Abstract
Current UK guidelines regarding clinical research on children permit research that is non-therapeutic from the perspective of that particular child. The guidelines permit research interventions that cause temporary pain, bruises or scars. It is argued here that such research conflicts with the Declaration of Helsinki according to which the interests of the research subject outweigh all other interests. Given this, in the context of clinical research, who is best placed to protect the child from this kind of exploitation? Is it the medical researcher, the child’s parents or the nurse advocate? This article describes the problem, possible responses to it, and closes with a consideration of, and rejection of, a defence of current guidelines that claims moral parity between clinical research and clinical education.
Keywords
Introduction
The focus of this article is on non-therapeutic clinical research involving children who are incapable of giving their consent to take part in it; the clearest example of such a child would be a neonate. There is no intention here to call into question the importance of clinical research on children per se. For reasons that have been fully explained elsewhere, 1 it is important that such research needs to take place. It is argued here that current UK guidelines produced by the Royal College of Paediatrics and Child Health (RCPCH) 1 conflict with the guidelines given in the Declaration of Helsinki (DoH). 2 It has been claimed that the current RCPCH guidelines are unethical because they treat the non-consenting child as a mere means, 3 but the case developed in this article takes a different tack. It tries to show that given the most plausible interpretation of the term ‘well-being’ as this appears in the DoH, the RCPCH guidelines fail to protect the well-being of the research subject who is incapable of consenting to take part in research. After establishing this claim, the discussion turns to focus on protection of the research subject in clinical research. Who can protect the child given that the guidelines do not? It is concluded that advocacy in this context consists in protection of the child from involvement in non-therapeutic research, which places the interests of others above those of the research subject.
The RCPCH guidelines
According to the current guidelines governing clinical research on children, which were produced by the RCPCH Ethics Advisory Committee, ‘A research procedure which is not intended directly to benefit the child subject is not necessarily either unethical or illegal’. 1 In the context of medicine and nursing in the United Kingdom, these are the guidelines regarded as authoritative, and in the context of clinical research, they cohere with the European Union (EU) Clinical Trials Directive. 4 It is common to describe research procedures that are not intended directly to benefit the research subject as non-therapeutic 5 and I will use that term here.
In the context of research, interventions are classified under three categories of risk by the RCPCH. Minimal risk procedures include questioning, observing, measuring, collecting urine and using blood, which has already been collected. 1 Low-risk procedures ‘cause brief pain or tenderness, and small bruises or scars…’. 1 And high-risk procedures include things such as ‘lung or liver biopsy’. 1
Contrary to earlier guidance produced by the British Paediatric Association (BPA), 6 the RCPCH guidelines are explicit that low-risk research interventions are not necessarily unjustifiable, as earlier guidance stated, but may instead be justifiable: They say ‘The most common example of such research involves blood sampling’. 1 Furthermore, ‘where children are unable to give consent…, their parents or guardians may consent to the taking of blood for non-therapeutic purposes…’ 1 So, the RCPCH guidelines, it seems, permit the involvement of children in non-therapeutic research, even though they cannot consent to take part in it. It is reasonable to claim the guidelines ‘permit’ such research since according to the previously issued guidelines of the BPA, such research was not permitted. Moreover, as noted, such research can include procedures that ‘cause brief pain…[etc.]’ 1 An attempted defence of the practice of venepuncture in non-therapeutic neonatal research was published some time ago in 1997. 7 The authors concluded that it was justified after conducting 92 venepunctures on 69 neonates. 7 We return to consider this study later in the article. But for now, we turn to contrast the moral standards present in the RCPCH guidelines with those captured by the DoH.
DoH 2
When one compares the RCPCH guidelines to those given in the DoH, the former can be shown to be unethically permissive. According to clause 6 of the DoH, ‘the well-being of the individual research subject must take precedence over all other interests’. 2 This seems a plausible moral principle partly due to the familiar idea that it is wrong to use others solely as a means to one’s own ends or to the ends of others. 8 But strictly speaking that Kantian moral imperative applies only to moral agents and so would not include neonates and others who are incapable of consenting to take part in research. So instead of appealing to that Kantian objection, I intend to focus on the term ‘well-being’ as this occurs in the above-mentioned guideline. As will be seen, when one does this, one can show that the RCPCH guidelines are unethical.
Notoriously, the concept of well-being is very hard to define, and there is a large and difficult literature on this concept. 9 But three main approaches can be discerned – a hedonistic theory, a preference satisfaction theory and an objective good theory. In relation to very young children, the most likely sense of well-being that is relevant will be a hedonistic sense. In this, assessment of whether an experience increases or decreases well-being will be determined by reference to sensory experiences. Crudely, one’s life goes better when one undergoes pleasurable experiences and goes worse when one experiences painful ones. In the other two main approaches – preference satisfaction and ‘objective good’ theories – crudely, according to the ‘preference satisfaction’ theory, one’s life goes better when one’s preferences are satisfied, and goes worse when they are frustrated. And according to the objective good theory, one’s life goes best when one achieves certain objective goods and goes worse when one does not achieve them. The preference satisfaction account is not obviously relevant to our discussion since the child need not be psychologically capable of forming the relevant psychological states – this was indeed the case in the neonatal research referred to above. 7 Furthermore, an ‘objective good’ approach does not seem to me to be relevant here since surely in order to attain such goods one must at least be capable of trying to attain them, and the neonate clearly is not yet capable of this. So the most relevant construal of ‘well-being’ in relation to very young children such as neonates is the hedonistic sense.
Having noted the complexity of the concept of well-being, consider again the nature of the intervention: The research procedure is non-therapeutic, the child has not consented, and the child undergoes a harmful intervention, albeit one of ‘low risk’. The primary objective is to benefit future patients. If one thinks about the child’s well-being from a hedonistic perspective, that is, from consideration of the kinds of experiences, which the subject will undergo, it is plausible to claim that undergoing painful experiences runs counter to one’s well-being. This follows because, as noted, one’s life goes better when one has pleasurable experiences and goes worse when one has painful ones. Therefore, when well-being is understood in this way, to cause pain to a person for the benefit of others is to place the interests of others above that of the research subject and therefore to contravene clause 6 of the DoH. 2
Perhaps, the non-therapeutic interventions could be justified by reference to the ‘greater good’, or by exploiting the idea that we should presume altruism in non-consenting research subjects (perhaps in the way we might argue for ‘presumed consent’ in the context of organ donation). The former claim is indeed advanced in the article by Shah et al. 7 After having quantified (allegedly) the degree of distress caused to the neonates by the venepuncture, the researchers conclude that ‘Thus even if all the infants who suffered any upset are counted as major adverse reactions and the unwanted outcome of cerebral palsy is avoided in only one child, the risk-benefit ratio is favourable’. 7 So here it is claimed that the distress to the neonates is outweighed by the benefit of preventing just one case of cerebral palsy.
The authors seem not to notice the fairly radical implications of such a claim. It implies that it is justified to conduct research on any population of patients who are incapable of consenting to be involved if the likely benefits of such research outweigh any distress endured by the research subjects. In other words, to place the interests of others above that of the research subjects – which is of course to transgress the DoH clause quoted earlier. The whole point of the DoH is to stress that the well-being of the individual cannot be trumped by appeals to the greater good. To act in such a way is to give precedence to ‘other interests’ (DoH), namely, the interests of those future patients who stand to benefit from the research. So justification by appeal to the ‘greater good’ is not successful.
Having set aside the justification on grounds of ‘greater good’, what of the justification based upon a presumption of altruism. To be sure, this would need a long period of discussion and consultation before it could be adopted. It is not plausible to presume that it can be applied to some of the most vulnerable groups in medical research (i.e. children incapable of giving consent). But further, when one considers that, by definition, altruistic acts exceed what is expected by ordinary moral standards, 10 any attempt to presume this applies to a whole population simply in virtue of their age would be incoherent. Any radical change of this kind would need a phased introduction beginning with those who are least vulnerable, it would not be defensible to expose the most vulnerable to risk for the benefit of future patients. To do so would be a clear violation of clause 6 of the DoH.
Harris and Holm 11 defend a weaker position than that just described. They argue that since there is, plausibly, a moral obligation to take part in clinical research, we should presume that children would want to take part as opposed to presuming ‘moral turpitude in our children’. 11 But in response to this weaker version of the ‘presume altruism’ idea, it is important to distinguish the fact that one has an obligation to do something, from the question of whether or not one will do what one is obliged to do. Since in our own case one cannot assume one will do what one is obliged to do, one cannot legitimately conflate these in the case of subjects incapable of choosing. Once again, to do so would be to ride roughshod over clause 6 of the DoH. Also, it should be said that Harris and Holm seem to restrict their conclusions to ‘children’, as opposed to ‘very young infants’, 11 and so there is even less reason to suppose their proposal to be plausible in the present context.
Finally, one might claim that subjecting the neonate to such harms is no different from causing pain to a child in that child’s longer term best interests. Thus, for example, one might justify inoculating a child in their best interests in spite of the pain caused by the needle. But such a claim is not morally on a par with the kind of case under discussion here. We are discussing research in which the neonate undergoes harms for the benefit of others, not for their own benefit. So the relevant moral considerations in the context of non-therapeutic research cannot legitimately be compared to those which arise in the case of childhood vaccination for example.
Having justified the claim that the RCPCH guidelines are too permissive and exceed what is permitted by the DoH, perhaps it might be complained that the RCPCH guidelines are not representative of ethical guidelines in relation to children and, therefore, that other guidelines will cohere with the values set out in DoH. But when another influential set of guidelines are examined, this proves not to be the case.
Council for International Organizations of Medical Sciences guidelines 12
The Council for International Organizations of Medical Sciences (CIOMS) guidelines, if anything, are more permissive. When there is ethical and scientific justification to conduct research with individuals incapable of giving informed consent, the risk from research interventions that do not hold out the prospect of direct benefit for the individual subject should be no more likely and not greater than the risk attached to routine medical or psychological examination of such persons. Slight or minor increases above such risk may be permitted when there is an overriding scientific or medical rationale for such increases and when an ethical review committee has approved them.
12
The passage here reminds us of the distinction between therapeutic and non-therapeutic research and, as with the RCPCH guidelines, it is explicit that non-therapeutic research on those incapable of giving informed consent is justified (hence the reference to ‘research interventions that do not hold out the prospect of direct benefit for the individual subject’). It is permissible on the condition that such interventions do not expose the research subject to risks of harm that are greater than those present in ‘routine medical… examination’. 12 Thus, given that venepuncture is an example of a routine medical examination, it is plain that taking blood from a non-consenting patient would be permitted according to these guidelines. Moreover, these guidelines are even more permissive than those of the RCPCH since they permit research interventions which involve a greater level of risk than that which is present in routine medical examination.
Having shown that the RCPCH guidelines are unethically permissive according to the standards set out in DoH, it follows that the CIOMS guidelines are too since they are even more permissive. So if the analysis presented so far here is correct, it looks as though the ethical guidelines that are supposed to offer protection to vulnerable research subjects fail to offer such protection.
Parents
Given that the guidelines we have considered do not provide sufficient protection for children – in comparison with the standards captured in the DoH – and given that parental consent is needed before children can be researched upon, 1 one might think that sufficient protection can be supplied by their parents (or guardians). But is this the case? We should note that there are certain constraints on the extent of parental consent. Recall that the moral foundation of obligations to seek the consent of patients stems from the importance of respecting their autonomy, and so one might infer that the moral foundation of any obligation to respect parents’ views about what should happen to their children is simply respect for parental autonomy. 10 However, parental autonomy is not unlimited, and nor should it be, thus for example, the wishes of parents who are Jehovah’s witnesses and who seek to deny life-prolonging treatment to their child can be overruled in the best interests of the child as these are seen from a secular perspective. This legal position reflects a plausible moral position according to which if one person seeks to exercise their autonomy in a way which seriously impugns the autonomy of another, then there is no obligation to respect the autonomy of that person. Also, the idea that parents are best placed to give consent to medical treatment for their child (proxy consent) need not imply that this is justified in the context of research. These are two importantly different contexts and so it cannot be assumed that the same moral norms are appropriate in each. In asking for parental consent for treatment what is being requested is permission to act in ways which are primarily for the good of the child. But in requesting proxy consent for permission to involve the child in non-therapeutic research, one is requesting permission for that child to be used for the benefit of others. This is much less obviously justified. Specifically, if the analysis given above is correct, and if the research involves subjecting the child to ‘low risk of harm’ as defined by the RCPCH, 1 it is not justified since it involves placing the interests of others above the well-being of the child.
In their discussion of surrogate decision-making, Buchanan and Brock 13 argue that the decision-maker should ‘choose what will best serve the patient’s interests’. On grounds of consistency of usage, it is reasonable to assimilate best interests and well-being in this context. So sticking with a hedonistic construal of the latter, just as to undergo painful experiences runs counter to one’s well-being, so it would run counter to one’s best interests. Given this assimilation of the two concepts (which in any event are closely related in fact), our earlier assessment of the RCPCH guidelines shows that it cannot be in the research subjects’ best interests to undergo low-risk harms in non-therapeutic research. This follows since to subject the child to avoidable painful experiences runs counter to its best interests and that is what occurs in such research. So we have strong grounds to reject the suggestion that it is ethically justified for the parents to consent to the child to be part of non-therapeutic research which incurs low-risk harms.
In addition, other concerns focus on the factors that may affect adversely parents’ decision making in this context. Thus, for example, in many cases, parents are extremely keen to co-operate with the research team. In their research in this area, Shilling and Young
14
make the comment that … half [of parents agreed] that if [their child’s doctor] suggested their baby should enter a trial they would agree [(p.3). And] parents and children were more likely to respond intuitively and emotionally to research information than to systematically attempt to consider and weigh all the available information [p.4].
This kind of empirical finding raises important questions about the extent to which decisions by parents are competent ones. Other research reminds us that ‘many patients and parents perceive a clinical trial as a form of treatment, rather than an experiment’ – so-called therapeutic misconception. 15
So, given that the RCPCH and CIOMS guidelines on research ethics seem unethically permissive; and given that there are good grounds to query the defensibility of parental proxy consent, is it the case that the research subject can rely on the nurse for protection of their well-being in clinical research?
Nurses
It is reasonable to think of the role of protecting the interests of children in the context of research as part of the advocacy role of the nurse. The Nursing and Midwifery Council (NMC) code 16 mentions advocacy in the following clause: ‘You must act as an advocate for those in your care, helping them to access relevant social care, information and support’. In the research context, the nursing guidelines that concern ethics of research on children refer nurses to the relevant medical guidelines. Thus, in the Royal College of Nursing (RCN) document Research ethics: guidance for nurses, 17 there is a section on ‘Research involving people who are vulnerable’, and with specific reference to non-therapeutic research, readers are referred to the Medical Research Council (MRC)’s guidelines of 2004. 18 When one looks at these, one is led back to the original RCPCH guidelines in which, as we know, low-risk research is not regarded as unethical. (On this specific point, the relevant page is p.20 of the MRC document where it implies that research involving ‘more than minimal risk’ (i.e. research which is low risk) is permissible.) So it does not seem as though one can point to any distinctive nursing ethical advice on the issue which we are discussing.
Before commenting further on the issue of the nurse as advocate, it is worth just reminding ourselves of the complexity of the relationships, which are present in this context. The nurse is a clinical colleague of the medical team, and as part of his/her duty is there to support the child and the parents. If the nurse thinks the child’s interests are being jeopardised, this implies that other parties are being negligent or at best insufficiently sensitive. So advocacy in this context may be a difficult task, which calls for a great deal of tact.
Furthermore, any claim about who is best placed to perform the role of patient advocate begs a definition of advocacy and defining this has not proved to be an easy task. It is common to distinguish three types advocacy, which focus either on rights, autonomy or best interests. 19 Of these, it is the latter which looks the most relevant to our current problem, and indeed, it is that which the NMC code seeks to emphasise ‘Advocacy is concerned with promoting and protecting the interests of people in the care of nurses and midwives, many of whom may be vulnerable and incapable of protecting their own interests’. 20 So the advocate should try to protect the best interests of the child. For our purposes, and to be consistent with the terminology used above, as before I will conflate the terms ‘well-being’ and ‘interests’. Thus, if an act compromises one’s well-being, it will at the same time compromise one’s interests. So, the nurse (or whoever is advocating for the child research subject) must ask themselves whether it is in the best interests of the child to take part in the research. So can it be in a child’s best interests to take part in non-therapeutic research, which involves infliction of ‘low-risk harm’? We addressed this earlier and concluded it could not be since it conflicts with the subject’s well-being (where this is understood hedonistically). So it seems to follow that the patient advocate would fail in the advocacy role if she allowed the child to take part in non-therapeutic research, which involved low-risk harm as defined by the RCPCH.
Nothing we have considered so far has proved sufficient to defend the position set out in the RCPCH document from the criticism developed above. And if the criticism is accepted, then research on children that involves low risk of harm is not ethically defensible (if one accepts the ethical standpoint of the DoH).
Of course, as with any argument, there are a range of responses that can be made to it. So I will now turn to consider an apparently powerful argument against the line developed in this article. However, I also hope to show that this does not successfully rebut the position developed here.
A moral parallel with the educational context
The article by Shah et al. referred to above states that the venepunctures in the research were carried out by ‘two junior medical staff ’. 7 This serves to remind us that in teaching hospitals, it is necessary for inexperienced health-care professionals (HCPs) to do procedures that they may never have done before – such as carry out venepuncture on a neonate for example. Only by practising on current patients, they can develop and eventually acquire the skills, which will benefit future patients. As this kind of procedure can be extremely difficult, it is likely that the early attempts by the trainee to perform the technique will cause some distress to the young child and of course to the child’s parents if they are present. Surely, it may be claimed that this situation is morally on a par with non-therapeutic research on neonates or young children more generally. In both kinds of situations, patients are used in such a way that they are exposed to low-risk harms for the benefit of future patients. So if we are to view such research as unethical, by parity of reasoning we would also have to conclude that ‘using’ non-consenting patients in teaching hospitals is also unethical, and therefore, it will prove impossible to train HCPs and help them develop these important skills. Given that the position defended in this article has such an implication, it may be argued that it leads to such an absurd conclusion and so can be rejected.
So since medical training is, plausibly, considered to be a morally desirable, and since it is morally on a par with non-therapeutic research on non-consenting subjects of the kind we have been discussing, it follows that such non-therapeutic research is similarly justified. In both kinds of case, we find that a child is exposed to ‘low-risk’ harms (those which cause ‘brief tenderness or pain…’ 1 ) and that this is for the benefit of future patients. In the teaching context, because of the opportunity to ‘practice’ on current patients, which is open to the inexperienced practitioner, it is more likely that future patients will benefit from the expertise acquired by practising on previous patients. This is the case in spite of the fact that the ‘practice’ attempts exposed the patients to harms deemed low risk by the RCPCH.
But are these two situations morally equivalent in all relevant respects? There are important similarities: especially the ‘using’ of current patients to benefit future ones, and the fact that in both kinds of cases, the patients incur harms for the benefit of others. However, there is surely one important, morally relevant difference between the two types of case. In non-therapeutic research, by definition, there is no intention to benefit the research subject, only future patients. But when the junior HCP seeks to take blood from the neonate for the first time, the primary purpose of the intervention is to promote the well-being of the patient. As a consequence of this, a happy side-effect, so to speak, the HCP acquires relevant skills and techniques that can be used to benefit future patients. But nonetheless, the primary motivation in such a situation is to act for the benefit of that specific patient, not future patients. So there is an important difference – at the level of intentions – in the two kinds of case. And I submit that this is a morally relevant difference. It is captured in the distinction between being used as a means and being used merely as a means. To use the patient in research is to use them merely as a means. To use them to develop one’s skills is to use them as a means. Such a distinction is indeed one which is recognised in the DoH since in the therapeutic encounter the well-being of the patient is ‘taking precedence’, 2 but in the research context, it is the interests of future patients that take precedence. As explained above, it cannot plausibly be claimed to be in the non-consenting patients’ interests to incur pain for the benefit of future patients given the account of well-being, which was described earlier. So I think the critical line developed in this article is not vulnerable to this response.
Conclusion
So I conclude that the case is successfully made and call for a revision of the RCPCH guidelines because they are too permissive. Also, given that the RCN guidelines on research ethics simply refer readers back, eventually, to the RCPCH guidelines, this article should prompt a review of those guidelines if the argument presented here has been successfully made.
Footnotes
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.