Is it ethical to enrol patients into a clinical trial without first obtaining their consent? Most people’s immediate reaction is “No”. But what should people consent to: “receiving treatment”, or “being in a trial”, or both?
The ethical principle behind requiring consent is autonomy or self-determination.
Here I want to consider the ethical arguments for and against recruiting competent adults to a clinical trial without first obtaining their consent.
Is consent needed for an offer?
An example of the sort of trial I am thinking about might help. Researchers wanted to know whether offering women a breast-screening appointment in the evening or at weekends, rather than only during office-hours, would improve screening uptake. All the women in the study were about to receive an invitation for breast screening as part of the routine programme. In one arm, women were given an office-hours appointment but told that they could change to an out-of-hours appointment if they wished. All remaining women were only given an office-hours appointment. Had everyone been told about the possibility of an out-of-hours appointment it would have been impossible to estimate the impact of offering such an option. Many women might prefer an evening appointment, but that does not mean that if offered one they would be more likely to attend – in fact slightly fewer women attended when initially given an out-of-hours appointment. Therefore, for the trial to be plausible the women had to be enrolled without consent.
There are several factors they may be important to why this trial was ethical to carry out without consent:
- The trial was of different offers – not different treatments
- No one was deprived of usual care – all women had an invitation to screening
- The research would have been impossible if consent was required
- Had some women been told that they were only being offered an office-hour appointment whereas others were offered an evening appointment they might have been upset
- A small increase in screening uptake in routine practice would result in hundreds of additional women avoiding dying prematurely from breast cancer.
In this trial the ethical harm of not obtaining consent was small, but the potential gains to society were great.
In English law, a mentally capable adult has an absolute right to choose whether to consent to medical treatment. In the breast screening trial mentioned above no women were deprived of the right to choose whether to have a medical treatment. Women were not even deprived the choice whether to be screened. They were only deprived of the right to choose what type of invitation they would receive.
Similarly, in another trial, women who had not attended their breast screening appointment were randomised to either receiving a standard reminder (to make an appointment for breast screening) or a second timed appointment. The only medical decision the woman needed to make was whether to attend breast screening. She had fully autonomy in making that decision. The lack of consent applied only to the external environment over which we have little determination – we cannot in general determine the wording of an invitation we receive likewise whether certain advertisements are placed on buses that we may see.
Do you need to provide consent if you will be in the usual care arm of a trial?
Suppose there are insufficient resources to treat all patients who might benefit with a new treatment. Is it ethical to allow clinicians to offer the new treatment to half of their patients and not to mention the experimental treatment to others? Would it be more or less ethical to select half at random? The principle of justice argues in favour of randomisation, but bizarrely no one seems to mind a clinician using an arbitrary rule to select patients to be offered a place in a trial, but many people say it is unethical to randomly select patients without mentioning the trial to the others.
Ideally, patients receiving the best evidence-based care and serving as controls in clinical trials should consent to being in the trial, but in practice many ethics committees allow trials in which such controls are unaware that they are part of a trial. The harm of not obtaining consent is considered relatively small. So, if there are good reasons for not obtaining consent and if the potential benefits are large, such trials may be considered ethical.
The flexible sigmoidoscopy screening trial offered those randomised to the intervention a new form of screening (while the controls did not receive any offer). Individuals in the intervention arm had to consent before being screened. But those in the control arm and those not in the trial were not given the choice to be screened by flexible sigmoidoscopy. As a direct result of this trial, flexible sigmoidoscopy is now routinely offered as a screening test within the UK.
One might ask whether a patient has a right to an intervention, even if medical experts consider it to be of no benefit?
For many experimental drugs it is not possible to receive the experimental treatment unless one takes part in a clinical trial and is randomised to the intervention arm. The principle of autonomy with respect to the novel treatment does not apply to controls – they can accept or decline the treatment offered to them, but there is no option for them to accept or decline the novel treatment. For such individuals, the principle only applies to whether they should have autonomy over whether they could be randomised; or autonomy over whether data about them can be used in medical research. There is no principle that says that a patient must be allowed to choose an expensive treatment of no proven benefit.
Are there circumstances in which consent should not be sought on the grounds of beneficence?
It is sometimes argued that telling such patients about a trial and emphasising their poor prognosis might be harmful to the patient. In such circumstances, the principle of beneficence might be invoked to argue that ethically those randomised to being a control should not be told about the trial.
Zelen design
Trials in which patients are randomised and enrolled before consenting are known as Zelen designs. The design is named after the Harvard statistician, Marvin Zelen, who proposed the design in a paper in the New England Journal of Medicine in 1979. As Dr Zelen wrote “Patients … in the second group are asked it they will accept the experimental therapy; if they decline, they receive the standard therapy.” He goes on: it “has the advantage that, before providing consent, a patient will know whether an experimental treatment is to be used.”
Cluster-randomised trials
Some people consider Zelen designs to be ethical only if patients are randomised as part of a group. Such studies are called cluster randomised. For instance, one might develop a class-room intervention aimed at reducing the uptake of smoking, or at preventing violence. Schools, as opposed to the individuals would be randomised to offering the intervention or not. Other studies randomise GP practices: all patients in one practice would be treated one way, all patients in another practice would be treated in a different way. In other studies, the unit of randomisation might be a village, or everyone seen in a given week.
Consent in cluster randomised trials raises all sorts of issues. Can a GP consent to the practice being randomised (on behalf of her patients)? If individual level consent is sought, does consent have to be unanimous or is it okay to be based on a majority? Individuals will always need to consent before receiving any form of treatment, but the randomisation might be done before consent is obtained. People are mostly happy with Zelen-type designs in cluster randomised trials – everyone is randomised, but only those offered the novel treatment or screening test are asked to consent before receiving it.
It does not make sense to me to say that it is okay to randomise without consent in a cluster randomised trial but not in an individually randomised trial. Which brings me back to my opening question, ‘Is it ethical to enrol patients into a clinical trial without first obtaining their consent?’ Whenever, a researcher wants to randomise without consent it is necessary to weigh-up the pros and cons. What are the harms associated with such a design in the particular trial? What makes it very difficult (or impossible) to answer the research question in a trial in which consent is sought up front? What are the potential benefits of conducting the research? Depending on the answer to these questions, there is no reason not to permit a Zelen design – there are numerous examples of their use having benefitted society.
The views expressed are those of the author. Posting of the blog does not signify that the Cancer Prevention Group endorse those views or opinions.