Doctors and patients have spent a decade on the “exciting” end of a triple pendulum, living the random changes in direction that come from those incoherent pressures.
Abolishing the public body that called itself “NHS England” is a good idea in general.
Speaking in Parliament, Wes Streeting said: “NHS England will be brought into the Department entirely”. Entirely is the key word, with consequences for decisions that have not yet been made. The abolition is a change that everyone thought beneficial but few expected before Wes Streeting first announced it two weeks ago and which surprised almost everyone when Keir Starmer re-announced it on Thursday.
The body known as NHS England has conflated itself with the NHS in England which caused public confusion – they are very different things. This confusion is one reason we have referred to the overlapping behemoth as the Department of Health in England for some time.
Formally creating the Department of Health in England is a simpler and less dysfunctional of what is de facto the case today. NHS England takes instructions from DHSC and does what it thinks it should do within those constraints – the laws of unintended consequences abound. It’s not that the new Department of Health in England will necessarily make any better decisions, but it also doesn’t have to, if only because accountability will be tractable and improvements possible.
We expect social care responsibilities to also shift around after the Spending Review, as the new Department of Health in England will want to use Palantir to manage social care, in addition to the 3 DH priorities of analogue to digital, hospital to community, and prevention.
New Directions on Data
To move data around the existing system, DHSC currently writes Directions to NHS England.
But a public body can not direct itself – it can not bind itself for the simple reason it could equally easily unbind itself. When DHSC gave NHSE a Direction, NHSE had to follow it (DHSC doesn’t).
When E ceases to exist as an independent entity, the legal force of the Directions falls away, and because of the way the current legislation is written, s255/s259 of the Health and Social Care Act 2012 will cease to be operable unless there is still a separate public body to Direct. The approach used when NHS Digital was eaten by NHS England doesn’t work unless two bodies remain.
There are two choices:
Secretary of State can ask Parliament to pass new legislation reworking all data flows in the NHS to be under the control of Ministers and the Department, leaving everyone’s medical notes at the Secretary of State’s whim;
A new public body is formed, perhaps an NHS Information Centre, to act as data controller and centre of information governance (to avoid politicians being in charge your medical notes)
Although Streeting told Parliament: “NHS England will be brought into the Department entirely”.
The way the NHS England Federated Data Platform (Palantir) has been designed and build has given E joint data controllership of all data in the system. When E uses Palantir to decide which patient gets the operating theatre, that is because E wants to be able to change that clinical decision, and now it will be Secretary of State involving himself (or deciding not to which is also a political decision).
E has also been working with the NHS to work out the economic benefit of treatments (paras 3+4), to allow deprioritization of treatments which don’t benefit the economy, and use Palantir to prioritise diseases of the rich.
As DH eats E “entirely”, then those will all become politically responsible decisions, rather than a public body doing what it thinks it should be doing. Institutionally, the staff of E conflated the two.
Passing legislation gives the first opportunity to place the National Data Opt Out on a statutory footing. To give it the teeth that NHS E took away in implementation, and to reflect the rights of a citizen that NHSE outright ignores. It’s E’s official position that the way to express the Data Protection Act “right to object to unnecessary processing” is to express a National Data Opt Out, but then E does not apply NDOO when doing “unnecessary processing” citing loopholes in the definition of NDOO. DHSC has washed its hands of E’s decisions for years, now it will be responsible, and will have to address the blatant illegality that it will inherit (and upon which the Good Law Project have a legal case). Or DHSC can simply recognise MHRA got their implementation right as part of DHSC.
An effective Department of Health in England offers benefits to everyone – it will give a coherence to decision making that has been missing since Lansley left DHSC, and his successor expected to govern and have responsibility for delivery (NHSE as a model worked for Lansley because he wanted power without responsibility).
DH and E are flailing in a mess of E’s own design. NHS E’s indirection meant they used data on people who had objected, but then couldn’t use it on those who did want data used – generally making everyone unhappy.
When NHS England eventually opens OpenSAFELY to non-covid research, many of the challenges of GP data can be solved in a manner that is consensual, safe, and transparent, and the remaining issues can be addressed (but there’s always someone who pees in the pool).
Now DH/E can do things, respecting patients’ decisions will mean those people who don’t want their personal data used for purposes beyond their direct care would have confidence that it isn’t used for purposes they have objected to, and those who want data to be used can see that it was used for the benefits advocates claim (including growth).
Innovation and Growth
By the end of this Parliament it will likely be feasible to cure cancer in your cat, making it politically untenable not to equally cure it in you. The outgoing NHS England would have been a barrier to delivery of that improvement.Creating a functional Department of Health in England is an opportunity which could make practical in 2026 the kinds of improvements that were not feasible in 2024.
Part A of our innovation note from January is minimal largely because anything had to be acceptable to DH and deliverable by NHSE and those often fundamentally conflicted. Now there should be some increased coherence. Part B will be easier to write.
It doesn’t mean better decisions will be made, but it increases the odds and that helps patients.
When the think tanks move beyond self-congratulatory pieces about their foresight, maybe they could think about what the new DH/E should do with that capacity.
The Department of Health in England has the power to take functions from NHS England starting tomorrow, but in practice it will take some time. High priority functions are likely to go first, which will be a way to tell the priorities of the new Department of Health in England.
NHS England itself can only be abolished by primary legislation, but it can be hollowed out into a shell of its former self by administrative decree. Going from the hundreds of functions today down to one last function is trivial under section 103(3) of the Health and Social Care Act 2022. It merely prevents the eventual last remaining final function of NHSE from moving. That last remaining function does not have to be material and can be new – it could be a webinar on progress since the Darzi Review and an annual report, and the legislative box is ticked for everything else to have moved. Going from one to zero requires primary legislation long after the Department of Health in England has taken over.
NHS England reaps what it sows. It is culturally fitting that the powers used by it to devour other public bodies will now be used to devour it.
NHS England’s cultural blindspot was that it was “the NHS” manifest, and that no one could do to it what it did to others. That blindspot leaves a few messes to clean up.
The loopholes in the National Data Opt Out used to be a long way from the Secretary of State’s view, whether the opt out implemented by NHS England does what Secretary of State said it should becomes a lot more testable. DH knows NHSE sells identifiable patient data, yet has never been forced to do anything about it. That luxury goes away.
Wes Streeting is also proposing a “patient passport”, a single central government owned health and care record run by the Secretary of State. The initial indications from the Department of Health in England are that you’ll have no choice about having a central care record, there will not be any notification to you , let alone any choice. All your medical notes – the why of prescribing – are proposed to become the official property of Wes Streeting.
The way Palantir has been implemented for the Federated Data Platform today has placed NHSE as a joint data controller over everything. Post-abolition that appears to be going to the Secretary of State – i.e. every data flow in Palantir for every hospital, ICS, and nationally, will now have political responsibility.
The use of algorithms in FDP to prioritise/delay care would be a political act in law as well as in political terms.
Wes wants to run AIs over the records nationally and directly and push your behaviours via the NHS App. This prioritises politically driven health decisions over your own doctor’s clinical decisions about what care is usually best for people like you, and then your doctor discusses with you what you personally want to do. Such decisions have a habit of cutting through – Wes Streeting’s line on “Jabbing overweight people” has been in the BBC Newscast opening credits for months now.
The old NHS Digital and before that the NHS Information Centre existed for a reason – to separate the process of accessing your medical records away from political control.
The end state, of consolidating your entire medical record in one place, under the control of the government of the day, is a radical shift in the status of your medical notes.
medConfidential’s newsletter goes out around once or twice a year, and you can subscribe using the box on every page on our website. This is what most people subscribe to.
Nothing is changing about our occasional newsletter which is where important things go, gets emailed out by us, and is also posted to our news page. That newsletter is what most people should subscribe to.
Most medConfidential output appears on our news page as articles. In the past, they have been tweeted and posted to facebook. The consequences of recent changes in the social media ecosystem, primarily twitter/X deprioritising posts with links, means they are now also posted to bluesky.
Starting now, we add a new email update list which gets the post to this news page after they come out here. If you want email notifications of new blog posts, we now have a medconfidential substack mailing list which contains a copy of what’s posted here (or a link to briefings published on our website, depending on what happens). In a time of rapid change to the Department of Health in England, it may be somewhere we put drafts of the pamphlet.
In a couple of hundred words on the last Friday before Christmas, a short piece talked about what the recent HDR/Sudlow Review means for you and how the Government thinks it will use your medical history and your medical notes in ways you may not entirely like. This blog post is a little longer than that – we look at what they’re already doing and how.
“NHS England is working with the Office for National Statistics to assess the economic benefits of several health interventions including talking therapies, bariatric surgery, treatment for endometriosis, and the NHS Type 2 Diabetes Prevention Programme.
The analysis will also cover the impact on waiting times, employment rates and earnings while feeding into work by the Office for Budget Responsibility and the government on labour market effects.”
DH/E and HMT refuse to say whether the instruction to prioritise public funds for economic growth, and the instruction “not doing things” that don’t grow the economy, apply to the NHS. NHS England clearly thinks they do. When reviewing the project, NHS England’s Independent Advisory Group on Data (AGD) asked if HM Treasury should be named as a joint data controller for this project (section 5.1.10). In order to prioritise treatments that assist the economy in a fixed NHS budget, the Government will also have to deprioritise treatments that help patients but don’t measure as helping the economy.
Your health records linked to your tax records is how that is measured, and they’ve already started. The HDR/Sudlow Review covers their perspective on how that should be done and what the Office of National Statistics should do with data. ONS convened the review because ONS has a major problem – they culturally insist they have the support of the public for what they do with data, but they measure that support by response rates to their surveys, and public support has collapsed to only 20% and falling.
In response, like other abusers of power who want to feel they were thinking like Obi Wan, ONS found it easiest to act like Vader.
The context: All your medical history and medical notes all in one place to be used however the Governments wants
The proposal for a Single Central Care Record has been announced and re-announced several times now as Wes Streeting’s “big idea” to help the NHS. It will put your entire medical history inside Palantir/FDP where DH/E can run AIs over it to do “population health management” and where access can be offered to your medical notes anywhere the NHS logo is seen – not just any doctor you see, but any doctor you could see (including private doctors and hospitals). Will any creepy single doctor anywhere in the NHS be able to look up your full medical history including all medical notes? Probably.
At this time we don’t know whether you will be forced to have one of these records, or what transparency will be around where your record is accessed. Given how DH/E are talking about the record, you are likely to be forced to have one as it is for direct care but it is replacing existing systems which do have an opt out, so who knows. The Department of Health in England has repeatedly sided with creeps over their victims.
When thinking what data the HDR/Sudlow Review covers, it is not just your hospital records, but all your GP records and all your written notes, and anything any other part of government (DWP, HMRC, DfE, etc) may know about you, all linked together and available to civil servants with no obligation to disclose what it’s used for unless someone in Parliament asks.
One argument made throughout is that data is only used for “public good”. There is a vast difference between what the Government of the day thinks of as “public good” and what is good for you. Austerity was a public good according to the Government of the day.
When a pregnant woman shares her heart rate readings 24/7 with her doctor for her health, it is the position of HDR/Sudlow that the Department of Health in England should be able to sell those data so they can be mined by anyone who pays the access fee. It is the position of DH/E that there is no opt out for you from the sale of that data – the choice is literally your data or your life…
HDR/Sudlow in brief: Clubcard Culture sends more data to more places more recklessly
As described on page 167, Biobank wants the Department of Health in England to reuse data that was collected with the promise of “pandemic only”, and give it away to Biobank and others to hand out like any other NHS data. Buried on pg 102 is the desire to collect and share data from your smartwatch and on pg 103 your loyalty card. In practice, that would be the NHS collecting your smartwatch and loyalty card transactions and placing them in Palantir for Biobank and others to copy and rifle through.
One slide from the launch sums it up – it shows Clubcard data and smart watch data being copied to the NHS and onwards to research and commercial use – but that slide doesn’t mention the very controversial HMRC tax data that ONS and NHS are using to change waiting list priorities. Funny that…
Recent research shows that the smartwatch on your write can help diagnose a range of conditions earlier, but that level of detail can also be used to
HDR/Sudlow: Biobank as the role model for more recklessness
Proposing that more data goes to Biobank to go to racists doesn’t grow the economy for everyone, it just grows some bank balances. Biobank believes that’s ok (still) because they pay Biobank’s fees.
Speaking at the launch, the report author said the goal was applicants getting data in “days” and proclaimed Biobank have “one of the best systems” for giving data out rapidly (without sufficient checks on who they gave it to).
The Department of Health in England are auditing Biobank, but it’s unclear whether that audit will cover Biobank processes which led to the coverage in 2024, or whether the audit will be narrow and restricted only to questions around insurer access by the Guardian in 2023, possibly also excluding the story from 2022 about Biobank selling data to China (nothing seems to be happening about our 2024 followup to that – Biobank ignored our questions). Biobank’s main concern seems to be whether they have received money from applicants, only checking that the applicant isn’t on Biobank’s list of known racists to reject – a list that will always be incomplete.
The NHS application form is 30 pages, the Biobank version is only three pages. Biobank continue to insist that nothing in those 27 pages matters (and nothing in their contract with NHS England requires them to ask anything in those pages), but the checks are so laxthat Biobank didn’t spot an applicant was a bunch of eugenicists operating out of the same fake office as QAnon sites and other scams. That’s the approach HDR/Sudlow and ONS wants to all public sector data.
Biobank repeatedly argues that eugenicists keep trying to get Biobank data, and when Biobank catches them they’ll say no (pity about the cases they’ve said yes to). Biobank claim not to understand the approach of making multiple requests until it works, yet Biobank keep repeating their demands that data they should get Biobank think they deserve without following the rules, and will keep stomping their little feet (and getting meetings with Ministers) until they get what they want – the ripping up of promises to patients.
HDR/Sudlow’s friends shouldn’t have to keep their promises?
Unless patients can see how data is used, the government will get lobbied to break every promise they’ve ever made to patients whenever the day ends in “y”, because no patient can tell the difference. This report is another example of that – HDR/Sudlow Review agrees with Sudlow’s previous lobbying job that their friends should get a free pass out of their obligations.
Following the Review’s recommendation, DH have announced that they’re going to give the “pandemic only” GP data to Biobank and others under terms that are not limited to the Covid19 but will be for whatever Biobank, Our Future Health (and others) think they can do usefully with the data in future (and only they get to decide that, no one else has a say)
HDR/Sudlow: only HDR’s Sockpuppets should be listened to
The Review says some things about consultation and engagement, picking winners from amongst consultation respondents by choosing those deemed a “good” consultee because they agree with what HDR/Sudlow wants to hear. HDR has form for this in designing an supposedly “open” call for applications in which UKRI will channel public funds to the “single collaborative bid” HDR chooses to support (the bid of their friends)
Unfortunately and embarrassingly for the authors, the HDR/Sudlow Review was so distracted by ensuring the Review only included the views of HDR that it forgot to suggest anyone else be invited to the group that supposedly should write the rules. The Review was so far in an echo chamber it forgot all it was listening to was itself.
According to HDR/Sudlow (pg 163), those that “would be well positioned to lead on SDE standards” is only “HDR UK [legal entity: HDR], ADR UK (partner with HDR but legally part of UKRI who fund HDR), the UK Health Data Research Alliance [legal entity: HDR], and UKRI’s Data and Analytics Research Environments UK (DARE UK) programme[legal entity: HDR]” and no one else. We have covered at length the ongoing subversion of safeguards by HDR as they continue their cash to cronies grants process and closed shop.
DH/E may be told they’re hearing from four organisations, but all bar one are sockpuppets controlled (in the data controller sense and in the practical sense) and owned by HDR where all the people in the room report up to HDR’s leader Andrew Morris and follow his party line.
Similarly, many of the public engagement groups of which HDR/Sudlow speaks supportively – PEDRI [legal entity: HDR], DAREUK [legal entity: HDR], etc are again simply HDR hiding behind another logo. The one exception (UseMyData) have staff funded by HDR and other staff paid by NHS England to do engagement work – entirely legitimate but difficult to be considered independent from HDR’s desires around NHS data.
Biobank and HDR want all data the NHS has, and they act as useful idiots for others who want that data for their own gain, including the Secretary of State who has an idea…
The Review was written to justify what was already being done
The NHS England announcement gives four examples where DH/E knows that there is a clear investment case to be made for funds, which is good for the bits of DH/E managing those waiting lists, but in a fixed budget imposed by HMT, that takes cash away from other treatments.
Focussing on economic benefits means “economic growth” supplants clinical decision making and prioritisation. The more say HMT has, and the more control the Department of Health in England takes, the less choice your doctors have.
The obesity work has already been done in the pilot, and now will be repeated with the other three areas (the work is being done by statisticians who like comparability). Ian Diamond talked about the work at the launch of the Sudlow Review where he disclosed that the existing work was done by ONS linking together taxpayer records with their health data, something the HDR/Sudlow review seems to have obfuscated in the Review itself.
All your health records and all your tax records linked, for departments to do with whatever they see fit. Once they’ve done these four areas, it is bureaucratically indefensible for not comparing all treatments across all of the NHS – what happens if there’s one that offers more benefit? But it will also show all treatments that have patient benefit but not measurable economic benefit. According to ONS (item 5.1 in April), the organisation making these decisions should be ONS, not the NHS. ONS makes decisions about the value of a statistical analysis, without any regard for the impact on patients or health.
ONS can do this because when NHS England gave them data for statistics purposes, NHS England chose not to respect the National Data Opt Out because it wasn’t identifiable data that would be linked onto your tax records (supposedly). Once ONS got the data for something, it can reuse it for anything. The NHS England Advisory Group on Data was not entirely happy (section 5.1 in October) but NHS England doesn’t appear to care.
Despite the use of tax records being discussed at the launch of the Sudlow Review, the diagram on page 123 of the report notably omits HMRC taxpayer data being in the plans outlined by the Sudlow Review. The mention on page 90 is so opaque you could be entirely forgiven for missing that your tax records are being linked to your health records and data mined in ways you have no awareness of (while academic projects are publicly disclosed, projects internal to government are kept secret as clubcard culture has taken over the “digital centre of government”).
The HMRC/health data linkage is not used to give or refuse care – it is a model used to say what types of treatments help economic growth (and so should be treated faster), and which don’t (and so people can wait longer), in general terms, for people like you.
Simultaneously, DH/E is centralising waiting list cleansing and prioritisation in Palantir means DH/E will be centrally setting the criteria for who gets the limited resource of an operating theatre or specialist care, replacing centrally what is currently done by the doctors in your hospital making decisions based on clinical need.
Like equivalent analysts at US health insurers, ONS will claim their findings don’t affect your care, but it is intended to affect the priorities of care provided to people like you for care you may need in future. If the care you need doesn’t promote economic growth then being rich won’t help you, and if you happen to have a disease of the rich then your poverty doesn’t matter. It’s all about “in general” not you specifically. But, diseases of London have a bigger impact on the economy than diseases of Blackpool or the North, so waiting lists in London should be prioritised according to the logic that the previous government used to start the NHS down this analysis path.
The Everyone Database: names, addresses, dates of birth, and all the identifiers used for you across Government linking all the data government has
I’ve previously sat in a UKRI meeting where a Professor spoke at length that he was not suggesting creating a population index or population register, while his slide behind him said he was, and he wanted it to be based on the NHS patient register. They had invited no one from the NHS to that meeting – they weren’t aware (until we told them). “Population Research UK” (another HDR sockpuppet) seems to have got them to do it anyway. If the meetings of that have started, has anyone from ONS noticed that everyone they’re talking to is basically Andrew Morris in a disguise?
On page 123 the Review discloses what the Database is, and at the review launch, National Statistician Ian Diamond disclosed ONS had already built the “demographic index” and are using it as the link between tax and health records in order to do the analysis of which care to prioritise and which to deprioritise (ie cut).
The inclusion of justice data is particularly troubling – given it is a database of the victims of crime, not just perpetrators, mostly because HDR/Sudlow didn’t bother to think of those in any database as people with rights or concerns (or think of them as people at all).
The National Statistician has a choice
Speaking in 2021, Sir Ian Diamond, National Statistician said:
“There’s no god given right for us to have data. There needs to be a really sound public good reason for collecting data, and using data, and people need to feel absolutely comfortable that their data are being used properly and kept securely and in a way that satisfies all forms of privacy”
Does the National Statistician still believe that? If someone isn’t comfortable with that, for demonstrably good reasons, does the National Statistician want to use their data anyway? How does someone opt out?
At the HDR/Sudlow Review launch, he repeatedly talked about data being “used properly”.
Does “used properly” include the economic analyses that DH/E have announced being done on the data of people who have opted out of their data being used for purposes beyond direct care?
Does “used properly” include using the tax records of people who are given no choice about that at all?
As currently built, the ONS Population Index includes the NHS identifiers of everyone, but could easily have a second field which is only the NHS identifiers of people who have not got a national data opt out, and it is that second field which is used for research purposes. Until DSIT gives a “rest of government opt out” the “NHS national data opt out” is available, especially for NHS uses. The existence and scope for abuse of the Population Index is an entirely different problem (noting also that the spooks want access to everything any entity in the UK uses for anything). What happens when Home Office staff walk across the corridor from their open plan office into the ONS open plan office in the same building and want special access to data?
Some of the other simpler issues can resolve themselves – transparency will come to ONS either voluntarily or from outside. ONS can request a second dataset to which the National Data Opt Out has been respected, which is used for all research proposals (as AGD recommended); or as the HDR/Sudlow Review suggested for Biobank and the Department of Health in England, do promises only apply when it’s easier to keep them than not? Will ONS treat the census the same way?
ONS claims some abstract sense of “public support” for what they do, but they want data this way because public support for their work, in the very real measure of response rates to their flagship survey, is only 20% and falling. ONS claims of public support do not stand up against their own response rates.
For those who have followed medConfidential’s team since the 2000s, the institutional funders who focussed on these topics under the previous labour government seem to have moved on to other areas. We remain grateful for individual donations.
The reasoning in our short note on business models for data from 2019 holds up pretty well. Policy was a mess then and it’s a mess now as nothing substantial has changed. The re-announcement of existing work provides an opportunity to look at what happened to the examples we cited in 2019. Then it was speculation, but it played out in ways that were predictable in 2019:
What failed:
Deepmind/RFH is a mostly-forgotten debacle buried in the vaults of denial at Google Health (which no longer exists either).
Sensyne went bust, and the NHS equity stakes got wiped out – “diluted to oblivion” as Lord Drayson put it (after he left) still insisting (Q24) his model worked (the company went bust).
Whatsapp/Facebook – continue as AI enshittification continues to play out.
Orth.AI has rebooted as “nAItive” with the son of a Oxford Professor still in charge, denying any wrongdoing/nepotism and not explaining who didn’t know what when about the deals which resulted in the company having a copy of hospital records it tried to train models on.
Dr Foster ended their deal with Imperial and went their own way having been eaten by Telstra.
“Rather than focusing on speculative business models, OLS should be attempting to deliver commodity pricing for all innovations, as fast as possible.”
“The primary measure of success should be net cost to the entire NHS and Social Care, rather than to any individual budgetary silo.”
“All uses of data controlled by NHS bodies should be made available to the patients involved, via NHS.UK”
Successes listed have come from rapid competition to allow a diverse and commoditized supply chain. Where monopoly or profiteering were supported, temporary success was temporary and then they failed.
In the US, “Cost Plus Drugs” is the radical pharmacy undermining US problems of price gouging, but that generic-replica approach can be applied to algorithms and diagnosis tools using AI which may become available worldwide. When companies think they can make profits off a published algorithm, their profit is a negative elsewhere
The interesting case of the Deepmind/Moorfields project – 5 years on
The Deepmind/Moorfields project is the most interesting case study as it went both ways. DeepMind published their discoveries openly and freely in Nature, to wide acclaim and understanding, a model repeated in other projects that got recognition.The publication of that research means that it was available for anyone else to copy and build upon at will – the substantive knowledge was public and free for all.
Today, if you walk into an opticians, they’ll have new and better machines that can have incorporated that research to do more and tell if your sight is at risk and help you take early preventative steps, or emergency measures, to avoid going blind. There are people walking around today who would be blind without those innovations published freely and openly by DeepMind/Moorfields. This is an unalloyed good for which the benefits can be calculated (but as far as we are aware, this hasn’t been done).
However, to someone sitting in an office in Moorfields, they are disappointed that none of that benefit to people’s sight is attributed to Moorfields. Moorfields themselves have spent the intervening years trying to figure out how to charge exactly the right amount of money across different NHS balance sheets to satisfy DHSC guidance, and those other trusts have been careful to make sure they didn’t get charged too much, so nothing happened. People can see, but Moorfields (and Wes Streeting’s reannounced approach) want some cast to drop into a Moofields bucket every single time. Moorfields spent the time since 2018 arguing with others about how much that should be, and got precisely nowhere. How much is it worth to save your sight? Are you willing to pay that much to save someone else’s sight? Moorfields can’t agree with anyone else what those two numbers should be. The benefits to the public purse are massive, but Moorfields entirely disregard that and look only at their own accounting. The good thing is that the research was open and so no one had to pay Moorfields anything and people’s sight got saved anyway. That’s how progress should work.
The DeepMind/Moorfields project is how benefits can be demonstrated and realised, and simultaneously demonstrates why all too often they aren’t.
“Pharma Bro” 5 years on
We illustrated page 1 of our 2019 piece with the classic picture of the “Pharma Bro” who made money hiking the costs of non-generic drugs simply because he could, and then went to jail for fraud. He’s now out of jail but remains banned for life from the pharmaceutical industry (at the time of writing – he may get that restriction quashed by Trump2).
Seeing that as a constraint, he pays for AI compute time to build models of novel pharmaceuticals and releases them onto the internet for free, undermining the industry he’s banned from. If the drugs work in testing, there is prior art to undermine a profiteering patent (in his theory, in practice, who knows?)…
The same approach can be taken to get commodity pricing for algorithms. If one bit of the NHS thinks it will make money from selling what is publicly available, there should be a small prize fund to reward bored PhD students who replicate it and give it for free to all the other NHS Trusts and healthcare organisations for them to test internally and use internally. The US model is how to scale that.
In the NHS, any money one organisation makes from selling innovations will largely come from the budgets of other NHS entities and so the “innovation” income lauded by Streeting is simply taking money out of other NHS budgets. It is not quite a zero sum game, it is worse than that as overheads (tax, transactions, admin, interest, VC payback, etc) will eat some at every stage. If you get £1 profit into an NHS budget, it’s probably £3 or more cost to later NHS budgets.
2019+5: Still “Between Goat Rodeo and Black Elephant”
Five years after our original draft, there are a few new examples, but the thesis from then held. While the examples of barriers and profiteering lauded in 2019 have largely collapsed, as those who profited from the collapse remain in denial (Q24, theranos). Those who offered commodity pricing all survive.
The Office of Life Sciences and Department of Health in England thinking on business models continues to bounce “Between Goat Rodeo and Black Elephant” based on the interests of the day, while DeepMind/Moorfields deserve recognition (and calculation) of how many people’s sight was saved as a result. Just because it doesn’t appear on an NHS balance sheet doesn’t mean it wasn’t a benefit.
Done responsibly, correctly, without profiteering, diagnostic algorithms may be one day seen like vaccines, which are currently the only “prevention” that is so effective many forget what has been successfully prevented by innovation.
As Labour discovered, the Department of Health in England will conjure reasons to reject something they don’t want to do by only speaking to those who will agree with them.
Buried in the National Data Guardian annual report is the outcome of some work by the Department of Health in England on whether to tell users when their GP records accessed for direct care from hospitals and other care settings – they’ve decided not to do it. DH/E never talked to us about this work on a topic we’ve been working on for the last decade.
If you listened to Wes Streeting’s speeches, you’d be forgiven for hearing that anyone reading your GP record from outside your GP was impossible. In practice, it’s routine.
Summary Care Records, Shared Care Records, and GP Connect all already exist, and are used to help people most of the time, but are also abused by creepy single doctors to look up the records of women they want to go on dates with, or used by stalkers to read what their victims told their doctors about their fears and health conditions – that last link being the first time we’ve seen a disciplinary hearing cover these topics. The doctor was struck off.
Item 6.7.4 of the NDG annual report tells you why the Department of Health in England chose to do nothing – it shows DH/E looked at telling users when/where their GP records are accessed for direct care today, and decided protecting patients from creeps employed in the NHS is too “technically and legally” complex. What that means in practice is DH/E would have to cooperate with GPs to show patients where/when DH/E had facilitated abusive access to GP records, and DH/E has decided it doesn’t want to. DH/E has legal responsibility for those abusive access, and has decided that the best way to behave is to keep secret from you the evidence of how your record has been abused, so you can’t complain because you don’t know, and GPs can’t hold DH/E to their agreement because they don’t know. The NDG says her “observations” are that it’s humans using systems in ways they can and which no patient can easily detect.
You can carry on reading into section 6.8 and mentally substitute the various NHS bodies/roles with “Met Police” equivalents from the Sarah Everard case, or the many many other cases where institution denial reigned supreme and innocent women paid the price. How many ghouls have looked up the GP records of victims of crime? No one will ever know because, like the police (until recently?), DH/E has sided with the perpetrators they employ against the victims.
The Department of Health in England is aware of the benefits (to them) of sustaining that ignorance, as item 6.8 of the NDG annual report says the view of DH/E and the NDG on how to resolve it is that: “the public need to be assured that deterrents and sanctions against improper use are meaningful and effective to deter such abuses occurring” (but when they do occur they’ll be covered up so occurrences can be dismissed as rare – which makes the defence rather pathetic). Most victims do not have the evidence to make a legal complaint, and without prior police involvement that evidence will not be made available to them.
The Department of Health in England insists on marking its own homework on access by creeps, but it has so little confidence in its own efforts it will never tell you the truth about the results. As NDG says “bad actors can and do significantly undermine public trust”, but it is the facilitation of the coverup that is the systemic flaw that undermines all honest police officers NHS staff. Individual bad actors will always try to undermine trust, but when guardians institutionalise a conspiracy of silence over bad actions, then the bad actors are seen to embody bad institutions.
In other entirely expected acts of duplicity and secrecy, while NHS England has previously said it would publish the Data Protection Impact Assessments for all parts of the Palantir procurement and all the “use cases” in the Federated Data Platform, they haven’t. The “Privacy Enhancing Technologies” contract has no published DPIA, and the use cases are all being withheld so the Department of Health in England doesn’t have to explain why what they do is different to what they said they would do – coverup is the norm.
DH/E sides with creeps and abusers because that gives them a quiet life, not hearing from anyone that might challenge their decisions to do nothing. (That culture repeats with the proposed reuse of “pandemic only” data governance for non-pandemic uses merely because DH/E don’t want to have the short conversation about doing it properly when they think they can just do whatever they dictate instead).
The new government’s “vision” for their time in Government is to take ownership of all your medical records, including all your written notes, make them all available to creepy single doctors anywhere the NHS logo is seen, to feed them to AIs, and to sell them for economic growth. Inspired by the chatGPT output of the Blair Institute, Wes Streeting’s position is you’ll have no choices in any part of that, and they might even keep secret from you whether it’s happening.
Merry Christmas from medConfidential and best wishes for 2025. We’ll be here.
Update January 2025: Government has now answered a written question on this topic which suggests that if you want to know who has accessed your GP records, NHS England and NDG believe you should do a Subject Access Request to NHS England (for Summary Care Records), to your GP (for some of GP Connect), your local ICB (for your Shared Care Record) and every hospital and care provider in the country who may have creeped on your records. It seems both the Department of Health in England and NDG have sided with creeps against victims. The NDG annual report says “trust can only be maintained if health and care organisations, professional regulators, the ICO, and potentially the police respond seriously and comprehensively to breaches of confidentiality by staff” which are fine words unfortunately disconnected to the actions of those bodies.
Many of the cases say doctors didn’t know about a condition that would be clearly recorded in the Summary Care Record, or Shared Care Record, or in GP Connect, or any of the data systems designed to share data around the NHS. These existing systems are already there. But the response of DH is to say a new system might be ready by the mid-2030s – which is entirely disconnected from preventing deaths.
The list of PFDs is very long, there were 11 new ones in one day last week, so the lists you see are cherry-picked as well as being terrible experiences. Because someone died, these are the cases that get investigated. How many other cases are there where no one died so no coroner looked?
If missing access to records were really the life-saving priority claimed, a 3 months away solution is to tell all patients where/when their record was accessed via the national systems. So all the routine cases of non-access, the many more that don’t end with a coroner, can see when their record was not accessed when they believe it should have been. The living can ask why the frontline didn’t check these systems; the dead need coroners.
Your case shouldn’t have to be in front of a coroner to know whether your records weren’t read – by then it’s too late for you. We focus on scope for abuse of Wes’s plans, seeing where and when your records were accessed by creeps and stalkers, but it will be far more common to see that you interacted with the health system and your records from elsewhere were never checked.
If NHS England continues to refuse to tell patients – as they have refused for a decade, perhaps the GP Profession can have their IT providers do it instead, with a monthly audit automatically going into your documents via prospective access.
A central care record is not a necessary prerequisite to preventing future deaths, but it does offer an excuse for the Department of Health in England to do nothing for another decade, while the PFD reports keep rolling in.
There is vast scope to use digital to do good – progress by HMRC since our 2022 paper on the paperwork of new parenthood has removed some of the barriers we outlined (some remain), but DWP still prosecutes people for the consequences of DWP’s own service design failings, and other parts of the state prosecute for information not disclosed to citizens at all. DWP now wants access to medical records to measure and justify their own policy positions (item 6.3.4).
The “tell us once” service has long been constrained by shifting political priorities. Carers allowance was redesigned by DWP but the new processes didn’t account for consequences that could lead to prosecution. Did anyone go to prison because a 2011-2015 flagship changed only the easy parts of a service? Service design can ignore edge cases or complexity in order to meet a HMT business case or a Ministerial announcement that will move fast and break people. So a summary of our questions:
Which other parts of partly digitised public services are prosecuting honest and blameless citizens for institutional failures of service design?
Who in the hierarchy should bear responsibility for the failures that resulted in honest people receiving criminal records? Does the board feel no one should?
What happens when a “tell us once” style service conflicts with the primary legislation?
In northern Rail or elsewhere, a lawyer will have said the self-serving change was fine as it satisfied some “user need” – and like others we deal with, the consequences on people were never considered simply because the action is in line with “policy”. “🎼That’s not my department says Werner Von Braun🎶”
Data and UKSA Statistics Assembly
There are few limits on data analyses in Government, and fewer limits on what public services can do with digital dark patterns. Some civil servants believe they deserve more attention than citizens, and civil servants can be sent to endless meetings: the suggestion of “User Needs*” aren’t enough, promises have to be kept if they are to mean something. One of the first actions of this government was to tear up promises made about pandemic data.
Against this backdrop, the UK Statistics Authority are running a statistics assembly and asking for submissions about “user needs” (no asterisk). We said:
While ONS and the statistical system assume that statistical data for policy making and raw data being made for decisions are entirely separate functions, and they are in government, to the citizen the effect is indistinguishable. The research paper advocating cuts in benefits can have a direct impact on their reduction, even if authors hide behind the comfort blanket of “policy” and “research”.
DSIT’s role model
Government often assumes that its data is perfect and accurate, treating it as an “official truth” regardless of actual reality. In this context, data serves only to reinforce institutional fictions, perpetuating a narrative among civil servants. The consequences can be severe: if you’re just £1 over or under a limit you are a criminal. This mentality has its roots in the Home Office’s hostile environment towards migrants and continues to spread to other areas of citizen interaction with government.
Data isn’t enough – an essential component of the system must be digital services and digital service design. Governments prosecute people for what they type into forms and apps, yet changes to those forms can occur as unpredictably as a developer’s whim on any given day.
“Imitate Tesco” might be the personalisation vision from some within DSIT, but when the Tesco app decides it should recommend something to you, there’s no way to tell it not to – if the reason you regularly buy a product leaves your life, Tesco will keep reminding you they are gone indefinitely, with no way to tell it you no longer need mandatory and intrusive reminders to buy senior cat food, or baby food, or your ex-partners favourite treat. Your only option is to shop elsewhere. Various silos in DSIT see that as their role model.
Tesco will share data with Governments if required or encouraged (or Tesco gets a better deal on something as a result), and citizens have no choice. If you shop at Tesco, having a Clubcard is decreasingly a “choice” because of the price differences. Clubcard culture was brought into Government by the CDEI (as was – while everyone welcomes Responsible Technology Adoption, it is the Irresponsible Technology Adoption that causes the problem; and all adopters believe, of course, that they are ‘being responsible’).
The same way more facts turbocharge racism, smarter data risks turbocharging institutional intransigence.
Institutions extend “counter fraud” activities irrespective of outcome. Failure does not deter as more intrusive searches are justified under the guise that fraud hasn’t been found; and finding fraud also justifies expanding the powers. Ultimately, such activities can become overly focused on speculative pursuits rather than concrete results, devolving to little more than ghost hunting and unicorn farming.
When a hospital makes a decision that it’s not in the public interest for the hospital to try to recoup costs from someone nominally “chargeable” for that care, then the hospital will tell her, but the Department of Health in England wants to copy the data so it can later make a decision to reclaim the money anyway, even if the patient was told that they were not liable for it. The first that anyone would hear about it might be when the Home Office rejects them at a border decision because the data goes from one bit of Government to the other without any knowledge of the patient – because the Department of Health in England doesn’t have an easy way to talk to individual patients to talk about charging.
That’s the sharp end of digital and data in government, but no one sees it as their responsibility.
ONS say they want to demonstrate good practice, but instead the Integrated Data Service perpetuates the secrecy by hiding how data is used from those to whom ONS at one time felt they were accountable: the public. Parliament suggested they change it (paras 100/102), the new government may well decide to double down on secrecy.
[to come in November] Response to the unpublished HDR/Sudlow Review (chatter suggests it will argue in favour of the lead author’s prior job of getting more data for biobank to give to racists, and the desire of HDRUK’s various sockpuppets to misuse ”smarter data”).
Population health management will be the focus of the “neighbourhood health service” strategy [announced tomorrow as we publish this]
These different strands will start to overlap and merge over time – fitbit data provided to care for people will be copied to the Department of Health in England who’ll give a copy to DWP so DWP can design policy and future sanctions regimes.
Pregnant women will be required to share data with the app if they want NHS care, and that data will then be sold on. In many cases the “choice” in Wes Streeting’s NHS will become “your data or your life”.
Government believes that any data that is available anywhere should be available everywhere, reflecting the turbocharging of the database state using techniques unimaginable when labour last left power, and they will leave a powerful set of tools for the next government. Wes Streeting wants to believe he’ll be Secretary of State forever, neglecting both his own desire for promotion and forgetting that all governments end.
