care.data

There are currently proposals in the English NHS which involve the making available of patients’ health and social care records across the health service.

health records

This is now possible because these records are held in electronic form and can easily be duplicated and shared. One obvious advantage of such a process is that when people need treatment or care those delivering that treatment or care will be easily able to access the relevant records and draw on the information in them to tailor the treatment or care to the needs of the patient. However there are issues:

Anonymity – when records are shared then others can get at them and there is a risk of confidential data falling into the wrong hands. Data protection legislation and processes are designed to stop this happening but it is always a potential problem.

The use of the records by private corporations to:

(a) engage in research motivated only by a concern for profits – big pharma.

(b) potentially identify individuals in terms of risk factors and then charge insurance premiums based on that individual information – insurance companies.

(c) target individuals  marketing private health care – private health care companies.

All these are real issues but there are other positive aspects apart from the value to patients of their records being available to those treating them.

Traditionally clinicians developed their understanding of how to treat patients based on experience, both their own experience and the transmitted experience of medical educators and senior colleagues.  So the evidence related to particular patients and what worked for those specific people. Indeed the core traditional method for teaching medical students was the demonstration of a case in terms of diagnosis, treatment procedures and prognosis.  The turn to Evidence Based Medicine, welcome as it is in principle, has tended to downgrade this kind of knowledge derived from experience and to privilege knowledge derived from Randomized Controlled Trials  (RCTs) of interventions. However, there are problems with RCTs. First, no RCT can inform a clinician about what will happen if they deliver a particular treatment to a particular patient. All RCT knowledge is about what will happen to a large population of patients, on average. Second, RCTs are based on a simple and even simplistic notion of cause / effect relationships. They are not able to deal with complex and / or multiple causation – with confounding factors and multiple effects.

One way to look at health care records is to regard them as representations of the experience of clinicians in treating patients drawn together for very large numbers of cases and describing the health trajectories of those cases. They add together experience without losing sight of the cases as individual patients. Of course it is not practicable to deal with individual data per se but it is possible to classify and explore trajectories in terms of highly differentiated intervention / outcome relationships. The data mining tools which enable us to do this, many but not all based on neural net approaches, were developed for commercial use in relation to targeted marking. Just as Amazon’s records enable it to make suggestions to you or your use of a Nectar Card enables retailers to target you in marketing, so your health care records can contribute towards the development of treatment strategies which have been demonstrated as effective not just across a large undifferentiated population but for people very like you yourself.

This raises the question of who the health records actually belong to. My own firm view is that anyone who treats you has the right to use the experience gained in treating you to benefit others they may treat in the future. The use of anonymized case based health data is simply a process of generalizing experience. It is quite different from the very real issues of informed consent required of those who participate in experimental procedures. This is the solidaristic view. Regarding your health records as your property is individualistic. Of course all research must work with anonymized data. Your have a right to privacy. You do not have the right to your health record as private property. That is the view of possessive individualism and should be rejected out of hand by any socialist.

Of course what is required to make this work in an acceptable way is a proper National Health Service, something which just about still exists in Scotland and Wales but not in England. Moves by health trusts to treat clinical data as ‘intellectual property’ are even more to be rejected by socialists than the notion that anonymized treatment data is the property of the patient in the first place.

So there are real possibilities in the pooling of health data but also real risks. As ever the devil will be in the detail.