I looked in two previous posts at the background theory [1] to commercial uses of our data, then, the background to my concerns of commercial use with data intermediaries. [2] This is now part three, my glimpse into commercial use in real-world practice. It’s become rather a saga.
Here’s the short version: “In general commercial uses of data, I am increasingly learning that if you don’t pay for the product, you are the product. We need to shout a bit louder, that we are not a product for sale. It’s not only that there is an increased risk in a move of our health records from binder to byte and broadening access to them. We take issue with the change of approved purposes from care, to commercial use.”
At the Health Select Committee on July 1st, [3] I believe Sir Manning misses the key issue the public has with care.data and health record sharing, when he gave a response to Q562 to David Tredinnick MP:
‘We made big mistakes over the last 10 years’
“I am saddened by some of the comments that have been made this afternoon about the lack of trust and also by the impugning of our motivation. […]
We made big mistakes over the last 10 years, and we have a once-in-a-generation chance to get it right. I am absolutely clear that we have to engage the public in an open debate about the balance of risks and benefits. There will always be risks with data. There were risks with the Lloyd George envelope; notes were lost, they flew and went all over the place. There will always be risks, but those risks and the benefits are both enhanced by the technology.”
Whilst I applaud Sir Manning’s apology, and his call for open debate, I think he misses here the fundamental point of disagreement the public has with the HSCIC current practice. Selling our health data.
It’s not only that there is an increased risk in a move from binder to byte and broadening their access. We take issue with the change of approved purposes from care, to commercial use.
And these commercial (ab)uses in current form must stop if we are to trust the governance system in future.
Health Records for Commercial sale
HSCIC currently sells our health records for commercial purposes, to intermediaries with commercial re-use licenses, and had no consent nor our permission for this in the past, it continues to do so in the present and appears to have no concern or intention to stop doing so, for the future.
Mr. Kelsey added at the HS Committee,
“We have a very big job to do, and I hope that you will hold us to account in delivering it.”
To which I can only reply, it is you who say it. But who is accountable? The Open Debate which Sir Manning calls for has not been taken up by NHS England. We are told this is a programme of national importance, one which Mr. Kelsey has repeatedly said, including to the Health Select Committee previously, on which the entire future of the NHS depends. Why then, no national discussion, no news since the pause and a focus on updated communications of the current plan. The current plan with flaws in consent collection, scope determination, confusion of purposes.
There are so many ways this could be improved and gotten right, but not by November and without public debate.
How can you insist a programme so vital for the entire future of the NHS yet encourage no public discussion? This seems to be a theme in NHS England recent programmes. [4] The decision to outsource the GP support services was taken in private sessions, not available to the public like the rest of the Board Meetings [5]. Other programmes, pilot and actual plans for implementation go on without public discussion.
There’s been no apology for the data sharing policy developed since 2010 which has encouraged commercial trading and enabled this erosion of security, confidentiality and trust in the data management system of our nation’s health records. No one at the Department of Health has said, we got this policy wrong. No one at NHS England, the same people if under a different label. Poor Sir Manning at the Information Centre who carried out their policy, has been left to say there were ‘big mistakes’ made. But not by him since July 2013.
Trust and care.data off course
That our trust now lies in tatters, is not the fault of the Health Select Committee member to whom Sir Manning says, he is saddened and disappointed. It’s not Joe Public’s fault who had no idea this was going on, until six months ago. Where did these policies and plans since 2010 come from? Where did the use of our data go so astray and why is flagship care.data now so terribly off course? Mr. Cameron outlined it in 2011. What happened in the three years?
Health records for sale
As I wrote in a previous post,
“Some of that data goes back into our health market as business intelligence, both for NHS and private use, for benchmarking, comparisons and making commercial decisions. In our commissioning based marketplace, this re-use of data is now becoming normalised.”
But should it be normal that our medical records are for sale?
When celebrity Michael Schumacher’s notes are for sale, [6] being offered concretely to the media, we all see that is wrong. Just imagine 70 million copies of Schumi’s record, each with our own name on it, being offered to anyone outside of those who need it for our care. Offered to these commercial for-profit data intermediaries. It’s not a theory – this is what is happening to our records, today. Don’t accept the ‘anonymised’ statements, they’re simply not true. Identifiable data and pseudonymous data has been sold. The register confirms it, and that was only a 10% sample.
“To earn the public’s trust in future, we must be able to show that our controls are meticulous, fool-proof and solid as a rock.”
I think banning data sharing for commercial use and re-use would be a good start.
What is it to be used for and why?
When we think of our health records being used by others, we need to separate the uses of the data, in order to understand different ways it is used, who uses it and why. Data once it is processed becomes knowledge which is used as Business Intelligence. It is common in discussion to conflate use in care with care.data. It’s even in the name. But the uses of care.data are secondary. Not to be used by clinicians caring for us, not replacing hospital notes to give to consultants when we are referred for a hospital stay. Not providing discharge papers. It’s only approved for commissioning and sketchily [imo] approved for risk stratification. [ref p.5 ] [8]
care.data extracted from GP surgeries, is not even approved for research purposes, but to read all the recent debates you’d think research depended on it. Research using GP extracted patient data, is not an approved use of care [dot] data. Research using GP extracted patient data is not an approved use of care [dot] data. Repeat, ad nauseaum.
What is already being done, and what is used legitimately i research such as public health (albeit without our past knowledge or consent), is with our hospital data, HES, SUS, Mental Health data, usually with CAG review, and through 251 approval sometimes through DAAG review at HSCIC – it is available and is on sale to all sorts of other non-care providers. And that is planned to continue.
The records extracted so far, when not used for research appear in recent years increasingly used for comparison, the concept of ‘ranking and spanking’ professionals and providers of healthcare. They are also used in commissioning, payment validation and understanding costs and spending. But beyond that, there are all sorts of others who still come under the umbrella of ‘health purposes’ but don’t directly benefit the NHS or individual patients. What is their demand and what are they being supplied?
In the newly created NHS marketplace, customers at individual level are patients, or at a market level they could be any part of the healthcare buying structure, a GP practice, a Clinical Commissioning Group, a Hospital Trust.
The challenge of any demand and supply chain process, is that you need a market willing to pay at the price you are prepared to sell. And you need to offer what they want to buy. For that, the buyers must see a value in the data they want to obtain. Where is the value for these areas of use: Generic NHS Business Intelligence, Generic Commercial Intelligence and Pharmaceutical intelligence?
Health records as Business Intelligence
Some companies take data and process it before selling it to NHS and other health providers in England. This provides a third party service and skill set which the HSCIC nor the NHS Trust for example, has themselves, such as IMS Health.
So business intelligence used for the benefit of the NHS, makes sense and is necessary to a greater or lesser degree depending on your attitudes to comparison websites, green/red flagging professionals and commissioning. Benchmarking was provided by Tribal until that part of their business was bought out by Capita.
These companies’ experience and market is healthcare. The kind of knowledge they can give to the NHS is highlighted in their case studies.
So for clinical care, and for commissioning at individual organisations, these tools are clearly useful and use individual patient level data. [9]
Al sorts of other places and individuals perform these services. They include a wide range of commercial organisations, small and large.
Health records as Commercial Marketing Intelligence
Commercial buyers however, can include wanting data for identity verification, fraud prevention and background checks. Services such as Experian offer. These may be what the loose definition in the Care Act would say are now banned, but are they? What is to say that a company which offers the use of private health services, healthy eating or pharmaceutical marketing is not providing information to others, for the promotion of health?
“Experian employs more than 12,500 people in 34 countries worldwide, supporting clients in more than 60 countries. Annual sales are $3.1 billion (£1.7bn/ v2.5bn).”
Identity verification can be done, matching data across a biographic footprint, ” in databases, established for 45 million UK citizens and hold in excess of 1 billion records.”
“Experian public sector currently works with 380 plus local authorities, 52 police and investigatory bodies, as well as central government agencies including DVLA, HMRC, DWP and the Cabinet Office.” [10]
There is clearly a lot of data sharing in the public sector, about which we may understand very little. But mostly the buyers of data want to sell something. Companies buy lists of people to use in marketing campaigns, who might be interested in what they’re selling — and companies also want to learn more about their current customers.
This is where I find the level of detail and what is done with our data, more than a little freaky.
Every UK consumer is classified into one of 22 types, aggregated into six groups. The 22 types are linked to six decision-making styles, providing insight into consumers’ motivations when using different media and the processes they go through in deciding about products and services.
“TrueTouch is built using over 700 individual data variables. These are chosen for their ability to illustrate an individual’s range of behaviours in relation to media consumption, including use of different channels, responsiveness and exposure to media. These are distilled into two core data sources: Quantitative data Experian’s UK Consumer Dynamics database compiles information on all UK individuals, their demographics and lifestyles, attitudes and responsiveness to media. It includes known data on demographics and lifestyles from publicly reported sources such as the ‘edited’ electoral roll, company directors, shareholders and council tax, as well as Experian’s proprietary lifestyle information taken from its programme of consumer survey.” [11]
I don’t know what segment I am in. But I know that I will have data stored in many of those different data sources they mention. So do they actually know more about my habits and inclination, that I have self-awareness? If their tool has over 850 million input sources which they process, it’s more than likely. 34 million email addresses, 20 million mobile phone numbers, 49.7m names and addresses.
Experian may well have much of this data from the electoral roll (unless like me, you opted out of these uses) but in the HSCIC January-April 2014 register of releases [7] data was given to Experian for use in Mosaic. (see July – 132kb right of page)
“Mosaic is Experian’s powerful cross-channel consumer classification designed to help you understand the demographics, lifestyles, preferences and behaviours of the UK adult population in extraordinary detail.” [12]
That they understand and track my behaviours probably better than I do, and at such detailed level, I find surprising and invasive. In fact, I find it threatening in a similar vein to the visceral reaction that the Facebook experiment generated this week online.
As SF Gate reported,
“Using unsuspecting members as human guinea pigs is repugnant. And when the biggest social network on the planet does it, can its leaders be trusted with their own technology?”
This idea that just because one can and the technology permits it, does not mean that one should. It just feels wrong to find out others may manipulate our thinking and behaviours in such a targeted way. Just as Experian does with consumer data:
“Within rural areas we are able to pick out the individual households that are likely to be commuting to towns and cities nearby…”[12]
Individual households? Understanding my behaviours, gives them information which they use to nudge or influence my decision making. Understanding our behaviour ‘in extraordinary detail’ helps companies market and sell more to customers.
There are other re-uses even for health purposes, which seem less transparent and more about us as general consumers, rather than for our health. For example, the use of HES data is in social marketing targeting:
“In this way, companies who process data such as Beacon Dodsworth received data in the last year and offered it for commercial exploitation by others “HES data may be used by pharmaceutical companies “to improve [their] social marketing / media awareness campaigns”. Others included OmegaSolver and Harvey Walsh.”
These companies have re-use licenses for data. what that means is better explained here by medconfidential. [14]
How will HSCIC know how data will be used after release and how will it be audited and how often? When it comes to human tissue, the HTA only audits tissue banks in the UK once every three years. That’s a long time in between audits if something has gone horribly wrong in best practice.
Health records as Commercial Pharmaceutical Intelligence
To global pharma it is again not the data itself which is of value, but in the knowledge it reveals. The pharma business intelligence. It can show at an individual level what is being prescribed or show any gaps it reveals, which will allow pharma, to address ‘unmet clinical need.’ The data already compares hospital prescribing and reports make recommendations used by NICE on what drugs to use and recommend. My concern is that to treat the worried well who have cash to spend, will deflect attention from the needs of the sick and poor and that even if only at postcode level, we will be targeted for pharmaceutical marketing.
“The parties will initially look at how anonymised, integrated health data can be used to identify unmet clinical need in patients with diabetes. In the UK, diabetes affects approximately 2.9 million adults overall, with more than 90% of these patients having type 2 diabetes. This makes diabetes one of the most common chronic medical conditions and represents a significant strain on U.K. health services.”
Astra has another Memorandum with IMS Health. So we, whose data it is, have zero transparency and can request no accountability for the use of our data once it has left the HSCIC.
And it matters because when there are data breaches in these companies, we should know whether our data has been involved.
In January 2012 AstraZeneca signed a three year partnership with IMS MOU [16] and stated it builds on AstraZeneca’s existing ‘real-world’ data and research partnership with HealthCore in the US, the health outcomes research subsidiary of WellPoint. Wellpoint which had a massive breach a year ago, July 2013. So how do we know where our data was stored, and if it were involved or not? Here is what pharma use data for, to analyse “unmet clinical need.”
“The partnership with IMS Health will give AstraZeneca access to pre-existing anonymised electronic health records, which include clinical outcome, economic and treatment pattern data. In addition, the companies will jointly develop a customised research and data analysis platform. The information will provide a deeper insight into how medicines that are already on the market are working in real-world settings across Europe, painting a picture of unmet needs …”
We can look at this more than one way. Some feel strongly commercial use should exclude Big Pharma. On the one hand, the State and Government does not own manufacturing of drugs nor medical products. Though we used to do both. Recently, that we did own, has been increasingly sold to commercial buyers or venture capitalists.
The State and pharma work together, often through University research, to create future health solutions, drugs and the drive towards personalised medicine and diagnostic tests. When companies which own our data are sold and bought internationally what happens to our data they own? Boots Alliance bought data from HSCIC, and they are about to be bought by US Walgreens. So many questions.
Those more informed than me will know all about the challenges of pharmaceutical companies, the patent cliff, mergers and diversification. IP, diagnostic tests and generics in the market. Big Pharma and the State are working together in much research to find solutions and discoveries to current and future medical issues.
How far does cooperation stretch and when does it become inappropriate? Is commercial interest supportive of State practice or driving decision making policy? Should commercial companies fund any costs at our NGOs? And do those which buy the most data, get a bigger slice of the influence of what conclusions reports using the data, reach? Whilst there is a public move to #Alltrials I believe we should demand #Allreports in the public interest as well. I would like to have transparency at HSCIC how their reports are funded, when working with partners which are frequently commercial pharma partnerships.
Mr. Hunt recently defended to the Health Select Committee the reasons why a commercially supported pharma lobbying group was used to advise on the NHS Commissioning plan – the Specialised Healthcare Alliance. Supported by 14 pharma companies, these corporate members are contributing £12,000 each towards the costs of the Alliance for 2014.
Are we really seeing transparency on who is driving change in our health service?
The Richness of our records open for Exploitation
The value of Big Data is only extracted by exploiting its richness. And these days, with mobile phones, social media and shopping habits tracked by the minute, the average citizen like me, it seems can’t easily avoid being part of it, whether we want to be or not.
But if we don’t even have the right to control and own our data and we can’t control the knowledge generated from it, how can we control who knows what about us and what they use it for? If we’re unaware of its existence, how can we understand its impact on our life to make free and uninfluenced choices in what we buy, for example? Or understand how we may be segmented and discriminated against. And this is aside from the assumption that the data held is accurate and that as a result, no mistaken judgements are being made about us.
As for our health data, how can we control its use by these massive data managers if we don’t even know who they are at the end of a chain of re-use licenses?
Put Business Intell, Commercial Intell and Pharma together
The vast amounts of data already held and analysed to the nth degree by these data intermediaries, means that making even more data available to them is going to increase the segmentation and risk of identification. They already have data on individuals and is it not enough that they make analysis at household level as shown by Mosaic? Individual health level data seems that they could put a final piece in the puzzle and know exactly who in which house had which ailments, their lifestyle risk factors could be refined and these data brokers would be able to look inside our very bodies.
One which fits data together, we do know from the HSCIC data release register, and press reports in March, is Harvey Walsh. The company tracks individuals pathway data, over time and the website now says:
“Harvey Walsh use non-sensitive and non-identifiable HES data for patient pathway mapping that is used by the healthcare industry with the NHS to improve the quality of healthcare management and service delivery by better understanding how patient cohorts move around the healthcare system.”
[Harvey Walsh’s system] “AXON holds non identifiable and non-sensitive HES (Hospital Episode Statistics) data and other sources of data including GP Practice Prescribing, QOF, Demographic and NHS personnel data sets.”
Data snapshots combine to give a Picture over a Lifetime
So now, not only can these companies understand us in infinite detail, but can do so over our lifetime. We are tracked over time and anaylsed not as a snapshot, but as a living album of snaps, moving across time. They know what we do commercially, in our lifestyle and how it interacts with our health and what may affect our consumer habits and help nudge our decision making. Put them together, and it starts to feel like I’m on The Truman Show.
I’d like to know though, once the data is processed, what happens to the new combined knowledge set, it creates? The original raw data as extracted may not be given to others, but is it the same product and protected, if it now shows up as a small piece, in a bigger jigsaw?
Omega Solver took their product offline this year, after privacy campaigners identified the risk of identifying individuals.
Acxiom as a world data leader example, is a company which provides consumer data and analytics for marketing campaigns and fraud detection. Its databases contain information about 700 million consumers worldwide.
“For more than 40 years, Acxiom has been a leader in harnessing the powerful potential of data.”
It seems others share my concerns, as this article on how data brokers use of our data is creepy, from Julia Angwen showed up in my alert feed this week, and another in ProPublica from last September. As she says,
“Commercial data has become a honeypot that government likes to dip its hand into.”
You can see more on this, in her interview with PBS News:
Our lifetime data is attractive to commercial marketing and all sorts of organisations who wish to understand us and sell to us. The one purpose, possibly the least trusted I have not really touched on. Hospital records have been shared with insurers and used for refining policy. Records have been sold to re-insurers, even since January 2014. And these insurers mine and use data much more deeply than we want to imagine. In fact, as I finish this I see the FT front page tomorrow carries a current story how insurers trawl our Big Data.
HSCIC Data Sharing Agreements will prevent Data Merger?
IMS Health UK & Ireland’s general manager, Michael Sanvoisin shows that exploiting the different data sets ‘out there’ in Big Data, is kind of the whole point. [17]
“The smartest use of data will be the effective combination of all the various sources of open data and patient information services available in the marketplace, augmented by companies’ own internal information and data from other reliable and reputable sources.”
IMS Health is working in partnership with the MHRA – and in particular the clinical practice research datalink (CPRD) – to help the UK increase its capabilities to build cohorts of patients for clinical trials. This has led to the linkage of IMS Health’s Hospital Treatment Insights (HTI), the aggregation of HES and prescribing data, to the CPRD. This powerful linked dataset enables the identification of specific patient cohorts and allows companies to monitor patient flow between primary and secondary care. IMS Ardentia’s Costed Care Pathways (CCP) sequences clinical events together with detailed financial information to give a longitudinal view of a particular patient care pathway.” [17]
When these global companies have in addition, bought data from HSCIC, where is the transparency for patients to know what internal practice at these private companies prevents all data becoming one Big Data set, in identifiable or pseudonymous formats, and sold or shared onwards with others?
The Recent register states explicitly, that IMS will not do this, that the data will not be sold onwardly, but how about the knowledge they create from it?
IMS Health works in partnership with pharma for example:
“ANDromeda is an engagement tool enabling greater market access with a tailored need across all functions within pharmaceutical companies.
And in the UK, are involved in work shaping our health market: “that may involve looking at how primary care organisations operate or focusing even closer on area-level commissioning, such as GP consortiums.”
Where is our Data being Used?
“The effective combination of IMS Health’s proprietary data assets, in addition to the vast swathes of open data being made available, can help inform key strategic decisions for both the NHS and pharma. Moreover, it can drive an increase in joint working towards shared benefits and therefore transform healthcare services in the UK and beyond.”
“in the UK and beyond.” So I ask myself, which countries outside the UK have received our medical records? Remembering that non-US citizens have no privacy rights in the US, if it landed there, we can say good bye to ever getting control of that knowledge back again.
Indeed HES extracts have been given to places in the US, specifically the University of California, the FOI request I got back confirmed. The Partridge Report contained two examples of data which has gone to Kyoto University. Yes, Japan. And remember, if the data is completely aggregated and anonymised it’s not included in these registers, because it is open, green data. So what exactly went to California, Japan and who knows where else. No one knows 100%. The Report only sample tested 10% of all releases.
IMS received 251 access (which is required for confidential data without consent) for identifiable data extracted from hospital pharmacy systems, sent to HSCIC and linked with HES (hospital records). The main customer for these products will be the pharmaceutical industry. (Lines 101-2).
IMS Health is massive, as is the global health data they hold.
On the IMS One intelligent cloud, the company connects more than 10 petabytes of complex healthcare data on diseases, treatments, costs and outcomes to enable our clients to run their operations more efficiently.
Drawing on information from 100,000 suppliers, and on insights from more than 45+ billion healthcare transactions processed annually, IMS Health’s 9,500+ professionals drive results for over 5,000 healthcare clients globally. Customers include pharmaceutical, medical device and consumer health manufacturers and distributors, providers, payers, government agencies, policymakers, researchers and the financial community.
Another user of our data is Optum UK (formerly United Health Group, and if that sounds familiar it was Simon Stevens [18] last employer). I wonder for example, does that mean it is also used by Optum Insight in the US? This presentation by Christopher M. Blanchette, shows different data providers of ‘RWE’ real-world evidence and where their data is sourced.
If international companies have NHS England patient data and re-use licence, is it likely in to have been exported around the world or how can we know in which locations it is used? I want to know how often data is given directly to International companies? How often is data given to companies in the UK, who have foreign centres outside the UK, which would routinely share that data with their central systems and therefore export it? It is a basic right of data management to require fair processing for identifiable data, to know who has it for what purpose.
How do we protect consumers’ concerns?
And as US Commissioner Julie Brill’s report shows, in the States there are concerns how this data is used and they are acting on it. Are we doing the same here?
Dr.Neil Bhatia in Hampshire, a GP who founded the non-commercial website care-data.info, asked HSCIC in an FOI request for the data *about him* which was released to these type of intermediaries. He was told, the data controller, the Health and Information Centre, does not know. And he can’t ask for what data is held in pseudonymous format – even though the data is pseudonymous with a key to make it linkable with new identifiable data coming in, so to me, that makes little sense. It is by its nature, re-identifiable.
But if HSCIC won’t release it in a Subject Access Request (SAR), we can then only surmise, whether our individual data was contained in bulk data transfers. So from the released data register, we should look at what types of companies are using pseudonymous (so called ‘amber’ data), and assume our own data was indeed included.
Overseas Data Distribution and Protection
care.data, it was said at the Health Select Committee meeting by Mr.Kelsey in March, was only for use in the UK but the HES/SUS data application form includes a field for use overseas. So, does that mean policy for export has changed for all data, or should they have spoken more precisely meaning only that “GP data extracted in care.data” was only to be used in the UK?
Because IMS, again, already has access to primary data from CPRD and secondary care data according to line 10 from HES. And whilst, it states “[Note added 28/3: The data are onwardly released only in aggregate form] I am curious – where does ‘onward’ mean? There is no Ltd. on the company name, no territory or geography indicated in the register. So if data is released to an American firm, should we assume it sits on US servers and is accessed directly by their US staff? Does onward only restrict them from giving the raw, identifiable data they received, to others outside IMS? Is it availble in non-aggregate form inside the whole of the IMS system? I, in the general public, can’t tell from the register and IMS is hardly going to tell me. We should be able to find out. I’ve found it a challenge, and my FOI request to HSCIC [14] to find out what data may have been given to US or Asian organisations, was tougher than my entire lifetime of dental appointments combined. It shouldn’t be difficult. Patients should be able to easily ask, to whom did you give my health data and where, for what?
Do we know enough about the plans to use and commercially re-use our data for commercial ‘health purposes’ as being broadly defined in the Care Act? If not, patients should be asking. GPs don’t have time.
Why does it matter? Because legal jurisdiction of data is still (perhaps outdatedly) physically geographic at least in aspects with which I am familiar. When working on global implementations of confidential employment data, we had to gain legal advice from each territory submitting data, on how we should legally properly manage data from over 50 countries in the world and its access by regional and global teams in the US, Europe or Asia. And on simple terms, we should always handle, process and use data in a way the individual expects and feels common-sensed appropriate to the purpose for which it was submitted. British citizens are not protected by US privacy laws because they apply only to US citizens.
“Existing laws do not sufficiently address data brokers’ handling of sensitive data in marketing or risk mitigation contexts,”
says Julie Brill’s statement. Well they don’t protect us Brits, at all, so I want to know if it’s being used abroad.
Few in England, will expect their data to have been made as freely available at identifiable individual pathway level, as it appears to have been in recent years. Do I at least have the chance to protect my children’s future data privacy, if not my own now?
Surely we can trust Data Protection Laws?
Because of the legal status of data which is deemed “de-identified” or “anonymized”, it is claimed they don’t violate our rights to health information privacy – Data Protection law accords us only the right to fair processing, not to prevent its processing, due to the the Health and Social Care Act 2012 which requires its extraction — but if it’s possible to re-identify longitudinal data sets – and if the whole point of getting these data sets together is to combine them, surely common sense would say, it may be legal, but that doesn’t make it right. There are other DPA expectations which HSCIC also fails to meet. The Minimum data required, for example. Deletion. Accuracy. I am guessing that every single one of the eight Principles have been broken by our data extracted before the HSC Act 2012. Yet, everyone seems to be ignoring this.
When it comes to Data Protection, identifiable data is treated differently from anonymous data. Amber individual level ‘pseudonymous’ data, is not the same as aggregated anonymous statistics and the care.data privacy impact assessment [19] confirms the risk of re-identification, yet the data is being treated as if it is anonymous. I can’t believe people working in the field believe themselves these data groups should be looked on as being equal. In my opinion, it’s not so much a case of wearing rose-tinted spectacles, it’s more like a blindfold on the wise monkeys; hear no evil, see no evil. [20]
I can quite clearly state on behalf of many, we feel that our rights to privacy have been and continue to be violated, no matter what the letter of the law says.
Whilst HSCIC may see only its own data sharing practices in a silo, that’s not how the impact of its sharing works in real life. It’s a join the dots between different data sets from different sources.
Can Good Governance Give us Confidence?
We are told that data-sharing agreements make it illegal for the data to be combined with other data held by the recipient, to make it identifying. But if the Data Controller doesn’t know what data the company already has, and doesn’t even keep track of what data has been given to them already, it must be impossible for individuals within these massive corporations to know the impact of adding their piece of the jigsaw puzzle. Over time, they will not track either, what from their company has already gone into creating the Big Data picture.
We could only rely on release controls and good governance, but for the past ten years reported in HSJ and the Partridge Review, it appears some datasets have been inappropriately shared without audit, which would have spotted the mistake. Governance is simply inadequate. In my opinion, not with malicious intent. Rather, simply, the data sharing strategy has been too fast for its own good practices to keep up. Now, it has to catch up fast.
As awareness increases, so too is the push back on the privacy grab. How do we feel about losing our individual rights, the removal of confidentiality and consent, the right to freedom from cold-calling, and to know who has our data for what reasons. And do we feel the same if we lose those rights in the name of commercial or public interests?
The British public is pushing back on banking failures and resents increasingly to see the minority of individuals benefiting commercially at the expense of the many. We resent the paternal state definition of the ‘Public Good’.
Private vs Public Good?
Business benefit vs the wider public good is in some ways an intangible and unquantifiable, but a debate which needs revisited on an ongoing basis says Dr. Mark Tayor when it comes to health records’ data privacy. [21]
The public interest considered by CAG in reviews of data release applications, must consider protecting both the public interest in research access to confidential patient data and the public interest in a confidential health service. Add to that the public interest of providing a national health service, and its safe to say ‘the public interest’ will be hard to satisfy for all of the people, all of the time and will be subjective.
“that the purpose for which the data will be used should be in the public interest and for the provision of health and care services; [and] that any approved processing must respect and promote the privacy of patients and care service users… ” (Hansard, 10 March 2014, Col.137)
Perhaps even more subjective, is the atmosphere of public interest and how interested the public is, in how how level decisions affect us on the ground. Certainly, Snowden and other data sharing revelations have coloured the muddy backdrop of how our data is gathered and used by others, and increased calls for transparency.
The Department of Health will be furious with the Home Office I expect this weekend, as they triggered a massive outcry over the perceived lack of transparency and scrutiny afforded to MPs and civil society over the Data Retention and Investigatory Powers Bill. Even Radio 2 gave it 20 minutes coverage. [22] (From 01:36.40) This kind of governmental out-of-touchness with the public and the perceived desire to hide something in the rush to the new legislation, is what undermines trust in all areas of the public-state relationship.
It implies a paternal notion, of “we know best, so just trust us little children.” Well, that ain’t gonna fly. Seahaven is not “the way the world should be.”
Patient empowerment to own our Health Records
This flawed process, within and beyond NHS data sharing, has also created a sense of loss and disempowerment. Whilst presentations are all about ‘patient centred’ care, and ‘personalised medicine’ sounds so about the individual patient, it seems safe to say patients have been left out of the digital decision making and sharing how those decisions will affect the public on the ground. This for care.data, should have been central to plans to ensure support and success. There are still unfilled positions supposed to be filled by patient organisations or patients on the tech board.
It seems endemic to new programmes too. Or have patient organisations been widely involved in the genomic plans for the nation and not told us? Unlikely.
The talk thus far, does not match the walk. Knowledgable patient involvement is as desired by some of those leading parts of NHS patient engagement, as a chocolate teapot is useful. One is documented having said on another programme, “this was not a suitable point for patient involvement.” Either you want patients involved or not. Involved means from the beginning. Not as the decoration at the end, a way to tick the engagement box.
The notional idea of patient empowerment in this programme is tokenism, if the most basic principle of care, the only thing I can control in my consultation – my patient confidentiality – is treated with such little respect.
Is the public good really defined and does it outweigh the private good and our long established rights of consent and confidentiality? Does it vary depending on circumstance and if so, who decides?
It certainly doesn’t seem to be us, the patients in healthcare. Nor as citizens in any other field of our personal data.
If you don’t pay for the product, you are the product
In general commercial uses of data, I am increasingly learning that if you don’t pay for the product, you are the product. Maybe we need to shout a bit louder, that we are not a product. We do not all want the knowledge of our health & lifestyle to be for sale.
We’ve got used to these third party uses through the recent media revelations and the acceptance that current Government seems to be prepared to sell anything the State has in its possession. I wonder how representative that is of what the people would choose to do?
So at the risk of repetition, let’s not forget the basics:
1. Our health records are for sale without our consent. 2. These companies are some who have bought data in the last year. 3. At these prices.The list of past customers in the Partridge Review of those who received data before April 2013 shows the extent of what was hidden from us for twenty years.
Should we be asking, what may be hidden still?
By stretching the scope of the potential discussion around the ‘industrialisaton’ and use of our health records for secondary purposes, we must not normalise the basics which we at first, found so surprising. We need to get them fixed first. Then, only then, will patients be willing to look at broader future scope. If I can’t trust you to manage my hospital record when I broke an ankle, why would I want to trust you with my genomes in future? It reveals a complete disconnect at NHS England level with the public in care.data thinking.
Come back to reality and listen to patients’ real concerns. We don’t want our data given to third parties, these data brokers and intermediaries or to continue re-use licenses. Even if it’s for ‘the promotion of health’ the purposes in the Care Bill.
And honestly? NHS England and the Department of Health shouldn’t want that acceptable in policy either, because they need to know who has our data, to govern it to make sure it is acceptable. As Sir Nick says in his report, the future data governance must be:
“meticulous, fool-proof and solid as a rock”
One more big mistake in who received our data in the future, and all cards will be off the table. For this to work, you need to properly manage it. And all this at the time where NHS England has now decided to outsource population wide databases, through the Steria outsourcing. Ha. Get that outsourcing security wrong, and for all your future programmes, as Truman would say, “Good morning, and in case I don’t see ya: Good afternoon, good evening, and good night!”
In the words of more Americans for whom I have a respect & love of their self-determined own words, Simon and Garfunkel, ‘Slow down, you move too fast.’
Julie Brill’s Statement made a recommendation in the US:
“A second accountability measure that Congress should consider is to require data brokers to take reasonable steps to ensure that their original sources of information obtained appropriate consent from consumers.”
We should feel that we consent to this mining of our health, wealth and lifestyles and know what is done with that knowledge. I feel disempowered because in finding out how my health data is used, I’ve discovered a brave new world of how my personal data is used. By commercial business. By Government. By suits and wonks as may be nicknamed. I am not equipped or informed enough to understand it all, but I’m doing my best to find out.
We need to trust in the people who manage these systems, who drive the policy and who advise the two, to work together and make technology work well for the rest of us. It should work well with privacy and security, and functionally.
Patients must speak up and Ask Questions
Patients must start asking more questions about these commercial uses and re-use licenses, because whilst the commercial intermediaries may access data for the purposes permitted in the Care Act, we are not a partner in patient engagement. Our data is being mined in the name of NHS improvement. Our samples being gathered in the name of science.
We are the product for sale. Our name, and everything else about us.
*****
image source: Time cover 2011
[1] Previous posts: Commercial uses of care.data and
[2] care.data use with intermediaries
[3] Health Select Committee July 1st, 2014: http://data.parliament.uk/writtenevidence/committeeevidence.svc/evidencedocument/health-committee/handling-of-nhs-patient-data/oral/11192.html
[4] NHS England Board meetings – to outsource support
[5] Decision to outsource Primary Care support services
[6] Schumacher’s health records stolen http://www.mirror.co.uk/news/world-news/michael-schumachers-medical-files-probe-3823793
[7] The Partridge Review Summary
[8] Risk Stratification guidance issued to CCGs
[9] IMS Health Ardentia http://www.imshealth.com/deployedfiles/imshealth/Global/Content/Technology/Technology%20Platforms/Ardentia/Ardentia_Royal_Free_PLICS.pdf
[10] Experian Public Sector http://www.experian.co.uk/assets/identity-and-fraud/authenticate-for-public-sector.pdf
[11] Experian’s Truetouch http://www.experian.co.uk/business-strategies/truetouch.html
[12] Experian -MOSAIC http://www.experian.co.uk/marketing-services/knowledge/case-studies/mosaic-case-studies.html
[13] HSCIC Data Register of Releases
[14] Medconfidential on commercial re-use licenses and Omega Solver https://medconfidential.org/2014/commercial-re-use-licences-for-hes-disappearing-webpages/
[15] Astrazeneca MOU with HSCIC – http://www.astrazeneca.com/Research/news/Article/121204-astrazeneca-to-collaborate-with-the-hscic
[17] IMS Health using NHS patient data http://www.imshealth.com/deployedfiles/ims/Global/Content/Solutions/Healthcare%20Analytics%20and%20Services/Healthcare%20Outcomes/IMS_HTI.pdf
[18] HSCIC FOI Request on data sharing with the US & Asia
[19] The Eight Data Protection Principles via ICO
[20] care.data privacy impact assessment http://www.england.nhs.uk/wp-content/uploads/2014/01/pia-care-data.pdf
[21] Previous post: The Partridge Review
[23] Data Retention and Investigatory Powers Bill. Radio 2.
****
See Wired: for an overview of care.data in general http://www.wired.co.uk/news/archive/2014-02/07/a-simple-guide-to-care-data and what good things should be taken from it when flagship care.data goes down http://www.wired.co.uk/news/archive/2014-03/06/care-data