Intended for healthcare professionals

Education And Debate

BMJ's present policy (sometimes approving research in which patients have not given fully informed consent) is wholly correct

BMJ 1997; 314 doi: (Published 12 April 1997) Cite this as: BMJ 1997;314:1111
  1. Jeffrey S Tobias, clinical directora
  1. a Directorate of Cancer Services, University College and Middlesex Hospitals, London W1N 8AA


    Few if any issues engender such passionate–often acrimonious–disagreement among clinicians, ethicists, statisticians, and representatives of patient groups as does the continuing debate about informed consent and clinical research trials. In the blue corner: clinicians and biostatisticians keen to “move the field forward,” so to speak, and answer as quickly as possible the research question currently under investigation. In the red corner … just about everyone else. Anyone left in the centre? Only the hapless referee, in this case the somewhat perplexed journal, whose editorial board–constantly hectored from both sides–somehow has to give all parties a decent airing and ensure fair play.

    Those arguing in favour of fully informed consent as an inviolable rule (except, perhaps, in very special circumstances) often point out the essential, non-negotiable nature of a patient's right to autonomy and self determination. Quite rightly they remind clinicians that patients now wish to participate in decisions concerning their own management, to a far greater degree than ever before. Indeed, over the past decade, the move towards fully informed consent for all participants in clinical trials has become increasingly difficult to resist and is now formalised in various guidelines.1 However, neither lawyers, ethicists, nor medical scientists have so far agreed precisely what this term actually means–though it is generally held to imply a full declaration of the competing treatment options for any patient participating in a clinical research study, particularly one which involves randomisation between two or more treatment options. Together with the full description of treatments, there should be an explanation of the possible side effects of both new and standard therapies and a clear explanation that the “choice” of treatment is no choice at all–in the conventional sense–but is no more than a computerised flip of the coin.

    Most clinicians recognise that the anxious patient sitting opposite them in the consulting room requires both reassurance and a clear exposition of what needs to be done to provide a cure.2 However, an increasing degree of frankness on the part of the doctor, for the most part laudable and constructive, may also cause considerable distress to patients who would prefer to be directed rather than participate as equal partner. For clinicians who genuinely believe in evidence based medicine and recognise the central role of randomised trials, it is the need for explaining the randomisation concept, coupled with a detailed account of the shortcomings of standard treatment, which jointly symbolise the difficulty of the task: how to put these points across to a frightened patient in a highly charged atmosphere, with limited time available yet so much ground to cover and so many questions to answer. As Souhami and I have previously pointed out, many doctors repeatedly faced with this difficult task will not surprisingly decide that for them the game is simply not worth the candle.2 Hence the lamentable record in Britain of poor patient recruitment even where excellent clinical trials are on offer. British clinicians certainly don't seem to be signed up to the proud Harvard Medical School slogan, “Clinical research is an obligation not an option.”

    Embedded Image

    Doctors' concerns about their patients' anxieties in these circumstances were supported by the findings of an Australian study which compared two methods of seeking consent for clinical trials of different standard treatments for cancer: an individual approach at the discretion of each doctor, or a policy of total disclosure of relevant information given both verbally and in writing.3 This study found that although patients having total disclosure became more knowledgeable about their illness and treatment, and about the research aspects of what was proposed, these same participants were less willing to enter as subjects for the trial and had a significantly higher anxiety score. As many clinicians had expected, there are clearly trade offs to be made in the amount of information patients are given before consenting to studies, at least in the field of cancer. Detailed information, given indiscriminately, resulted in a more knowledgeable yet more reluctant and anxious patient. What is more, the ethical position of clinicians who decide, for whatever reason, not to inform patients about appropriate clinical trials for their particular condition has increasingly–and rightly–been questioned.4

    Concerns for patients' rights

    Although ethicists, counsellors, and other commentators argue their case–as research clinicians do–with the best possible intentions and concerns for patients' rights to information (and retention of as much control as possible in the face of serious illness), an atmosphere of mistrust has clearly developed. For example, last year the BMJ published a randomised study of psychological support for patients undergoing breast cancer, in which they were randomised (without informed consent) to receive routine care from ward staff, or with interventional support from a breast care nurse, a voluntary organisation, or both.5 Yet after publication of the paper, one distinguished member of the journal's editorial board felt moved to write that “the hospital ethics committee was surely at fault in allowing the research to proceed in contravention of the Nuremberg code” and even complained at the fact that two of the authors of the study were related.6 The authors of the study were clearly concerned to assess the potential benefits of an expensive and labour intensive form of intervention, and the journal felt the paper important enough to publish with a commentary regarding the ethics of clinical research without patient consent.7 But the letters to the editor were heated and even produced friction among the trialists, with a published reply from one of them as dissenting author.8

    In my view, the origins of this mistrust stem largely from a single source of disagreement: the passionate belief of those who insist that the individual patient in the consulting room should be the sole focus of concern for the doctor, and those who feel–and are prepared to say publicly–that they owe a duty not only to the patient sitting opposite but also to society at large which, with an equally urgent passion, has charged us to get on in all haste and find that cure. No point in pinning one's colour firmly to the fence: I'm for the latter group. This does not in any sense mean that the clinical trial is more important than the patient sitting so anxiously in the waiting room. A kind and caring approach to patients should always be the sine qua non of the doctor-patient relationship, as Sally Magnusson reminded us in the Christmas issue of this journal, even when (especially when) there is little that can be done.9

    But a proper respect for the patient's individual circumstances inevitably leads the research clinician to a varied set of approaches. The highly informed, articulate 39 year old journalist with a small but operable node positive breast cancer may be a candidate for several randomised trials and is likely to need a full, frank discussion with total disclosure of not only all the available treatment choices, but also the limitations of current treatment. In the enthusiasm to engage this intelligent and questioning patient in a proper dialogue, the chief danger generally lies in forgetting that above all she's a patient and, instead, falling into the trap of conducting a two way research seminar rather than a kindly and courteous consultation. On the other hand, and often at the other extreme of the social spectrum, the patient so characteristic of the clientele in a head and neck cancer clinic is much more likely to be male, older, far less educated, an enthusiastic consumer of cigarettes and alcohol: in short, someone quite unused to being “in control” of his own circumstances. Such patients are often homeless or struggling in an inner city hostel to retain what they can of their dignity and self respect. A cool and dispassionate discussion about the current research study (at present a trial essentially addressing the question of whether or not to offer chemotherapy in addition to radical surgery or radiotherapy) may be highly inappropriate since it pays no attention to the circumstances and culture–and, dare I say it, the need–of this particular patient. Indeed, as Brewin has pointed out, it may be better to consider that doctors participating in randomised treatment trials should not be thought of as research workers at all (in the normal sense of the word “research”) but simply as clinicians with an ethical duty to their patients “not to go on giving them treatments without doing everything possible to assess their true worth.”10

    Practical difficulties with informed consent

    Quite apart from the difficulty with randomisation–such an elegant, reliable, sophisticated concept to the research clinician, but so brutal and harsh from the patient's viewpoint–it is the nearness of the consent discussion to the diagnosis which causes greatest concern, together with the patient's perception of the intensity of the threat. Imagine yourself (this is often worth doing: after all, we're all of us either patients or potential patients) in the shoes of the thousands of patients taken each year to hospital with severe chest pain and acutely aware that this could be a fatal heart attack. We now know (through well conducted randomised clinical controlled trials, of course) that clotbusting drugs such as streptokinase play a valuable part in recovery; but would you really wish at this moment of crisis to be faced with a medical registrar keen to treat you properly but equally aware of the need to gain your informed consent before randomising you to one or other of the appropriate treatments? It's not that you're no longer competent to take it all in, but simply that there are likely to be other concerns on your mind–to say nothing of the need to feel full confidence both in the judgment and technical competence of those looking after you.

    As Collins and others have pointed out, at the time when the key studies addressing this issue were taking place, the differing ethical requirements (relatively low key in the United Kingdom but far tougher and with more constraint in the United States) led to a greatly differing recruitment rate (6000 patients from Britain compared with 400 from the United States despite an approximately equivalent degree of apparent interest by cardiologists in the two countries).11 In turn this led to a compelling statistic: if the United States had recruited as fast as Britain then the trial would have ended six months earlier, and since the eventual results transformed medical practice (improving the treatment of several hundred thousands of patients a year worldwide), that six month delay meant about 10 000 unnecessary deaths “directly due to whatever it was that slowed recruitment” in the United States. It should at least be a matter of some concern when what is judged ethical in one civilised society is dealt with so differently in another.

    Ah yes, the proponents of universal informed consent might reply, this is just one of those “special circumstances” which we all agree should be exempt from the usual rules. But how then might we go on to define these circumstances further? I have previously tried to divide or classify studies in oncological practice (at least those involving studies of new types of chemotherapy) into those which might or might not require fully informed consent.12 As others in similar or analogous situations have discovered, it is not always easy to recognise the differing circumstances which might demand full, partial, or non-disclosure when the study in question is randomised.13

    In the case of cancer trials, highly refined studies investigating technical differences between the two arms of treatment may be reasonably straightforward, in the sense that the patient will realise that the difference between the treatments represents only a relatively minor point of detail–not too alarming. On the other hand, where the treatment options are startlingly different the situation is altogether more charged. It can be extremely unnerving to discuss, for example, the possible use of chemotherapy in cancers such as those of the cervix or head and neck, in which we don't yet know for sure whether such treatment is genuinely valuable or simply meddlesome; with full disclosure of options, one finds oneself explaining carefully the pros and cons of the new treatment, then randomising half the patients to the control (the current “best buy”) treatment, to be met later with a disappointed patient who often feels “let down” by the loss of perceived benefit from the newer treatment (chemotherapy) which, naturally enough, in previous discussion had been described as “promising.” This often leads the doctor towards a rather shabby display of back pedalling in which the possible advantages of the chemotherapy are “talked down” and perhaps the side effects “talked up.” 14

    Still more difficult were the studies undertaken a few years ago to try to establish whether or not mastectomy for breast cancer–the traditional treatment during the first half of this century, hallowed by tradition but never validated by science–was tested for the first time against less mutilating surgical alternatives. The outcome of these studies, showing no clear superiority for the traditional approach,15 has proved hugely influential; yet it is hard to envisage how a strict and honest adherence to principles of fully informed consent could have been possible. I don't know about the American studies, but it certainly proved impossible in Britain. Although it was described as “the breast cancer trial that everybody needs but nobody wants,” 16 the Cancer Research Campaign, which supported and paid for the study, had to accept that recruitment was impossibly slow as a result of the disinclination of even the most committed trialists to put their patients through the rigours of informed consent.

    Yet partial disclosure17 or disclosure of the facts of the randomised study only to some (usually half: those in the “new treatment” arm) of the participants, is clearly regarded as an ethical minefield, making it unattractive to many clinical researchers and almost all health ethicists. Although it protects the right of patients not to be allocated novel treatments which are not yet fully established (and might never be) and ensures reasonable recruitment for clinical studies, it is generally rejected by hardliners as unethical since it denies the right to autonomy and self determination to each and every patient in the study–even though carrying the obvious and humane advantage of sparing all the patients treated to the best of current standards (the control arm) the anxiety of knowing that further improvements or refinements in their treatment are still urgently required. In my view, these benefits represent substantial gains for the individuals concerned and for any group of patients with the same illness, since valuable academic information might well flow from the study.

    Perhaps a still more helpful approach would be for patients to be informed at the outset of their treatment that several clinical and laboratory studies (some randomised, some not) might be in progress during their illness; might they be prepared to offer “blanket” approval here and now, accepting that the doctor would always act in good faith and be prepared to explain further any unconventional or novel treatment, if required, at any future point? In childhood leukaemia, for example, it is already commonplace for pretreatment blood samples to be stored and used later for laboratory tests not available when the sample was obtained. I greatly dislike the current trend towards ownership and commercial exploitation of medical samples–blood, tissues, cell lines–and admire the altruism behind this type of donorship. Shouldn't medical material be treated just as a personal letter might be after it has been posted through the letterbox slot–no longer strictly yours, even though you created its content in the first place. The posting is a consignment to a higher authority. Once again, Brewin has provided an elegant and clear headed argument designed to protect patients and at the same time allow sensible research studies to be conducted without unnecessary constraint: “The idea that the mere fact of randomisation always requires special informed consent–with all its disadvantages and potential for causing misconception and anxiety–is surely illogical. A doctor in his normal practice, giving treatment without randomisation, is trusted to choose from several options, even though there may be no way that he can be sure which is best. Why should we not also trust a doctor who submits such options to randomisation, while taking full responsibility for the suitability of each? Are the two situations really so different?” 18

    Buried in the quotation is that small but compelling word “trust.” Not really a word at all: a concept, a philosophy. Somewhat outmoded, certainly unfashionable. Yet most patients, it seems, still trust their doctors; and for their part, most doctors are well aware of the responsibilities that the trusting patient-as-supplicant brings to them. The correspondence pages of this journal are fine and lofty places to discuss these issues in a detached and intellectual manner–but do they approximate closely enough to the demands of the real world, in which the doctor must somehow juggle the multiple responsibilities of expert, humane, and above all respectful support for the patient in his consulting room with the wider healthcare concerns and requirements of society as a whole?


    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.
    18. 18.