5.5 Ethics in paediatric palliative care

A starting point for medical ethics is that there is potentially an imbalance of power between physician and patient that should be redressed by ‘rules of engagement’ setting out how a doctor should behave. That conception encompasses the idea of a moral obligation for persons to be fair to one another, and indicates that one requirement of fairness is that the strong should show special consideration to the weak. It is especially relevant to children needing palliative care, made multiply vulnerable by youth, small physical size, illness, and motor and cognitive impairment.

Ethical codes have elaborated on those basic ideas, through the Hippocratic Oath and principle of primum non nocere to the four-principles approach of Beauchamp and Childress (2009). The focus of medical ethics was how doctors should behave towards patients; the need for it also to account for a responsibility to society is recent.

The relationship between physician and child and family largely defines the scope of paediatric medical ethics. Its function can be expressed in several ways. From the perspective of child and family, it is an attempt to ‘level the playing field’ in relationship with the clinician. To society, it ensures doctors are not isolated from contemporary mores, including notions of how health-care money should be spent. From the perspective of the clinician, however, the task of medical ethics is analytical; to help resolve moral quandary and guide decision-making in patient management through an appeal to what is morally right, in clinical situations where rationality and medical competence alone do not provide a single clear answer.

This chapter sets out current approaches to ethics in paediatric palliative care, suggesting that none is completely satisfactory; a coherent, specifically paediatric account is needed.

Two paradigms dominate contemporary ethical dialogue in children: rights-based arguments (United Nations General Assembly, 1989), and the ‘four-principles’ approach (Beauchamp and Childress, 2009). Rights-based claims are child specific (Goldhagen, 2003; Paul, 2007; Archard, 2010; Alderson, 2011; Streuli et al., 2011) but often not argued from an ethical premise. The four-principles approach flows from understandings of ethical theory, but is not designed for children. Each usefully informs ethical discussions in children, though neither represents a complete account.

Rights-based dialogue provides a legalistic rather than ethical account of how a clinician should treat a child. Nevertheless, the idea that children have rights at all is a powerful and foundational ethical claim. Modern concepts of children’s rights come largely from the United Nations (UN) Convention on the Rights of the Child (CRC) (UN General Assembly, 1989). The Convention is intended to give force to a child-focused legal agenda (Goldhagen, 2003); it does not provide any coherent ethical account of childhood or attempt a moral argument concerning the nature of rights. It invokes the earlier preamble of the UN Universal Declaration of Human Rights (UN General Assembly, 1948) in which rights were held to be self-evident, to apply to any human, and to be inherent in the individual. By applying this definition to children, the CRC insists there is no distinction between the moral values of child and adult. That is an important assertion. Children are often considered ‘works in progress’ whose moral value is as potential, rather than current, persons.

The CRC indicates that children are also distinct from adults in morally relevant ways. Their developmental nature imparts special value. It is their nature to be cared for in a family; society’s responsibility to children is in supporting the family in caring for the child. These claims are in the preamble of the convention, identifying them as foundational.

As a matter of jural logic rights exist as the corollary of duties and must inevitably restrict the influence of states; even democracies (Hohfeld and Cook, 1978; Lazarev, 2005). That power concerns rights sceptics (Hafen and Hafen, 1995; Guggenheim, 2005). A lesser claim for rights (Alderson, 2011) as aspirational statements of moral intent, allows rights-based arguments more helpfully to inform an account of ethics in children.

Thomas Beauchamp and James Childress (Beauchamp and Childress, 2009) famously articulated four principles: autonomy, beneficence, non-maleficence. and justice. Autonomy is the patient’s own capacity to decide what medical treatment is done, beneficence is the responsibility of the clinician to do what is good for their patient, non-maleficence is the duty to avoid harm, and justice the responsibility of clinicians to participate in designing and maintaining a health-care system that is fair.

There may be no obvious way to resolve conflicts between the principles:

Applying the four principles, it is possible to construct:

This is not a shortcoming of the approach, but an illustration of the limitations of its scope. Principlism does not represent a single coherent ethical theory, but a practical heuristic for busy clinicians. The four principles represent a practical summary of complex philosophical deliberations, for use ‘in the heat of battle’, rather as a formulary summarizes practical prescribing but does not replace a sound knowledge of pharmacology.

In paediatric palliative care, however, there are important weaknesses in principlism as a result of the failure of some of the principles that lie behind it to account consistently for the nature of children. Autonomy is problematic in relation to children, accounts of beneficence and non-maleficence do not consistently distinguish between the interests of children and those of families, and establishing justice is complicated in a society composed mainly of adults. The status of the child in principlism is often unclear. Children ideally need a coherent ethical approach that is consistent with their nature as moral beings that are both distinct from, and have much in common with, adults.

Personhood is often held to be a function of rationality, and the status of children in many ethical theories remains uncertain. Children probably rationalize differently from adults, but that is only determinative of rationality if the definition is limited to the way adults do it. Since even some non-human animals reason (Glackin, 2008), it seems unlikely that adult rationality is the only form of rationality there is among humans. In his book Practical Ethics, utilitarian philosopher Peter Singer (2011) asserts that infanticide is permissible. The infant, he says, lacks reason and therefore awareness of self over time. Being killed cannot rob the infant of anything she can truly be said to possess. Given modern understandings of infant neurodevelopment (Gopnik, 2007; Bauer, 2008), Singer’s claim is improbable. The idea that infants cannot reason is perhaps a contemporary parallel to earlier ideas that they cannot experience pain; a convenient and widely held belief about children that simplifies decision-making for adults but is ultimately untrue. Nevertheless, the moral status of children who are genuinely non-rational (such as those with severe cognitive impairment) presents paediatric clinical ethicists with a challenge.

Autonomy is often taken to mean freedom, in the sense of having no restrictions on actuating one’s will. In practice, a child is multiply restricted in the influence he or she can exert. Many limitations are inherent to the nature of the child; children are physically small, and often lack verbal skills simply by virtue of their age. The child given less influence because she is unimportant inevitably becomes increasingly unimportant because she cannot exert influence, and lack of autonomy is self-perpetuating.

Properly understood, however, autonomy describes the freedom to decide for oneself which set of restrictions to acknowledge. That more realistic definition does not exclude children so clearly since children often manifest autonomy by allowing parents to make decisions on their behalf, seeing themselves in a reciprocal and balanced caring relationship with family members (Carnevale et al., 2008). Ambiguities in the nature of children’s autonomy are particularly significant in palliative care. Cognitive impairment and severely restricted physical mobility are usual. Expectations that parental autonomy will be respected, as well as that of the child, interpose a layer of ethical complexity. Even caring parents may, rationally or otherwise, choose what is less than ideal for their children (Brierley et al., 2013). Disentangling what parents feel their child would want from what parents want for their child is morally relevant, but not always possible.

Limitations on understanding autonomy are relevant in considering the ethics of euthanasia and physician-assisted suicide legislation that extends into childhood or young adulthood. Adolescents may appear more autonomous than they are. A perfectly competent and rational adolescent may request euthanasia for reasons that have little to do with a considered desire to end life. The high frequency of risk-taking behaviour and suicide during adolescence (Self and Findley, 2010) illustrates that an actual desire to die is only one among many influences. Even if it is conceded that people who ask for it should be killed, it is not clear that it is possible, in practice, to distinguish clearly between an adolescent’s request for death and (for example) a request for more attention, or a protest against parental restrictions. The assumption that there is always a straightforward relationship between what people say, and what they actually mean, is even less reliable in adolescents than it is in adults.

The concept of autonomy, then, is foundationally problematic in children. Autonomy requires some degree of control over one’s own life. For all children, it is more limited than it is for adults, for various reasons of which some have no moral relevance. It is over-simplistic to suggest a child inherently lacks autonomy, and ethics therefore need account only for the wishes of parents. Nevertheless, a child is inherently in relationship with a family. To deny the importance of that relationship, and the authority it gives to parents to speak for their child, would be to deny something central to the ethical questions clinicians need to ask about children.

Some sceptics of the concept of children’s rights argue that it creates a separation between the interests of child and family that is spurious (Hafen and Hafen, 1995; Guggenheim, 2005). While that overstates the case, it is true that a child’s moral deliberation normally occurs in a context in which family is extremely influential. The child will find it hard to articulate views of interests that differ from strongly expressed views of parents, potentially allowing coercion. Where differences are expressed, the medical team may need to arbitrate between them, or else privilege one over the other.

The concept of ‘best interests’ is especially problematic in children who are cognitively impaired. There are no objective measures of a good outcome. Survival probability is typically irrelevant, since death in childhood is expected. Other measures are subjective, requiring the child to be verbally and conceptually capable of articulating issues that relate to existential as well as physical components. In practice, this is difficult, and clinicians are often drawn instead to an account of ethics that allows us to regard a child’s parents’ interests as inseparable from those of the child. It is perhaps as a reaction to this problematic assumption that the modern children’s rights movement has developed.

It is important to individual families to know that a health-care system is fair, and to that extent, justice is appropriately considered one of the four medical ethical principles. In other respects, it appears out of place. Where autonomy, beneficence, and non-maleficence all relate primarily to decisions made in the context of a relationship between an individual child and an individual doctor, the principle of justice refers instead to ‘states of affairs’ (Smart and Williams, 1973). Typically, justice arguments are rooted in utilitarianism; that the ethically right decision depends on its impact on the whole community. While the other three principles focus on the needs of an individual patient, justice makes individuals subordinate to the needs of society.

Few clinicians would disagree with the conclusion here that the child should not be ventilated. The ethical reasoning also appears sound. It is, however, cruel, unreasonable, and coercive to expect parents to sacrifice their child to the well-being of others (American Academy of Pediatrics, 1996). Such an expectation flows from an ethical understanding that is hard to argue convincingly without ignoring the child’s nature as a dependent being in the care of a family. Relationship has moral relevance in ethical deliberation around specific quandaries that is not easily accommodated by most forms of utilitarian consequentialism.

The relationship of physician and child is also of ethical relevance Bartholome, (1985, 1996; Ramsey, 2002). In considering how to design a fair system of health care, it is a legitimate ethical starting point that all patients, real or hypothetical, are equally important. In the case described earlier, however, considering the impact on hypothetical ‘other children’ is outside the scope of this physician’s relationship with this particular child.

Clinicians may need to make ethically relevant decisions in both modes. They will need to make some decisions in the context of their relationship with individual patients, and others in developing or maintaining fair systems. This is not inconsistent. It may even be ethically necessary; tension between the needs of the individual actual patient and those of hypothetical other patients is inevitable given limited health-care resources, and dialectic is integral to a consistent ethical approach, not antithetical to it.

Deontological arguments flow from an assumption that there is an absolute morality that is equally relevant to all people, in all situations, irrespective of the consequences. There at least three possible sources for an absolute moral code. One is the authority of God through religious scripture. Even in secular societies those codes are, it can be argued, authoritative at least to the extent that they are rational. Most scripture was, however, not designed primarily as an exposition of a consistent ethical code, and is often elliptical and/or subject to interpretation. Ethical accounts derived from religious understanding are rarely completely agreed, even among those who share that faith.

A second source of an absolute moral code might be the universe itself. Ancient and medieval understanding was that the universe was rationally ordered, and coherence with that rational order was the yardstick by which the rightness of behaviour should be measured. So, for example, the ethical rules by which children should live and by which parents should care for them both derive from the inherent nature of family.

A third authority might be that of pure reason. Kant (1785/2005) attempted to show that there were certain ethical premises that were beyond debate because for reasonable people they were logically inescapable. Some, he argued, were always true, irrespective of context (categorical imperatives) while others were contingent on the individual’s prior beliefs about context (hypothetical imperative). Applying this to children is not straightforward; Kant considered the moral law to be co-extensive with rationality, and since he does not obviously consider children rational it is not clear in Kant whether a child has valid status as a moral being.

In practice, the outcome of an ethical decision can be worse by obeying a rule than by breaking it. Consequentialists argue that this is an inherent weakness in ‘rule-worship’ itself, and that the only morally relevant aspect of an action is found in its consequences (Smart and Williams, 1973). Consequentialists differ about what outcomes should be measured. Libertarian consequentialists judge the moral ‘rightness’ of an action by the extent to which it contributes to freedom. In medical ethics, utilitarianism is the best-known form, and the most widely recognized outcome measure is the effect on ‘the sum total of human happiness’.

Some important ambiguities restrict the usefulness of utilitarianism in practice. The sum total of happiness could relate to a great deal of happiness on the part of a few people, or moderate happiness on the part of many. The nature of happiness itself is not agreed, and may not be measurable. Evaluating happiness is particularly problematic in children needing palliative care, who are often non- or pre-verbal and cannot articulate feelings or preferences. A close correlation between happiness and observable function or cognitive ability is sometimes assumed (Shaw, 1977) but there is little reason or evidence to indicate that life quality is influenced in any readily predictable way by physical condition, or closeness to death.

Utilitarianism does not concern itself with the happiness of the individual, except insofar as it contributes to the sum total of human experience. A central principle is that the happiness of each individual person is equally important—or, perhaps more accurately, equally unimportant—to the agent making the ethical decision. A paradox in most forms of utilitarianism is that utilitarians can be concerned for the total happiness of society, while remaining indifferent to the happiness of all its individual members.

Finally, utilitarian consequentialism does not acknowledge any ethical relevance for intention. This again makes it problematic for most quandaries in medical ethics. A doctor trying to decide on a particular course of clinical action has to decide whether to do a particular act or not to do it. The quandary defines, and is defined by, the agent’s intention. Utilitarianism does not provide a mechanism for that decision, because it claims that ethical ‘rightness’ depends on something entirely isolated from intention; that is, outcome. The physician cannot know the outcome at the time the decision needs to be made. She can only make a judgement on the basis of knowledge and prior experience. That informed judgement is what constitutes intention in the context of medical decision-making. The fact that utilitarianism cannot accommodate intention—still less complex intentions as in double effect—illustrates an important limitation of its scope.

Although deontology and consequentialism might appear to be opposites, they have two important things in common. Both are ‘absolutist’ approaches—they claim unlimited scope to provide moral values that are always and everywhere right. Both also ultimately elevate something abstract over real individual human beings. In the case of deontology, it is a concept of duty. In the case of utilitarian consequentialism, it is a concept of happiness or some other descriptor of persons. Since neither duty nor happiness can have any meaning independently of individual persons, neither deontology nor consequentialism can, on its own, provide an adequate account of medical ethics, which relates primarily, though not exclusively, to decisions in the context of relationships between individual persons.

In contrast, virtue ethics (Macintyre, 2007) has its roots in Aristotle’s understanding that an action is right if it is done by a virtuous person for the right reason. Aristotle locates the quality of virtue, not only in the ethical rightness of an act, but in the character of the moral agent whose act it is. Virtuous people, and ethically correct acts, are mutually defining in the same way that force and acceleration are mutually defining in Newton’s second law of motion. One cannot argue from virtuous act to virtuous person; it is simply impossible for either to exist independently of the other.

An understanding of the ‘virtuous doctor’ emerges from the Hippocratic oath and underlies professionalism; that is, what constitutes good medical practice (General Medical Council, 2006). To be professional is to behave in a certain way that is, as a corollary, the way in which a professional behaves. Virtue ethics provides a coherent and consistent account of ethical behaviour that most other theories cannot, but it does not provide an obvious decision mechanism for resolving individual ethical quandaries. Virtue ethics can therefore shape expectations of professional behaviour from doctors, but may not in its own replace an analytical approach to clinical ethics.

The principle of double effect (PDE) is an acknowledgement of the possibility that a clinician can foresee a range of outcomes from a single act, but intend only a subset of them. It is logically inescapable, and the basis on which good clinical decision-making is made in many specialties, including palliative medicine (Fohr, 1998). Despite its apparently self-evident nature, the PDE is considered by some simply to be a ‘fudge factor’ that allows physicians to feel comfortable making decisions at the end of life that, in reality, hasten or cause death. There are several reasons for this, of which this chapter will consider only two: it is sometimes misunderstood, and its self-evident first premise requires certain further qualifications that are not agreed. Those include, in particular, conceptions of proportionality, and the extent to which a single act can be defined within, and reliably distinguished from, a series of acts.

A common misunderstanding of the PDE is that it legitimizes equally all outcomes that are foreseen. In the earlier example, relief of pain is ethically permissible, while inducing apnoea is not.¹ It is sometimes thought that the PDE permits the overdose because the ethically problematic outcome is simply a secondary consequence of treating pain. The PDE does not allow this, however, for an important reason; the dose that is suggested is disproportionately high for the ethically permissible outcome. The dose required for pain in an opioid naïve child would be a tiny fraction of this (1 mg/24 hours). The dose suggested cannot be justified on the grounds of the permissible outcome.

The illustration shows that some element of proportionality must be shown between the act and its outcome in order to judge the intention. In palliative medicine, that proportionality can be considered in many ways. One is the reasonableness of the dose. Prescribing morphine analgesia is ethically justified only to the extent that the dose is appropriate for the degree of pain (Hain and Friedrichsdorf, 2012).

A second measure of proportionality is to consider ‘directness’ of the intention. Any action leads to a series of consequences that are increasingly remote from the agent’s intention. It is not, as Glover humorously suggests (Glover, 1977, p. 89), a question of how long after the act its consequences manifest. The question of directness relates rather to the disposition of the agent; what result the physician expects from his or her action. Except to purist consequentialists, there is a morally relevant difference between intending A while recognizing that B is a consequence (that is, willing A but permitting B), and intending A and B equally. Temporal remoteness may be incidentally relevant, especially if it allows other agents to intervene to change the consequence as a result of their own intentions, but it is not of itself a measure of directness.

Proportionality can also be seen as a function of probability. Apnoea is likely at the dose suggested in the earlier example, but not inevitable. Pancuronium, however, guarantees it; a physician prescribing pancuronium during extubation intends apnoea and the PDE does not apply.² Conversely, anaphylaxis is a foreseen event for most drugs, but its rarity makes it obvious that a physician does not usually intend to cause death by prescribing it. Probability, as well as reasonableness and remoteness, are relevant to intention.

Problematically, the PDE demands that a single act can always be considered the sum of several smaller acts that are discrete. It is hard to find a valid basis for such analysis. The act of giving morphine could be said to begin with the injection, in choosing not to give naloxone to reverse it if apnoea occurs, or it could be considered inseparable from the overall act of caring for a dying child. According to utilitarian consequentialist understanding, the PDE is meaningless since it gives different ethical value to acts with the same outcome. Although utilitarians therefore dismiss the PDE as without moral meaning (Glover, 1977; Harris, 1985), it is equally rational to suggest that this is a further limitation of consequentialism in medical ethics.

The fact that application of the PDE in practice is not always straightforward does not, however, undermine its central thesis. It is a matter of everyday reality for paediatric palliative care physicians that it is possible to foresee consequences of an intervention that are not its aim. It is inconceivable, without re-evaluating the whole task of medical ethics, that the intention of the physician is not relevant. The ethical ‘rightness’ of an intervention with such multiple outcomes must rely on some concept of proportionality, measured as reasonableness of the dose, remoteness of the intention, or probability of a consequence.

The Royal College of Paediatrics and Child Health (RCPCH) (2004) in the United Kingdom recognizes several archetype situations in which medical treatment aimed at sustaining or saving life might be withheld. The guidelines represent a practical tool for considering the quandary, rather than a coherent ethical approach. Withholding and withdrawing are not the same decision, but if both serve the best interests of the child they are logically equivalent ethically.

Withholding or withdrawing treatment from the child who is brain dead or in a permanent vegetative state is permitted, not because the child is no longer of value, but because she no longer has interests and the residual interests are those of others, so that a beneficence argument need take only those into account. This situation therefore permits withdrawing or withholding, but it does not mandate it; for example, the family’s interests may be best served by continuing. The justification for withdrawing in the no chance situation is not merely that the child cannot recover but, again, that the child’s interests are best served by not prolonging treatment.

The same can be said of the unbearable situation. The term ‘unbearable’ is inherently subjective. There is an ethical duty on those caring for a child to explore with him what he feels about what is going on. That is not the same as securing the child’s permission for treatment to be withdrawn. Although it is helpful if the child expresses a clear view, the unbearable situation does not preclude the possibility that carers will need to act as advocates, as well as proxies, for the child.

The no purpose situation appears to presuppose a provisional value for the child’s life. What follows clarifies, however, that the purpose in question is that of treatment; that is, the benefit of that individual child. The no purpose situation is, again, a special example of beneficence, but unlike the unbearable situation it assumes the child cannot participate in the discussion. The responsibility of carers is therefore a little different; rather than giving a voice to his wishes, it is to make a wise and compassionate decision about what is best for the child.

Palliative care foundationally recognizes that death should sometimes be allowed to occur, even where it is technically possible to delay it. All the RCPCH ‘situations’ are best considered illustrations of the same principle of balancing burden and benefit, distinguished by what is at stake, whose interests are relevant, and the extent to which the child’s own estimate of ‘bearability’ can be solicited. The only indisputably relevant ethical question in deciding whether treatment should be offered to or withheld from a child is whether the intervention is likely to do more good than it does harm to that individual child. That balance needs to take into account much more than the duration of life, including dimensions that are difficult to measure and/or are inherently subjective. A child’s disability is relevant to the extent that it impacts on those dimensions.

This chapter has outlined some theories and paradigms in ethics of care at the end of life, and some special challenges of considering them in children’s palliative care. Paediatric palliative care clinicians face all the moral ambiguities of their colleagues working with dying adults, and often have to turn for ethical guidance to sources that were not intended for that purpose. Of those, the most important currently are the four principles approach, and the UN CRC. The first was developed for adults and is often unsuitable for children, particularly because of its emphasis on autonomy. The second is a series of axiomatic rights claims that, while they make some important fundamental ethical claims about the nature of children, are not obviously justified philosophically. Neither on its own provides the kind of coherent account needed by those working in children’s palliative care.

Because it does not accommodate the moral relevance of the relationship of clinician and child, the scope of utilitarian consequentialism (in its purest sense) in paediatric palliative care should be limited to development of fair systems, rather than to specific clinical decisions about individuals. Children are vulnerable to the misapplication of utilitarian ethics to individual patient quandaries because they are typically unable to express clearly what they want and their removal from the world, particularly as infants, does little to detract from the sum total of human happiness.

Many of the specific ethical quandaries faced by clinicians working with dying children have their counterparts in the adult specialty. It seems likely that, as it becomes technically possible to resuscitate children at ever earlier stages in fetal development, and to keep children alive but suffering for ever longer periods of time, it will also become increasingly important to be able to address those quandaries in a way that is specifically paediatric. To do so requires that we do more than extrapolate from existing accounts such as rights or principlism; paediatricians and bioethicists must develop an ethical account of childhood itself that is consistent, coherent, and complete.

The author is most grateful to Dr John Perry, Lecturer in Theological Ethics School of Divinity, University of St Andrews, for his comments on the manuscript.