Intended for healthcare professionals


Informed consent: edging forwards (and backwards)

BMJ 1998; 316 doi: (Published 28 March 1998) Cite this as: BMJ 1998;316:949

Informed consent is an unavoidably complicated issue

  1. Richard Smith, Editor
  1. BMJ

    News p 955 Education and debate pp 10001011 Letters p 1022 Personal views pp 1026-1027

    The issue of informed consent within medical practice, research, and publication is coming increasingly to the fore as the balance of power in the doctor-patient relationship tips towards patients. Last week Britain's General Medical Council heard a case in which a paediatric cardiologist was accused of going beyond the consent that he was given to treat a child. The child died, and he was found guilty of serious professional misconduct and erased from the medical register for six months (p 955).1 When, last year, we published a cluster of articles asking whether we should decline to publish studies where patients had not given fully informed consent we prompted a flood of correspondence. We received over 50 letters, most of them argued with unusual care and clarity. Authors split down the middle between those who argued that we should always insist on informed consent (except in very limited circumstances) and those who thought that there were occasions when we need not. Today we try to advance the debate by publishing further responses to last year's debate, including some from patients' representatives. Within the broad context of informed consent we also explore the particular issue of consent for publication of material that emerges from the doctor-patient relationship.

    Informed consent in research

    In our first cluster Len Doyal made the case for insisting on informed consent with only a few narrow exceptions,2 while Jeff Tobias argued that the BMJ should sometimes publish papers that did not include fully informed consent.3 Both reflect on the subsequent debate, but neither has changed his position (pp 1000,1001).4


    Previous articles and comment on informed consent are available on our website (see Collections)

    Mary Warnock, a philosopher who chaired Britain's Committee of Inquiry into Human Fertilisation, argues that “theprinciple of non-exploitation has come to seem to many to be by far the most important moral principle that should govern research using human subjects” (p 1002).4 She thinks it a “misuseof words” to suggest that not obtaining informed consent in itself constitutes a harm: “sometimes it amounts to exploitation, sometimes it does not.” She encourages editors to continue to live in a morally hazardous world, to shun dogma, and to follow a prayer from Hertford College Chapel “todistinguish things that differ.” This encouragement is hard to resist because morally hazardous worlds are, I believe, right and proper for journals. Dogma is not only dangerous but also boring. We are in the debate not the certainty business.

    We have specifically asked patients' representatives to contribute because patients' voices were not being heard—because the BMJ is read mainly by doctors and other health workers. Heather Goodare argues that we should take a strong line and reject all studies that do not include informed consent (p 1004).4 Lisa Power asks us to consider the broader issue of patients in planning research and thinks that “anyhard and fast rule that the BMJ made about publication would probably have to be broken at some point” (p 1003).4

    In a separate article Richard Lindley argues that researchers should be educating the public about trials and that “weintroduce a new type of card—the randomised controlled trial card—to be carried by people who understand randomised controlled trials and wish to be considered for future appropriate trials” (p 1005).5 David and Solly Benatar from South Africa attack both Len Doyal's position and that adopted by the ethics committee in Natal that approved a trial that did not have informed consent (p 1008).6 In a personal view Josephine Venn-Treloar describes how she felt abused by undergoing an investigation without consent (p 1027).7

    None of this provides a simple solution to our dilemma: rather, it complicates it further. For now we are continuing our pragmatic policy of considering each case on its merits, and we have ourselves conducted studies on papers submitted to us without seeking consent from either authors or reviewers (and been criticised for it). Our next steps are to hold a conference in London (see accompanying note) and then to invite a small group of representatives of all views to advise us on what policy to adopt. If, as seems likely, they cannot agree, then we will decide our own policy and announce it to readers. Any policy we adopt will, of course, be reviewed.

    Consent and publication

    While continuing to swither over the broad question, we have advanced on the particular question of consent for publication of material that emerges from the doctor-patient relationship. 8 9 Now we are proposing to retreat—a little. It used to be, and in many cases still is, that medical journals and books were relaxed about publishing material that emerged from the doctor-patient relationship—pictures, radiographs, case reports, or whatever. Weak attempts were made to anonymise the material, but generally nobody was worried. Then editors and others began to receive complaints, and we realised that anonymity is impossible to guarantee (particularly to the patient himself or herself). The inevitable logic was to move to informed consent for all such material, and that is the position adopted by the International Committee of Medical Journal Editors.10 Now Britain's General Medical Council is adopting the same line. Proposed guidelines state: “You must obtain consent from patients before publishing personal information about them as individuals in journals, textbooks, or other media in the public domain, whether or not you believe the patient can be identified. Consent must therefore be sought to the publication of, for example, case histories about, or photographs of, patients.”

    Catherine Hood and others show in detail how consent can be obtained from patients in 85% of cases (p 1009).11 There are, however, a growing list of cases where patients have been distressed by information about them being disclosed without consent, and at least one case led to charges of serious professional misconduct.12 Those doctors were not found guilty, but under the new guidelines a similar case might result in the charge being sustained. The BMJ has recently been embroiled in a further case.13 We published—without revealing the patient's name and with written consent—a radiograph and photograph of a patient who had been attacked with a machete. Later, when the case came to court, the pictures were reproduced in most of Britain's national newspapers and on television. Journalists had made a link between the case and the BMJ pictures. All but one of the newspapers reproduced the pictures without written consent, and we have complained to the press and broadcasting regulatory bodies. Our argument is that by reproducing these pictures without consent the media have invaded the privacy of the patient, undermined the doctor-patient relationship, and made it less likely that patients will consent to have material about them published in medical journals.

    David Bullimore accuses us of hypocrisy and naivety in relation to this case (p 1022)14: hypocrisy because we placed the pictures on our website and had obtained inadequate consent, and naivety for not recognising that journalists would make the connection between the case and the pictures. Our response is that material on the website is copyright just as in the paper journal and that we published the material without a name attached. This case has, however, prompted us to start asking patients to sign specific consent forms that give information about the BMJ. The form is available on our website (www.bmj. com), and we will modify it in the light of readers' and patients' comments.

    The GMC's proposed guidelines are brief and clear, but they may oversimplify, be hard to implement, and undermine scientific publishing. Particular—and unresolved—problems arise, for instance, with the publication of family trees. Information, sometimes very sensitive, may be given about large numbers of people, and some of those people may not know that they have a particular genetic trait. Is consent required from everybody? In the process of obtaining consent might people be given information they would rather not have? Series of cases also present a problem. We published a series of cases of patients who had recovered after being diagnosed as being in the persistent vegetative state.15 In one case permission was denied, causing a critic in JAMA to ask whether “ajournal that knowingly omits scientific information from a report because of the lack of consent [can] still be called a scientific journal.” 16 The implication that science may demand that patients' rights be overridden is perhaps unfortunate, but that author attacks the editors of JAMA for declining to publish his paper on an outbreak of drug resistant tuber- culosis because patients had not consented. Few if any ethical rules can be absolute, and a case may arise where editors would choose to publish without consent “inthe public interest.” Certainly there are occasions when doctors break confidentiality in the public interest.

    Publishing informationthat emerges from the doctor-patient relationship

    Our general policy is that we require written consent from patients to publish material that emerges from the doctor-patient relationship. This is because the doctor-patient relationship must be confidential and because attempts to anonymise information about patients may fail. In papers describing recent experiences with patients consent will thus always be necessary: thus, in almost all scientific papers consent will be needed.

    Sometimes, however, it may be possible to publish material about patients—particularly general anecdotes—without consent. We cannot produce completely specific guidelines on this subject, but the decision depends on balancing the importance and interest of the information against the likelihood that a patient might be damaged.

    Publication without consent may be acceptable in the following cases.

    • The patient is long dead and has no living relatives.

    • The interaction with the patient was long ago—perhaps more than 15 years.

    • Because the interaction was long ago and the patient was elderly or terminally ill, the patient is likely to be dead.

    • The piece is to be published without the authors' names attached, making it unlikely that anybody could identify the patient.

    • All extraneous information that might help identification is excluded. We must be careful about removing information from scientific papers because it is difficult to tell what is important, but these “let outs” will rarely apply to scientific papers. They are more likely to occur with fillers or stories in essays.

    • Even if the patient were to identify himself or herself, the events described are unlikely to cause offence. We must remember, however, that it is difficult to know what will cause offence: some patients will be offended simply by the fact that the information they gave to their doctors was published without consent.

    • Sometimes authors—particularly Soundings authors—fictionalise material: they mix stories from different patients together. This is not acceptable in fillers because people read these as true. It may be acceptable in Soundings columns, but the author should make clear that the account is fictionalised.

    Similar problems arise over confidential inquiries into patient deaths. This methodology began in Britain with maternal deaths and has been extended to surgical and other deaths. The information that arises is extremely valuable but has so far been published without consent from surviving relatives. Will the GMC allow these to continue? Almost by definition, these are identifiable cases.

    Relaxing our absolutism

    We have also been criticised for becoming too absolute in our rules. James Rankine bemoans on p 1026 the fact that a personal view he published in the BMJ in 1994 would not now be allowed because of the problem of consent.17 The fillers that we publish on doctors' interactions with patients are popular with readers, and many make an important point. Yet many describe events that happened years ago and where the patients are almost certainly dead and their relatives untraceable. Should we reject these because we don't have consent? We have been doing so, but we think that we have gone too far. So just as the GMC is introducing clear but strict rules we are proposing to soften ours. The box contains our proposed guidance, and we welcome readers' comments. In essence, the guidelines ask authors and editors to balance the importance and the interest of the piece against the possibility of harm to patients.

    The BMJ and the Committee on Publication Ethics are hosting a conference in London on Friday, 15 May on informed consent in research, teaching, and clinical practice. Contact the BMA's conference unit. Tel 0171 383 6605. Fax 0171 383 6663. Email eoliver{at}

    This continuing debate over informed consent illustrates clearly that most ethical conundrums don't submit to simple solutions. Doctors are practical folk who like to get on with things, and many will be frustrated by the expanding complexity of this debate. But doctors will have to learn to inhabit the complicated world in which philosophers feel comfortable. Clearly ethical training is important, which is why our surveys of readers' wants always show ethics second to education. We are trying to oblige.


    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.
    6. 6.
    7. 7.
    8. 8.
    9. 9.
    10. 10.
    11. 11.
    12. 12.
    13. 13.
    14. 14.
    15. 15.
    16. 16.
    17. 17.
    View Abstract