Care.data: doubts about opting-out

“Prevent the Sale of NHS Patient Records to Drug and Insurance Companies”: the online campaigning organisation 38 Degrees has a new cause, and I’ve been exhorted daily by them and others to opt out of the care.data scheme. NHS England plans to harvest primary care data for research unless patients object individually. The start-date has already been put back by six months, in response to the medical establishment’s justifiable concerns about protecting confidentiality, while hasty ministerial assurances are being given that anonymity will be protected.  I know that many people are worried about the risk of being identified from their data, but personally, not being a celebrity or a criminal, I can’t get worked up about that - I simply assume that there are no secrets from a state that can monitor my emails, phone calls and banking.  I rely entirely on the anonymity that derives from being a mere drop in the ocean of Big Data.  The latest revelation that another public dataset, containing statistics about stays in hospital, has been uploaded to servers outside the UK, shows how hard it is to control information once it is digitised.

For me opting out from the care.data scheme is not a straightforward decision - I am an avid consumer of research data and have used many large survey datasets for teaching and research.  It’s always been easy to access national survey data at no cost for academic purposes; the UK has a wealth of datasets that are nationally representative, up to date and full of fascinating insights into society, health and public attitudes.  The explosion in electronic collection and storage has increased the volume of data in the public domain beyond imagining. So it seems hypocritical to use data and then refuse to allow one’s information to be included in the planned health dataset. Besides, the more people who opt out the more the quality of the dataset is impaired.

The crucial difference appears to be that care.data is being established at a time when the NHS is being eclipsed by new organisations whose interests are primarily commercial. Neil Bhatia is a GP and ‘Caldicott Guardian’ meaning that he has responsibility for information governance and patient-identifiable data protection within an NHS provider. He tells us on the website http://www.care-data.info/ that patient-identifiable information is already being extracted from hospital records by the government’s data clearing house, the Health & Social Care Information Centre.  This will eventually be merged with the care.data drawn from GP practices “and subsequently released, in various formats, to organisations within and outside of the NHS, for the purposes of administration, healthcare planning and research”.

It appears then that care.data information will be sold to commercial organisations, which makes it different from the 5,000+ government datasets that are publicly available on the HSCIS website and other government sites.  Jeremy Hunt’s promise (Guardian Society, Feb 28 2014) that care.data will not be sold to insurance companies is little consolation; it can still be sold to market research organisations, pharmaceutical companies and commercial providers of health care. I think that they will gladly pay for such a dataset, which will be vastly superior to almost any other similar database in existence (possibly outside Scandinavia), as it would contain information about the medical needs and services used of virtually a whole population, cradle to grave.  If insurance companies want information from it, for instance for actuarial purposes, they need only commission a third party – such as a university or policy think-tank, to run the analyses for them.  The dataset will have the potential to be linked to other datasets on income or lifestyle, so they can tailor private health insurance premiums to where people live and work, for example.

Is that a deal breaker for me, personally? I’m still not sure. Like many middle-aged patients, a few years ago I was invited to participate in the UK Biobank study.  This has worthy objectives as expressed on its website: “UK Biobank is a major national health resource, and a registered charity in its own right, with the aim of improving the prevention, diagnosis and treatment of a wide range of serious and life-threatening illnesses.” I went along to my Biobank session with a righteous glow that I would be contributing to the greater good.  The study was explained and the interviewer asked me to give a signature consenting to my data being used by any relevant organisation.  I asked for his assurance that this data, freely given, would not be sold on but the interviewer was flummoxed, clearly unprepared for this challenge and could not give me such a guarantee. I departed, feeling frustrated and disappointed, yet, when I visited the Biobank website recently, I saw that the users of the data appear to be exclusively medical academic researchers from all over the world - at least, those are the users shown on the website. They do pay to use the dataset, but the costs are reasonable administrative ones in my opinion. So I appear to have made the wrong decision in withholding my data, lacking sufficient information about Biobank at the time.

To tell the truth, the researcher in me finds the prospect of the care.data dataset appealing – I’d love to get my hands on it for research, and therefore I would feel guilty about withholding my personal data.  Yet, in effect, care.data will privatise medical records by placing them behind a pay barrier - records that have been amassed over decades by the NHS, funded entirely by taxpayers, generated and curated by public employees and entrusted to a state-run health service.  The terms of the project do at present not rule out the commercialisation of this information for private profit. I currently believe that I should opt out from care.data - not because I am particularly worried about confidentiality, but because it amounts to another nail in the coffin of the NHS, a stage further in the great NHS give-away which I abhor.  But if that pay barrier were removed, and all the care.data, suitably anonymised, were accessible as a public good at no cost, like so many government datasets, then I would probably change my mind.