The limits of law at the limits of life: to treat or not to treat?
The limits of law at the limits of life To treat or not to treat? … the answer which will be given in relation to a particular problem dealing with a particular set of circumstances, is a much better answer than an answer given in advance. The difficulty in this area is that there are conflicting principles involved. The principles of law are clearly established, but how you apply those principles to particular facts is often very difficult to anticipate. R v Portsmouth Hospitals NHS Trust ex p Glass [1999] 2 FLR 905 at 911A–B (Lord Woolf MR) The fundamental question posed in the hearings pertaining to David Glass, Charlotte Wyatt and Terri Schiavo is essentially: to treat or not to treat? The cases do, of course, have their differences: Mrs Schiavo’s previous status as an autonomous adult opens up the possibility of her having made provision for an incapacitated future, and the perceived interests of each of these patients might differ according to such factors as diagnosis, prognosis, and their family’s perceptions and preferences. Chapters 2 and 3 reveal that not only might different legal frameworks be available for different incapacitated patients, but also the same legal principles might point in different directions for different patients. Such differences might be entirely appropriate if we accept that each patient is different. But problems arise when the different frameworks and assessments fail to qualify as rational law, in the sense defended in Chapter 1. Central to this idea is law’s action-guiding function: rules that contradict one another, or cannot be known, or cannot be performed will all fail to pass muster. Unfortunately, as we will see throughout this chapter, there are numerous such failings in the law pertaining to the provision and cessation of life-supporting treatment, which reveal what Annas has described as ‘the limits of law at the limits of life’.1 Remarkably, even the judges appreciate the problems, with those in Airedale NHS Trust v Bland [1993] 2 WLR 316 detecting problems with – but simultaneously affirming and applying – a set of rules they saw as ‘misshapen’, ‘irrational’ and ‘illogical’.2 In what follows we will look at these problems in detail, focusing particularly on the ‘best interests’ standard, since this underpins the legislation pertaining to both adults and minors and dominates the jurisprudence on withholding and withdrawing treatment. The first problem with the cases which hold that it can be permissible to withhold or withdraw life-supporting treatment – be they those pertaining to children or those pertaining to incompetent adults – is that it is difficult to see how these decisions can occupy the same terrain as rulings which insist that failure to care for people who are incapable of caring for themselves can amount to a homicide offence when such a person dies. The latter rulings are well known to criminal lawyers. In Gibbins v Proctor (1918) 12 Cr App Rep 134 the defendants were convicted of the murder of Gibbins’s daughter, after 7-year-old Nelly starved to death. Darling J in the Court of Appeal upheld both convictions. Proctor ‘had charge of the child’, so it was her duty to ensure that Nelly was ‘properly fed’ and given ‘medical attention if necessary’.3 However, a murder conviction was appropriate because the jury found that ‘she had deliberately withheld food’.4 Gibbins may not actively have inflicted harm on his daughter but ‘he knew that Proctor hated her, knew that she was ill and that no doctor had been called in’.5 The jury accordingly found that he satisfied the mens rea (or ‘mental’ element) of the offence, that is, intention to cause death or grievous bodily harm. As this precedent demonstrates, parents and those in loco parentis evidently owe a duty to care to children in their charge; failure to fulfil such a duty, coupled with an intention to end life or cause grievous bodily harm, will amount to murder. The duty to care need not be restricted to parents, nor the right to children who fall beneath the age of majority.6 Where there is a failure to fulfil such a duty, whosoever it is who happens to come under the obligation, then the criminal law has an array of offences at its disposal, to deploy as circumstances dictate.7 We might presume that cases like that of Gibbins v Proctor, in which there was a direct intention to cause harm, will be rare. Other charges might nevertheless result where the mens rea is different, such as where a defendant is grossly negligent in failing to satisfy the duty. R v Stone and Dobinson [1977] QB 354 tends to be cited here. Stone and Dobinson were convicted of the manslaughter of Stone’s (adult) sister, Fanny, who suffered from anorexia nervosa. As in the previous case, the couple’s appeals against conviction were dismissed. Lane LJ talked in terms of ‘a reckless disregard of danger to the health and welfare of the infirm person’,8 but nowadays the question is more likely to be whether the defendant was grossly negligent to such a degree that a criminal sanction is warranted.9 The message from the criminal courts seems clear: there is a duty to take care of those incapable of so doing. Failure to oblige, such as through a failure to feed, can lead to the dock – but not, it seems, if the defendant happens to be a doctor; at least, this looks like the message from cases like that of Anthony Bland. Anthony’s judges did recognise that the law of murder presented a substantial obstacle to the removal of his feeding tube.10 Controversially,11 the judges suggested that his doctors would have the guilty intention.12 This certainly seems to suggest that they would be guilty of the same crime as Gibbins and Proctor. Lord Mustill drew a direct parallel: ‘Of course, the cases are miles apart from an ethical standpoint, but where is the difference on the essential facts?’13 The first step towards clearing the murder hurdle was to find that the withdrawal of clinically assisted nutrition and hydration (CANH) (via a naso-gastric tube) did not amount to a (positive) action but rather to a (negative) omission.14 Like some philosophers,15 the judges detected a degree of difficulty with this conclusion, with Lord Goff acknowledging that ‘some positive step’ was involved.16 But Lord Lowry conveyed the mood of the majority when he suggested that to find otherwise, and thus distinguish between withholding and withdrawing treatment, would be ‘illogical’.17 The next step was to find that Anthony’s doctors would no longer be under a duty of the sort which bound Gibbins, Proctor, Stone and Dobinson. The duty question appeared to turn on the finding that CANH, by which nutrition and hydration are delivered by artificial means, amounted to ‘medical treatment’ (or, at least, part of Anthony’s ‘care’).18 This too is controversial. John Keown, a vocal critic of the ruling, asks simply ‘what is being treated?’19 This manoeuvre nevertheless enabled the judges to invoke the Bolam test in determining what would be in the best interests of the patient.20 This in turn allowed the doctor who sought to withdraw treatment to claim the backing of a responsible body of medical opinion. But this too is highly contestable. Sheila Maclean objects that the introduction of Bolam means that the decision rests on ‘what some doctors actually do’21 and Lord Mustill saw no reason why the doctors’ viewpoint should be decisive.22 More troubling was Lord Browne-Wilkinson’s suggestion that doctors might justifiably differ about the prolongation of life.23 As Kennedy and Grubb appreciated, if ‘best interests’ are determinative: ‘Surely continuing or withdrawing is in the patient’s best interests: not both!’24 Even leaving aside the deference shown to medical opinion,25 the judicial attempts to grapple with the best interests assessment in the case were scarcely successful. Lord Goff, for example, saw fit to conclude that the nasogastric feeding could be withdrawn since it was no longer serving Anthony’s best interests because its continuation was ‘futile’.26 Yet, if this truly had been the case, then the situation would never have come before a court.27 In any event, the judges concluded that the doctors’ duty to treat had dissolved, leaving them entitled – maybe even obliged28 – to discontinue treatment, or else they would risk committing a different crime, this time of assault.29 Despite the Law Lords’ misgivings about any precedent they might have been setting, the logic of Bland must apply beyond the relatively narrow confines of adults in a VS. The ruling certainly gives the most comprehensive coverage of the limits of the duty to treat and the boundaries of the criminal and civil law. Remarkably, one can trace the origins of the distinction being drawn back to R v Arthur (1981) 12 BMLR 1, in which the paediatrician was acquitted of attempting to murder John Pearson. There too one finds a judge apparently striving to carve out an exception for the doctor in the dock, as revealed in his eagerness to depict Arthur’s conduct in the passive terms of a ‘holding operation’ (a term otherwise unknown to the law) and his scant reference to the alleged attempt to suppress the child’s appetite.30 But, just as with Bland, it is possible to take exception to this exception, given the conflict with established precedent on the duty to preserve life.31 The summing-up in Arthur’s case is also susceptible to challenge because it is clearly at odds with the decision in Re B (a minor)(wardship: medical treatment) [1981] 1 WLR 1421, which had been decided only months before but which did not appear to feature in the criminal trial. These cases certainly generate conflicting guidance about the duties owed by parents and doctors to infants with Down’s syndrome: Alexandra, who was the subject of Re B, had been judged by her doctors to be in a worse condition than John Pearson, but her doctors were obliged to ensure that she would live, while Dr Arthur was apparently free to leave John Pearson to die, at the request of the boy’s parents. Gunn and Smith have sought to argue that the two cases can be reconciled: each case acknowledges that a lower duty of care is owed to ‘abnormal’ infants,32 and in both cases a decision taken jointly by parents and physicians to leave an infant to die is deemed acceptable (or, as it was put in Re B, ‘entirely responsible’33). The different results might simply be explained by the fact that the judges in Alexandra’s case, in loco parentis, chose ‘responsibly’ to exercise their discretion differently, that is, in favour of life.34 This argument does not wholly convince, however. The quality of life thresholds are still drawn at different levels and, insofar as the law purports to protect all regardless of disability, there remains the conflict with hitherto unassailable principles about the duty to treat, feed and otherwise preserve and protect the lives of incompetent individuals. This dispute might be best consigned to the history books, since Re B seems to emerge victorious and few believe that Arthur has any value nowadays as a precedent. But Bland undoubtedly remains, and so too does the difficulty with squaring that decision from the civil realm with prior precedent from the criminal courts. Perhaps, in contrast to the opening of this section, it is best to say that they do not occupy the same terrain at all, and so we should consider them – and their ensuing principles – separately. A perfectly clean separation eludes us, however, given the civil courts’ willingness to settle criminal questions.35 Furthermore, as Brownsword argued, the convenient creation of categories cannot convincingly cover up inconsistency.36 Nevertheless, it is certainly apparent that decisions (not) to provide life-support are today most often viewed through a civil lens, and in particular one associated with medical law, in which the focus is upon the best interests of the patient.37 Even when we restrict our attention to civil cases including Re B and its successors, we find a body of jurisprudence that fails to measure up to the standards of rational law. Not only are fine lines drawn in an effort to avoid the appearance of conflicts in the case law, but also the objective rules with which the legal enterprise should be associated become increasingly difficult to detect. Too often one is left wondering precisely what the law in this area requires and how its edicts can be achieved. Leaving the anomalies of Arthur aside, the legal framework governing contemporary decisions (not) to treat critically ill children can be traced back to Re B. It is here that the central (nowadays ‘paramount’) concern with the best interests of the child is emphasised and we find Templeman LJ’s much-quoted statement that treatment might legitimately be abated where ‘the life of this child is demonstrably going to be so awful that in effect the child must be condemned to die’.38 From this line of thinking comes the idea, in Re J (a minor)(wardship: medical treatment) [1991] 1 FLR 366, that the child’s quality of life is a key component of the balancing exercise that must be conducted, in which consideration should be given to whether the ‘life would be so afflicted as to be intolerable to that child’.39 Subsequent cases nevertheless confirm that any reference to ‘intolerability’ should not detract from the central concern, which is with the best interests of the child.40 Whilst the presumption should be in favour of sustaining the child’s life, this can be rebutted where to do so is not in the child’s best interests.41 That determination looks like a matter for the doctors and parents to decide together, but clinical opinion seems to carry particular weight,42 notwithstanding the judges’ reassurances that emotional and other features enter the equation.43 According to the Court of Appeal in its judgment in the case of Charlotte Wyatt, the cases may be difficult but the ‘intellectual milestones’ just outlinedare ‘simple’.44 A rational system of law – in the sense defended in Chapter 1 – should yield consistent answers to cases like these. The judges agree: ‘They are always anxious decisions to make but they are invariably eventually made with the conviction that there is only one right answer and that the court has given it’.45 Unfortunately, the decided cases do not wholly support this conviction. The 1997 case of Re T (a minor)(wardship: medical treatment) [1997] 1 All ER 906, in which the court declined to authorise a life-saving liver transplant, is undoubtedly the most starkly aberrant decision, since it appears to be completely at odds with the decision in favour of life in Re B. Two-year-old T was residing abroad with his parents, who concluded that he should not be subjected to the transplant, which would require his mother to return to England with him for the treatment. At first instance Connell J found the mother’s refusal unreasonable but the Court of Appeal rejected this approach and allowed the refusal to stand. As Lyons sees it, ‘The court preferred that the child should have a short happy life rather than (most likely) a significantly longer life involving medical intervention where the level of happiness T might attain would be speculative’.46 He goes on to question whether the decision therefore necessitates a wholesale rethink of the permissibility of liver transplants for children: following the logic of Re B, the message from Re T appears to be that such procedures should cease as they condemn the recipients to ‘demonstrably awful’ lives. Of course, the vast majority of such patients are likely to experience the opposite, that is, a vast improvement in the quality of their lives. Butler-Sloss LJ nevertheless sought to distinguish the two decisions on the basis that T’s surgery would be more severe than Alexandra’s and would require considerable after-care.47 Few find this attempt convincing, with Michalowski pointing out the considerable commitment required by the carers of children with Down’s syndrome.48 Indeed, this is certainly not the logic adopted by Holman J in the 2007 case of A, when he authorised a bone marrow transplant in the face of the parents’ objections that A had endured enough.49 Perhaps safer, then, to view Re T as an anomaly, dictated primarily by the trial judge’s erroneous fixation on the ‘reasonableness’ (or otherwise) of the parents’ position – and no doubt also influenced by the fact that they happened to be ‘trained as health care professionals and are both experienced in the care of young sick children’.50 The apparent conflict between Re C (medical treatment) [1998] 1 FLR 384, in 1998, and An NHS Trust v MB [2006] EWHC 507, in 2006, might not be explained away in such terms. The children in question were suffering from the same condition, spinal muscular atrophy. In each case the parents wanted treatment to be provided, but only M’s parents had the backing of the court. Holman J felt that distinctions could be drawn: C had only months left, whilst M’s prognosis was far less certain and might have meant survival for a few years; M also enjoyed greater awareness than C, and was (again unlike C) wholly reliant on the breathing support. Perhaps, then, there is a case to be made for the different results issued in these cases. Yet, such a conclusion seems to require the drawing of some very fine lines, against which Brownsword advised. The difficulty here, in terms of law’s capacity to guide action, should be apparent: had M’s parents consulted the law books they might well have thought their case was destined to founder; and C’s parents might also have believed they had legitimate grounds for complaint. The doctors, meanwhile, might have had cause to question the practical effect of the decisions here, as well as in cases such as Re B; notwithstanding the judges’ insistence to the contrary, such rulings must amount to ordering the doctors to provide treatment.51 Similar difficulties arise in the jurisprudence as it pertains to incompetent adults. Once again, the best interests test is central and, as developing case law and the Mental Capacity Act 2005 make clear, this will usually (save in the case of the VS patient) involve a balancing exercise comprising not only medical factors but also any wider issues that should have a bearing on the patient’s welfare.52 Here too the judges think that the application of the ‘best interests test ought, logically… give only one answer’.53 But here too their optimism is misplaced. Consider, first, the case which immediately succeeded Bland, Frenchay NHS Trust v S [1994] 1 WLR 601, involving the patient known as S. The Law Lords in Bland were acutely uneasy about the ruling they were making and its possible ramifications, hence their call for parliamentary input and insistence that future such cases be brought before the judges.54 John Finnis wondered what would count as a ‘similar’ case.55 Indications in Bland suggested that some of the Law Lords were resistant to extending their reasoning beyond patients who were irreversibly insensate, and specifically those in a PVS.56 Yet, whilst Anthony Bland’s brain had apparently liquefied, S seemed more like the patient with ‘glimmerings of awareness’ to whom Lord Mustill had referred.57 This did not encourage Bingham MR to authorise the reinsertion of the feeding tube, even though, as he openly admitted, ‘the evidence in this case is not as emphatic and not as unanimous as that in Bland’s case’.58 Nor was the judge too perturbed by the lack of an independent investigation by the Official Solicitor on S’s behalf, despite the ambiguity in the case and therefore the arguable need for such scrutiny. Bingham MR seemed to think the court would not need to be involved in every PVS case but the requirement does remain, even following the 2005 Act.59 Subsequent decisions, accompanied by Practice Notes issued by the Official Solicitor, helped to redirect the judges to the original parameters of Bland, so the case of S is probably (like Re T) something of an aberration. Indeed, the judges have since acknowledged that some early cases of alleged VS might actually have involved MCS.60 However, this is not to deny that other unclear and somewhat contradictory cases have slipped through in the wake of Bland, such as Re D (medical treatment) [1998] 1 FLR 411, in which the doctors could not agree on D’s diagnosis.61 Brown P felt that he was not extending the law in authorising the removal of treatment but, as Caroline Sawyer spotted,62 The Guardian’s typographical error inadvertently conveyed the confusion, when it described D as ‘not not in a permanent vegetative state’.63 There are, of course, many adult patients that the law would consider incompetent whose condition is emphatically not one of PVS. The difficulty with this undeniably large class of patients is that it is not always apparent how decisions should be made in their cases or, indeed, why apparently fine lines should be drawn according to their age (minor or major?) or condition (PVS or not?). A case such as Re R (adult: medical treatment) [1996] 2 FLR 99, from 1996 (thus some considerable time before the Mental Capacity Act 2005), illustrates the potential difficulties. In that case one finds the judge – Brown P again – seeking to apply the jurisprudence pertaining to critically ill infants to a 23-year-old patient ‘in a low awareness state’.64 Now, of course, recourse can be made to the Mental Capacity Act 2005 but this leads one to wonder why the determination of best interests required by that Act should differ from that required by the Children Act 1989? There might well be good reasons for making such a distinction. Nevertheless what should be apparent from the cases so far cited is that the judges (and perhaps also the legislators) have not always successfully defended the distinctions they seek to make. And perhaps these difficulties of application stem from the principle they must apply. It is to the best interests standard, as a legal rule, we should therefore turn. So far we have considered some problems of formal rationality and in particular the inconsistencies inherent in treating arguably comparable cases differently and departing from seemingly established principles. The source of these problems might not be the judges; it might be the principle – or principles – with which they must work. Fuller argued that law must be general, in the basic sense that to qualify as law there must be rules. As Miers and Twining put it, a ‘rule’ can be understood as ‘a general norm mandating or guiding conduct or action in a given type of situation’.65 Does the best interests standard qualify? Returning to the jurisprudence concerning children, the judges might insist that they reach their decisions ‘by the application of an objective approach or test’66 but there is a palpable lack of sufficiently general rules and, in their place, a strong emphasis on the particularities of the case at hand. There is certainly no attempt at providing a definition of the best interests or welfare of any child, let alone one who is seriously unwell. The merits of the law in this area have been commented upon in a lively exchange between Bridgeman, Baines and Lyons, who notably come from different disciplinary perspectives (law for Bridgeman, paediatric intensive care for Baines and Lyons, who at the time of writing were both working in academic bioethics).67 Although the standard initially looks objective and (as Choudhry says) ‘sensible’,68 Bridgeman accepts Baines’s point ‘that there is no one objective, best interests’ but she insists that the courts have ‘established the principles and approach to be adopted, establishing a very clear framework for assessment’.69 She proceeds to list the main messages one can glean from the judges’ aforementioned ‘intellectual milestones’.70 To these milestones we might usefully add Morris’s detection of certain influential factors, such as the judges’ suggestions (in cases like Re C (a minor)(wardship: medical treatment) [1990] Fam 26) ‘that intellectual impairments are seen as especially burdensome’.71 None of this will persuade Baines, however. For him, no matter how purportedly clear the framework, this is not the same thing as identifying an objective standard; it involves describing what will be treated as objective, rather than defining what is actually so. According to Baines, the judges are merely drawing pictures of unicorns, but there will be no way of saying whether their depictions are accurate, because unicorns do not exist. Put more directly: there ‘is no clear calculus’ and nor can there be, given the plurality of ethical world views that can and do exist.72 Baines is scarcely the first person to detect these difficulties; others, like Herring, even invoke the same terminology of the missing ‘calculus’.73 The clearest insight undoubtedly comes from Gillian Douglas, who believes that ‘uncertainty and inconsistency may be both the greatest strength and greatest weakness of the “welfare principle”’.74 Herring picks up on the positive: he concludes that the opposition often levelled at the principle is more likely to be directed at its application in particular cases and contexts.75 The strength surely resides in the flexibility the principle allows, and thus its capacity to respond to a vast array of situations that might arise in the life of a child.76 Of course, the problem with this is that it can allow too much or even too little, and it becomes increasingly difficult to label this standard a ‘rule’ in the sense Fuller would have envisaged. Similar things can undoubtedly be said in the context of incapacitated adults, not least because the same terminology is invoked. Once more, there is no definition and no definitive or exhaustive list of factors to which reference must be made in determining the patient’s best interests. Coggon finds no single test – neither in the cases succeeding Re F (mental patient: sterilisation) [1990] 2 AC 1 nor in the 2005 Act – but instead ‘a bundle of guiding factors rather than an expression of one concept’.77 The courts appear openly to admit this.78 And so, again, the cases are seen as fact-specific, in which conflicting accounts of what is ‘best’ for the patient can be accommodated, and so a multitude of results can be supported; here too, then, the inherent flexibility carries both promise and peril. Given the elasticity of the best interests standard, one is left pondering whether it has the potential to achieve its task as a legal principle, that is, to guide people’s behaviour. Coggon seems to think so: ‘As a construct, it is a system of action guiding principles’.79 However, I can detect numerous ways in which the principle and its progeny fail Radin’s requirements of ‘knowability’ and ‘perform-ability’. The problems essentially boil down to a lack of clarity and predictability. If we start with ‘know-ability’, then this requirement encompasses Fuller’s call for constancy, clarity and consistency, as well as due publicity and congruence between the law-on-the-books and the law-in-action. To some extent these requirements are being satisfied: it is, after all, relatively easy for the law-in-action to comply with an inherently nebulous standard. More substantially, however, the degree of obfuscation should give cause for concern. The problems begin with the differences in jurisdiction dictated by the age and condition of the patient and (it has sometimes seemed) the skills (or whim?) of counsel. Even after the introduction of the Children Act 1989, and with it the welfare principle, some cases continued to come before the High Court under its inherent jurisdiction, according to which the child would be made a ward of court and the decision based on the child’s best interests.80 The point need not be laboured since the distinction between the two jurisdictions and especially the respective criteria could be considered superficial. In Re A (children) (conjoined twins: medical treatment) [2000] 4 All ER 961, Ward LJ quoted Lord Hailsham LC: There is no doubt that, in the exercise of its wardship jurisdiction the first and paramount consideration is the well being, welfare or interests (each expression occasionally used, but each, for this purpose, synonymous) of the human being concerned.81 Nevertheless, there will be important differences between the two, such as the potentially longstanding ramifications of making a child a ward of court.82 To this jurisdictional distinction might be added the purported differences, noted earlier, between the handling of those adult patients in a PVS and those not, despite the apparently uniform use of the best interests test as nowadays interpreted in light of the 2005 Act. More troubling is the lack of clarity about what this test is and what it requires in any given case. It is not insignificant that Charlotte Wyatt’s plight came before the courts on at least 11 occasions.83
4.1 Conflicts in the clinic: to treat or not to treat?
4.2 Taking exception to exceptions: a duty to treat?
4.3 In the patient’s best interests?
4.3.1 Treating like cases alike?
4.3.2 Objective rule(s)?
4.3.3 Know-ability and perform-ability?