1. In this paper, we use “well-being” since, as a value, it is explanatorily prior to the principles of beneficence and nonmaleficence, which are derived from it. In fact, principlism might be partially responsible for the prevalence of the asymmetry we oppose between refusals and requests, since it focuses on the duties of the healthcare provider instead of the patient. Nevertheless, our argument is the same if principles are used instead of values, so readers are invited to substitute according to their preferences.

2. We have chosen this example of a Request due to its relevance to the case we selected in Refusal. But it is worth highlighting that many other examples frequently discussed in the literature involve requests that do not involve termination of life, but rather drastic changes to one’s body—for example, those who experience body integrity identity disorder (BIID). For recent discussion, see Gibson, RB. Elective impairment minus elective disability: The social model of disability and body integrity identity disorder. Journal of Bioethical Inquiry 2020;17(1):145–55CrossRefGoogle ScholarPubMed.

3. See Gert, B, Bernat, JL, Mogielnicki, RP. Distinguishing between patients’ refusals and requests. Hastings Center Report 1994;24(4):13–5CrossRefGoogle ScholarPubMed. We have chosen the examples in Refusal and Request because of their salience, but we intend our argument to apply generally in healthcare. Other types of refusal should be respected, and they are, as respect for Do Not Resuscitate orders shows. Exactly what types of requests should be respected is, of course, a major question. We have argued elsewhere that autonomous requests for physician-assisted death are justified (see Davis J, Mathison E. The case for an autonomy-centered view of physician-assisted death. Journal of Bioethical Inquiry 2020; 17: 345–356). Our aim here is not to offer a complete account of the view, but, less ambitiously, to show that respecting requests is justified.

4. This approach is particularly common in the euthanasia literature. For example, L.W. Sumner says, “The arguments from well-being and autonomy provide the basic justificatory framework for assisted suicide and voluntary euthanasia.” Sumner, LW. Assisted Death. New York: Oxford University Press; 2011:91 CrossRefGoogle Scholar. Jurriaan De Haan writes, “while the patient’s autonomous request for euthanasia is a contributing factor indeed, his condition is an enabling/disabling factor. Despite the patient’s autonomous request, euthanasia is only permissible if the patient’s suffering is unbearable and hopeless.” De Haan, J. The ethics of euthanasia: Advocates’ perspectives. Bioethics 2002;16:154–72, at 169CrossRefGoogle ScholarPubMed.

5. The most common justification is that it violates the principle of nonmaleficence. Most physicians and bioethicists, for example, oppose amputations of healthy limbs for people with body integrity identity disorder. Arthur Caplan said, “It’s absolute, utter lunacy to go along with a request to maim somebody,” thereby appealing to nonmaleficence, though he also questioned whether someone requesting amputation could have capacity (https://www.salon.com/2000/08/29/amputation/).

6. This is not to say that the burden of proof is on the asymmetrical view, or that asymmetries are inherently problematic. However, if a survey of explanations produces no compelling reason for the asymmetry, and if there are reasons for symmetry, we have reason to believe that our intuitions for the asymmetry are mistaken.

7. Given this, our argument overlaps but is different in scope and approach to Flanigan’s, Jessica in Pharmaceutical Freedom (Oxford University Press; 2017)CrossRefGoogle Scholar, in which she argues that people have a right to self-medication by way of a similar pair of cases comparing refusal to treatment access. Flanigan’s argument is, roughly, that broader access to pharmaceuticals ought to be permitted, whereas our argument is not, in the first instance, about what is permitted (or prohibited), but rather about how the various principles ought to be ordered when making such determinations in clinical contexts.

8. Those who might wish to impose further necessary conditions on an autonomous decision are invited to add these conditions to the description of the initial case. In most cases, the argument we make here will still go through. If not, then this might suggest that such a view of autonomy is so restrictive as to rule out far too many plausible cases, which would give us good reason to reject such a view.

9. Philosophers disagree on the extent that autonomy is part of welfare or distinct from it. For examples of theories of welfare that include autonomy, see Mill JS. On liberty. New York: Norton; 1859/1975; Kymlicka, W. Liberalism, Community, and Culture. Oxford: Oxford University Press; 1989 Google Scholar; Liberalism, Wall S., Perfection and Restraint. Cambridge: Cambridge University Press; 1998 Google Scholar; Sumner LW. Welfare, Happiness, and Ethics. Oxford: Oxford University Press; 1996. Our view is that at least some of the time autonomy is distinct. This is because it is possible for someone to autonomously do what is prudentially bad for her by, for instance, acting altruistically. Therefore, “living according to one’s values” is not always synonymous with “promoting one’s welfare,” which means that there is some important sense in which not getting the blood transfusion is bad for the patient. Her death is a great loss to her, but her values mean that she orders some things above the badness of death.

10. This is complicated by the fact that, in many cases, physicians defer to substitute decision-makers for deciding which treatment the patient ought to receive (or not). For a discussion of some of these difficulties, see Buchanan, AE, Brock, DW. Deciding for Others: The Ethics of Surrogate Decision-Making. Cambridge: Cambridge University Press; 1989 Google Scholar. For a discussion on end-of-life substitute decision-making, see Appel, JM. Trial by triad: Substituted judgment, mental illness and the right to die. Journal of Medical Ethics 2022;48(6):358–61CrossRefGoogle ScholarPubMed.

11. The conditions of informed consent are generally agreed upon, though in practice there will be tough cases when determining capacity. See Faden, R, Beauchamp, T. A History and Theory of Informed Consent. New York: Oxford University Press; 1986 Google Scholar; Beauchamp, T, Childress, J. Principles of Biomedical Ethics. 4th ed. New York: Oxford University Press; 1994; Beauchamp T. Autonomy and consentGoogle Scholar. In: Miller, FG, Wertheimer, A, eds. The Ethics of Consent. New York: Oxford University Press; 2010 Google Scholar; and Berg, J, Appelbaum, P, Lidz, C, Meisel, A. Informed Consent: Legal Theory and Clinical Practice. 2nd ed. New York: Oxford University Press; 2001 CrossRefGoogle Scholar.

12. Examples both for and against include Howe, EG. Should military care providers force-feed detainees? Military Medicine 2015;180(12):1203–4CrossRefGoogle ScholarPubMed; Howe, EG, Kosaraju, A, Laraby, P, Casscells, SW. Guantanamo: Ethics, interrogation, and forced feeding. Military Medicine 2009;174(1):iv–xiii Google ScholarPubMed; Rubenstein, L, Annas, G. Medical ethics at Guantanamo Bay detention centre and in the US military: A time for reform. Lancet 2009;374(9686):353–5CrossRefGoogle ScholarPubMed.

13. Those who value the sanctity of life over autonomy, as some defenders of force-feeding do, will likely also be against accepting Refusal. For the reasons we give above, we think this position is mistaken.

14. For a view of the former sort, see Savulescu, J, Momeyer, RW. Should informed consent be based on rational beliefs? Journal of Medical Ethics 1997;23:282–8.CrossRefGoogle ScholarPubMed For a view of the latter sort, see Muramoto, O. Bioethics of the refusal of blood by Jehovah’s witnesses: Part 1. Should bioethical deliberation consider dissidents’ views? Journal of Medical Ethics 1999;24:223–30CrossRefGoogle Scholar.

15. Our aim here is not to defend this distinction. (Indeed, we disagree between ourselves on the extent of its application in moral theory.) Our point is simply to show that this distinction cannot explain why we should treat Refusal and Request differently.

16. For our purposes, what is relevant is that the act becomes permissible. Whether or not there is still something harmful about it—that is, whether there can be “consensual harms”—depends on one’s definition of harm. Because our point stands either way, we will not try to settle this issue, but see Feinberg, J. Harm to Others. Oxford: Oxford University Press; 1984 Google Scholar; Lazar, S. The nature and disvalue of injury. Res Publica 2009;15(3):289–304 CrossRefGoogle Scholar; Hanser, M. The metaphysics of harm. Philosophy and Phenomenology Research 2008;LXXVII(2):421–50CrossRefGoogle Scholar.

17. For those who think being consensually punched is still a harm, the symmetry point also applies in Request and Refusal. That is, if the physician unjustly harms the patient by acceding to the request for assistance in dying because dying is a harm, the physician also harms the patient in Refusal by letting her die.

18. Rachels, J. Active and passive euthanasia. New England Journal of Medicine 1975;292(2):78–80 CrossRefGoogle ScholarPubMed.

19. McMahan, The Ethics of Killing Chapter 5, Section 2

20. We have argued elsewhere that assisted dying is a legitimate goal of medicine Mathison E, Davis J. Value promotion as a goal of medicine. Journal of Medical Ethics 2021 47(7): 494–501.

21. One might object here that we are not plausibly talking about “care” in any standardly accepted sense in Request. On this view, the patient does not need this treatment and so does not have anything resembling a positive right to it. But this definition of care builds the well-being condition into it, which seems too tendentious for this discussion. Ultimately, we must either understand “care” in a way that is sufficiently broad such that it allows the requested treatment in Request to fall under it, in which case we can continue to discuss whether or not the positive–negative right distinction is of use in distinguishing Request and Refusal, or else understand “care” so narrowly that it rules out requests like Request. If we take this latter route, we have identified an important distinction between Refusal and Request—namely, that Refusal (and not Request) involves an important right. But we would still need a further argument to show why this right is enough to reorder autonomy and well-being. And, for reasons related to those we go on to discuss here, we doubt that this will be successful.

22. For a sustained argument in defense of this general idea, see Woollard, F. Doing and Allowing Harm. Oxford: Oxford University Press; 2015 CrossRefGoogle Scholar.

23. Kass, L. The end of medicine and the pursuit of health. Public Interest 1975;40:11–42 Google Scholar, at 12.

24. Davis J, Mathison E. The case for an autonomy-centered view of physician-assisted death. Journal of Bioethical Inquiry 2020;17:345–356; Mathison E, Davis J. Value promotion as a goal of medicine. Journal of Medical Ethics 2021;47(7):494–501.

25. Notice that some of the most popular defenses of abortion in the philosophical literature are autonomy-based. See, for example, Thomson, J. A defense of abortion. Philosophy and Public Affairs 1971;1(1):47–61 Google Scholar.