Informed consent in medical research

In the issue of 12 April 1997 the BMJ invited comment on the acceptable limits of informed consent in medical studies. In view of the large correspondence this generated, we invited the two original commentators, Len Doyal and Jeffrey Tobias, to revisit the subject. We also invited comments from three people who are not doctors, researchers, or medical ethicists: two of them represent the views of patients and potential patients

Informed consent—a response to recent correspondence

Editorial by Smith and Personal views pp 1026-7

The publication of the debate between myself and Jeffrey Tobias about the acceptable limits of informed consent in medical research has generated an immense and varied number of letters to the BMJ.1^–4 This in itself is gratifying, whether or not correspondents agree with my arguments. It provides ample evidence of widespread and serious deliberation about the moral boundaries of the rights of participants in research.

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

• Download figure
• Open in new tab
• Download powerpoint

Many correspondents either explicitly or implicitly endorse the hard line that I take in my paper on the right of competent people to an acceptable level of information before agreeing to participate in medical research. Other contributions confirm my emphasis on the moral importance of the principle of informed consent but, in light of the highly specific circumstances where I argue that the principle must be qualified, question the degree or clarity of my own commitment to it. What is important here is our shared belief in the moral imperative of respecting human autonomy in almost all circumstances.

I still disagree with those authors who argue that it is not necessary to obtain informed consent if this will lead to the methodological compromise, or possible cancellation, of potentially beneficial studies involving clinical interventions that carry minimal risks. What these correspondents either fail to recognise or to take seriously is that to fail to respect the autonomy of competent people is to inflict harm on them that is just as morally unacceptable as direct physical or mental harm. To do so rejects the letter and spirit of the Helsinki Declaration—the “interests of the subject must always prevail over the interest of science or society.” Simply to assert that the declaration is wrong in this regard—without even attempting to rebut counterarguments, which, for example, I outline in my paper—is to embrace the dogma of scientific progress at any price. When human autonomy and dignity are at stake the cost of such progress is too high.

Some correspondents simply misunderstood or misread my paper. For example, Naomi Pfeffer and Priscilla Alderson maintain that I somehow claim that research may be done on children without parental consent.5 In the relevant section I specifically state, “Informed consent should always be obtained from someone with parental authority.”1 Similarly, Pat Soutter suggests that the HIV study of Satish Bhagwanjee and colleagues, which did not obtain informed consent from patients for seropositive testing, conforms to qualifications of the principle of informed consent that I outlined in my paper.6 7 It does not. I specifically exclude all studies in which there is an intent to contact subjects in the future, an inevitable consequence of the HIV study in question since it was designed to inform patients later that they had been tested.

This same mistake is made by Paul Little and Ian Williamson, who suggest that arguments in my paper are consistent with randomised trials without consent.8 It is true that I do morally defend some epidemiological research that is based not on direct patient involvement but on medical records—provided, among a long list of other things, that, again, there is no anticipation of further contact with the patients concerned.1 Yet Little and Williamson try to defend their position with reference to the merits of an antibiotic study in which patients were directly involved without obtaining their informed consent. Then, through making this fact clear in their letter, they go on precisely to initiate further potential contact with these patients. We can only speculate about the patients' potential distress and anger when they read or hear about this self confessed violation of their autonomy. This is the danger: patients may well (and do) find out about such abuse through, among other things, talking to other patients. Then utilitarian justifications can blow up in the face of those who use them to justify disrespect for human rights.

Failure to respect the autonomy of competent people is to inflict harm on them

• Download figure
• Open in new tab
• Download powerpoint

The most puzzling response of all to my paper was that of Michael Baum, a surgeon for whom I have great respect.9 Professor Baum seems to want it both ways. On the one hand, he draws an analogy between the moral appropriateness of conscription in warfare and the “responsibilities” of the lay public to participate as subjects in medical research in the “war against cancer” (and presumably other disease). On the other hand, he never really comes clean about what he proposes to do if members of the public do not live up to his perception of their responsibilities. If, ultimately, he accepts their right to refuse to participate then he agrees with me that they should be given enough information to do so on an informed basis—and does so despite my “absolutism,” “uncompromising zeal,” and professional life in an “armchair” on a “veranda.” If he rejects this right—as some of his comments and his agreement with Jeffrey Tobias's paper suggest—and really does support the quite extraordinary idea of conscription then let him say so and try morally to defend himself. It will take more than ad hominem arguments to do so successfully.

Changing the BMJ's position on informed consent would be counterproductive

Any author would be gratified by an overwhelming postbag in response to a provocative article—provided, of course, that not all the voices are raised in condemnation. Fortunately, however, it is clear even from the titles of the letters published by the BMJ 17 May and 26 July 1997 that a wide variety of views persists. On the one hand, titles such as “Doctors are arrogant to think they need to debate issue of patient consent”1 and “Lack of respect for patients in medical research may reflect wider disrespect in clinical practice”2 provide a clear and unambiguous view. But on the other, “Ethics committees and the BMJ should continue to consider the overall benefit to patients,”3 “Consent is not always practical in emergency treatments,”4 and “Let readers judge for themselves”5 offer a more relaxed view. As Little and Williamson point out,3 writing from a department of primary medical care, “adopting an absolute ethical view in open trials ignores the realities of—and would undermine the ability of research to inform—normal practice and thus could ultimately harm patients, including those who agree to take part in trials.”

As one of the protagonists of the debate, I am greatly concerned by many of the specific issues raised by correspondents. As well as the problem of, for example, emergency medical situations, the issue of risk of bias raised by a senior statistician6 is of particular importance since well conducted randomised trials tend to form the most influential basis of today's evidence based medical practice. Added to this, we have a past chairman of a research ethics committee at one of London's most prestigious research hospitals pointing to the wide disagreement as to which clinical situations require trial without fully informed consent—reminding us that “no one can claim to have a monopoly on deciding what is ethical.”7

Equally difficult is the argument—supported by preliminary data—that many patients may not digest information sufficiently well to permit a genuinely informed level of consent8; at the very least, it is clear that many patients in this study by Montgomery et al had no recollection whatever of consenting even to a course of radiotherapy—a consent which, we are assured from the article, had most certainly been given. If, as I believe, fully informed consent can sometimes be needlessly cruel,9 what is the point of insisting on it in all cases when about a quarter of patients (judging by Montgomery et al's study) cannot even recall being told about common side effects of treatment when all had been provided with this information?

As I pointed out when first setting out my stall, one of my chief anxieties concerns the somewhat old fashioned concept of doctoring in its traditional pastoral sense. While applauding the use of evidence based approaches and recognising the need for powerful trials to generate essential information, I do, nevertheless, feel a responsibility of equal importance—to act as patients' adviser, counsellor, advocate, and support. With many sophisticated patients, well informed and willing to enter into a robust two way dialogue, the medical scientist occupying a fair portion (I hope) of my brain can take the lead. For the majority, however—less educated, less well informed, and less able to marshal their arguments—a somewhat more directive or (without being pejorative) “paternalistic” approach will often be far more appropriate, and gratefully received. As Dr Thurstan Brewin, past chairman of Health Watch points out, “Those who want the BMJ to take a rigid view should spend a day in a ward full of elderly people. They would probably find many who, though far from being mentally incompetent, are at times confused and forgetful. What could be more unrealistic than to refuse to recognise this for fear of being called patronising? … Some people underestimate the harm that can be done to many sick patients when fully informed consent for every trial is sought, no matter how tense or difficult the situation.”10

I willingly give Ms Hazel Thornton, chairwoman of the Consumers' Advisory Group for Clinical Trials, the final word.11 As she clearly explains, her group “works directly with the professions … [and] identifies an urgent need to advance public education about clinical trials. Concepts such as randomisation, risk perception, and probability are poorly understood.… Such cooperation … will create a different attitude to research, which will be seen not as an imposition but as an activity to which we all have a responsibility to contribute.” Her letter, entitled “We all have a responsibility to contribute to research,” echoes my own view that both doctors and patients have much to gain from this type of partnership and that overzealous directives attempting to monopolise the moral high ground will surely prove counterproductive. The BMJ would be unwise to stifle important research by confining too closely the outline, structure, and phraseology of trial consent—details that are far better left to the originators of the studies and their local ethics committees.

Informed consent—a publisher's duty

Informed consent has become a shibboleth: you cannot be a respectable member of the medical research world unless you invoke the concept and accede to its demands, nor can you be a respectable publisher of research papers unless you ensure that your authors have clean hands in this regard. Informed consent is also, and perhaps more urgently, required in the case of medical and surgical procedures; but it is in the context of research requirements that the following remarks are offered.

The concept itself is not wholly simple. Questions may be raised about what counts as full consent or sufficiently informed consent, especially in the case of subjects who may find the idea of randomisation difficult to grasp or who may have problems, as we all do, with the calculation of risk. I believe, however, that we should not make too much of these difficulties, which are inherent in the nature of medical research and which can be minimised by tactful and sympathetic dialogue with potential subjects. The central moral problem, however, is concerned with the possible exploitation of the subjects of research. For research, including clinical research, is aimed, not at the good of the individual patient, but at the production of medical knowledge, which is for the good of society at large (although the individual patient may benefit from it by chance). This is the difference between research and the use even of innovative treatment for an individual patient.

In a research programme the subjects are being used as a means, not as an end in themselves. To treat someone merely as a means is widely agreed to be a moral evil, a breach of the “categorical imperative,” on which the very possibility of morality was held by Kant to depend. Philosophy apart, to make use of people, especially when they are not aware of what is going on, is generally agreed to be wrong. This evil is removed if people offer their services voluntarily. They then become willing partners in a joint enterprise rather than mere tools in it. Since they are free to decline to take part, their power of choice has not been overridden. They are being treated as befits a human as opposed to any other animal. The moral principle involved here is often referred to as the principle of autonomy. I prefer the more precise title of the principle of non-exploitation. Since it is especially easy to exploit the helpless and incompetent—those who, though human, seem to have little power of understanding or making a serious choice—the principle ought to be considered scrupulously in the case of such people. However, if research into the very conditions that produce such incompetence, such as Alzheimer's disease, is to continue it may be necessary to resort to consent by proxy. It seems morally important that such consent should be sought.

The principle 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. This is understandable on historical grounds: there are far too many cases, in the second world war and, sadly, more recently, of whole populations of people being damaged or destroyed as victims of research programmes about which they were ignorant or had no choice. The relevance of history is that it causes people to deploy the “slippery slope” argument—if once the principle of non-exploitation is allowed to be breached where will it end? To which the answer implied is that it will end in horrors such as were revealed at Nuremberg.

However, the slippery slope is a weak argument (though it exercises an enormous power over the imagination) in that there is no logical connection between allowing the principle to be breached in some cases and allowing it to be totally forgotten. The argument relies on a poor view of human nature: “Give them an inch and they'll take an ell.” Biological and medical scientists are especially suspect these days, and this arises from the power of the slippery slope. lt is crucial, therefore, that in this context editors should keep their heads and differentiate between different cases in which the principle has been breached.

There is all the difference in the world between, on the one hand, extending the use of anonymous data, collected for a particular study, to a further, previously unthought of, study and, on the other hand, the randomised testing of drugs in the treatment of a specific disease. In the first case there is no question of harm accruing to the subjects, and thus the use of the word “exploitation” is an exaggeration. It seems to me a misuse of words to suggest that not obtaining informed consent in itself constitutes a harm; sometimes it amounts to exploitation, sometimes it does not. Nor does it seem that the use for research purposes of discarded or unwanted tissue is exploitation—though there exists a lack of clarity about the relation between an individual and his or her body parts, which ought to be remedied. The matter becomes critical when a pharmaceutical company may make vast profits from the use of, say, a spleen that has been removed from the body of an individual. Does the person have property rights over something that was once, in some sense, his or her property but is so no longer?

The conclusion is that editors must try, in the words of a prayer much used in Hertford College Chapel, “to distinguish things that differ.” This makes the editorial function hazardous, with editors potentially subject to accusations of failing in their duty to ensure the moral respectability of research. But any other policy seems to me to rely on a dogma—that there are no other principles worth considering in the ethics of research except the principle of non-exploitation—and to rely also on an exceptionally wide and unrealistic view of what counts as exploitation.

Trial subjects must be fully involved in design and approval of trials

Reading the BMJ debate about informed consent and publication recently, it seemed to me that there was a basic flaw in the premise. Instead of “Why?” I wanted “How?” If informed consent is about the dignity and empowerment of trial subjects and the genuine participation of patients in our health research, then how can this be maximised throughout the trial process? If we look at the overall issue—the involvement of patients or potential patients—rather than the single aspect of informed consent we can begin to treat the disease rather than arguing over the symptoms.

I do not believe that you can obtain better practice about informed consent merely by making a rule about publication. There will always be some people prepared to obtain such consent technically without any real commitment to its spirit, because all they see it as is a signature at the bottom of a form and not a partnership. This is not to impugn the motives with which they entered research, but lack of time and money and urgency of need can put pressure upon the best of intentions. Of course, there are trials in which informed consent cannot be obtained, as Len Doyal outlined, and any hard and fast rule that the BMJ made about publication would probably have to be broken at some point. But the onus of justifying failure to obtain consent should not arise at publication stage for the first time; questions should be being asked far earlier in the process.

To improve the practice of obtaining informed consent wherever possible there must be a number of changes in attitudes. There needs to be a greater emphasis in doctors' education on interpersonal and communication skills, and a greater willingness on the part of some trial investigators to involve nursing staff in communicating with trial volunteers; doctors are not the only people with a voice and a brain. Secondly, there needs to be an understanding that giving patients or potential patients some say in the design and approval of trials is a positive process and not just a hoop to jump through. This involvement can stretch from trial design to writing information sheets and sitting on ethics committees. Thirdly, the onus should be clearly on those designing trials to show, as part of their basic data, their process for subject consent and uptake, rather than on others to challenge them in retrospect.

Placing the subjects of a trial at the centre of the process is not an easy matter. It may need extra finance or education, or other forms of support, and it may take time. Sometimes, I agree, it is not possible because of the nature of the trial, but this should be the exception—the question about informed consent should always be “Why not?” rather than “Why?” In my experience, as a participant in a vaccine trial and as an activist pressing drug companies to talk with us about their trial designs, such involvement is always to the good. I can appreciate that it feels like a nuisance to people who have not had to consider us before, but it leads to better trials with better uptake and, of equal importance, to greater involvement of individuals in their own health.

By fostering debate about informed consent, the BMJ has already added more to this process than any simple rule would do. I hope that it continues to do so.

Studies that do not have informed consent from participants should not be published

In his editorial of 12 April 1997 the editor asks, “Should the BMJ reject all studies that do not include informed consent?”1 The simple answer is “Yes.” This is the stated policy of others that observe the “uniform requirements for manuscripts submitted to biomedical journals.”2 There is no good reason why the BMJ should not follow suit.

It is clear that the Declaration of Helsinki is no longer entirely satisfactory as a standard to which medical journals should adhere. The declaration is a watered down version of the Nuremberg Code, formulated after the trials of Nazi doctors who had experimented on concentration camp inmates during the second world war.3 The code states unequivocally: “The voluntary consent of the human subject is absolutely essential.” But the Helsinki Declaration introduced a section on clinical research which says: “If the doctor considers it essential not to obtain informed consent, the specific reasons for this proposal should be stated in the experimental protocol for transmission to the independent committee” (Clause II.5).

Lack of consent in cancer trials has long been a matter of concern, 4 5 and this clause could have been used as an excuse for not seeking consent from competent patients in recent examples of clinical research.6^–10 There is some evidence that not seeking consent, far from eliminating bias (which is usually the reason given), actually adds to it. Patients who find that others in the same category are receiving different treatment will want to know why.11 It is best to come clean at the outset: patients who discover they have been deceived lose trust in their doctors.

If the present debate leads to a radical rethink of the way clinical research is conducted, matters may improve. Researchers are ignoring a valuable resource if they do not consult patients in designing their trials in the first place. This can save time and money and lead to better outcomes.12 Also, “joint ownership of the work being done keeps patients involved, instead of isolating them.”9 There should be no more debate about the need to seek consent from competent patients. There are, however, some grey areas that need further consideration.

The Helsinki Declaration makes provision for cases of legal incompetence, or physical or mental incapacity, though national legislation varies, and there is a case for amending legislation when it is deficient, to make proper provision for proxy responsibility where appropriate. We cannot take it for granted that an unconscious person would have consented to a trial had he or she been conscious: indeed, we have a special duty to respect the rights of those who cannot speak for themselves. If a proxy for the patient cannot be found the research should not proceed. In an emergency the doctor's duty is to do his or her best for the patient in the light of current knowledge. The rights of children, too, need to be respected: in the words of Lisa Hammond, aged 15, “Society should accept people of all types, and respect everyone's right to make their own decisions once they have all the facts, be they adults or children.”13

There remains the matter of clinical audit and epidemiological research. We cannot assume that patients will not mind their data being used for such purposes. As Doyal observes, “Normally patients should give their explicit consent for their records to be accessed.”14 Moreover, these data must be anonymised: we cannot be sure that patients will not mind if researchers and civil servants (who could well be colleagues in the same office) see their clinical details. Researchers may have overstepped the mark in a recent breast cancer audit,15 by requiring personal data—including names, dates of birth, and postcodes—not from the patients themselves but from doctors and administrators. This sheds light on the uses to which cancer registries could be put and raises awkward questions.16 It seems that careful thought needs to be given to this matter, including the possibility of a standard question to patients at the time of treatment asking permission to review their records for research purposes. Some clinicians already follow this procedure.17 Patients are well aware of the importance of such research, and if it is conducted appropriately they could be enthusiastic participants. But their consent must not be taken for granted.

A further problem occurs with the use of stored human tissue. Donors of blood, organs, or cadavers usually give explicit consent to the use of their bodies for therapeutic purposes, medical education, or research, but patients who provide tissue specimens during the course of their own treatment normally do not. If any use of this material for other purposes is proposed, patients' permission (or that of a responsible relative) should be sought. There have already been examples of commercial exploitation and even attempts to patent such material: any possible profit should be used in accordance with patients' wishes. A moving story is told by Steingraber of the cell line MCF-7, widely used in medical research. The initials stand for Michigan Cancer Foundation, and the 7 for the seventh attempt to establish a self perpetuating stock of cells from the body of the patient. The woman was a nun, Sister Catherine Frances, who died in 1970.18 Would she have wished a donation by way of royalty to be made to her convent every time her cells were used? Was she asked?

A breast cancer patient expressed the dilemma to me as: “In Victorian times they got upset about body snatching. Now they steal bits of your body when you're still alive.” These issues need further debate, with members of the public and patients themselves taking a full part in the discussion.