

All of that is on the website.

These are the things within any contract that we would want to feel confident about before we take decisions, not because—

It does not worry you? It does not worry you that there is now a level of disquiet among the public? Trust in medical data is, as you know, held in higher esteem in people’s minds. This is the most sensitive data that it is possible to have out there.

Of course.

There are some people who have concerns about the company, and that is not material to your decision? Just to be clear, is that a yes or a no?

I appreciate that people will have issues, legitimate issues, raised about Palantir, but let’s park that, because when you say—

The answer is no?

I understand the concerns that people have, but this is not about transferring ownership of NHS data, because the NHS is the controller of data. That is what the public care about—who owns my data and who is going to use my data?

At the same time the NHS makes mistakes that the National Data Guardian has to go into, and Palantir does have access to some data.

It was not a mistake; it was a clarification of who has access. There are safeguards in place within the contract that shows you who has access for a tiny, limited period of time.

You said it was a mistake, not me.

No, I am saying that it was a mistake in terms of the language that was written about who has access, but the safeguards in place in the contract mean that you can audit and see who used what and for what purposes, on a time-limited basis. I want to assure the British public that their data is secure because the NHS is the data controller. The Federated Data Platform just brings existing data into one system.

I have one very quick final question, which is about the capability of people in ICBs in particular, given the 50% cuts. You heard earlier from Matt Hennessy—thank you to Dr Beccy Cooper for her questions—that there are potentially some people who have been let go who we desperately need to make sure that this is now a success. Minister, what is your plan to make sure that we have the right people in the right places to make whatever, whoever runs the Federated Data Platform, a success?

For the ICBs I have spoken to my chief executive, and he tells me why it is important. In the context of the cuts and the limited resources they face, this is hugely significant because they can make population-related decisions by having all of that information accessible in one place.

That was not my question.

What was your question?

How will you ensure that for the digital transformation and particularly the rolling out of the Federated Data Platform, whoever may run it, that you have the right people with the right skills in the right places? Because right now some of those people have already been let go.

In the ICBs?

Yes. Maybe you were not there for that line of questioning, but it was very clear from panel 1 that we currently do not have that. We do not have as many people who have the technical capability at the right level at ICBs, given the 50% cuts that your Department has asked them to make, to do this.

I would have to undertake to speak to my colleague Minister Kinnock about the capacity within the ICBs. I am sure that will be very different in lots of different places. You are right that we need to make sure that the right expertise and the resource is available so that they can utilise the FDP effectively and make decisions on population health in their regions.

I will ask you for a few clarifications about some of the testimony today. You talked about there being no role for preference or opinion and that it has to be based on performance. Is there no role in public contracting to look at issues such as whether the contractor meets fit and proper persons tests? Is there any role in NHS procurement for those sorts of assessments?

Yes, absolutely. There is a thorough procurement process where you have to be compliant with UK and international law. You have to be compliant with NHS governance. If there have been breaches by companies, that would have to be considered, especially when a supplier has many contracts across Government. In the end, you want to make sure that whoever you have as a supplier who has won a contract based on the terms and conditions, based on a robust process, has done so fairly and that it is transparent as a process.

Do you feel that any of the behaviours by Palantir that we have seen publicly reported breach those conditions about fit and proper persons?

I think the allegations that are out there, if they have broken international or UK law, absolutely should form part of any decision-making on any contract.

That would be part of future decision-making?

As with anything, we would have to give due regard if there was a breach of international law. What we cannot do is get into allegations.

We follow UK Government procurement rules and regulations. The Government commercial framework and Cabinet Office set them.

Do they consider consideration about fit and proper persons?

Yes, if there was a fit and proper angle to this, they would consider that.

You said you were happy about the rollout and the performance of the contract. We have heard, and it has been mentioned, that we are not at the 85% level of secondary care trusts adopting the Federated Data Platform by March. I think 168, and you suggested a higher figure, have signed the memorandum of understanding, but active use seems to be much lower than 85%. We have heard some concerns about different areas. You have challenged those concerns, but we as a Committee have heard that areas have concerns—Leeds and Manchester you have mentioned. Why do you think that performance has been behind where you had hoped on adoption, and why do you think that some areas are concerned and reluctant?

There have been the early adopters. The fact that we have—

But those who are not, why do you think they are not? Why do you think they are concerned?

We have 139 live, we have 170 signed up. If I can just give you some examples, Chesterfield has used it to improve theatre utilisation.

My question is about those who are not using it and who are not signed up, and why we are behind our 85% target by March and why they have concerns.

Different trusts will have different issues in terms of adoption, technology and systems, and there will be culture. When clinicians share best practice, that will encourage lots of other trusts to see what the benefits are. It is about making sure that we can widely share what the benefits are. If I had just read this, as opposed to going to visit, to see for myself—

To turn the question around, it is not understanding the benefits enough. The areas that are concerned, the areas they have not adopted, based on your responses, is that they do not understand enough.

No, some of this is about capability. They may not have all the IT infrastructure in the way some hospitals have; they will be at different places. It does not mean to say that the 170 that have signed up—they are clearly indicating that they want to work and they want to be able to have this and roll it out.

On that point of signing up being an indication of choice—it was mentioned that there is no mandation—to press on that point, planning guidance published in October 2025 said that trusts should sign a memorandum of understanding. Is that planning guidance “could” or is it “should”? Because our reading is that it is “should” and they have to. Are you saying that they do not have to sign that memorandum of understanding, that it is completely up to them?

There are different systems. Where there are already good systems, they do not have to, but there are different tools within the FDP and at the moment different trusts are using different parts of those tools.

It is “could” not “should”?

To clarify, we did say “should” in the planning guidance.

It is not “could”?

It is not “could”. If I can just—

Why does what the planning guidance says differ from what you are saying to us today?

The planning guidance was issued at a point in time, and, as we have said in this Committee already, the FDP was not mandated. Not every organisation required the tools that we were offering, because, as you have said in this Committee—

I do not understand why the system does not think that, why it does not think that what you are saying today is the situation, based on planning guidance that said they should. Considering that the Government have decided to remove a lot of requirements in planning guidance and streamline the system of requirements, removing targets on vaccinations and other things that this Committee recommended to have a lean model for ICB leadership, one of the things you have said that they should do is this, but you are now saying that they should not do that. It is quite a confusing picture, is not it?

They could do that.

Just for absolute clarity, you asked an earlier question specifically on being behind on the 85%. The actual contractual target was 100 by the end of year 2. We hit 117 at the end of March, out of 134 acute trusts.

It was not 85%.

It is over 85%.

Isn’t that to be using it, not to have signed the memorandum of understanding?

That is using it, correct; 117 delivering benefits, as on our website.

We have 205 providers of secondary and tertiary care.

It is 134 acute trusts, which is physical health, and the rest are community and mental health.

So the target does not relate to all of them.

The 134 and 85% was clear in the target as being acute organisations.

So it does not apply to all. I am quite confused about what the policy is; I do not know about my colleagues. There have been reports down this track that senior NHS procurement officials have been warned about speaking out about Palantir. It has been reported in the FT that officials have warned staff about criticising the rollout. Have you heard that report, and do you recognise it? Are you aware of any of those warnings?

We are aware of the FT report, and at the time the Government CCO as well as our CCO made it very clear that people should come forward. If that is the case, please come forward and speak to us. That stands, and at no point have we been—

There was no warning or sense of repercussions communicated?

There has been no warning.

And that did not come from the centre, from your perspective?

That is correct.

Capabilities come up a lot as a challenge, and the Chair has pointed to some of the 50% headcount reductions at ICB and NHS England. The Committee has talked about this previously. Do you have a sense about the headcount level of digital staff at ICB level and at NHS England level? Is it higher or lower than when the contract started? Has capability increased or decreased since the contract started in 2024?

This is information that we need to get from the ICBs, because while they have been doing this merger, they will be able to provide where their capability exists. Many trusts already have good IT officials. However, as I say, where adoption has been an issue, some of the flags have been that they need to get in more staff who have the right skills to be able to roll this out.

You are not really sure what is happening at the ICB level?

Just to clarify what the Minister said, the model ICB that was published as part of the merger of ICBs set out a set of directions on where digital, data and technology staff should remain in the new architecture of the NHS. A lot of that was that digital, data and technology staff should reside in different places, which included regions as well as the trusts.

Would the capability and capacity be higher or lower than 2024 when the contract—

As has been made public, we have been reducing the size of the ICBs as well as NHS England.

Less capability.

There will absolutely be fewer people than currently available.

NHS England is currently merging with the Department of Health. Do you envisage there being more or less capacity in digital staff at the end of the merger?

The merger is being worked through at the moment. We do not have the level of detail of what that will look like at this stage.

It sounds like a key consideration.

Yes.

Looking at the fewer and fewer staff who are employed in NHS Digital, it seems, are you confident that they are the right people, the right grading, that we are competitively paying the staff? Looking at a data engineer role in the NHS at the moment, it is about £49,000, in the private sector it is upwards of £70,000, up to £100,000 in some roles in the private sector. How will we have the capability to create such capacity in the NHS if we are paying half as much as the private sector?

Public sector pay is competitive.

It is not competitive; it is half as much.

What skillsets you will need to roll this out and what support you will need will be different in different roles. You would have to look at what it is that each role entails, as opposed to making a blanket—

It may be different, but I am just comparing general data engineer roles; there might be other comparisons. Have you benchmarked with the private sector for digital staff at NHS central level and ICB level? Are you confident we are competitive?

At a central level we have benchmarked.

Are we competitive?

We are quite competitive. The role that you talk about, the entry-level role, which is an entry-level £49,000 role. I have a number of data engineers who work in my team, and they range from £49,000 to £90,000, so there is a good level of distribution in that area.

Just to finish, the NHS more broadly was pushing for more devolution to local decision-making. We have moved now from “should” to “could”. Apparently areas were not expected to have to do this, but generally that has felt like a mandation towards Palantir as a single provider in this space. Do you regret that movement? Is that counter to the broader trend of the NHS and local decision-making, and do you feel that we should be looking at regional procurement to support competitive operations in the digital space?

This is a national rollout. It is not unique to a particular trust. This is about a rollout that has to take place everywhere where possible where they do not have systems and so it makes sense. It is much more cost-effective—

There needs to be vaccination everywhere, but the message from your Department is that local decision-making matters, local commissioning matters.

I know, but it is value for taxpayer. At the end of the day, when you procure nationally, you will get a far better deal than if every single trust, as we see right now, has its own procurement and has to pay much more—

On the second point, should we be looking to regional procurement of digital contracts rather than single national contracts?

There will always be. Do not forget that we have the single patient record coming up. It is part of the Health Bill. We will be looking for capability across the country in terms of delivering some of that. While there might be some regional projects, this is a national rollout. This is digital transformation across the NHS, and it will be value for taxpayers’ money.

Thank you very much. Just be aware, it looks like we might vote soon. If that happens, please can we finish the series of questions? I promise that we will be let go in order to be able to vote. There are three votes, but with any luck we will get through the last set of questions.

I will be very brief. Just to note, I think that that “aha” that a few people said was a reflection of the fact that vaccinations are widely regarded as extreme value for money in terms of the public purse, but there is not the same emphasis on national rollout, despite our Committee repeatedly calling for that. That was probably some of the reflections there. On the rollout, where do you see the next steps are for the FDP? Are you concerned that there was a report in The Guardian—and I believe that this was the result of a Freedom of Information request—that DHSC officials briefed the previous Secretary of State that the public profile of Palantir is likely to make it harder to go further with the FDP and is also likely to make it harder to encourage the inclusion of GP data locally? Is there a worry about that public perception hindering rollout and hindering the FDP going further?

The platform is live over 139 trusts, and 170 have signed up. Of course we do not want a big transformational programme like this, which is about patient benefits, to incur any issues. I know that people have concerns about the supplier and Palantir, but let’s look at what it has delivered—100,000 patients have undergone more procedures.

My question was more towards the next steps. Where do you see the next stage going, and what is the medium-term vision for the FDP? Is that going to be hindered by this public perception?

Next steps is that, of course, we want much more adoption. As I say, different trusts are using different tools because of what they need. There will be tools that we want greater adoption in. The ICBs, especially with the cuts and issues with resources, means that this is quite a game changer when we are talking about vaccination rates, making population health decisions. It absolutely will help them direct resources. They have a lot of responsibility in local commissioning about where services go and what is the unmet need. This is important so that people can see the benefits of what the FDP is, and patients can feel it. I will bring in officials on the further next steps.

You made the vaccine point and it is pertinent, because through the FDP a few months ago, we started the work on the meningitis B vaccine programme. I do not know if you are aware as a Committee, but we run the covid, flu, monkeypox, MMR and meningitis B vaccine programme through the FDP in a matter of weeks. That is the capability we have, and that is the capability we have procured to ensure that we can stand up a vaccine programme in that way. Just to add further on where we are going next, as the Minister already said, we have significant penetration already in terms of number of organisations. It is about spread within those organisations to make sure that the capabilities are being maximised. We have talked about the cancer plan—or we have talked about the cancer services that we are supporting currently; 94,000 patients have already benefited through going through the cancer MDT, the Cancer 360 tool, and the single queue diagnostic work, which is live across every single hospital in Greater Manchester. Every NHS trust in Greater Manchester uses a single queue diagnostic tool as well as the patient stratified follow-up. That is happening now.

The question is where are the next steps? When you are talking about more penetration, when you are talking about bringing more trusts on board, there is a key about bringing primary trusts on board as well, as that is a big data blank. Are the concerns around Palantir delivering this hindering that work? Minister, with respect, it seems as if your own officials briefed the previous Secretary of State that this is a concern. Is it a concern?

There is a perception out there, absolutely, and part of the population will have some feelings on the supplier. We are concentrating on working with organisations to maximise the capability, and that is what we are trying to take forward with those organisations.

Are you working with organisations that are currently bought in with the 131 trusts, or are you wanting to expand the scope?

We already cover the vast majority of acute organisations, and we are trying to work with more community and mental health organisations currently. We are also working with 35 of the 36 ICBs, which are the legal entities, by the way. It is 35 of 36 legal entities that exist. We are working with those organisations to make sure that the capabilities already developed on the FDP can be maximised.

We do not want it to hinder. I know there have been legitimate concerns being raised, but it is important that more people understand what the FDP is and what it is being used for by hospitals and by ICBs. In the end, what patients want is that their data is not sold, is not being used by anybody else, is stored here in the UK and is owned by the NHS. It decides who has access to it and there are very sufficient safeguards where there is an audit process that shows you who used it, for what purpose. That is really important. We need to communicate that out much more.

There are still questions on that, particularly around some of the queries that my colleagues have raised. Very, very briefly, on learning points between the FDP and plans for the Government rollout of the single patient record, is there anything that you will be doing differently in the data privacy space?

Are you talking about the opt-out specifically? We have not started working on data privacy, and we do not know what it will look like. It is part of the Health Bill. We have started to put some principles together, doing some work on it. That will be a core part of making sure that people know that their data will be secure. It will be integral to that.

Would having a company like Palantir, which has a certain amount of negative publicity, shall we say—I will not go into the rights and wrongs of that—hinder a single patient record process being rolled out and see a large uptick in opt-outs, to the extent that you would consider them not being a partner in this?

It is a very, very different technical problem. If you come back to what the FDP is, it is mainly an operational data system to help run the acute system at the moment. Single patient record is about the patient’s individual clinical record. We would take a wholly different approach technically on how we roll that out, and in the process of doing that take a very, very good look at the data privacy and opt-in, opt-out elements of it as well.

Would you find a company like Palantir to be a decent partner for something like single patient record, if there were sufficient concerns from the public that you would see such a high uptick of opt-outs?

Currently, it is allowed to bid for work. There is nothing to stop them bidding for any Government work, let alone anything to do with the health system. It can bid for it, but ultimately it is quite a different technical problem. It is a different societal problem as well.

A different beast, which we may yet come back to.

We want a British company to come forward, and the capability is much different.

There were follow-ups but, Minister, you have been saved by the bell. Could I ask for an undertaking from you? We would love to put the questions to you in writing. Particularly, I had one about IP disentangling. I know colleagues were querying some of the penetration numbers, for example. We will write to you and ask for a response. Thank you for your time.