• euthanasia
• end of life care
• palliative care
• care of dying minors
• right to refuse treatment

DeMichelis, Shaul and Rapoport propose a framework for paediatric medical assistance in dying (MAID) based on their conclusion that it is ‘practically and ethically equivalent to other medical practices that result in the end of life.’1 We respond to two of the flaws in their argument. First, their construal of intent is inadequate, and they disregard standard moral rules that justify end-of-life decisions but need not entail MAID. Second, we question their claim that MAID alleviates suffering.

Their key assertion is that MAID is ethically equivalent to palliative sedation and withdrawal of care (WOC). In their view, all three involve actions that cause the patient to die sooner than if no such action were taken. Therefore, since all three come to the same end, death, they are ethically equivalent; and if sedation and WOC are ethically justified for children, so should be MAID. They doubt whether clinicians’ mere intentions can distinguish the procedures: ‘This empirical claim [that there is a moral distinction between intending and foreseeing death] is frequently made in the absence of evidence and also without explicit theorisation of why and how intention matters.’1 These arguments are not new, but DeMichelis, Shaul and Rapoport make little attempt at answering previous refutations, which we briefly review here.2–4

To begin, the so-called ‘empirical claim’ is not empirical, but philosophical. That aside, the assertion that there has been little theorisation about intention is simply false and undermined by the works of Keown, Finnis and Kaczor cited in the authors’ own footnotes. An enduring scholarly tradition has given many reasons why intention is crucial to ethics and law.4 5 Others have highlighted the influence of this tradition on clinical ethics, where the importance of intent has been repeatedly upheld by high courts and professional bodies and should be axiomatic.2 4 6 Common sense discerns a plain moral difference between a nurse administering a lethal dose of medication because she misreads an order, versus one who intends to kill her patient. While harms done unintentionally may be bad, harms done intentionally are both bad and immoral. Correspondingly, distinctions based on intent ‘form the basis of our criminal law’, as Justice Sopinka writes in Rodriguez v. British Columbia (AG), the Supreme Court ruling on physician-assisted suicide that preceeded Carter: ‘While factually the distinction may, at times, be difficult to draw, legally it is clear. The fact that in some cases, the third party will, under the guise of palliative care, commit euthanasia or assist in suicide and go unsanctioned due to the difficulty of proof cannot be said to render the existence of the prohibition fundamentally unjust.’7

Cursorily renouncing intention allows DeMichelis, Shaul and Rapoport to conflate all end-of-life decisions that may hasten death and ignore the well-developed but distinct ethical analyses for sedation and WOC. Accepting intention’s place in ethics allows one to understand these analyses and see how MAID is different.

First, DeMichelis, Shaul and Rapoport would seemingly have us believe that all palliative sedation is ethically indistinguishable (a consequence of their claim that intentions do not matter). One form of sedation is deliberate sedation to unconsciousness and death, which truly is morally equivalent to MAID.4 Yet if intentions do matter, one can be coherent in opposing MAID while accepting some forms of sedation.4

Sulmasy (among others) makes a distinction between permissible and unethical palliative sedation based on the rule of double effect (RDE), a rule physicians implicitly invoke whenever they prescribe a drug that has side effects.4 For example, when used to treat streptococcal pharyngitis, penicillin can cause two separable effects: killing bacteria and killing the patient via an anaphylactic reaction (figure 1). The physician intends to kill the bacteria, not the patient. Though the risk of anaphylaxis is foreseen, one judges the treatment reasonable or ‘proportionate’ (given the rarity of anaphylaxis, the likely harm of not treating). Additionally, the good effect of relieving the patient of bacteria is not mediated by the bad effect of relieving the patient of life. Significantly—contra DeMichelis, Shaul and Rapoport—should the patient die from anaphylaxis, no physician would think their prescription of penicillin was ethically equivalent to MAID, based on their instinctive understanding of the RDE. Accidentally triggering an allergic reaction is ethically very distinct from giving penicillin with the intent to cause anaphylaxis. Similarly, with permissible palliative sedation, sedation can be a foreseeable side effect of high doses of symptom-control medications, intended to preserve a measure of the patient’s health.4

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

‘Causal fork’ of the rule of double effect.4

WOC involves different ethical analyses from palliative sedation, and again the analyses need not entail MAID. The simplest justification for WOC is that physicians, respecting the patient as a rational self-governing person, can never force competent patients to use or to continue using any treatment, including life-sustaining treatment.

Physicians are obliged to respect refusals (to do otherwise would be battery). Yet we cannot stop here, for what may be said of the patient’s intention in refusing treatment, and the physician’s intention in discontinuing it? DeMichelis, Shaul and Rapoport find it ‘implausible to suggest that causing death is nowhere in the intentional landscape of a physician who withdraws life-sustaining treatment.’1 Further, they would have us focus on the ‘meaningful practical [and] ethical difference for the patient’, which we shall do (though the same ethical analysis can be applied to physicians).

Causing death could be in patients’ and physicians’ intentional landscapes in WOC, but it need not be.6 Patients need not intend every outcome in a causal chain.6 Imagine a patient with chronic pain treated with amitriptyline, who discontinues the drug because of intolerable side effects (overwhelming drowsiness, dry eyes, weight gain, sexual dysfunction and so on). He foresees but does not intend to experience increased pain; his moral aim is simply discontinuing the drug (figure 2).6 The difference between intending and foreseeing pain is the meaningful difference between being a masochist and a rational patient who weighs benefits and burdens.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

‘Causal chain’ of avoiding disproportionate treatment.6

Illustrated here is the principle that one cannot be expected to use disproportionately burdensome treatment, where the burdens associated with treatment outweigh its benefits, in the context of one’s condition.6 8 9 This concept applies to any treatment refusal.ⁱ For a patient on a life-sustaining ventilator, the burdens of being in the intensive care unit with a terminal illness may be disproportionately severe. He may reasonably decline further treatment, leading to WOC. Both the patient and physician’s intentions could be fulfilled once care is withdrawn (if the patient lives, neither has failed).ⁱⁱ By contrast, a physician administering MAID (or intending death through WOC) cannot be said to have fulfilled his intention until the patient is dead (if the patient lives, they have failed).6

Therefore, WOC differs from MAID in meaningful ways, including this: With WOC, patients and physicians need not aim at death; with MAID, they both necessarily intend the patient’s death.6 8 The distinction can be made ethically. It only ceases to matter if one already believes that intentionally killing oneself or one’s patient is permissible, or that intentions are irrelevant.8

In passing, we note an association fallacy in the authors’ citation of Carter to support their position:

‘[Supreme Court:] … There is no reason to think that the injured, ill, and disabled who have the option to refuse or to request withdrawal of lifesaving or life sustaining treatment, or who seek palliative sedation, are less vulnerable or less susceptible to biased decision making than those who might seek more active assistance in dying… (par. 115).’ Thus, [DeMichelis, Shaul, and Rapoport conclude,] the Supreme Court seems to support a view of MAID as practically and ethically similar to other medical interventions that result in the end of life…1

That conclusion does not follow. MAID and WOC may be similar in one aspect (both involve vulnerable patients), but that does not mean they are similar in any other relevant ethical aspect.

So far we have explored DeMichelis, Shaul and Rapoport’s inadequate analysis of intention, sedation and WOC. We conclude by challenging a key rhetorical claim that MAID is similar to palliative practices in that both ‘alleviate unendurable suffering’: ‘Rather than seeing the primary goal of MAID as ending life, this position sees the primary goal as ameliorating suffering.’1

This is a specious argument facilitated by the euphemistic term, ‘medical assistance in dying’. Semantically, ‘MAID’ is a general term, virtually synonymous with palliative care, and a universally compelling idea. Few, unacquainted with the term, would object to nurses’ or physicians’ ‘assistance’ at the end of life, as assistance is something healthcare professionals ought to provide throughout a patient’s life. Operatively, though, paediatric MAID is very specific: it refers to the intentional killing of the paediatric patient (by administering or prescribing a lethal dose of medication). The authors do not make clear how killing a child ‘ameliorates’ her suffering (vs simply ‘ends her life’). ‘Ameliorate’ implies that her condition has improved, so has clear meaning for palliative pain management and the like where the before and after are observable to us. Post-MAID, however, the child’s condition is dead.

Do the authors mean that the child is better off dead? Such a suggestion extends beyond mere medical claims into the metaphysical, relying on one’s philosophy of existence, or lack thereof, after death.10 Hamlet realises this unknowable: ‘To die, to sleep;/To sleep: perchance to dream: ay, there’s the rub;/For in that sleep of death what dreams may come/When we have shuffled off this mortal coil?’ (III.1). MAID, for all we know empirically, could lead to worse afterlife experiences than the ‘control group’ of natural death.10 Therefore, this construal of MAID, as being just another medical treatment that ameliorates suffering, is ungrounded.