Yang, Hannah --- "Acts, omissions, and the doctrine of double effect: the law and ethics of end-of-life choice in New Zealand and the United Kingdom" [2018] NZPubIntLawJl 14; (2018) 5 PILJNZ 97

	Home \| Databases \| WorldLII \| Search \| Feedback Public Interest Law Journal of New Zealand

Acts, Omissions, and the Doctrine of Double Effect: The Law and Ethics of End-of-Life Choice in

The current law relating to end-of-life choice in both New Zealand and the United Kingdom is such that one may refuse life-saving treatment, have life-sustaining treatment withdrawn, or have painkillers administered with the effect of hastening death. It is not legal to provide assisted dying. A survey of case law from both jurisdictions shows that ultimately two philosophical doctrines underpin this position: the act/omission distinction and the doctrine of double effect. Neither of these, however, is ethically sound. On closer examination, the act/omission distinction has no moral significance because any perceived moral differences between acts and omissions are really caused by a confusion between the omission itself and other factors commonly arising with omissions. The doctrine of double effect is equally problematic, as it is both inconsistent with the way criminal law handles intention regarding foreseen avoidable consequences, and misleading in the way it characterises such intention. As a result, the line drawn by the law on end-of-life choice is not adequately justified from a moral standpoint.

I Introduction

The purpose of this article is to investigate where the current law draws the line on various end-of-life choices and to examine the philosophical underpinning of the position. More specifically, I ask whether the acts/omissions distinction (AOD) and the doctrine of double effect (DDE) are ethically sound bases for the law on end-of-life choice.

* I am grateful to Professor Jo Manning and Professor Tim Dare for their feedback on prior drafts of this paper.

The rest of Part I deals with some preliminary matters surrounding definitions. In Part II, I give an overview of the current legal position in New Zealand and the United Kingdom on various end-of-life options. In Part III, I look more closely at the case law and show that in both jurisdictions, courts essentially appeal to two doctrines to justify the legal position: the AOD and the DDE. Finally, in Part IV, I evaluate the ethical soundness of these doctrines. Ultimately, I will conclude that neither of them is adequate to morally justify the law as it currently stands.

A Definitions

It is important to first define the ambit of this article. The title refers to end-of-life choice. By this, I mean decisions that lead to the death of patients suffering from serious and irremediable illness or disease, where the decision is for the patient’s benefit and where a second person other than the patient (presumably medical staff) is involved. I do not take this to be controversial. The illness component is to confine this discussion to a medico- legal context. The patient-benefit component is required to prevent the discussion spilling into other incidents of death such as those motivated by malice. Finally, the second person component differentiates it from suicide, where no other person is involved.

Occasionally, I will also use the term euthanasia, although for reasons that I hope will

become clear at the end of Part II, nothing should turn on this usage and where I do it will be for convenience’s sake only. Not only has the term fallen out of favour in end-of-life discussions, it is also an emotively loaded word, the precise definition of which is difficult to pin down without assuming those very distinctions this article intends to discuss. To frame it as the intentional bringing about of death, for example, automatically subjects it to the DDE; to call it killing implies a distinction from letting die; and to require a voluntary action would subject it to the AOD.¹As this article will tackle these issues of intent and actions directly, whether any option amounts to euthanasia is irrelevant, and I will not attempt to lay down a definition.

Nevertheless, it will be helpful to clarify some other terms. Euthanasia may be divided into three categories of voluntariness: voluntary, involuntary, and non-voluntary; and into two categories of activity: active and passive. Voluntary euthanasia is where a competent patient requests death; involuntary euthanasia occurs where a competent patient does not agree to death; and non-voluntary euthanasia applies to an incompetent patient who is unable to consent either way.²Active euthanasia generally describes death brought about by a positive action, for example the administering of a lethal injection, while passive euthanasia describes death brought about by an omission, for example by withholding treatment. Abstaining from resuscitating a patient in a persistent vegetative state where there is no advance directive, therefore, might be termed non-voluntary passive euthanasia; complying with a terminally ill patient’s competent request for a lethal injection termed voluntary active euthanasia, and so on.

II Current Law

The law in New Zealand on end-of-life choice essentially mirrors that of the United Kingdom. Both jurisdictions allow for three options which terminate life, outside of suicide.

2 Dan W Brock “Voluntary Active Euthanasia” (1992) 22(2) Hastings Centre Report 10 at 10–11.

The first flows from the right to refuse treatment; the second is the withdrawal of life support from an incompetent patient; and the third is the administering of a lethal dose of pain relief in the context of palliative care.

A Refusing treatment

At English common law, every competent adult patient has the absolute right to refuse consent to treatment, even if it may lead to their death.³This is irrespective of any reason for the refusal or lack thereof.⁴Practically, it is open to question how well this right is preserved, as ultimately it is up to the courts to determine whether a patient had capacity to refuse consent at all,⁵but this is an issue which falls outside the ambit of this article. Suffice to say that given the patient is competent, they will be able to refuse life-saving treatment.

Patients may also make advance decisions to refuse treatment under the Mental Capacity Act 2005 (UK).⁶Where an advance decision to refuse treatment is valid under s 25 of the Act, medical staff will not be liable for the consequences of withdrawing or withholding treatment to which the advance decision applies.⁷

In New Zealand, the right to refuse treatment is codified in three places. First, right 7(1) of the Health and Disability Commissioner (Code of Health and Disability Services Consumers’ Rights) Regulations 1996 (the Code) provides that services may only be provided where the patient “makes an informed choice and gives informed consent”. Secondly, the corollary of this right is manifest in right 7(7), which states that every person has the right to “refuse services and to withdraw consent to services”. This right, like that in the United Kingdom, can also be exercised as an advance directive under right 7(5) of the Code, “in accordance with the common law”.⁸Lastly, s 11 of the New Zealand Bill of Rights Act 1990 provides that “[e]veryone has the right to refuse to undergo any medical treatment.”

“Everyone”, however, does not mean literally anyone. Consistent with the Code, courts have interpreted the provision to apply to every person who is competent and has the capacity to make a rational and informed decision.⁹But provided this condition is satisfied, it would be lawful for doctors not to provide treatment to a patient who refuses consent, even where the absence of treatment would result in death.¹⁰The Court in Department of Corrections v All Means All, for example, declared that it would be lawful not to provide a hunger striking prisoner medical treatment as long as he does not give informed consent or there is an advance directive refusing consent, despite it being clear that such an omission would lead to the prisoner’s death.¹¹

10 It must be noted that this applies to competent adult patients. Where parents refuse to give consent for life-saving treatment for children, the courts have demonstrated a willingness to override parents’ wishes and place children under their guardianship so they may receive that treatment. See Re A [2013] NZHC 2154, [2013] NZFLR 451; Re M [2012] NZHC 1563, [2014] NZFLR 381; and Waikato District Health Board v L [2009] NZFLR 83 (HC).

B Withdrawing treatment

In both jurisdictions, where the patient is incompetent and no advance directive has been left, life support may still be lawfully withdrawn in certain circumstances.

In Airedale NHS Trust v Bland, a case which concerned the withdrawing of life support from a patient in a persistent vegetative state, the House of Lords held that such treatment could be withdrawn because its continuation was futile. ¹²As treatment was no longer beneficial to the patient, it could not be said that doctors had a duty to continue the treatment, and therefore discontinuation would not be unlawful.¹³Where treatment may provide some benefit, relevant considerations must be weighed to determine what is in the best interests of the patient.¹⁴

This best-interests test has now been codified in the United Kingdom under the Mental Capacity Act. Under s 1(5), any decision made for a person who lacks capacity must be made in their best interests.¹⁵What is in the patient’s best interests is to be determined with regard to “all the relevant circumstances”,¹⁶including:¹⁷

In some cases, it may well not be in the patient’s best interests to continue receiving life- sustaining treatment.¹⁸This was the decision reached in both Aintree University Hospitals NHS Trust v James and Re Briggs (No 2).¹⁹

In New Zealand the test is similar, although it is not codified. Initially, the Court in Auckland Area Health Board v Attorney General laid out four specific requirements for when withdrawal would be lawful:²⁰

These requirements were later curtailed and the scope for lawful withdrawal of life support from incompetent patients broadened in Shortland v Northland Health Ltd.²¹There the Court held that neither the third nor fourth requirement could be mandatory in

all cases.²²Whether consultation with an ethical body ought to be a requirement should depend on the facts of the case, and it would not be necessary in one which raised only medical, not ethical issues. ²³Consent of family, on the other hand, could not be a requirement for cessation of treatment as it would amount to giving family the power to require treatment.²⁴The present test, therefore, comes down to what is in the patient’s best interests and what is in accordance with good medical practice.

C Administering pain relief

In New Zealand, hastening the death of someone already suffering from an existing disease would usually constitute a killing for the purposes of culpable homicide under s 160(2) of the Crimes Act 1961.²⁵It was held in Seales v Attorney General, however, that acts or omissions in the context of palliative care which accelerate death would not necessarily be criminally culpable.²⁶Collins J held that a doctor will not be liable for the patient’s death resulting from a lethal dose of pain relief if the doctor’s intention in administering the dose is to provide palliative relief.²⁷This is subject to the condition that the administration is “reasonable and proper” for that purpose.²⁸

In coming to this conclusion, Collins J cited the “established rule” in England referred to by Lord Goff in Bland that a doctor caring for a dying patient may “lawfully administer painkilling drugs despite the fact that he knows that an incidental effect of that application will be to abbreviate the patient’s life”.²⁹This was because such an administration could “properly be made as part of the care” of the patient.³⁰

Hence in both jurisdictions, doctors will not be criminally liable for hastening a patient’s death where the administration of painkillers is for the purposes of palliative care.

D Assisted dying

Despite the above, it is not lawful in either the United Kingdom or New Zealand to assist in a patient’s death by administering a lethal injection or by giving the patient the means to administer such an injection.

In the United Kingdom, assisting suicide is an offence under s 2 of the Suicide Act 1961 (UK).³¹At common law, the English Supreme Court in R (Nicklinson) v Ministry of Justice also affirmed the position that “mercy killing” amounts to murder or manslaughter, regardless of whether it was done out of compassion and whether the person had consented to their death.³²

In New Zealand, Collins J held in Seales v Attorney General that a doctor administering a lethal injection would commit an unlawful act for the purposes of culpable homicide

29 Bland, above n 12, at 867 as cited in Seales v Attorney General, above n 26, at [105] per Collins J.

under s 160(2),³³and that the s 63 bar on consenting to death would prevent the patient’s consent from providing a lawful excuse. ³⁴At the same time, a doctor who does not administer the injection but supplies a patient with a lethal drug knowing they intend to use it to end their life would be aiding and abetting suicide under s 179, regardless of whether the patient’s decision was rational or not.³⁵

E A puzzle in the law?

To summarise, then, it might be said that in both jurisdictions voluntary and non-voluntary passive euthanasia are legal in limited circumstances, alongside a limited form of voluntary and non-voluntary active euthanasia through administering pain relief. Voluntary active euthanasia by lethal injection, however, is not legal.

One may, of course, object to this classification. Refusing treatment, withdrawing life support, and providing pain relief, one might argue, are not instances of euthanasia at all. Here we run into the problem of definitions referred to in Part I, as whether any of these constitute euthanasia will depend on how the term is framed. The underlying assumption in the objection seems to be that there is something inherently immoral in euthanasia, whatever it means, hence my desire not to attach the term to existing legal options. This is precisely the sort of emotive baggage I wish to avoid. The moral value of an activity does not depend on its label, but rather on the content of the activity itself, and it is this alone on which we must base our discussion. In other words, there is no magic to a term outside of the meanings we ascribe to it. As this article deals directly with the various end-of-life options by their content, to be further preoccupied with the adequacy or inadequacy of labels is apt to confuse.

Taking these end-of-life options by content alone, then, a puzzle seems to arise. The line drawn between what is legal and what is not appears to depend on arbitrary distinctions, or at the very least leads to some counter-intuitive results. A competent patient with a terminal illness and labouring under extreme pain may procure their own death by requesting withdrawal of their ventilation, while another equally competent patient in equally unpleasant circumstances would not have the same option if it so happened that their condition was one which required no ventilation. Furthermore, once a decision has been made to withdraw life support from an incompetent patient, it would seem irrational that death cannot be allowed to take place sooner rather than later. As Lord Browne-Wilkinson put it in Bland:³⁶

How can it be lawful to allow a patient to die slowly, though painlessly, over a period of weeks from lack of food but unlawful to produce his immediate death by lethal injection, thereby saving his family from yet another ordeal to add to the tragedy that has already struck them?

III Acts, Omissions, and the Doctrine of Double Effect

The purpose of this Part is to provide a descriptive legal answer to the issues raised at the end of the previous Part. We have seen that it is lawful to refuse life-saving treatment, to withdraw life-support from an incompetent patient in certain circumstances, and to administer a lethal dose of painkiller for palliative care, but not lawful to procure the patient’s death by lethal injection. Two explanations arise, and I will consider them each in turn through an analysis of case law.

A The AOD

Before beginning I wish to note the difference between the legal distinction and the moral distinction between acts and omissions. The moral distinction states that one is less morally responsible for an omission than for an act.³⁷The legal distinction is merely the fact that the criminal law generally regards the actus reus as being satisfied by positive acts only, whereas omissions will not satisfy the actus reus unless there is a specific duty to act.³⁸The legal distinction may be justified by the moral distinction, but conceptually the two are separable. In this Part, I refer solely to the legal distinction for the purposes of this descriptive exercise.

The legal AOD plays a crucial part in drawing the line between voluntary active euthanasia and other end-of-life options. In New Zealand, the behaviour element of offences is generally satisfied by a positive act only. Only in exceptional cases where there is a duty imposed will an omission to perform that duty result in criminal liability.³⁹These legal duties are found in ss 151–157 of the Crimes Act. In this way, it is possible to commit a homicide by omission. Section 160(2)(b) provides that culpable homicide can be a killing by omitting to perform a legal duty. A doctor may therefore commit a culpable homicide by withdrawing life support from a patient, causing their death, where they had a duty under s 151 to provide necessaries to that patient.

Omissions, then, are a much narrower category and therefore their requirements are more difficult to satisfy compared to acts, for liability depends on whether there is a legal duty. If the doctor is found not to have had a duty, then their omission will not be an unlawful one and so not amount to a culpable homicide. Where there is a positive act, however, it will amount to culpable homicide as long as the act is unlawful.⁴⁰

One might object that there is no discrepancy in standards between acts and omissions, for an act must still be “unlawful” and therefore goes through the same sieve, so to speak, as an omission. The number of duties that exist, however, is much more limited than the number of prohibited acts. “Unlawful act” is defined by the Crimes Act to mean “a breach of any Act, regulation, rule, or bylaw”.⁴¹On this basis, it was held in Seales that a specific offence under the Crimes Act would not be needed to establish an unlawful act for the purposes of culpable homicide under s 160(2)(a).⁴²The scope for committing

37 Ann-Marie Begley “Acts, omissions, intentions and motives: a philosophical examination of the moral distinction between killing and letting die” (1998) 28 Journal of Advanced Nursing 865 at 866.

38 AP Simester and WJ Brookbanks Principles of Criminal law (4th ed, Brookers, Wellington, 2012) at 45.

an unlawful act is therefore much broader than for an omission. To extend the metaphor, both acts and omissions may go through a sieve, but they certainly are not of the same mesh.

Given this, it is clear that the classification of a particular course of action as an act or omission may be crucial for liability. In the United Kingdom, the Court in Bland has classified the withdrawing of life support as an omission.⁴³While Lord Goff acknowledged that removing life support seemed to be better described as a positive act, he considered that viewed from a larger scale, it was no different from omitting to provide life support at all.⁴⁴Lord Browne-Wilkinson was of a similar opinion, noting that the same result could be achieved by installing a timer on the machine which required it to be reset every 12 hours, the failure of which must no doubt be considered an omission.⁴⁵As such, it was found that it would be legal to withdraw life support from Anthony Bland because there was no duty to provide him with continuing life support. ⁴⁶By contrast, Lord Browne-Wilkinson confirmed that “the doing of a positive act with the intention of ending life is and remains murder”.⁴⁷This position was adopted by the Court in Nicklinson where it was held that voluntary active euthanasia would amount to murder or manslaughter.⁴⁸

In New Zealand, the problem of categorisation has not been explicitly discussed, although it is evident that the same position has been implicitly adopted. In Auckland Area Health Board, a case concerning withdrawal of life support, part of the s 160 culpable homicide analysis focussed on whether the withdrawal would be a breach of the then s 151(1) duty to provide necessaries of life.⁴⁹The question was whether the doctor was under a duty to continue providing life support.⁵⁰It was therefore taken for granted that the withdrawal would be an omission, not a positive act. On the other hand, the discussion in Seales focused, among other things, on the meaning of “unlawful act”,⁵¹with Collins J finding that active voluntary euthanasia by lethal injection would amount to two offences, either assault under s 196 or administering poison with intention to cause grievous bodily harm under s 200.⁵²There is therefore implicit acceptance in New Zealand of the legal AOD when it comes to drawing the line between withdrawing life support on the one hand and assisted dying on the other.

B The DDE

Alongside the AOD, the DDE is also employed by the courts to justify the legal position. As a general principle, the DDE states that where a course of conduct has good and bad outcomes, the conduct is justifiable if the good outcome is what is intended, and the bad

43 Bland, above n 12, at 866 per Lord Goff, 881–882 per Lord Browne-Wilkinson and 898 per Lord Mustill.

46 At 858–859 per Lord Keith, 869 per Lord Goff, 885 per Lord Browne-Wilkinson and 898 per Lord Mustill.

outcome merely foreseen but not intended.⁵³The focus, therefore, is on the intention behind the course of action.

The effect of this doctrine is most keenly demonstrated in the matter of administering lethal doses of painkiller. Here, the AOD does not help, for such an administration cannot properly be considered an omission even if we accept that the withdrawal of life support can be. The only difference between the withdrawal of life support and the illegal administration of a lethal injection is in the doctor’s intention. The DDE was explicitly applied by Collins J in Seales, where he said:⁵⁴

... there is a morally relevant distinction between the intentional effects of a person’s acts or omissions and the unintended, though foreseeable, effects of those actions. ... Thus, if Ms Seales’ doctor were to administer a lethal dose of pain relief such as morphine to Ms Seales, the doctor’s actions may not be an unlawful act ... if the doctor’s intention was to provide Ms Seales with palliative relief ... even though Ms Seales’ life would be shortened as an indirect but foreseeable consequence.

It is legal to hasten a patient’s death by administering an injection where the intention is to relieve pain, but illegal to do so where the intention is to cause death. Although the doctrine is not explicitly referred to, we will see that the same form of reasoning has been applied by courts in relation to the other two end-of-life options.

On the question of refusing life support, the competent patient’s refusal is accepted not out of a belief that the patient should die, but out of respect for the patient’s right to consent or not consent to treatment. As Panckhurst J noted in All Means All, the medico- legal context in New Zealand is shaped by the Cartwright Inquiry in 1988, the New Zealand Bill of Rights Act, the Health and Disability Commissioner Act 1994, and the Code of Rights in 1996, all of which reinforce the importance of patient rights and the paramountcy of the need for informed consent.⁵⁵The right to refuse treatment is simply the natural corollary of this right, and is quite separate and distinct from any form of the right to die.

This logic was expressed plainly by Lord Goff in Bland, who stated that where a patient refuses treatment and dies as a result:⁵⁶

... there is no question of the patient having committed suicide ... It is simply that the patient has, as he is entitled to do, declined to consent to treatment which might or would have the effect of prolonging his life, and the doctor has, in accordance with his duty, complied with his patient’s wishes.

This principle was adopted by the Court in All Means All and was essentially decisive of the case.⁵⁷It is therefore legal to withhold life-saving treatment because the intention is not to cause death, but to respect the patient’s right to refuse consent.

A similar approach underlies the matter of withdrawing life support from incompetent patients, with the courts distinguishing between the question of whether treatment ought to be continued on the one hand, and whether the patient ought to die on the other. Thomas J, in Auckland Area Health Board, focused on the “narrower question” of whether

53 Susana Nuccetelli and Gary Seay “Relieving Pain and Foreseeing Death: A Paradox About Accountability and Blame” (2000) 28 J L Med & Ethics 19 at 19.

the doctor has a duty to continue treatment that has no medical benefit.⁵⁸The point was also made by Lord Goff in Bland, where his Lordship said:⁵⁹

The question is not whether the doctor should take a course which will kill his patient, or even take a course which has the effect of accelerating his death. The question is whether the doctor should or should not continue to provide his patient with medical treatment or care which, if continued, will prolong his patient’s life.

Again, it is legal to withdraw life support from an incompetent patient because the intention is not to cause death, but to act in the best interests of the patient. This position has been codified in the United Kingdom’s Mental Capacity Act, which provides that where the decision relating to the incompetent patient concerns life-sustaining treatment, that decision must not be motivated by a desire to bring about the patient’s death.⁶⁰

IV Evaluation

We have seen that the current legal position depends on two doctrines: the AOD and the DDE. Their relevance in relation to the different end-of-life options may be crudely summarised by the following table, where a tick or cross indicates whether the doctrine justifies the legality of the option in contrast with assisted dying:

	Status	AOD	Intention	DDE
Refusing life support	Omission	✓	To respect autonomy	✓
Withdrawing life support	Omission	✓	To discontinue futile treatment	✓
Administering painkillers	Act	✗	To relieve pain	✓
Assisted dying	Act		To cause death

What follows will be an examination of the ethical soundness of the two doctrines. A The AOD

I will proceed on the basis that the justification for the legal AOD is the moral AOD, that is, that one is less morally responsible for an omission than for a positive act. In considering the adequacy of the AOD, I propose a three-step analysis:

In search of the right theory, let us begin with a dictionary definition. The Oxford English Dictionary defines an act as “something done”. ⁶¹On the other hand, an omission is defined as “the non-performance or neglect of an action which one has a moral duty or

61 Oxford English Dictionary (online ed, Oxford University Press) at [act, n].

legal obligation to perform”.⁶²Clearly, this definition is inadequate as it does not fit the law. The withdrawal of life support is, strictly speaking, “something done”, for one must physically interfere to switch off the machine in order to achieve the result, yet we will recall that this strict approach was dismissed by the Court in Bland. Lord Browne-Wilkinson said that to classify withdrawal as an act “would be to introduce intolerably fine distinctions”.⁶³It would result in, for example, the failure to replace an empty drip feed bag being categorised as an omission, whereas the switching off of a ventilator would be classified as an act.⁶⁴

Nor does the dictionary definition fit our moral intuitions. It is simply not the case that all physical omissions are less morally reprehensible than acts. Under the dictionary definition, punching Donald Trump in the face would be an act, whereas failing to lift my foot off the acceleration and hit the brakes when I know I will otherwise plough directly into Donald Trump would be an omission. No matter how many problems we might think the latter course of conduct would tend to solve, I think we would all agree that the omission is no less reprehensible than the act. A more nuanced method of categorisation is therefore needed.

George Fletcher has proposed a test based on causation, which distinguishes between whether the behaviour causes harm, or merely permits harm to occur.⁶⁵The former would amount to an act, whereas the latter to an omission. This test seems to fit the law. Certainly, withholding life support is to permit harm to occur. The same may be said about withdrawing life support, as the life support artificially prolongs life where it would otherwise have ended earlier. Removing the life support would merely permit the harm that would have happened if not for the artificial interference.⁶⁶This would be consistent with the legal categorisation of both as omissions.

The theory does not live up to the third step, however, when faced with James Rachels’ scenario.⁶⁷In this scenario, Smith goes upstairs and drowns her six-year-old cousin in the bath because she stands to gain a large inheritance on the child’s death. In a parallel universe, Jones also heads upstairs with the same intention, but in this world, her cousin slips, hits their head and falls face down into the water just as Jones enters the room. Jones stands by watching, ready to push the child’s head underwater if need be, but the need does not eventuate. The child dies in both cases.

Using Fletcher’s test, Smith’s behaviour would undoubtedly be classified as an act, but Jones’ behaviour would be classified as an omission, as in her case, the child drowned by themself and Jones merely permitted the harm to occur. Intuitively, however, we are inclined to think that Jones’ behaviour is no less reprehensible than Smith’s. But why is this?

64 Andrew Ashworth “The Scope of Criminal Liability for Omissions” (1989) 105 LQR 424 at 437.

65 George P Fletcher “Prolonging Life: Some Legal Considerations” in Bonnie Steinbock (ed) Killing and Letting Die (Prentice-Hall, Englewood Cliffs (NJ), 1980) 45 at 49.

67 James Rachels “Active and Passive Euthanasia” in Robert M Baird and Stuart E Rosenbaum (eds)

The answer seems to lie in the fact that we consider that moral responsibility is linked to control,⁶⁸and that not all instances of permitting to occur are of equal moral status. A physical omission is only considered reprehensible when the person: ⁶⁹

In Rachels’ example, we consider Jones’ behaviour to be as wrong as that of Smith because Jones was expected to prevent the child from drowning and had the ability and opportunity to do so. ⁷⁰Had everything else been the same except that Jones arrived upstairs several minutes later, by which time the child had already drowned, her omission would not equal Smith’s act because she did not have the opportunity to act otherwise.

So much for the causation approach. Lastly, we turn to a test formulated by Douglas Walton, which draws the distinction not between acts and omissions per se, but between “making happen” and “letting happen”. Here, to make happen is to behave in such a way so as to exclude any development that is incompatible with the intended result, whereas to let happen is to behave in a way that will exclude some but not all possible developments incompatible with the intended result.⁷¹For example, to stab someone is to make the harm happen because the act of stabbing excludes all possibility that harm will not occur, whereas to walk past a drowning baby is to let happen because there remains the possibility that someone else will come and rescue the baby. The difference between this and Fletcher’s approach is that it is possible to make happen in Walton’s sense by permitting harm to occur in Fletcher’s sense, if the permitting excludes all other incompatible developments. Conversely, it is also possible to let happen in Walton’s sense by causing harm in Fletcher’s sense, if the act causing harm didn’t exclude the possibility of harm not occurring.

Applied to the present topic, it could be said that both withholding and withdrawing treatment would be omissions because in both cases the withholding and withdrawal do not preclude other possibilities that may cause the patient to live.⁷²In this way they could be said to let death happen. The doctor who administers the lethal injection, on the other hand, would make death happen because the administration of that drug would exclude all possibility of the patient living, justifying its classification as a legal act.

Prima facie this would fit the law. There are, however, several problems with this account. First, whether the approach fits the law depends on the questionable assumption that withholding or withdrawing life-preserving treatment would leave open possible outcomes other than death. No doubt, there are situations where withdrawing life support may not lead to death. Practically, however, it is precisely those cases where death is inevitable that lead to legal and ethical controversy and make their way to the courts, and it is precisely the futility of treatment that courts draw on when deciding to legalise the withdrawal. In these cases that matter, withholding or withdrawing treatment essentially excludes any possible outcome other than death. Walton’s approach would therefore have to categorise them as making death occur, rendering them legal acts, not omissions.

70 As a side note, this also explains why the law only prosecutes omissions where there is a duty: the duty is the law’s recognition that certain acts are expected, and so only omissions in light of such expectations are considered worthy of criminal sanction.

71 Douglas N Walton Ethics of Withdrawal of Life-Support Systems: Case Studies on Decision- Making in Intensive Care (Greenwood Press, Connecticut, 1983) at 234.

Secondly, even if the above objection should fail and withdrawal could rightly be classified as an omission, taking such an approach would fail to distinguish between the conduct of a doctor and that of an interloper. In Bland, Lord Goff said that the conduct of the doctor could be differentiated from that of an interloper who removes life support maliciously because the conduct of the interloper is an active interference in the doctor’s doing their job, and therefore cannot be categorised as an omission.⁷³Under Walton’s approach, however, the action of an interloper would be no different from that of a doctor because the test rests on the objective measure of whether the action or inaction excludes the possibility of other developments. If the doctor lets death happen by unplugging a ventilator and commits a legal omission, then so does the interloper who does exactly the same thing.

To this one might respond that the problem lies not in Walton’s test, but in Lord Goff’s reasoning. It seems odd that the same conduct done by one person would be an omission, but when done by another person would be an act. To maintain such reasoning would be to indulge in legal fictions.

If this is the case, then it seems to me that this would only be reason to reject the AOD and to classify the doctor’s conduct as an act. The alternative would be to allow the interloper to escape liability since they owe no legal duty, and this is clearly an incorrect result. I think there is, however, something to be said for Lord Goff’s reasoning. As Andrew McGee has observed, just because two people perform physically the same actions does not entail that they “[do] the same thing”.⁷⁴A Justice of the Peace and a layperson may both watch someone sign a document and then stamp and sign it themselves, but only in the case of the Justice of the Peace has the signing of the document actually been witnessed in any meaningful sense.⁷⁵The status of the person can affect the legal and/or ethical nature of their conduct, and so any attempt at categorisation which also aims to be morally significant must take this into account.

The danger of doing this and calling the doctrine an “acts/omissions distinction”, however, is that it would merely be a new doctrine, say, one of duties, and not a doctrine of acts and omissions at all. Indeed, this is the final objection to Walton’s approach. Drawing the distinction as Walton draws it essentially amounts to an acceptance that the AOD is not itself morally significant. Under Walton’s test, both acts and omissions in their strict sense may make happen and let happen.⁷⁶If I throw a ball into a crowd intending that it be caught, I perform a physical act, but I do not make it happen that the ball is caught even if it is, because the possibility was open that it bounce off someone’s head. Conversely, if I do not show up to teach my class on Saturday, I make an omission, but my doing so will almost certainly stop the class because there is virtually no chance that my students will somehow call in a reliever. Given the test is no longer contingent on acts and omissions, we must be given a good reason why we ought to draw the line here and not somewhere else.

Walton’s argument is that the passive course of conduct under his approach, or to let happen, is “safer” because it leaves open other possibilities.⁷⁷We have seen, however, that this is not an adequate reason for our purposes, as withdrawing life support does not

74 Andrew McGee “Ending the life of the act/omission dispute: causation in withholding and withdrawing life-sustaining measures” (2011) 31 LS 467 at 472.

necessarily leave open other possibilities at all. But even if the argument was sound, the point remains that the distinction would no longer be one of acts and omissions, but a different doctrine altogether merely dressed up as acts and omissions in the legal sense.

It seems, therefore, that the AOD is not a useful doctrine for morally justifying the position of the current law. The categories may be formulated and reformulated depending on what we find to be the appropriate course of conduct given the circumstances: there is no moral distinction between acts and omissions per se. Fletcher, for example, ultimately ended his analysis with the conclusion that what is legally permissible for doctors really depends on “what doctors customarily do”.⁷⁸Furthermore, on the matter of the interloper, PDG Skegg has written that the withdrawal of ventilation could be classified as a legal omission where there is a duty relationship, but as a legal act where there is no relationship.⁷⁹Skegg accepted, however, that this approach is not, on the surface, a particularly satisfactory one.⁸⁰It is not clear why the question of an omission should come after the question of whether there was a duty, much less be dependent on it, other than as an ad hoc solution to the interloper problem. The goal of Skegg’s exercise was merely to find a suitable legal train of reasoning to accommodate the desired result, not to justify the conclusion on moral grounds. As Glenys Williams observed:⁸¹

As withholding or withdrawing treatment is not considered to be killing, it is not judged as such. However, if we thought it was not acceptable to withhold or withdraw treatment, we would refer to it as killing.

The argument from the AOD therefore turns out to be a circular one, based on the implicit acceptance of one particular course of conduct and the rejection of another before the analysis has even begun.

To be clear, this is not to say that the legal AOD ought to be discarded in general. The reason, as Michael Tooley argues, that we hold different moral intuitions about acts and omissions is not because there is a moral difference between acts and omissions themselves, but because of other factors such as the motive of the person, the cost of alternatives, and the possibilities left open by an omission.⁸²It only happens that for the most part, those factors arise alongside an omission, causing a moral confusion as to precisely what it is that makes the conduct less reprehensible.⁸³Therefore, the legal AOD will generally catch the right conduct, and its advantage is that it will do so in an efficient way without requiring an impractical examination of all the relevant factors. Indeed, for reasons of individual liberty, it would be undesirable for every omission to be prosecuted as if it were an act.⁸⁴The point is that this convenient generalisation ought to only extend as far as these underlying factors are present. Beyond that, in hard cases where they may not be present, it would be fallacious to maintain the AOD blindly, as the distinction would then have lost its moral basis.

79 PDG Skegg Law, Ethics, and Medicine (Revised ed, Oxford University Press, New York, 1988) at 177–178.

81 Glenys Williams “Acts and Omissions in Treatment Withdrawal: Conceptual Problems and Policy Decisions” (2008) 39 Cambrian L R 75 at 89 (emphasis in original).

82 Michael Tooley “An Irrelevant Consideration: Killing Versus Letting Die” in Bonnie Steinbock (ed)

Killing and Letting Die (Prentice-Hall, Englewood Cliffs (NJ), 1980) 56 at 59.

We have seen that the AOD in of itself is not a particularly helpful moral distinction if we are to justify the current legal position on end-of-life choice. It also fails to account for the legal status of the administration of lethal doses of painkiller. What seems to have been missing from the above discussion is the important factor of intention, which lies at the centre of the DDE. We will remember that the DDE is essentially the claim that there is a difference between intending some result, and merely foreseeing a result as an unintended side-effect. More specifically, the DDE requires four things:⁸⁵

This would allow us to separate the doctor giving pain-relief from the doctor giving a lethal injection to cause death. To illustrate, consider Drs A and B. Patient P is terminally ill and wishes to die. Dr A would take active steps to administer the professionally appropriate painkillers should P need them, even if it accelerates death. Dr B would carry out P’s request when she requests it.⁸⁶In both cases the doctors would have done an act; in both cases the result would be the same: P is dead; and in both cases P would be alive had the doctor not behaved as they did.

Applying the DDE, Dr A passes the first requirement because relieving pain is a good act, whereas giving a lethal injection is not. On the second requirement, Dr A’s giving pain relief would cause the good result (pain relief) to occur first, and so it is not a consequence of the bad result (death), whereas in Dr B’s case, death must come first. On the third requirement, Dr A only intended the pain relief, while Dr B intended that P die. On the last requirement, Dr A would say that relieving pain is valuable enough an act that it outweighs death, while Dr B could argue that respecting P’s autonomy and wishes is also valuable enough to outweigh death.⁸⁷However, the DDE requires that all four conditions be met, and so on this account Dr A’s conduct would be permissible whereas Dr B’s would not.

So, prima facie, it seems that the DDE is a good justification for the current legal position. There are, however, several objections against adopting this as the moral rationale.

The main objection to the DDE challenges the claim that there is a morally significant distinction between intending something and deciding to bring about a foreseeable consequence. We have seen in Part III that the courts rely heavily on this distinction in order to justify the current position of the law. If this distinction is shown to be untenable, then it would be fatal to the justification on which the current law rests.

George Graham argues that something can be done intentionally, despite it not being intended.⁸⁸To show this, Graham draws a distinction between the foreseen and avoidable consequence on the one hand, and the accidental or unavoidable consequence on the other. ⁸⁹The doctor, X, who foresees certain death but nevertheless administers a painkiller is not in the same position as the doctor, Y, who accidentally or even negligently

88 George Graham “Doing Something Intentionally and Moral Responsibility” (1981) 11 Canadian Journal of Philosophy 667 at 667.

causes death due to, say, an incorrect label. Nor is Dr X in the same position as doctor, Z, who cannot choose not to administer the painkiller for whatever reason. In each case, Drs Y and Z may be said to have caused death unintentionally, but not so for Dr X who is fully aware of the consequences but nevertheless chooses to proceed.⁹⁰

The point here is that while the doctor who administers the lethal painkiller for the purpose of pain relief may not intend death, they nevertheless cause death intentionally. This means there is more moral responsibility attached to the supposedly justified course of action than the DDE would lead us to believe. The doctor who makes the conscious decision to take a course of action, knowing of the undesirable consequence and fully able to avoid it, cannot disclaim responsibility in making that decision as if their conduct was unintentional. To do so would open the floodgates for the justification of things we should not be able to justify, such as the example of the strategic bomber. ⁹¹The Protocol I amendment to the Geneva Conventions prohibits indiscriminate attacks, such as bombardment, which strike military and civilian objects without distinction.⁹²Under the DDE, however, foreseen civilian casualties sustained in an attack on a munitions factory would be excusable: the act of destroying a munitions factory is a good or at least neutral act, the intended good effect is to end the war sooner, the result is not a consequence of civilian deaths, and it may be argued that ending the war sooner outweighs or is not worse than civilian deaths. And yet surely this is the wrong result.

Related to this objection, Judith Jarvis Thomson considers the situation an absurdity which becomes apparent if we imagine a patient who is near death and whose pain cannot be alleviated other than by a lethal dose of morphine.⁹³The patient asks for the injection, and the doctor wonders whether providing it would be legally permissible. The DDE would force us to answer:⁹⁴

“Well, I don’t know. I can’t tell unless you tell me what your intention would be in injecting the drug. If you would be injecting to cause death, either as means or end, then no. But if you would be injecting only to cause relief from pain, then yes.”

In such a case, the “unintended” consequence is so certain and so closely related to the intended result that it would be ridiculous to separate them. To argue that because the consequence was not intended and therefore justifiable would be disingenuous and an abdication of moral responsibility.

Here one may respond that intention must surely be relevant when it comes to criminal liability. Crimes such as manslaughter and murder rest on the element of intent.

This is true, but it is a misguided response. The objection is not that intention ought to be irrelevant to moral and legal responsibility, but that where a consequence is foreseen and under one’s control, it should not be treated differently from a consequence that is intended. Where manslaughter is distinguished from murder, it is done so on the basis that death was inadvertent in the sense as distinct from a situation where the

92 Protocol additional to the Geneva Conventions of 12 August 1949 and relating to the protection of victims of international armed conflicts (Protocol I) 1125 UNTS 3 (opened for signature 12 December 1977, entered into force 7 December 1978), art 51(4)(c).

93 JJ Thomson “Physician-Assisted Suicide: Two Moral Arguments” (1999) 109 Ethics 497 at 514– 515.

consequence was foreseen and avoidable. This is not the sort of lack of intent that is relevant here.

In fact, if Graham is right that the doctor does cause death intentionally, this would be in line with the current position in the rest of criminal law in New Zealand, which lends support for an account of intent that does not distinguish between intended consequences and unintended foreseen avoidable consequences. Under the Crimes Act, culpable homicide will amount to murder where the offender kills any person in the pursuit of an unlawful object and death was foreseen to be likely, even though they did not desire to hurt anyone.⁹⁵At common law, the two kinds of intent are referred to as “direct intention” and “oblique intention”, and the latter may qualify as intention for the purposes of the offence if the consequence was foreseen to a sufficiently certain degree.⁹⁶As Fisher J put it in R v Wentworth:⁹⁷

Contract killers usually want money, not the death of their victims per se. Receipt of money is the ultimate, desired, consequence. Death of the victim is the incidental, perhaps regretted, consequence. If it is clear that the intended course of action will result in both, both are said to be intended.

This is an inconsistency that requires addressing. The law cannot on the one hand accept oblique intention where we wish the defendant to be convicted while on the other hand deny oblique intention where we wish the doctor to be exonerated, without also providing a principled reason to explain why the DDE applies in one case but not the other.

Here, one might respond that our present end-of-life choice context can be differentiated from other criminal cases because in those cases, it is inherently very unlikely that condition one (that the conduct itself be morally good or neutral) of the DDE will be met. For conduct to be allowed under the DDE, it must satisfy all four conditions. Therefore, there is no inconsistency, as these comparator criminal cases would not be excusable under the DDE in any case.

This response would be appropriate if it did not matter which condition of the DDE was operational in producing the outcome, but this is not the case. We have seen from case law that the issue of intention is what liability truly turns on—not the morality of the conduct. The DDE, or similar reasoning, is evoked solely for the purpose of distinguishing between different states of intention, not applied generally to all criminal cases in its complete form. Therefore, the inconsistency remains.

Lastly, throughout this discussion, it has been assumed that death is always necessarily a bad consequence for the purposes of the DDE, but it is not clear that this is the case. Where patients request aid in dying, they do so because they believe death is the better option compared to what they are currently or will be experiencing.⁹⁸If this is true, then for that patient, dying will not be a bad result, and so the DDE becomes irrelevant.

This statement will no doubt be repugnant to some. Life is universally considered one of the most basic and valuable goods. We must remember, however, that we are dealing with voluntary active euthanasia, where the competent patient makes the decision, and we cannot know what it is like for the patient to experience such a condition. Kim Atkins, whose article was cited by Butler-Sloss P (as she then was) in Re B (Adult: Refusal of medical

98 See, for example, Re B (Adult: Refusal of medical treatment) [2002] EWHC 429 (Fam), [2002] 1 FLR 1090 at [78].

treatment),⁹⁹calls this the “specificity of the first-person perspective”: we can gather as many facts as possible to try to understand what someone’s experience is like, but we cannot know because the subjectivity of experience is itself part of the experience.¹⁰⁰Atkins argues that this difficulty is heightened the more debilitating the patient’s condition, ¹⁰¹which seems right. The more serious the condition, the greater the gap between the patient’s experience and the experiences with which we are familiar, and the fewer the common features we can draw on in an attempt to understand that experience. It is, therefore, precisely because of how much we usually value life that we must take a request for death by a competent patient very seriously indeed, for we cannot ourselves know what it must take for such a basic instinct to be overridden.

One might think that this is a circular argument because it presupposes that autonomy ought to outweigh the sanctity of life, and that is not established. This objection, however, rests on a mistakenly binary view of respect for autonomy. It is not the case that either we do not respect autonomy, or the principle of sanctity of life is outweighed. Respect for autonomy is a matter of degree, and we need not think that the level of respect needed to recognise patients’ subjective views on the value of death must be a level which necessarily also outweighs the principle of sanctity of life. Whether it is outweighed, of course, is another question.¹⁰²For the purposes of this argument it is enough to say that once it is accepted that death may not be a bad outcome, the DDE becomes irrelevant.

This objection may, of course, be open to the deontological response that the goodness or badness of a result can be intrinsic.¹⁰³Therefore, even if death may be a good result for the specific patient, it is nevertheless objectively a bad result. This debate, however, opens a very large can of worms which extends beyond the purview of this article. I personally do not share the view, and I mention it only for completeness. The point remains that if one does not believe goodness or badness can exist independently from its effect on a particular thing, then it also leaves the DDE in a questionable position as a justification for the position of the current law.

V Conclusion

This article began with an overview of the law in New Zealand and the United Kingdom on end-of-life options in the medico-legal context. In both jurisdictions, it is legal for doctors to withhold and withdraw life-saving treatment from competent and incompetent patients, in the former case where the patient requests it and in the latter case where treatment is not deemed to be in the best interests of the patient. It is, however, not legal for doctors to provide aid in dying to a competent patient who requests such an option.

102 I would argue that the relationship is the reverse, and that recognition of the patient’s view on death gives us reason to then think that sanctity of life might be outweighed, as the patient’s view on death means that sanctity of life is less relevant to them than it would be for someone else. Therefore, it is not because we value autonomy over life that we recognise the patient’s view; it is because our existing basic respect for autonomy recognises the patient’s view that we might think autonomy should actually be valued over life.

An examination of case law revealed that this position crucially rests on two separate but related doctrines: the AOD and the DDE.

I have argued that neither provide a morally sound basis for drawing the line where the law current draws it. The AOD is defective as a means of justification because there does not appear to be a moral significance in acts and omissions themselves. Instead, moral responsibility for an act turns on other factors which may but need not be present in cases where there is an omission. One of these factors is intention, which is what the DDE seeks to rely on in order to distinguish voluntary active euthanasia from other end- of-life options. I have argued, however, that the DDE is also defective because it seems that carrying out a course of conduct that results in a foreseen and avoidable consequence is nevertheless done intentionally. Indeed, the DDE is inconsistent with how the criminal law would otherwise deal with oblique intent. Furthermore, the DDE relies on the assumption that death is a bad result, an assumption which is open to question.

The law, then, treats the legality of different end-of-life options in a way that is not ethically sound. To be clear, this does not entail that voluntary active euthanasia must be legalised. I have endeavoured to show that the justifications currently at play in case law are problematic. Certainly, law reform is an option, and this may well happen with ACT Party Leader David Seymour MP’s End of Life Choice Bill.¹⁰⁴There may be, however, other reasons which adequately justify the current position, whether they be policy considerations or a new theory altogether. Ultimately, whichever the approach, and on whichever side of the jurisprudential debate over the nature of law one might fall, I think we would agree that the law ought to be morally and logically defensible, and to this end it is crucial that the law be honest in its reasoning and rigorous in its self-examination.

Public Interest Law Journal of New Zealand

Yang, Hannah --- "Acts, omissions, and the doctrine of double effect: the law and ethics of end-of-life choice in New Zealand and the United Kingdom" [2018] NZPubIntLawJl 14; (2018) 5 PILJNZ 97

I Introduction

A Definitions

II Current Law

A Refusing treatment

B Withdrawing treatment

C Administering pain relief

D Assisted dying

E A puzzle in the law?

III Acts, Omissions, and the Doctrine of Double Effect

A The AOD

B The DDE

IV Evaluation

V Conclusion