How Not to Think About the Autonomy-Based Argument for Physician-Assisted Suicide


How do we define autonomy?

medical aid in dying, physician-assisted suicide

Dan Race/AdobeStock

The rationales for physician-assisted suicide (PAS) are usually thought to be consonant with the same ethical values that support a patient’s right to forego life-sustaining treatment: autonomy and beneficence. The idea behind autonomy is that rational individuals should be able to direct their lives how they see fit, all things being equal. PAS is permissible, then, because autonomous, terminal patients have the moral authority over their lives to make such a decision. The idea behind the value of beneficence is that life, though normally seen as a benefit, is often no longer seen as such to the terminally ill and suffering patient. When these patients see continued life as more burdensome than beneficial, their well-being is not realized by continued treatment, and they choose PAS over continued life.

The fear of losing autonomy at the end of life seems to be a main rationale driving many terminal patients to seek PAS. This is not to say that fears about unbearable suffering do not motivate many to support PAS—they do. But when we look at surveys of terminal patients seeking PAS in, for example, Oregon and Washington—2 of 11 jurisdictions in the United States where PAS is currently legal—we see that fears of losing autonomy and dignity are the primary reasons for seeking PAS.1,2 In Canada, where PAS has been legal since 2016, one study in British Columbia found that of 112 terminal patients who received PAS, most of them cited fears of losing control or independence as their primary reasons for seeking PAS.3 Concerns about autonomy seem to be the primary driver of support for PAS.

But the autonomy-based defense of PAS has its critics. Pies and Geppert argue that the autonomy defense for PAS is based on some myths.4,5 PAS legislation provides for patients, at best, pseudo-autonomy – something less than genuine autonomy. PAS laws, according to Geppert and Pies4:

…provide nothing remotely resembling “autonomy” for the patient, in either the procedural or personal sense. Patients who wish to avail themselves of prescribed, lethal medication must clear a number of procedural and administrative hurdles that depend entirely on the diagnostic, prognostic, and prescriptive authority of the patient’s physician. The controlling decisions regarding the patient’s diagnosis; the need for a consultant to confirm the diagnosis; the putative “terminal” nature of the illness; the completion of required certification forms; and, finally, the writing of the lethal prescription are all exercises of the physician’s autonomy.

There are several problems with this reasoning. First, instead of diminishing autonomy, these “procedural and administrative hurdles” that need to be cleared before a patient can receive PAS actually protect autonomy. Take, for example, the requirements for receiving life-ending medications under Oregon’s PAS law. The Oregon law requires, amongst other things, (1) that the patient make 2 oral requests, separated by at least 15 days, to their physician; (2) that the patient provides a written request to their physician, signed in the presence of 2 witnesses; (3) that there be 2 physicians—the prescribing and consulting—who confirm the diagnosis and prognosis; and (4) that both physicians determine whether the patient is capable, etc. What is the point of these requirements? They are safeguards against coercion and manipulation by others, and against the very patient’s “unsettled” or transitory preferences to die (which is the point of the different number and types of requests)—that is, safeguards to protect patient autonomy, not vitiate it.

But Pies and Geppert think that these safeguards are not sufficient to guarantee genuine autonomous choice. Most importantly, they claim that “current PAS statutes are generally not capable of assessing genuine autonomy” because they are incapable of assessing the patient’s family dynamics, that is, “the potentially coercive influences of those who might stand to gain emotionally or financially from the patient’s death.”5 Geppert and Pies are arguing that we cannot know if a patient seeking PAS has been pressured or guilt-tripped by a controlling family member to choose PAS—maybe because continued treatment is a burden on the family or because a family member wants their inheritance sooner rather than later. These kinds of choices are not fully autonomous, argue Geppert and Pies, and current PAS laws are unable to assess that such choices have occurred.5

This leads us to a second problem for Pies and Geppert. The difficulty of assessing genuine autonomy in patient choice is something we have decided to accept in other areas of health care and, most importantly for our discussion here, in matters at the end of life. Take, for example, requests for a do-not-resuscitate order (DNR). The DNR is made when a patient enters a medical facility, obviously before an emergency occurs. The physician writes the order only after conferring with the patient, the patient’s proxy, or the patient's family. Obviously, talking with the patient is very important, since the doctor must get a sense of the patient’s level of capacity. But how can we be certain that the patient was not pressured to request a DNR by, say, a controlling family member who would benefit by the patient’s demise? It is worth noting that in most states a DNR order has none of the safeguard requirements that, for example, Oregon’s Death with Dignity Law has in place. The typical DNR order only requires that the ordering physician—and only one physician is required—discusses with the patient their desire for a DNR, and that the patient gives informed consent. As noted, these safeguards, and many additional ones, are present in PAS laws. From a standpoint of protecting a patient’s autonomy, PAS seems safer—less prone to abuse or manipulation—than the common DNR order.

The same is true of patient decisions to pursue or forego life-sustaining treatment—a right enshrined in both law and clinical practice. Are our concerns that some patients may choose to forego life-sustaining treatment because they are pressured by an influential family member sufficient to limit or prohibit the right of patients to refuse medical treatment at the end of life? In fact, the practice of allowing patients to pursue or forego life-sustaining treatment is way risker than PAS. Due to the influence of financial pressures on many patients deciding to pursue or reject life-sustaining care, the patients making these decisions are at a greater risk than patients deciding to pursue PAS, since many patients are incompetent when decisions about continuing life-sustaining treatment are made, and many of them lack surrogate decision-makers.6 If Geppert and Pies are concerned that PAS is problematic because it requires only a superficial assessment of decision-making capacity, then they should be way more concerned that the assessments of decisions for DNR’s or to refuse life-sustaining care are even more superficial.

In a similar vein, Pies and Geppert express even more skepticism about our ability to evaluate whether a patient’s decision to seek PAS is genuinely autonomous. Not only do they think that no law has what it takes to sufficiently evaluate “family dynamics,” in order to determine if a decision to seek PAS was coerced or pressured by family members, but “Nor do the superficial assessments of decision-making capacity required under various PAS statutes address subtle, but important factors than can compromise genuine autonomy, such as the patient’s unconscious fears and fantasies.”5 (italics added) What are we to make of this? If we are going to appeal to unconscious fears and fantasies as possible autonomy-defeaters for patients requesting PAS, this would seem to equally apply to patients who request DNRs or choose to forego life-sustaining treatment. In fact, if a physician refuses a patient’s request for PAS, it is possible that the physician’s refusal is due to unconscious fears and fantasies. Should PAS legislation specify that if a physician is asked to help a qualified candidate receive PAS, then they must write a prescription for life-ending drugs, even if the physician does not want to? If the physician’s refusals are not truly autonomous but the result of unconscious fears and fantasies, then isn’t that something we would wish to avoid? This objection may seem silly or tongue-in-cheek, but it can also be seen as a reductio ad absurdum of Pies and Geppert’s skepticism concerning our ability to evaluate a patient’s autonomous choice for PAS.

Another problem with Pies and Geppert’s argument that PAS statutes provide only pseudo-autonomy is that they come close to contradicting themselves about the value or importance of autonomy. They complain that proponents of PAS are wrong in thinking that PAS laws provide autonomy for patients. If anything, they think that the real autonomy is with the physicians involved. If a patient is to obtain PAS, they must clear some hurdles “…that depend entirely on the diagnostic, prognostic, and prescriptive authority of the patient’s physician. The controlling decisions regarding the patient’s diagnosis; the need for a consultant to confirm the diagnosis…and finally, the writing of the lethal prescription are all exercises of the physician’s autonomy.”4 Pies and Geppert appear to be arguing that patient autonomy is very important, and one’s autonomy should be enhanced and respected. But PAS laws do not provide real autonomy—which is a good reason for proponents of PAS to be hesitant to support PAS legalization. This complaint against the autonomy-based defense of PAS sounds like a strong defense of the value of autonomy in health care decision making. But Pies and Geppert then dismiss autonomy as being just the result of some accidents of time and place. They claim that5:

…the overly individualistic and near absolutist concept of autonomy that reigns in Anglo-American law and bioethics is historically and culturally conditioned. There are many other types of autonomy…Personal autonomy does not resonate with many cultural groups... Indeed…individual autonomy is subordinate to other significant ethical values in many other ethnic and cultural groups. For example, in many traditional Native American and Latino cultures, familial or community autonomy is far more important, particularly when end-of-life decisions are involved.

What is strange here is that in criticizing current PAS legislation as not providing genuine individualist autonomy, Pies and Geppert seem to be endorsing individualist autonomy as an important medical value. Why else criticize PAS for not providing it? But then, 2 paragraphs later, Pies and Geppert dismiss individualistic autonomy—the autonomy driving current support for PAS, according to them—as historically or culturally conditioned, as relative.

The biggest problem for Pies and Geppert’s argument that PAS laws do not provide genuine autonomy is that they seem to accept a very naïve account of autonomy. When Pies and Geppert claim that patients seeking PAS do not have genuine autonomy because obtaining PAS “…is critically dependent on the authority of powerful others who must approve (or veto) every decision along the way,”5 we are left wondering how patient autonomy is even possible, since all patients—not just those involved in dire end-of-life decision-making—depend on a physician’s actions when seeking treatment. Having autonomy does not entail omnipotence in being able to satisfy one’s wishes and desires, though this is the conception of autonomy that Pies and Geppert think is underwriting the autonomy-based defense of PAS. “Rational autonomy,” they write, “is much more than being able to ‘do as you please.’ After all, an infant picking up his rattle is doing as he pleases, but he is hardly acting with rational autonomy.”5 Of course this is true, but Pies and Geppert seem to be attacking a straw man; they are employing a folk concept of autonomy—one which expresses an inchoate desire for freedom in one’s life, and this concept is not necessarily the concept of personal autonomy that motivates a desire for legalizing PAS.

Respecting a person’s autonomy means that we cannot use them as a means to an end—that is, we should not deceive them, decide what’s best for them, or impose our will on them. Autonomy requires agency—that one’s actions result from their own beliefs, values, and choices, not from someone else’s. When I act autonomously, I am the agent. When someone directs or coerces my actions through threats of force or potential penalties, I am no longer the agent of action, another person is. Autonomy also requires that my decisions are relevantly informed—that I decide with an understanding of options that are relevant to my choices. We cannot make sense of my actions being autonomous if I am ignorant of information that might make a difference in how I decide to act. This is one reason why lying or deceiving a patient about diagnosis or treatment is wrong. Patients deserve to have information that will allow them to decide, for themselves, how their lives should go.

I think something like this is the correct account of autonomy, and what we do not see in this account is the idea that one is autonomous only if they are able to “do as they please.” Is it true that because a patient must rely on a physician to receive PAS, the patient is not autonomous? Pies and Geppert seem to assume that the autonomy defense of PAS only makes sense if there are no external, unavoidable hurdles or constraints (eg, relying on the physicians to diagnose, give a diagnosis, and then agree to write the prescription, etc), that one must clear before obtaining PAS. But this cannot be correct. So much autonomous action requires us to surmount unavoidable, external hurdles to satisfy our desires—from the relatively trivial (eg, getting a pizza delivery, requesting that a mechanic fix your car), to the momentous (eg, refusing treatment, requesting PAS). Can we not make sense of any of these actions being autonomous just because there are unavoidable, external hurdles?

Pies and Geppert are asking the wrong question about autonomy. The question is not “Is one autonomous only if there are no unavoidable, external hurdles?” The correct question is “Given the unavoidable, external hurdles, do one’s actions reflect their own values, beliefs, and relevantly informed choices?” I assume that requests for PAS can be autonomous in this way. If there are myths or misconceptions underwriting the autonomy-based defense of PAS, Pies and Geppert have not identified them.

Dr Merrell is assistant professor of philosophy at Northwest Arkansas Community College.


1. Oregon’s Death with Dignity Act--2014. Oregon Public Health. 2015. Accessed February 21, 2023.

2. 2021 Death with Dignity Act Report. Washington State Department of Health, Center for Health Statistics. July 15, 2022. Accessed February 21, 2023. 

3. Wiebe E, Shaw J, Green S, et al. Reasons for requesting medical assistance in dying. Can Fam Physician. 2018;64(9):674-679.

4. Geppert C, Pies R. Two misleading myths regarding ‘medical aid in dying.’ Psychiatric Times. 2018;35(8).

5. Pies R, Geppert C. Physician-assisted suicide and the autonomy myth. Psychiatric Times. October 27, 2021.

6. White D, Curtis J, Lo B, Luce J. Decisions to limit life-sustaining treatment for critically ill patients who lack both decision-making capacity and surrogate decision-makers. Crit Care Med. 2006;34(8):2053-2059.

Related Videos
Dune Part 2
© 2024 MJH Life Sciences

All rights reserved.