Palantir wins the Federated Data Platform contract (predictably) 

Having had a few days to find some more surprises and omissions in the statement to Parliament and the press release confirming that Palantir won the £330m FDP contract, we understand why it is confusing. The goal is, after all, to copy everything into Palantir and only get support afterwards – and to not allow anyone to say no.

On Saturday The Times (incorrectly) said there is no opt out; a day later, the Sunday Times correctly said there is. Things are clearly in a mess, but it must be remembered that last week’s announcement was a start line for this project, not a finish line.

In brief: The opt out exists just as it has previously, and nothing changes until around March 2024. Everything else is a currently mess; we’re working on it.

Slightly less brief version: The National Data Opt Out exists – it works as it always has, though it could always be stronger. NHS England has not published enough details to know whether patients’ objections will be respected, whether and where it believes they don’t apply, and/or whether the Government will make everyone opt out again. This is because there are no details of individual data flows in the FDP as yet; maybe by March 2024 we will know more. Maybe.

Some have read NHS England’s briefing to argue strongly that patients cannot opt out; medConfidential and others, however, work for a world where patients’ wishes are respected. Join our mailing list to find out how it goes, or opt out at any time.

We will continue to scrutinise details. NHS England could still decide that people’s opt outs don’t apply where they should, and officials will probably try doing this – which is akin to making you opt out again. NHS England talks about its five high level “priorities”, but data choices and uses and paperwork is about very specific purposes – exactly what data, to and by whom, treated how and for what uses (e.g. “strike analysis”), with or without respecting opt outs? The devil is in all the details.

We do not yet know whether those who have already expressed their wishes will have to do anything else – NHS England has not yet said whether Government will make patients opt out again, but as of now, the opt out process is enough.

The Details Do Matter

The Government is as committed to giving Palantir £500m over seven years (the current contract is for five) as it was to HS2 and Net Zero. If you wish to opt out, we have a page on how to do that. Many details are still to be argued about by lawyers and others, but you currently have the same choices after last week’s announcement that you had before.

As an administrative body, not the NHS in England, NHS England (NHSE) can only make decisions for NHS England. It can try to bully NHS Trusts and Integrated Care Systems into going along with its decisions – seemingly with limited success – but, as separate legal entities, ICSs and Trusts get to make their own decisions. Sometimes NHSE says Trusts “can” choose to use it, sometimes NHSE says they “will”. The comments published by the Science Media Centre demonstrate the lack of support that NHSE tends to respond to with strong-arm tactics.

While it is good that NHS England has an “engagement portal” of sorts, it will be far better when there is meaningful content in it – and when that content is entirely accurate and complete. 

We note that NHSE’s public press release was less informative than its statement to Parliament. And while the press release talks about “Trusts and ICSs” having access “from spring next year”, NHS England itself is notably absent from that timeline. 

Our current understanding is that NHSE will get access to data at the same time as Trusts and ICSs, when the “Privacy Enhancing Technology” is available. Previously, officials had suggested NHSE would have access before that. Of course, there is no obligation on NHS England staff to be fully candid in meetings, and someone decided not to make this all clear in the press release. This kind of decision has caused and continues to cause many problems.

NHS England has until around March next year to get its story straight, and for its officials to be transparent about governance and patients’ choices – hopefully far better than they managed on Radio 4’s Today programme last week.

(For example, GP data could at some point be copied into Palantir for where you live, and then copied elsewhere, unless you have done the separate GP Data Opt Out. Despite asking repeatedly, details on this are unclear. The Palantir Federated Data Platform is the first time that National Data Opt Outs and GP Data Opt Outs will overlap in this way, and we still have no idea how this is going to be handled – based on past behaviour, NHS England will probably just try to grab the data, because that’s what they instinctively do.)

“Direct care”

The heavy lobbying from NHS England has focussed on “direct care”. While it is unclear what direct care the bit of the Government that calls itself NHS England actually does, NHSE’s FAQ initially said (and some versions still do):

“Can patients choose how their personal information in the FDP is used for their care?

The new FDP will be used by NHS staff to offer patients care.

So, it is assumed that patients have given consent to access their personal information.

If patients do not consent to their identifiable patient information being used to support their treatment, they should email for information about withdrawing consent.

If a patient does not consent, it might affect their care.”

Then they took that text away. 

NHS England was correct that, in some circumstances, you can object to aspects of care that you don’t want, for any reason you like – this is normally referred to as “patient choice” – but it has apparently been decided by NHSE that such choice is not here and not now, any more. Or, to put it most generously, there’s no clarity. Again. (This is becoming a theme…)

NHSE is only clear that the National Data Opt Out and GP Data Opt Out do not cover direct care uses, and it is critical that that remains true; were it not true, that would be a resignation issue.

That an administrative body which treats no patients and runs no hospitals wishes to define some (or any) of its activities as “direct care” as a figleaf for processing data it otherwise could not access is spookily reminiscent of the failed GP data grab of 2021 and the catastrophic programme before that.

Confusion about purposes beyond direct care

Many people have noticed that the National Data Opt Out (NDOO) is somewhat limited – and, if you rely solely on the way NHS England chooses to describe it, it appears far from effective. But it exists. And it does work. Some argue that it is useless; we and others work to make it stronger and better.

For now, if you have chosen to do a National Data Opt Out, it does as much as it can be made to do currently, and will hopefully do more in the future without you needing to take any further action. (The same applies for a GP Data Opt Out, if you have done one of those.)

The National Data Opt Out was created to avoid a proliferation of new opt outs as new data programmes came along. So while NHSE could say that the NDOO doesn’t apply in a particular scenario, under the law (see below) and in effect it would be saying, “there’ll be a different opt out for that”. In practice, this is untenable, and it is precisely the reason why the NDOO came about. A single National Data Opt Out will eventually cover all dissentable processing, but that will take more campaigning. And lawyers. And time.

The day before the contract announcement, on the same page as the text above which confirms that you can opt out of data use for direct care, the FAQ said:

“Can patients opt out of their data being shared in the FDP?

No. Patients can only opt out of sharing their identifiable data for research and planning.”

A week later, the answer to that question now provides a link to the National Data Opt Out – so the answer was clearly not ‘No’ but ‘Yes’. Additionally, while NHSE attempts to narrow the opt out at any opportunity, the choice of words both initially and now reflect NHSE’s persistent belief – since its inception, and the ‘bad old days’ of Tim Kelsey and – that patients have limited rights, and that if NHS England wants to do something with data, or Government wants to do something with data, then they can ignore patients’ express wishes.

The FAQ does not say that NHS England will respect all of the rights of a data subject – for example, the right to object to unnecessary processing – instead it simply ignores that any inconvenient rights even exist. Courts tend to take a wider view…

NHS England might tell people in future that there is or will be processing of their data to which they can object, but to which NHSE will not apply the National Data Opt Out. This would be creating (yet) another opt out, and making everyone have to opt out all over again. Alternatively, NHSE could apply the National Data Opt Out to all such processing, despite what it is saying today. 

These are the two choices facing Government – and it should be clear that this is a Government decision, not an NHS decision. And when that decision is made, either way, the paragraph will have to be changed again. The current version is less wrong than simply saying “no”, but it’s still not right. Yet.

medConfidential does not believe anyone should have to opt out again – we can’t make that a promise because that’s a Government decision – so you should be able to make your NDOO choice now and not worry; but if you want to keep informed, join our mailing list for more as we know it.

Making inaccurate statements to the public

The FAQ mess – that it publicly stated contradictory things between last Monday and Friday – means that someone, somewhere inside NHSE, agreed that both statements were accurate at some point. Who changed their mind?

The same FAQ has previously claimed, “The existing web copy was produced in conjunction with Med Confidential”. This was not true, and in the weeks it took NHS England to remove that entirely false claim, we were assured that the FAQ ‘had process to follow’, and that it couldn’t be done quickly, as everything was checked. We never received a proper answer as to how that claim came to be made – officials seemed to be relying on a meeting in August 2022 for the claim they made a year later, as if they believed they had learnt nothing in the interim period. (Perhaps they didn’t.)

Even if one were to accept statements that such occurrences were honest mistakes, how many other similar mistakes are there in NHS England’s decisions? And why is it that those mistakes inevitably tend to benefit NHSE and fit its intentions, and remove choice from patients? Every. Single. Time.

Honest mistakes would go both ways; systemic failings only benefit NHSE.

No meaningful check, no meaningful challenge

NHS England’s “Check and Challenge” group met for the first time on Friday, but is only scheduled to meet once every two months. That means its first substantive meeting will be in January, and NHSE expects to start using Palantir in March, possibly even before the second substantive meeting of the group.

Even if the best possible questions having been provided with perfect information were to be asked at the January meeting, officials could (entirely reasonably) say, “We’ll come back to you on that,” and NHSE could do whatever it wanted – Palantir could even start running before the group meets again. It should be noted that some decisions made at this stage are irrevocable, and some very hard to roll back once begun.

We don’t expect the “Check and Challenge” group will be allowed to do much of either checking or challenging, and to make doubly sure they don’t – and in stark contrast to previous debacles like – medConfidential were not invited to join it.

One “check” the new group might discuss is “pseudonymisation”, and the extent to which NHS England will once again ignore that pseudonymised data remains personal data – and that, as such, any unnecessary processing is dissentable. So the NDOO should be applied. If not, the “check” group will be allowing NHSE to create a new opt out.

The group is not expected to publish minutes or papers, so we shall be FOIing them regularly which is a waste of everyone’s time. It should be noted that since “new NHS England” was formed, it does not promptly respond to requests for such things – the “Digital Data and Technology” subcommittee of NHS England’s Board, for example, refuses even to disclose its membership. 

Palantir Access To Data?

We don’t know what data flows will be in the Federated Data Platform as yet, as NHSE hasn’t told anyone. And we don’t yet know who will be auditing NHS England’s setup of Palantir – but we’ll certainly have questions for whomever it is. All we can do in the meantime is look at the data flows that were in the precursor system to the FDP in January and July of this year.

Notable in January was “strike analysis”, for which NHSE apparently used spreadsheets instead of Palantir Foundry – an item which conspicuously disappeared from the list in July, despite other items being “included for completeness to show a reconciliation with the original list of purposes”.

The July list gives sixteen purposes described as “system admin”, and we don’t know who those admins are. (This is why purpose descriptions are helpful; a feature Palantir Foundry has by default, which NHSE chooses not to reveal.) 

Hopefully when NHSE publishes the full list of which flows of data are being used for what – something they could choose to publish today on their engagement portal – we will know, and so will you. NHS England has signed the FDP contract, and all we know is that it would be illegal for Palantir to use data in ways that NHSE doesn’t allow – the problem being we don’t know what they don’t allow, and still know precious little about what they do allow.

Hopefully NHS England is better at writing and managing contracts than they are at writing FAQs. We don’t think this is a severe risk, but NHS England has failed at managing so many obvious risks, and so many bland reassurances have been shown to be untrue, that further bland reassurance at this point is clearly insufficient. 

If NHSE wants to use NHS patients’ data, it must provide full information about all of  the data it wants to use, how it will use it, and precisely what for, and for each flow, at each stage, either respect opt outs or explain clearly (and accurately!) why not.

The NHS procurement was done carefully, and had to be, despite some interests wanting to cut corners. But the contract was awarded, the facade fell, and Palantir sent round an unpublished briefing saying how wonderful Palantir is – ending with a quote from NHS England saying, “you have to liberate the data, and as we have done that”… 

Join our mailing list for more information as we know it.

Below is the text put together in pieces as we went through the documents for the first time:

According to the written statement to Parliament, Palantir won the £500m contract. The announcement is a start line, not a finish line. NHS England will now have to keep their contradictory promises. We’ll be here.

The Government is as committed to giving Palantir £500m as it was to HS2 and net zero. You still have choices.

If you wish to opt out, we have a page on how to do that (there are details to be argued about by lawyers), but you have the same choices today that you had yesterday.

NHS England can make a decision only for NHS England. NHS England can bully Trusts and ICSs into going along with it (seemingly with limited success), but as separate legal entities, ICSs and Trusts get to make their own decisions.

It appears Trusts/ICSs will not be able to start using it until “spring next year” (April?), because that is when the “privacy enhancing technologies” start, but NHS England will use the platform immediately, with or without those technologies.

NHS England’s FAQ used to say that opt outs don’t apply, but it now says they do. The day before the announced, The FDP FAQ started the answer to a question about opt outs applying with “No.”, and now is much of a vague yes. NHS England can still decide that the opt outs don’t apply where they should. We’ll continue to scrutinise details, because you shouldn’t have to opt out again. NHS England talks about 5 high level “priorities”, but data choices and uses and paperwork is about very specific “purposes” – what data, to what point, for what purpose (e.g. “strike analysis“), with or without opt outs? The devil is in all the details.

We do not yet know whether those who have already expressed their wishes will have to do anything else – NHS England have not yet said whether Government will make patients opt out again, but as of now, the opt out process is enough.

NHS England has until around March to get their story straight, be transparent about governance and patient choices, better than they managed on radio4; NHS England’s “mistakes” only seem to make life easier for NHS England…

NHS England gave a chosen few organisations different briefings on the 20th November, ahead of the announcement on the 21st, and we don’t yet know what else they’ve mislead people on, in the same way the FAQ used to say “no” to opt outs, and which now confirms you can opt out. We’ll update this page as we have more considered views.

Join our mailing list for more as we know it.