medConfidential’s GP data grab Situation Reports are a series of updates sent to stakeholders; this one is public.

The long delay was inevitable

The announcement of the short delay in June to 1st September was largely due to NHSx and DHSC thinking they understood their mistakes; as the GPDPR Data Provision Notice has now been withdrawn, and any new DPN will have process to go through, GP data collection can now begin no earlier than the 2nd September.

The next announcement, of a longer delay, will mark the inevitable realisation of the magnitude of these past mistakes – a delay already referred to by the former Secretary of State in his last speech at the despatch box, where he said:

It will take some time to move over to the new system, hence I have delayed its introduction, but we have also made that delay to ensure that more people can hear about it.

Both the Secretary of State and David Davis MP also entirely agreed in that debate on the risks of dissemination. It is therefore clear that the (very welcome) commitments on the use of Trusted Research Environments must apply to hospital data, e.g. HES, as well as GP data.

This realisation may yet come slowly. On HES, it may take a legal opinion quoting the Secretary of State’s speech, next to the ICO’s guidance on UK GDPR and DPA 2018, next to current DHSC policy that requires NHS Digital to disseminate the sensitive, identifiable personal data of every hospital patient in England – even if they have dissented – thousands of times a month. 

We understand it will be difficult to decide today, that from tomorrow HES is identifiable special category personal data, when the data was disseminated yesterday (and for years before).

The best time to have complied with the UK’s 2018 Data Protection Act was in May 2018; the second best time is now.

Sequencing of Events

While the delay was announced so the Trusted Research Environment (TRE) could be built to the satisfaction of research, there is now time to do everything in the right order. Hopefully.

NHSx may have gotten to choose the starting point but, as the Health and Care Bill demonstrates, it missed the boat. The headline focus of the Bill, Clause 1, formally re-names NHS England, but nowhere in the Bill does NHS Digital get a re-name. Perhaps DHSC expects to use its new powers to abolish NHS Digital – thereby abolishing the statutory safe haven? That is untenable.

There is, however, still time for the proposed legislation to be amended to resolve some critical data trust issues. The Bill should, for example, have a hook to put the National Data Opt-out onto a statutory footing – so patients can know and have confidence in what the rules are, so the profession all know what the rules are, and so the various national bodies know what the rules are – and so that everyone knows how those rules can be changed (in either direction) in future.

As the use of GP data evolves, there should be discussion as to whether the National Data Opt-out (NDOO) should apply to data leaving GP systems and going to NHS Digital, or not. If the conclusion is that it will not, then the Type 1 GP data opt-out must live on. If the NDOO were to be clarified in legislation to have the same effect as the current GP opt-outs, then Type 1s could effectively be deprecated for all but the most critical concerns – for a statutory opt-out is much better than a non-statutory one.

Hospital data

All of the examples given in David Davis MP’s adjournment debate were to do with hospital data, and the Secretary of State agreed on the risks of disseminating patients’ identifiable GP data, explicitly stating his intent that “The dangers that come with the dissemination of pseudonymised data are removed.”

So why is NHS hospital data not also being made ‘TRE-only’ from summer 2021 onwards? 

If NHS Digital and NHSEx wish to demonstrate to the GP profession (and to patients and the public at large) that the TRE-only approach will work, the most straightforward way to do so would be to show it working for the hospital data NHS Digital already collects – with a variety of researchers and, say, NHS England’s ‘Data Services for Commissioners’ Regional Offices (DSCROs) demonstrating good use of it. 

Such a transition should also make the DSCROs and other ‘DHSC / NHS family’ users far happier, as they will be getting both a much better data analysis environment for their ongoing work, while increasing safety as well. 

As the Health and Care Bill puts obligations on Integrated Care Systems to ‘use more data’, such patient-level data usage should also all be in formally NHS-accredited Trusted Research Environments – initially NHS Digital’s, also ONS’s or Genomics England’s. (‘Five Safes’ TREs are entirely achievable, but some will claim they meet the standard when they do not. Hence the need for formal, likely mutual, accreditation; trust in all being dependent on the weakest link in the chain.)

Communications

The need to communicate directly to the entire public actually makes other problems easier to resolve; with the data opt-out definitions written down in legislation, what is left for debate (as was the case in 2014) is exactly what text will fit on two sides of A4 – the text for the opt-out / opt-back-in form being derived from the legislation itself. 

This process could start with the last consensus draft of the care.data Advisory Group letter because, as a public advocate of the programme said, GPDPR is care.data.

NHS Digital, NHSx, and the new power to amend legislation

While NHSx may choose how many (NHSx-liveried, crowd pleasing…) elephants are in the ‘tech vision’ parade, it continues to be NHS Digital that has to follow it around with a shovel. And whoever holds the shovel will forever be in tension with those who want more elephants.

Many of the persistent problems around data are the result of such tensions, not necessarily the organisation itself that is making a decision. The same criticisms of NHS Digital would apply to the cancer registry, which learnt the hard way that giving data to a “causes of cancer study” is not such a good idea when the study is run by a tobacco company.

Someone has to enforce the rules that DHSC advertises as “strict”; that is currently NHS Digital.

It is not NHS Digital that decides what data uses there could be – it responds largely to requests. Sometimes it recognises that a request is valid but that an analysis would be better done by someone else. (A “causes of cancer study” is not inherently a bad thing.) But, as a result, NHS Digital gets a reputation for saying no to people – mostly because few notice the thousands of data file releases it does make every month.

It is, and should be, the job of a statutory safe haven to have a deep understanding of what is possible, what is legal, and of the necessity of keeping promises to patients. (Keeping promises not being a recognised strength of this Government.)

Any body fulfilling the role of safe haven must be transparent about where data goes. NHSEx have been actively dishonest in that regard, and – even if that was initially a mistake – have then explicitly refused to correct the record, and have repeated the dishonesty.

Differing interests may not like individual decisions that NHS Digital takes, medConfidential included – but what must be recognised and emulated is that it tells the public what those decisions are and why they were made, and people can know what we don’t know.

With DHSC and NHSEx, however, the cronyism and corruption of the Government’s approvals processes means not only is there no picture of what we don’t know, there appears to be an explicit desire to make sure no-one knows. 

Perhaps we are being unfair on the ‘organisation’ behind the first version of the NHS COVID-19 app that barely made it to pilot stage; the group which pushed GPDPR forward against expert advice, and which vetoed suggested improvements of GPDPR before it collapsed; the outfit that misleads stakeholders on what it publishes; which simultaneously added domestic vaccine passports for users of the NHS app, and which (still) expects NHS patients to hand an unlocked smartphone to the border guards of a hostile nation,  but we believe one’s actions speak for themselves.

(Of course, the person who signed the GPDPR Direction got promoted shortly thereafter. When NHSx is abolished, NHSx policy functions should really revert to DHSC – not because any particular incumbent has any particular talents, more because officials always move on.)

If NHSx – or any actual NHS bodies, for that matter – wish to be seen as trusted, they must show themselves to be trustworthy. Downgrading the statutory safe haven and/or transferring its statutory powers without reference to Parliament is unlikely to help in this regard.

New Secretary of State and Life Sciences 

To push our earlier analogy, some of Sajid Javid’s team will be retracing the path of Matt Hancock’s elephant – with a shovel to clean up those emissions that still litter the building.

And those whose ‘Vision’ is less rose-tinted will recognise the “alignment” claimed in the restated Life Sciences strategy was prompted more by the overwhelming necessity of combating a common enemy than any real change in institutional politics or public attitudes. It is notable also that the stakeholders on whose data the Vision depends, the public, get short shrift in a document whose focus is to “deepen collaboration and trust between Government, the NHS, and the [Life Sciences] Sector”.

Aspirational statements for “the full support of patients, the public and NHS, and must build trust into [the Vision’s] delivery” are hard to square with the far more clearly-defined intent that: “governance and oversight of NHS health data must be simplified to drive research and innovation”.

We welcome the commitment to consensual, safe, and transparent data infrastructure for a 21st century health and care system; as we have been saying for years, a modern TRE for research and all other secondary uses is inevitable. The best time to have started was in 2013; the second best time is now.

Available next steps:

• Available next steps to safely make GP data available for research and planning 
• Opt out review note (June) and available next steps on opt outs (July)
• Briefing on the Health and Care Bill

The countdown has already begun. The Government’s plan is to copy your entire GP medical history – including all the most sensitive parts – and make it available for third parties to apply for and buy access. Even though Matt Hancock Directed it to happen he’s not going to tell anyone about it. Neither will Boris. 

Details are still a bit sketchy; critical documents like the programme’s Data Protection Impact Assessment aren’t written published yet, and some of wha’s being said to patients seems… contradictory. The promise that “you can opt out at any time”, for example, doesn’t fit with the fact that once your data has been copied, it will never be deleted. 

We will provide more information as it emerges, but for now…

If you’re wondering why you haven’t heard about this, it’s because they haven’t told you! The Government are playing the odds that you, your family and your friends and colleagues won’t have noticed some information they buried on a website, or the handful of tweets they’ve made.

While you can opt out, they’ve made it deliberately confusing and difficult. Unlike the single form medConfidential provided at the beginning of the previous attempted GP data grab in 2014, you must now use several:

• The most important one is the ‘Type 1’ opt-out form – this is the only opt-out that will prevent your GP data being copied to NHS Digital, and then onwards. If you haven’t done so already, you need to fill one in and send it to your GP practice within the next six weeks. (If you opt out after 30 June, your GP data could be copied from your practice at any point and, once copied, it will never be deleted.)

• If you opted out from care.data in 2014, your opt-out will still be valid. Your own GP practice may still be using a form from that period, but it is completely different from the ‘National Data Opt-out’, which used to be called a ‘Type 2’ opt-out. Bottom line, if you opted out of care.data in 2014, you should be OK for now.

• The National Data Opt-out, introduced in 2018, limits NHS Digital from selling access to some of your data in some circumstances. (They still sell it 85% of the time.) This opt-out process is supposedly ‘digital first’ but in 2021, for people and families with dependents, NHS Digital’s process still involved multiple PDF forms – which we’ve combined into one. And because the process is so overly complicated, we’ve created a page to help guide people through it: https://medconfidential.org/2021/children/ 

If you don’t opt out before 30 June 2021, the Government will take a copy of everything medical that ever made it into your GP record, throughout your whole life – apart from some limited aspects of information around gender recognition or IVF treatment (if you have received any).

The first upload will be of your entire GP medical history to date – which will happen as soon as your next GP appointment, possibly even before that – and then there will be daily updates thereafter, to copy every new thing that is recorded about you.

There will doubtless be much more to come, but the headline is this: 

The Government intends to take YOUR entire GP history, and isn’t even planning on TELLING you that you have a CHOICE, i.e. by writing you a letter.

This time they’re not even sending out a junk mail leaflet, but they might do some tweets and social media ads. (As we write this, their YouTube video appears to have fewer than 260 views.) 

To summarise, the process to dissent fully from the copying and then further use of information from your GP record for purposes beyond your direct care is as follows:

• Give a Type-1 form to your GP, for your whole family’s GP records;
• Do the online National Data Opt-out (NDOP) for yourself;
• Download, fill in and e-mail the multi-page NDOP form for your children.

There should be better options, but this Government apparently doesn’t want to listen.

medConfidential has been fighting to ensure every use of patients’ data is consensual, safe, and transparent since 2013. We aren’t there yet, but there’s every reason to believe that if enough people take action, we will get the protections you and your family deserve.

This Government sees so little value in any form of protest that it is trying to ban it through legislation. And its Ministers’ (and senior officials’) view of profit seems to be that ‘any means are acceptable’. Indeed, in a global shortage in the throes of a pandemic, politicians picked their friends to profit off the NHS to provide PPE for nurses. 

Why would anyone believe they wouldn’t seek to profit in exactly the same way from your health data?

If you believe they can be persuaded to change course, or if you simply want to be kept informed, please join our mailing list for more news. And don’t forget to forward this e-mail to your friends and loved ones, who may wish to make their own choice before the end of June. 

Our thanks to those who have donated, especially monthly, over the last couple of years – we kept your money in a pot for times like this. We’re currently using it to pay for sending out forms to those who don’t have printers, and for some other things that we might ask you for help with, next time.

---

In order to register dissent for your children’s medical records to be used for purposes beyond their direct care (and why):

1) Protect your GP data: fill in and give this ‘Type 1’ form to your GP practice [PDF] [or MS Word version] – this form allows you to include details for your children and dependants as well. This is the most important step; the Type 1 opt-out is the only opt-out that will stop NHS Digital extracting your GP data.

2)  If you want to stop your non-GP data, such as hospital or clinic treatments, being used/sold for purposes other than your direct care (e.g. for “research and planning“) you must use this process: 

• • If you have children under 13, you need to fill in this form [PDF] and e-mail or post it back to NHS Digital – this form works for both you and your children.

• • If it’s just for yourself, use NHS Digital’s online ‘National Data Opt-out’ process – this only works for individuals aged 13 and over.

• • If you have an adult dependant for whom you have legal responsibility, you must use this form [PDF] and send it back to NHS Digital on their behalf

If you don’t have a printer

If you don’t have a printer, and can’t fill in the electronic forms above, you can e-mail children@medConfidential.org with your postal address and how many people you need forms for, and we will post you copies of the GP paper forms, for free, no questions asked (also tell us if you have children under 13, or the online hospital data service hasn’t worked for you, and so you need the hospital data form as well).

We will, of course, only use your details to send you the forms you want and we will delete them as soon as we have done that. (medConfidential is registered with the ICO to process personal data in this way.) If you can afford to make a small donation to support us in offering this service to others, we have a donation page where others have donated so we can send the forms to you for free.

---

Why does DHSC put you (and other families) through all this hassle?

The pandemic changed many things, but until 23 April 2021 and for much of the last year, the forms to express dissent for your children and other dependents were missing from the NHS Digital website. Instead, for up to a year it said:

The explicit decision back in 2018 that there would not be a digital route for families with children came back to bite those who had to implement it.

Postal processing being temporarily unavailable might be considered understandable at the height of a pandemic, though it was clearly someone’s decision to remove the links from the website to prevent new processing.

This may not have been a concern in and of itself, but buried in NHS Digital’s Board papers is a statement that the Direction and Data Provision Notice for “GP Data for Planning and Research” was due to be published “to enable collection to commence (March 2021)”, according to the papers for NHS Digital’s Board meeting on 31st March 2021. That the new children’s opt out form did not appear until three weeks later suggests the discriminatory and inappropriate approach to dissent continues. 

Before 2016, when the dissent process had been considered primarily from a patient’s perspective, the way to opt out was to give one piece of paper to your GP, which covered your entire family – and then the NHS would deal with any complexity. It did, and it worked. Over 1.2 million people used that process, which was possibly more than some would like…

In 2018, NHS England and the Department of Health and Social Care made a series of choices about how the National Data Opt-out Process (NDOP) would work; the effect of each of those decisions made it harder for someone to express their wishes: 

• By cherry-picking who was invited to meetings, and on a narrow reading of the Data Protection Act 2018, one of those choices was that if you are over 13, you must do it yourself online. (The people to whom NHS England chose to listen at that point were those who believed your GP shouldn’t be an interface between you and ‘the NHS’.) 

• Another such choice was that the databases used to validate that you are who you say you are online were not to be used to check if you had children who lived at home and who were registered with the same GP.

• In fact, the decision was then made that there would be no digital process for children at all – parents’ and carers’ choices for any of their dependents would have to be done via a form sent in the post. First that form had 8 pages, then 7, now 4, and (finally!) the option to send forms by e-mail, after feedback about the punitive nature of the process.

As a result of all this complexity, the GP opt out form – which, prior to 2018, used to work for your entire family for all NHS records – still works for your entire family, but only for your GP records. And since NHS England and DHSC chose to create another process, you now have to do that too!

The process doesn’t have to be this complicated. Most families have children registered with the same GP as the parent, living in the same house, so the NHS identity checks for adults should cover their dependent children. There will always be exceptions, families with more complex situations – the current PM’s for example – which is why a paper form is a necessary backstop. But having no web process at all starts to look more like a procedural punishment for families.

The Secretary of State or NHS England could have said that the process for dependent children (and other dependents) should be the same as for adults. Instead they shifted the bureaucratic burden from the NHS onto patients and families, in the hope you would care less about your GP data than their cronies who wish to buy it.

To register dissent for your children’s medical records to be used for purposes beyond their direct care (and why):

1) For your and your dependants’ GP data, give this ‘Type 1‘ form to your GP.

2) For your children’s and your own hospital and other non-GP data, fill in and e-mail or post this ‘National Data Opt-out’ form to NHS Digital.

3) Not forgetting that for your own hospital and other non-GP data, individuals aged 13 or over can also opt out (or opt back in) online.

One thing the NHS bureaucracy likes more than anything is having the same acronym to mean two different but similar things. In addition to Summary Care Records (SCR, SuCR?), which have existed since 2007, NHSEx now adds ‘Shared Care Records’ from 2021.

As explained to the Public Accounts Committee on 17 September 2020 by Matthew Gould, CEO of NHSX, Shared Care Records (SCR, ShCR?) should:

“…allow patient data to flow safely and appropriately between different care providers, not just in health but also in social care.”  – Question 46

It is not that NHS England / NHSx have access to data today that is necessarily the problem; it is what they will do with it tomorrow – and whether they will keep their promises (cf. Test & Trace DPIA, transparency court case, ‘contracts for cronies’, etc. , etc.) 

What is needed for Shared Care Records to work?

Patients’ and service users’ data flowing along their care pathway for the purpose of their direct care is a worthy and worthwhile goal, and one that medConfidential has supported since its inception.

If Shared Care Records are to be successful in practice, and not repeat the dead-ends of LHCRs, that success requires they must do an number of things:

1)  The Shared Care Record is claimed to be for direct care only, and a good Shared Care Record will indeed be for direct care. A bad Shared Care Record will be used for lots of other things that it does not tell you about. NHS direct care services are too often seen as a ‘Christmas tree’ off which to hang things; ShCR cannot be a means by which sensitive health records leak out of the NHS via a back door.

• What happens across the boundaries between administrative areas (ICS, ShCR, or other)? Do ShCRs facilitate care for those who live in one area but, e.g. visit A&E in another?

• Do (creepy single) doctors get to look up the records of women they’ve met on dates – or anyone in the country – without disclosure or recourse? Are meaningful measures in place for those who are affected to know that their records were accessed?

2)  Shared Care Records’ existence and purpose(s) must be properly communicated to the public, before they are used, (unlike LHCRs) which means:

• Clearly and publicly stating what they will and won’t do, and explaining the rights and choices people have;

• Writing to all those who will have ShCRs created – including service users – not forgetting or otherwise disadvantaging families with children, and not just to those who may have previously opted out of LHCR sharing and/or SuCR.

3)  Ensure Shared Care Records can be seen as trustworthy, on an ongoing basis. From the point they are introduced, a record of every access of a ShCR must be made available to the patient or service user via their new NHS Login, cf. Data Usage Reports.

• If a provider is capable of connecting to a person’s Shared Care Record, it must also be capable of recognising and respecting that person’s confidentiality and consent choices; if it cannot do both, it must not be permitted to do either. 

• There is no excuse to ‘retrofit’ later; GDPR requires that all data processing is lawful, fair and transparent – and the data flowing through Shared Care records is, by definition, identifiable individual-level special category personal data.

• Confidentiality and consent choices should be managed centrally, ensuring that system-wide rules and IG are applied consistently and effectively. If local areas / ICSs manage their own dissent processes and someone moves to another area, will they have to dissent again? How will anyone know?

4)  To the extent that any data contained within Shared Care Records is extracted, copied or otherwise processed (e.g. ‘anonymised’) for any secondary uses, this must be done either within the statutory Safe Haven (i.e. NHS Digital) or under its Information Governance processes, which confer (joint) data controllership. Anything less than this would be a retrograde step and, as the failings of consent and governance processes around individual LHCRs have demonstrated, will compromise public trust.

• Notably for social care data, the regulator (i.e. CQC) cannot also be the Safe Haven: the incentives would be completely perverse. We understand a reluctance to put a body named ‘NHS Digital’ in charge of (adult) social care data, and share concerns about the ‘medicalisation’ of social care. That challenge is, however, an improvement DHSC and the NHS have to learn if health and social care are ever to be properly integrated.

If Shared Care Records cannot meet these conditions then they will be unfit for purpose, and will have been commissioned badly at huge cost to the taxpayer and to the reputation of the NHS.

LHCRs largely failed; will the lessons be learned?

From NHS Data Day, October 2019

Commercial re-use of data

If there are to be secondary uses of data within the Shared Care Record, then plainly the National Data Opt-out must be made statutory, it must work properly for everyone (including families), and must be made readily available to all before any secondary uses go live.

• The deadline for system-wide implementation of NDOP has been extended repeatedly from the original compliance deadline of 31 March 2020, to 31 September 2021;

• Meanwhile, LHCRs and CCGs have struggled to interpret and in some cases properly apply correct and appropriate Information Governance for NDOP – a situation that cannot be permitted to continue beyond the pandemic.

Even if there were to be zero secondary uses of data flowing through the Shared Care Record (ShCR), there would still be the issue of Summary Care Record (SuCR) opt-outs. For if someone has objected to just a summary of their record being shared, how can it be assumed that they will accept the wider sharing of their ‘whole’ care record?

Where there is legislation, therefore, both the National Data Opt-out (NDOP) and a ShCR opt-out must be made statutory; and these must be made readily available to all, and must be respected by all across the whole health and social care system.

“Selling the benefits” is no longer enough if you are also selling the records.

Details documents

• Data flows of COVID-19
• Post-COVID Landscape
• Available next steps for a trial of a Post-COVID Summary Care Record
• Why a single cloud EPR cannot work
• What’s not in the The NHS Whitepaper, plus:
  • Specific note on the Direction powers in the white paper
  • Specific note on Available uses of the Testing infrastructures

As all of the future emerges, the graphics below from various NHS presentations show the thinking that went into them, and the direction of travel:

‘Shared Care Records’ for secondary use?

Data for direct care doesn’t stay for direct care. A series of slides over years show the embedded view of NHS bodies which all want data to flow beyond direct care to commercial companies, and for planning, policy and commissioning decisions.

• It all begins with a lifelong (“longitudinal”) record and “maximising the use of data”:

from: https://digital.nhs.uk/blog/transformation-blog/2019/so-what-is-a-local-health-and-care-record-anyway

• For the purposes of policy-making, planning, commissioning and near ‘real-time’ surveillance of individual-level patient data, explicitly intended long before COVID:

From: https://hscic.kahootz.com/connect.ti/PubNHSDDTSF/view?objectId=10508916 

• Of course, there is also research and commercial exploitation; discussed well before COVID…

From NHS Data Day, October 2019

• …and during the pandemic as well; here’s the Minister discussing legislation to align with Big Pharma interests:

From Baroness Blackwood’s roundtables with Pharma, June-July 2020

ALL of these interests (and many more, e.g. Big Tech, Big DNA…) will seek the data once it has been created.

Details documents

• Data flows of COVID-19
• Post-COVID Landscape
• Available next steps for a trial of a Post-COVID Summary Care Record
• Why a single cloud EPR cannot work
• What’s not in the The NHS Whitepaper, plus:
  • Specific note on the Direction powers in the white paper
  • Specific note on Available uses of the Testing infrastructures

[The 2020 update to our ongoing series on data usage reports (2014… 2021)]

The need for and consequences of data usage reporting is something medConfidential has worked through for a long time.

You have the right to know how data about you is used, but what does that look like in practice? We’ve mocked up a data usage report for the NHS, and the equivalent for Government – but what about the analyses that are run on any data? What should responsible data analysts be able to say (and prove) about the analyses they have run?

The new, eighth Caldicott Principle is “Inform patients and service users about how their confidential information is used”. In future work we will look at how this goes beyond existing legal requirements under the 2018 Data Protection Act, what Data Usage Reports (or Data Release Statements) should look like to the NHS in 2021, and what patients should see. For now, though, we want to take a look at the other end of the process.

Analyses, Analysts, and their readers

Public bodies (and indeed everyone) buying AI and Machine Learning products need to know what it is they are buying, and how it has been developed and tested. Ethically, they must be able to know the equivalent of “This was not tested on animals”, i.e. “No data subject was harmed in the making of this AI”.

We covered a lot of the procurement side of this in our recent work on AI, data and business models. But that raised a question: what is it that procurers should ask for when procuring data-driven products and services? And what does good look like – or, at a bare minimum, what does adequate look like?

At the most practical level, what should someone wanting to follow best practice actually do?

And just as importantly, who should do what?

In a world of the Five Safes, Trusted Research Environments (TREs) and openSAFELY, and as the role of independent third parties becomes increasingly viable, those who wish to follow more dangerous ‘legacy practices’ with data will be unable to provide and evidence equivalent assurances – and their offerings will therefore be at a significant disadvantage in the market.  

A trustworthy TRE records exactly what data was used in each analysis, and can report that back to its users and to those who read their analyses. Academic journals often require copies of data to be published alongside an academic paper, which is not possible for health data (and if someone were to make that mistake would be catastrophic), but this certificate could act as a sufficient proxy for confidence and reproducibility.

If you are running the data ‘in your own basement’, there’s no way for anyone to know what you did with it beyond simply trusting you. In health analyses and with health and care data, that isn’t enough – and it should certainly not be the basis for procurement decisions.

So, as before, we decided to mock something up.

Trusted Research Environments which facilitate transparent data assurance like this, and which automate the provision of evidence of compliance with the rules – Data Protection, Information Governance, Equality, or otherwise – will be offering advantages for their users over those which do not. And any TRE that does not report back to its users how its safety measures were used will clearly not be helping its users build confidence in the entire research process.

While they may claim to be “trusted”, organisations that fail to provide every project with an ‘Analysis and Input report’ cannot be seen as genuinely trustworthy.

[2021 blog post in the series]

Across Government data and digital is too often used solely to help civil servants to make decisions, rather than benefitting all stakeholders.

There is little sign this inequity will be addressed under current structures or priorities, but as Government thinking evolves around the structure of the new Information Commissioner, CDEI should also be fundamentally restructured so as to receive and consolidate a much wider range of inputs – including lay members (DHSC’s former National Information Board had six, for example). Without wide-ranging input, data in Government shall continue to make rookie mistakes such as those of the ONS / GDS Data Standards Authority.

There are alternatives to creating many ‘pools’ of data around Whitehall and simply hoping no-one makes a mistake. Built for the pandemic, and with appropriate governance within and beyond it, the model of openSAFELY could apply across the rest of Government – especially for monumental failures like the National Pupil Database at DfE.

While it is self-evident that the vision of the forthcoming National Health Data Strategy should be to maximise the health of patients within the NHS, the vision of a National Care Data Strategy is less clear. Is it only to maximise the health and health outcomes of those to whom care is provided, or do quality of care and quality of life have other dimensions? Whatever is decided, as the Health and Care systems move towards integration, those two goals must align – but it shows how far apart things are that to talk of (the state of) Health and Care data as if they are even remotely equivalent is quite clearly nonsense.

As the pandemic has brutally illustrated, there is no data strategy for social care – and no evident plan to move towards one. Given every journey must begin with a single step, something like this might work.

Health and Care ‘moving parts’

Whenever NHS legislation is next put to Parliament, the National Data Opt-out should be placed on a statutory footing. Aside from guaranteeing patient choice and underpinning trust, this will provide proper democratic scrutiny of official choices such as the one which the National Data Guardian highlighted in her recent annual report – page 11, right column – where NHS Digital, NHSX, and DHSC decided it wasn’t in their interests for patients to see how data about them is used. Should government attempt to defend that position, when the push-poll and focus group used to come up with it are more widely known, the u-turn will be more embarrassing than fixing it now. 

In a similar vein, transparency on access to patients’ details via APIs (whether new ones for COVID-19 or pre-existing ones, such as for the Summary Care Record) would also begin to address the ‘creepy single doctors’ problem that has been exacerbated by the widening of access in a time of reduced oversight. And that some in Government still wish to use patient records for funding and “decommissioning” decisions (para 2) is unlikely to be wise.

Government argues that new business models are the way the NHS and the Life Sciences Industrial Strategy will get them out of the hole they’ve created. Trillion-dollar tech fantasies abound. But while the conflicts of interest amongst advocates of this strategy are clear, whether it will work is not. 

National Data Strategy (outside of health)

The NDS is a “pro-growth” data strategy, which is an entirely appropriate mission for DCMS – but it creates a fundamental conflict of interest in its sponsorship of the ICO as regulator, and its role in choosing the replacement for the current Information Commissioner. For this if not other reasons as well, the ICO should move back to being Departmentally sponsored by the Ministry of Justice, to underline the fundamental importance of following the law and to ensure the principles of justice apply to all data use, as well as to quasi-judicial decision making by the regulator.

A data strategy for the UK should first and foremost respect the rights and freedoms of every data subject, and aim to provide the greatest net-benefit to the whole of the UK – yet there is no compelling vision in the strategy; no clarion call. There is also no testable hypothesis in the strategy, by which its success (or otherwise) can be known. It is likely no single vision acceptable to all stakeholders could have got through write-round – not least because the unreformed, institutionally-ignorant Home Office will not accept a nuance that is in the public’s interest (for example, PHE / Test&Trace / police data sharing).

As written, there is no explicit difference in the National Data Strategy between personal data and data about objects. Lacking specificity, much of the strategy that is intended for one could be used for the other, thereby creating effects entirely unintended by the authors. A recent misstep by the new “Data Standards Authority” illustrates the sort of harm that can be caused when ‘generic intent’ overrides substantive nuance.

After a summer tainted by “mutant algorithms” in education, nothing would say understanding data less than NHS England agreeing to run the COVID-19 vaccination database off a 66 million row Excel spreadsheet. (While a single worksheet can have a million rows, losing 65 million people should be relatively noticeable – plus they know to look… now.)

Enclosures:

• Dear Prime Minister
  • Do This First: Creating the Data and Information Commission 
• Social care 
• National Data Strategy response – top sheet
  • National Data Strategy detailed read through
  • The choices of the Data Standards Authority
• Offline harms 2
• PHE letter
• Net Assessment of pre-Covid analysis, which predates openSAFELY
  • OpenSAFELY in the rest of Government