Publishing information about patients

Time to change from guarding anonymity to getting consent

Last week three doctors appeared before the General Medical Council, the body that regulates British doctors, charged with serious professional misconduct because they had published a report on a patient without, it was claimed, gaining adequate consent from the patient (p 1245).1 The report, published in the British Journal of Psychiatry in 1993,2 concerned three patients with bulimia nervosa who had bled themselves. The paper came from Aberdeen, and the local newspaper picked up the story. It published only details that were in the case report (which included, for instance, “Ms C is a 26 year old preregistration doctor”), but a friend of one of the patients was able to identify her from the newspaper report. The patient made a complaint to the General Medical Council, saying that, although she had consented to the use of her case for teaching and research, she had not consented to its being published in a journal.

Despite the fact that the General Medical Council did not find the doctors guilty of serious professional misconduct, this case illustrates radical changes in expectations on guarding patient confidentiality in published reports; and these changing expectations reflect broader changes in the relationship between doctors and patients. Medical journals and textbooks have been published on the assumption that they will be read largely by doctors. This may explain why editors have been lax in allowing the publication, without gaining consent from patients, of material from which it is likely that some people will be able to identify the patient. In particular, patients are upset when they find out that their cases have been presented without their consent. Patients have contacted me after identifying themselves in reports published in the BMJ and elsewhere, and the BMA has also heard from patients who have been profoundly distressed by the publication of material about themselves that had not been sufficiently anonymised.

Both the General Medical Council3 and the International Committee of Medical Journal Editors (the Vancouver Group)4 have produced guidance on this issue. The council states that where, for the purposes of medical research, there is a need to disclose information which it is not possible to anonymise effectively, every reasonable effort must be made to inform the patients concerned, or those who may properly give permission on their behalf, that they may, at any stage, withhold their consent to disclosure.

The Vancouver Group advised in 1991 that “patients (and relatives) have a right to anonymity” and that, “if identification of patients is unavoidable, informed consent should be obtained.” The emphasis of editors' policies has, however, been on achieving anonymity rather than gaining consent.5 Authors have not usually sought consent from patients for the publication of case reports, and editors have not insisted on consent being sought. Yet case reports are usually published precisely because they contain something new, making it probable that patients, their relatives, and others (including, sometimes, highly inquisitive journalists) will be able to identify the patients. And pictures of patients with their eyes covered by black bands continue to be published despite our having known for a long time that such bands are wholly useless in protecting anonymity.6

We have deceived ourselves on our ability to guarantee anonymity, and we must recognise that medical journals and textbooks are now regularly consulted by patients, journalists, the police, tax collectors, and many others. Publication in a medical journal, particularly journals like the BMJ with a wide international circulation, is effectively publication to the whole world. Last week the Vancouver Group revised its guidelines (p 1272),7 and the emphasis is now on obtaining consent rather than on trying to achieve anonymity. The editor of the British Journal of Psychiatry has also--after taking legal advice--introduced new guidelines that say: “If an individual patient is described, his or her consent should be obtained. The patient should read the report before submission.”8 The new Vancouver Group guidelines emphasise that the consent should be in writing and the published article should record that consent has been obtained.

Authors and editors must thus ensure that patients have given their consent to publication whenever there is a possibility that the patient may be identified, and we have to recognise that patients may be recognised in most case reports and in some papers that contain information on groups of patients. The BMJ now gets written consent from patients or their guardians for all the pictures published on the Minerva page, even when they show only radiographs or pathological specimens. Very few patients refuse. We also try (although we don't always manage this yet) to mention on the page that we have obtained the patient's permission. The BMJ also gets written consent whenever we think that patients might be recognised in any published material, and over the years we have progressively lowered our threshold for seeking consent. We don't yet insist on written consent from patients for all case reports, but we may eventually do so.

This week's case before the General Medical Council also shows that doctors must be very clear about what the patient has consented to: vague statements about the information being used in teaching and research will not be enough. Patients must be told that the material will be published, and ideally they should see a final draft of the paper to be published. Indeed, one of the expert witnesses in the case argued that doctors and patients might negotiate over what appears in the final version. Editors, who often want substantial changes in papers, must also be part of that process.

Those who dread (as I do) the arrival of more process and bureaucracy may think that we are going too far. Surely, the argument goes, publication of reports on patients is essential for medical education and the development of medical knowledge, and the small risk that patients may be identified is outweighed by the overall benefit. But some patients are very distressed by discovering that material has been published about them without their consent; medical journals and textbooks are read by non-doctors; and most patients will give consent, which means that the advance of medical knowledge will not be impeded. We have to do better.