Facing Requests for Physician-Assisted Suicide: Title and subTitle BreakToward a Practical and Principled Clinical Skill Set

PUBLIC INTEREST in physician-assisted suicide (PAS) in recent years has prompted intense professional and social deliberation on the topic and has resulted in the US Supreme Court's consideration of the matter.^{1 - 2} The Supreme Court's unanimous decision to uphold the right of states to prohibit PAS nevertheless allows that states might also permit it; indeed, Oregon recently voted to legalize it under some conditions.^{3 - 4} So the debate was not ended by the Supreme Court decision but rather moved into a state-based forum. Whether opposed to or in favor of PAS, every physician is now squarely faced with it, and 1 question confronts the whole profession: how should requests for it be approached?

This article proposes for professionals a stepwise approach to requests for PAS. By contrast to some other policies and guidelines that seek to permit but regulate PAS,^{5 - 6} this protocol seeks instead to understand and treat root causes of the request. Most of these steps reflect the view that yes or no responses are insufficient for those who request PAS. The approach aims to represent consistent application of a set of clinical skills that may not resolve the debate but should resolve a majority of requests by patients or others for PAS. While the steps draw on established clinical skills, taken together the approach is a new synthesis and as such invites future research and clinical evaluation. Several of these steps and the philosophy of this approach may also be applicable to other end-of-life situations, for instance, a request for withdrawal of a life-sustaining intervention.⁷

There are quite a number of arguments, by now well rehearsed, that claim to justify positions for or against PAS.⁸ Through the ages this debate has never been stably resolved. By contrast, justification for this article's proposed approach needs just 2 generally uncontested arguments: that individuals can reject unwanted interventions and that suffering individuals have a legitimate claim to comfort care. Some of the tumultuousness of the current debate is avoided by not relying on the confusing doctrine of double effect (in which an act with one intent is justified even though it has a known additional and otherwise unjustified effect) and by avoiding erratically defined terms such as terminal sedation that can cover a mixture of ethically quite different categories (eg, pain control vs euthanasia).^{9 - 11}

The right to be free of unwanted intervention has to do with bodily integrity and freedom from assault. The right is deeply rooted in ethics and the law. It is consistent with principles of autonomy but can be derived in other ways as well, such as in religious explanations invoking the sacrosanct nature of human beings. The right to be free of intrusion justifies many different norms. For instance, in general society it justifies prohibitions against trespass and violence. In medicine it is this right more than any other justification that underpins the doctrine of informed consent. This right also convincingly justifies the withholding or withdrawing of unwanted medical intervention. The right holds even when the intervention being withdrawn is life sustaining, because unwanted intrusions are always an unacceptable violation of integrity, and perhaps especially so when the person is dying. This is so whether the intervention is a minor or major invasion, for instance, artificial nutrition and hydration, use of antibiotics, amputation, or mechanical respiration.^{12 - 16}

Relief of suffering is a time-honored professional obligation that is prominently treated in the Hippocratic oath.¹⁷ The duty to relieve suffering is one reason why compassion is a key virtue in a physician. Justices O'Connor and Breyer have argued the point differently and more strongly. Focusing mostly on pain relief, they suggest that there may be a right to relief of suffering.^{1 - 2} Whether the argument is posed as a physician's duty or a patient's right, the need for relief of pain and other forms of illness-related suffering is clear.

The right to be free of unwanted intervention and the obligation to provide comfort care are both different from the right to induce desired death. Some commentators have inferred a right to induce desired death from case law involving intervention withdrawal.⁸ But this inference is not a necessary part of the cases' justifications since the right to be free of unwanted intervention can by itself morally justify withdrawal of an intervention. This right to nonintrusion does not suppose, imply, or deny a right to achieve desired death. Logically, practically, and ethically, declining unwanted intervention is different from demanding any wanted intervention. Since nonintrusion and relief of suffering belong in different categories from any right to actively and intentionally secure death, with or without assistance, extension from either of the former to the latter is not automatic. Absent such automatic justification by its underpinning principles, PAS is not a part of this approach.

The next question is how can these 2 accepted claims to nonintrusion and to relief of suffering guide a practical approach to PAS? The answer is this: by their full implementation.

Full implementation applies to general directions for addressing shortfalls in end-of-life care as well as to specific clinical approaches. In the relatively short time since life-sustaining intervention has become possible,¹⁸ elements of both the legal and the medical professions have been working to secure appropriately constrained application of technology.^{15 ,19 - 20} Even though the right to be free of unwanted life-sustaining intervention is now established in case law, too many people still have insufficient opportunity to exercise their right, and the use of unwanted life-sustaining intervention has persisted.^{21 - 22} Confusions, for instance about what is legal and ethical, need redress with education.^{23 - 25} Skills in advance care planning and palliative care techniques in this intervention-oriented era need full implementation.^{22 ,26 - 30} Health care delivery systems still need to develop so they can support palliative care needs. The fact that many motivations for PAS are more psychosocial than physical in nature points to cultural expectations and to communities' care for the dying as areas for change and improvement.^{31 - 34} Societal aspects to end-of-life care policy must also develop to support community involvement.

However, more than general directions, specific application of these principles to PAS requests is relevant here. Physicians need to be ready and able to respond appropriately, using the principles of nonintrusion and comfort care, if someone in their care expresses interest in PAS. Such requests have occurred since antiquity, and physicians today must be able to provide a fully professional response. Tragically, and oddly, physicians do not routinely receive professional training to prepare them to respond to a patient's request for PAS.

Figure 1 depicts an 8-step clinical protocol for evaluation of and response to patients who request PAS, applying only the widely endorsed principles of a patient's right to be free of unwanted intervention and the health professional's obligation to provide comfort to the suffering patient. To attend effectively to issues involved in PAS requests, the approach does not describe but rather refers to and assumes knowledge of palliative and hospice philosophy care. Two steps involve proper evaluation of the patient for depression and competence; 2 steps involve interventions as well, namely, to discuss care plans and to evaluate and treat suffering; 2 further steps require full deliberations that meet informed consent standards and involvement of consultative assistance. A seventh step focuses on proper implementation of the care plans, ensuring in particular that no unwanted life supports are continued. It delineates both how to support a patient who chooses to cease oral food and fluid intake and how to ensure that options are available to manage any crescendo suffering in the last phase before dying. In accord with use of nonintrusion and comfort as the only justifying precepts, there is no step for eventual PAS; rather, there is a step on how to conduct a conversation in which the prior approaches are offered in preference to PAS.

The first step requires an evaluation for depression and other psychiatric conditions to help distinguish a rational request for PAS from a request resulting from disordered thought, especially if the condition is due to a treatable condition. Depression is an established cause of requests for PAS.³³ Further, clinical depression is an underrecognized illness in general³⁵ and likely an even more underrecognized complication of terminal illness, being too often cast aside as one of the common manifestations of terminal illness or as a normal response to facing death. Depression should be treated, whether it is triggered by a situation or is a preexisting condition.

The second step requires evaluation of the patient's decision-making competence. In the absence of this competence to make end-of-life decisions, by no reasonable standard would a request for PAS be considered valid. Further, a health care proxy is necessary for decision making if the patient is not able to participate on his or her own behalf. In the absence of reversible mental incapacity, a proxy must be sought.

If the patient is competent, the third step is a discussion about the patient's goals for care. This is a necessary part of all clinical care, especially for the terminally ill, but is placed here in the protocol as an intervention. If the patient is not competent, these goals (and each further step) are discussed with the patient's proxy. An advance care planning discussion that includes immediate care goals and goals for care in other possible illness scenarios may be an effective method of providing a forum in which to respond to the patient's root fears in a careful, sensitive, and efficient manner.³⁶ Patients who fear loss of control can see the extent of control afforded by withholding or withdrawing interventions and can secure the physician's commitment to honor those preferences. Patients who fear pain can choose aggressive comfort care. Patients who fear becoming a burden to family and other caregivers can learn about hospice and home care. They can learn how withdrawal and withholding of interventions can limit burdensomeness by allowing natural death.^{3 ,33} Patients who fear abandonment can see evidence in the conversation and the planning that abandonment is not likely. Patients who fear indignity can discuss what indignity is to them and find ways to minimize or avoid those situations.

The fourth step involves evaluation of and response to the patient's present physical, mental, social, and spiritual suffering—the patient's "total suffering."^{34 ,37} Physical symptoms in terminal patients vary widely. Whether dyspnea, anorexia, pain, or other symptoms are the sources of physical suffering, frequent and regular inquiry and proactive treatment can be effective. Sources of psychological suffering may include the need for attention or for a meaningful purpose in life, difficulty adjusting to serious disability, or difficulty in relationships. In the social context, sources of suffering may include insufficient support from family or community or cultural approaches to death that leave the patient feeling ashamed or unwelcome. Spiritual suffering is a common element of dying. Psychological, community, and spiritual support should be arranged whenever appropriate whether from the patient's community or care facility. Altogether, this fourth step of the protocol often involves a host of specific evaluations and interventions. It may include a great part of the full discipline of hospice and palliative care and almost always needs effective teamwork with the family and community as well as with various professionals. Reference to this care as a single step is not intended to diminish its extent but to emphasize its scope and purpose in this protocol; namely, a clinician should keep in mind the possibility that a request for PAS may evidence a failure in this area of competence.

The fifth step requires provision of full information and deliberation with the patient. This is a matter of basic clinical standards. All major decisions should be made with full information and after proper discussion. Two areas of inadequate information and discussion are noteworthy. First, many patients (and physicians) are not well informed about the kind of palliative or hospice care possibilities that are or should be available to them.²³ Second, some who consider PAS are assuming that it is a suffering-free mode of exit from this world. This is not assumable. Even among patients who do not have moral qualms about the act, some have described it as existentially wracking, as have some family members and some physicians.^{25 ,38}

The sixth step is for the physician to consult and involve professional colleagues. Decisions at the end of a patient's life are too important and too complex to be made and acted on alone; they need full collegial and consultative support. This applies even to specialists in end-of-life care. Requests for PAS are anxiety provoking, and many professionals avoid the open collegial discussion that has long been one of the profession's traditional quality assurance mechanisms.³²

The seventh step is to be sure that care plans are being followed and, in particular, to remove all unwanted life-support interventions and to secure maximum relief of suffering. Life-prolonging interventions that are unwanted may be technical and invasive or simply oral medications. Each unwanted intervention that must be removed should occur with suitable informed consent and with full efforts to alleviate associated suffering in any of the physical and nonphysical spheres.^{9 ,39 - 40}

In the course of this seventh step the patient may decline not just artificial nutrition and hydration (a possibility that should have been affirmed during advance care planning in step 3) but also regular oral intake.^{9 ,41 - 42} While explicit refusals among competent patients are rare, they do occur, and they have been honored by the founders of and authorities in the hospice movement. Few data are available on prevalence of patients' declining regular oral intake, but in the 30-year history of St. Christopher's Hospice, 2 such cases have occurred. Both patients' refusals of oral intake were honored. One patient persisted in this choice and died comfortably. The other patient changed the decision, taking food and water again, and also died comfortably later on (C. Saunders and N. Sykes, oral communication, December 22, 1997).

As a safeguard against a situation more akin to a hunger strike than a terminal care decision, only patients who have lost hunger and thirst as a consequence of their illness should cease oral intake. While this criterion may be difficult to ascertain in some cases, it is similar to that of intent; in both cases the distinction remains crucial. Cessation of oral intake in this protocol is not about a choice to die but about nonintrusion. Terminal patients should not be forced, but food and water should always be available. As a further safeguard the choice to refuse food and water should be limited to terminal patients. A practical definition of terminal for this step in this protocol is the same as that used for hospice eligibility, namely, a physician-certified prognosis of 6 months or shorter in a patient with a noncurable illness.

The decision to cease oral intake has 2 features that make it justifiable by the principle of freedom from unwanted invasion. Most importantly, it is a decision about what goes beyond a patient's bodily boundaries. What goes into a person's mouth is a matter for a competent individual to determine as much as any decision about medical intervention. Force feeding of the terminally ill is rejected by the principle of bodily integrity. This is especially clear if illness has resulted in absent hunger or thirst. Evidence for suffering due to dehydration and energy deprivation is lacking for terminally ill patients, and some suggest that not eating or drinking is, to the contrary, a positive experience.⁴³ The symptom of oral dryness is not well treated by intravenous hydration and is more adequately treated by local oral care. Further, the decision remains under the patient's control over a period of more than a day or so, provided that food and water are available should the patient have a change of heart.

This seventh step, with all its component decisions, should always be revisited periodically. This normal part of all proper planning and care should help prevent needless fears and help foster good decisions. Maintaining frequent visits to the patient is an essential part of this step. Professionals should not abandon the dying. Even when no decision or action is necessary, the physician's presence remains an essential component of comfort to the patient and family.

Symptoms may crescendo as the patient dies. The option of opioid or anesthetic coma for patients in unrelievable pain should have been explored in the third step of advance care planning and may need to be discussed again if the patient is in great discomfort despite full use of other treatment approaches. The term terminal coma is avoided here because of the confusion in the literature about how it can be used with intent to euthanize. Anesthetic coma is taken to mean induction and maintenance of coma for patients with pain that can be controlled in no other acceptable fashion. It should not be used as a vehicle to remove life support while the patient is unaware. Following coma induction there should be no further actions to hasten dying, and there should be suitable monitoring to maintain the level of coma, preventing either anesthesia-induced death or unwanted return to consciousness. While patients may continue for days in this state before dying, the patient is not suffering. No further action to hasten death is warranted according to this approach, since the principles of using no unwanted intervention and of providing relief of suffering are not stretched in this approach to justify further intentional acts to secure death. So, by this protocol's justification, patients may die while under anesthetic coma, but death should not be attributable primarily to the anesthetic coma.

Finally, what if there is a rare patient who reaches step 7 and who is disinclined to cease oral intake, is ineligible for or wants to continue life-sustaining intervention, and is ineligible for or declines anesthetic coma but persists in a rational request for PAS? Physician-assisted suicide is declined in this approach, because it is not justified by the principle of nonintrusion or by the obligation to relieve suffering, the 2 claims on which this approach is founded. Physicians need to be comfortable and skilled in explaining why PAS is to be avoided. The patient's goals of, for example, securing comfort, dignity, control, or nonburdensomeness should be affirmed, and approaches that are delineated in the above steps should be put forward again as alternative routes to the patient's goals. This conversation may share much in common with step 3, in which structured deliberation and advance care planning occur. For many patients it is also appropriate to explain the principled nature of the approach. Palliative care and hospice-philosophy care with full implementation of withdrawal or withholding of life-sustaining intervention and comfort care measures offer the patient almost all, if not all, that is fair and legitimate in his or her request for help. Under such circumstances, the gain to a terminal patient is at most very small from using PAS rather than cessation of life support or oral intake along with anesthetic coma if physical suffering warrants it. Further, and not least, in all states but Oregon, the law still requires that the request be declined.

Since family members incline toward PAS more often than patients do, it is also essential to know how to approach family requests for it.⁴⁴ Essentially, the steps are the same, and they need to be followed with the same compassion and respect. However, the order may vary and the responsibility for evaluation and treatment may fall to the family members' own physician. The first step may be to involve the patient (preferably with the family there) in a structured advance planning discussion, emphasizing the need to honor the patient's wishes. This intervention can have the same relieving effect on many families that it can have on patients.⁴⁵ If this does not secure resolution, it is necessary to establish and address the root cause of the interest. This is likely to be the desire to end the patient's suffering or the burden of care. Attention to patient suffering and discussion of the family member's perceptions can resolve the latter; involvement of visiting nurse and community support resources to reduce the burden of care can help the former. Evaluation for depression and provision of full information are also necessary steps in some cases.

Physicians may also incline toward PAS or euthanasia for their patients.^{31 - 33 ,46} The risk of transference is serious,⁴⁷ and careful adherence to the protocol's standards is necessary. Above all, no action should be taken in the absence of consultation with colleagues. If needed, the physician should secure training in or assistance with end-of-life care to redress inadequate knowledge or skill in the area, which may motivate a predisposition to PAS or euthanasia.⁴⁸ Occasionally such inclination occurs in the absence of patient requests.³¹ In such cases, professional intervention may be necessary to prevent the physician from acting against the standard of patient autonomy. The physician may need to remove himself or herself from the case, providing for suitable alternative care for the patient.

If existing rights had been fully understood and their proper implementation was readily available, many cases in which PAS has seemed most justified, although still heartwrenching, would be less persuasive. The cases put before the Supreme Court to argue the need for PAS are illustrative.^{1 - 2} William Barth was suffering such severe diarrhea that he produced some 20 stools per day. By declining artificial or oral hydration, Barth could have had a natural, controlled dying in a fairly short period. William Barth probably was experiencing considerable total suffering, and relief was certainly needed. Recourse to PAS, however, seems less rational with the availability of such options as refusing oral intake. Similarly, Jane Doe, whose thyroid cancer interfered with oral intake, could have opted for either anesthetic coma or terminal dehydration to provide high degrees of comfort and control over dying.⁴⁹ Perhaps PAS seemed preferable because permissible end-of-life care was seemingly unreliable. George Kingsley did not have specific physical symptoms but feared suffering prior to death, as a friend of his had. A plan for terminal care structured by a trustworthy care team working with Kingsley could have reassured him that withdrawal and withholding of intervention and aggressive symptom management could afford him a well-managed and dignified exit. In other cases, such as metastatic cancer with spontaneous bone fractures, aggressive pain management may ultimately result in anesthetic coma and offer thereby a similar degree of relief from physical suffering as induced death but without the higher degree of moral questioning associated with PAS.^{49 - 50}

This article has argued that the right to be free of unwanted intervention is strong but recently established in the field of end-of-life care and insufficiently implemented. Similarly the obligation to provide comfort to the suffering is an ancient professional obligation that is not well implemented in this technologically powerful era. The right, on the other hand, to induce desired death is not an incremental extension from either nonintervention or comfort care but a different kind of claim, so it is not invoked. Properly implemented, the proposed approach could, even in difficult and persuasive cases, afford requesting patients their goal almost as well as or better than PAS. Future evaluation of this and other approaches will determine their effectiveness.