Paul Gibbons, FOI Man, has blogged about data-sharing, questioning whether an over-cautious approach to sharing of health data is damaging. Paul says
What I’m increasingly worried about is what appears to be a widely held and instinctive view that any sharing of personal data – and even data that has been anonymised – is necessarily a “bad thing”.
I’ve got to say, in all the time I’ve worked in the field of information rights I’ve never come across anyone who actually thinks that, let alone articulates it (in my experience the only people who say it are those who seek to misrepresent it). The Data Protection Act 1998 (DPA) and EC Directive 95/46/EC to which it gives effect do not act as a default bar to sharing of data. There may be circumstances under which compliance with the law means that sharing of personal data cannot happen, but the converse is true – there will be times when sharing is lawful, necessary and proportionate.
Paul’s prime example of what he sees as (to adopt the title of his piece) “a disproportionate fear of ‘Big Brother’” preventing us from seeing the big picture” is the “predictable outcry” about the care:data programme, whereby the Health and Social Care Information Centre will, through the exercise of certain provisions in the Health and Social Care Act 2012, extract enormous amounts of health and social care information from local systems to centralised ones. The first step in this is the GP Extraction Service (GPES) whereby information relating to medical conditions, treatments and diagnosis, with each patient’s NHS number, date of birth, postcode, gender, ethnicity and other information will be uploaded routinely. The information will then be made available to a range of organisations, sometimes including private companies, sometimes in ostensibly anonymised, sometimes in identifiable, form, for a variety of purposes. This will happen to your medical records unless you opt-out (and if you think you’ve already done so, you probably haven’t – those who objected to the creation of a summary care record will have to go through another opt-out process). And this week we were informed that there will be no national campaign to alert patients to the GPES – the responsibility (and liability) will lie with GP practices themselves. (Anyone wanting to understand this complex and less-than-transparent process must read and follow the superb MedConfidential).
I accept that, on one view, this amassing of health and social care data could be seen as a good thing: as Paul suggests, medical research, for instance is a hugely important area. And the NHS Commissioning Board identifies the following desired outcomes from care:data
– support patients’ choice of service provider and treatment by making comparative data publicly available
– advance customer services, with confidence that services are planned around the patient
– promote greater transparency, for instance in support of local service planning
– improve outcomes, by monitoring against the Outcomes Frameworks
– increase accountability in the health service by making data more widely available
– drive economic growth through the effective use of linked data
But how realistic are these? And what are the attendant risks or detriments? Paul says
central medical records for all NHS patients…would mean that when you turned up at a hospital far from home, as I have done myself, doctors would have access to your medical records and history. Believe me, when you are in pain and desperate to be treated, the last thing that you want to do is to answer questions about your medical history
With great respect, the ideal of a centralised system whereby medics can provide emergency treatment to patients by accessing electronic records is never going to be more than a myth. Put another way – would Paul be happy trusting his life to the accuracy of an electronic record that might or might not say, for instance, whether he is allergic to aspirin? Treatment of patients is a matter of diagnosis, and emergency diagnoses will never be made solely, if at all, on the basis of records.
Security of information, and risks of identification of individuals are other key concerns. Paul says Daniel Barth-Jones identifies “deficiencies in [reidentification] studies” but I think what Barth-Jones is actually arguing is that the risks of reidentification are real, but they must be accurately reported and balanced against the likelihood of their happening.
But ultimately I have two major conceptual concerns about care:data and what it implies. The first is that, yes, I am instinctively distrusting of agglomeration of sensitive personal data in identifiable form in mass processing systems: history has taught us to be this way so I don’t see this, as Paul appears to, as a “fashionable” mistrust (and, for instance, the Joseph Rowntree Foundations’ exemplary Database State report is now over six years old). The second is that patient-medic confidentiality exists, and has existed for a very long time, for a reason: if patients are not certain that their intimate medical details are confidential, they might be reluctant to speak candidly to their doctor. In fact, they might not even visit their doctor at all.