General principles

It would be good, therefore, to start with some general principles, and from there consider particular applications. A variety of approaches to ethical analysis exist; one of the most recognized is based on prima facie principles.[1] These include autonomy, beneficence, nonmaleficence, and justice. “Autonomy” (from the Greek for “self rule”) requires both liberty and the capacity for intentional action. “Nonmaleficence” is the avoidance of doing harm, and “beneficence” is the obligation to do good. “Justice” in this context is usually concerned with the fair distribution of resources. In applying these principles to ethical analysis, they prove most helpful in characterizing an ethical dilemma as a conflict between two principles, both of which we would wish to honor. Therefore, a wish to honor a patient’s autonomy, for instance, could conflict with our wish to provide beneficent treatments or minimize avoidable harms. Because autonomy in particular can weigh heavily in moral dilemmas with elderly patients whose capacity may become impaired, it is useful to remember that autonomy means not only “self rule” but is also based on “respect for persons.” Thus we may be able to find ways to respect an individual’s autonomy even as their capacity for decision making may diminish.[2]

Although principles are very useful for framing an ethical problem, they may be less effective for determining a proper resolution. Other philosophical systems, such as deontology, consequentialism, or virtue-based ethics, may prove useful here. Deontology proposes to search for the right rules, which, if followed, will lead to inevitably right outcomes. Another approach, consequentialism, will look to a good outcome as the indicator of a morally good action, and is often formulated as doing the greatest good for the greatest number. In recent years, a revitalization of virtue-based ethics places the focus on the character (and virtues) of the one doing the action, as well as the intention underlying that action. This is far from an exhaustive list; other methodologies are often employed, ranging from feminist, care-based, narrative, communitarian, to theological ethics, all having their uses, as well as their strengths and weaknesses. In reality, most individuals other than philosophers, when approaching an ethical dilemma, do not typically employ a single methodology for every case, but may adapt their thinking to the case at hand, often using a somewhat blended approach.

The patient as decision maker

When faced with the need to make decisions that are both medically and morally challenging, the ideal situation would involve a patient who is an autonomous adult with full decision-making capacity. In many cases the patient’s full decision-making capacity seems self-evident. For doubtful cases a variety of evaluative instruments have been proposed.[3] It should be remembered that capacity varies not only with age and state of health, but also is specific for certain types of decision making. Therefore, one might have capacity (a medical judgment) for making treatment decisions for oneself, even after the same patient has been declared incompetent by a court (a legal determination) for handling his or her own financial affairs. Unfortunately, patients’ capacity to make their own decisions is often not called into question until and unless they disagree with the recommendations of their physicians. A determination that a patient no longer has decision-making capacity should follow the procedures prescribed by hospital policy or state law. Most frequently the determination may require the concurrence of two physicians, and may even require that one of them be a psychiatrist.

For those patients who are deemed nonautonomous (i.e., with diminished or absent capacity), there are still methods to honor their autonomy. Foremost among these is the application of their previously indicated wishes. These may be no more than verbal declarations, but in the best case they would come in the form of a legally binding advance directive. Advance directives typically come in two categories, a living will, specifying in advance those interventions that the patient would reject or accept, or a proxy appointment. A living will may have the disadvantage of too much specificity, such as listing specific medications or even dosages, or too much generality, listing interventions that are irrelevant to the patient’s present circumstances, while not accounting for those that are. Either in conjunction with this or standing separately may be the appointment of a surrogate decision maker, sometimes in the legal form of a durable power of attorney for healthcare. More recently patients have been filling out a Physician Orders for Life-Sustaining Treatment (POLST) or Medical Orders for Life-Sustaining Treatment (MOLST) form where state law permits it. These forms are an approach that takes the items often found in a living will, puts them in the form of signed physician orders, and mandates their acceptance in the hospital or nursing home charts and therefore are immediately in force. Other patients may fill out a values inventory as part of their advance care planning. (See Chapter 59 on advance care planning.)

When a proxy decision maker has been appointed by the patient, his or her obligation is to use the patient’s values in determining a proper course of action. When a surrogate or proxy has not been appointed by the patient, some state jurisdictions list a legally binding hierarchy of relatives who would be empowered to make decisions on behalf of the patient. Unfortunately, nothing about a legally mandated hierarchy ensures that a specifically ordered relative would have knowledge of the patient’s values and preferences or even be inclined to employ them. In those less common cases where there is no morally or legally valid proxy and no indication of the patient’s prior preferences or values, decisions can only be made using a “best interest” standard, frequently after the appointment of a guardian by the court. Attempting to apply such a standard can be particularly problematic. Some have suggested an actuarial approach to predict patient preferences;[4] others suggest that the patient’s apparent acceptance and comfort with their current status should be considered more determinative than their previously stated wishes, as a truer reflection of respect for persons.

Goals of care

All medical decision making should focus on, and be in accord with, goals of care that have been determined by the treatment team, and agreed to by the patient or the patient’s surrogate. These discussions should be routinely triggered when discussing prognosis or treatments with low probability of success, or patient’s hopes and fears, or any time the physician can anticipate a remaining lifespan likely limited to 6–12 months. A model for this decision-making interaction has been described as “Beneficence in Trust.”[5] In this model, the physician and patient work together for an outcome that is mutually beneficial, and is both right (the medically appropriate action) and good (the desirable action, according to the patient’s values). The physician’s expertise is crucial in determining what is medically right, and therefore making an appropriate recommendation. Yet, we must remember that nothing about a medical education makes the physician’s judgment about what constitutes a good outcome for a particular patient take precedence over the patient’s own values. Although focus must remain on the patient’s good, the locus of the conversation may vary widely. Ideally such conversations would be included in some fashion in every outpatient encounter. In reality, a detailed discussion rarely takes place in this setting, even when the patient has a progressive, potentially terminal disease. It is a difficult conversation at best, made more problematic by the time constraints in an outpatient setting, and unfortunately represents uncompensated time when many physicians are held responsible for all their billable minutes. A less desirable location is in the lawyer’s office; although patients should be encouraged to have a last will and testament and any other necessary financial arrangements, this is not the optimal setting for discussions about their end-of-life care. In this setting, conversations which are primarily medical not legal may result in an advance care document that is more legally than medically appropriate. When a patient is admitted to the hospital or a nursing home, especially when they had been recently hospitalized for a severe progressive illness, a discussion of end-of-life care becomes more urgent. If the patient is facing a poor prognosis with the prospect of severe suffering or imminent death, or themselves have brought up hospice, palliative care, or a desire to die, such a conversation is crucial. Inexplicably, even these conditions do not always trigger timely discussions.

Actions and intentions

A discussion of goals may lack clarity when the motives or intentions behind certain actions are not made explicit, either by the physician or the patient. It is true that not every bad outcome is associated with bad intentions; in fact, this is not usually the case. Good actions can also lead to a bad outcome that may be foreseeable but unintended, and can be justified by applying the Principle of Double Effect. According to this, a morally good or neutral act may be justifiable even when a bad outcome can be foreseen, if certain conditions are met. Only the good outcome must be intended; the bad outcome, even if foreseeable, must be unintended, and the bad act cannot be the means to the good outcome. This must take place in a situation where the risks are proportional to the good that can be achieved. (See Box 60.1.) The classic example is the use of morphine in sufficiently high doses for pain relief in a terminal patient, even when there is a foreseeable risk of an earlier death to the patient due to respiratory suppression. The intention here is pain relief for the patient, but not pain relief by means of terminating the patient’s life. The true intention is made clear when escalation of the morphine dose is stopped at the point of pain relief, but at a point that does not necessarily result in an accelerated death to the sufferer.

Other axioms can also prove useful in seeking appropriate goals, such as those regarding quality of life, and withholding and withdrawing treatments. The first guideline is that quality-of-life judgments should be made according to the individual’s own value of their life and is not to be a judgment made by a third-party. This will have a great bearing and even change the nature of discussions about the benefits and burdens of a particular treatment. Although it is not uncommon to hear someone say, “I wouldn’t want to live like that,” the important point is not whether you would want to live like that, but whether the patient is willing to do so. We are notoriously bad at predicting the tolerability of conditions experienced by others; only they can ultimately make a proper judgment about whether their life is worth living and medical interventions are worth enduring. Also, it is routinely accepted that both the withholding and withdrawing of life-sustaining treatments are morally equivalent. Although they may at times feel different to the involved parties, the comparability of such actions in a dying patient is found not only in their similar outcomes, but also in their equivalent intentions. Both are aimed at allowing the patient to die without an unwanted prolongation of their dying by burdensome but ineffective life-sustaining interventions.

Actions that result in death

Finally, the question must be asked: Are there goals or intentions that should be considered morally impermissible? Most commonly, these would arise in a discussion of killing versus allowing to die. Most recently, the discussion has extended to the morality of euthanasia and physician-assisted suicide. Let us consider each of these in turn. The distinction between killing and allowing to die has enjoyed a long tradition in medical morality. The direct and intentional killing of one’s patient has been discouraged or forbidden in the Hippocratic and Judeo-Christian traditions throughout the centuries. Some confusion or even denial regarding the difference has played a significant role in the discussions of legalization of euthanasia and physician-assisted suicide (PAS). It is important, therefore, that if the distinction is real it should be understood clearly for what it is and what it is not. Some would argue that the difference is based on active versus passive “killing.” This will not take us far, because the act of turning off the respirator in a terminally ill patient, a classic instance of allowing to die, is clearly active and not passive. Causation and intentionality have also been considered important, where the death is expected to be the result of an underlying illness. In order to clarify the differences, these definitions have been offered: “Killing is an act in which an agent creates a new lethal pathophysiological state with the specific intention of causing a person’s death” and “Allowing to die is an act in which an agent removes an intervention, which is opposing a pre-existing fatal condition, or does not begin that intervention.”[6] The second definition was phrased in that way to acknowledge that allowing to die could occur with the intention of causing the death (passively) or without directly intending the person’s death. Under this concept, all killing would be wrong, but also some allowing to die could be wrong, when it is done with the intention of making the individual die. For those who do not intend the death of the patient in a withdrawal of support (which allows them to die), there may be no discernible difference in the action itself. If I unplug the ventilator, how does anyone know what my intentions were? I might be intending a patient’s death or I might not. It is in this ambiguity between concrete actions and discernible or opaque intentions that the confusion about the real difference is found. It might even be difficult for the physician performing an “unplugging” to be certain of his or her own motives. Nevertheless, there are some indicators that may clarify the different intentions. For instance, if the patient is taken off the ventilator but then begins to spontaneously breathe, does this make the participants think it was a failure? If the life-sustaining treatment is withdrawn without the subsequent death of the patient, is the next question asked, “Well, now what can we do next?” Wouldn’t these reactions be indicators of an intention not to simply remove an unwanted burdensome or ineffective intervention? Doesn’t it begin to look like the burden to be removed is the burden of the patient or the patient’s life?

In traditional medical ethics, intentionally killing our patients is considered morally wrong, and therefore the distinction between killing and allowing to die has been important. In the present era, the challenge lies not only in maintaining the distinction, but in countering a proposed refutation of the immorality of killing itself. Although the traditional proscriptions against directly killing patients (euthanasia) or supplying them the means to kill themselves (PAS) are still maintained in most locales and by most physician groups,[7] this is no longer universally true. Following practices first started in the Netherlands and Belgium, at least three states in the United States have legalized PAS through a popular referendum or legislative activity, and others through judicial opinions. The stated justification is usually one of compassion (i.e., relieving a patient of the burden of their suffering), which then extends to relieving them of the burden of their life. In the United States, only PAS has been practiced legally, as is the case in Switzerland. In Belgium and the Netherlands, both PAS and euthanasia are now options. At the present time, these options are limited to actions by physicians. Although this may be seen as an attempt to maintain societal control, no serious discussions have taken place as to why these acts should be restricted to physicians. In fact, in venues where these actions are not legalized, they still have been done by nonprofessionals, sometimes with the guidance of publications from the Hemlock Society or Compassion and Choices. These groups may provide detailed instructions about medications to use, doses to ingest, and even supplementary actions to ensure completion of the desired effect (i.e., tying a plastic bag around one’s head).[8] Some physicians who are willing to participate in PAS are less willing to take an active role in directly euthanizing a patient, with objections that are not restricted to its illegality in the United States. However, with growing acceptance of PAS, there may be growing demands for direct euthanasia by physicians. This is almost inevitable, not only because of some reports of messy attempts by non-physicians, but also because not everyone can maintain the capability to participate in PAS as a patient. In fact, if members of the public and the profession come to see this as a benefit, as a good thing, to be offered by the profession, then its extension from PAS for the terminally ill, to euthanasia or PAS for a variety of nonlethal conditions in other patients is logically inevitable. In fact, this is no longer conjecture; there are multiple examples already of this sort of “mission creep” – the Swiss organization Dignitas has assisted in suicides already for patients whose problems ranged from uncomplicated old age to loss of good looks. In the Lowlands, deaths have been arranged for depression, hearing loss, and anticipated bad prognoses for infants and children.[9, 10]

This has inevitably resulted in reaction from those members of the profession who do not see this as a good thing.[11] Those arguing against it point out that euthanasia/PAS are not normally required for beneficent care. Pain relief is almost always possible, and suffering which may be psychological, emotional, or existential is not best treated by causing the patient’s death. Moreover, what appears to be an autonomous action on the patient’s part could easily devolve into a sense of obligation or expectation, if one’s continued life is sensed to be burdensome to others, either financially or emotionally. Such a shift in traditional medical practice diminishes the value of palliative care, as it diminishes the self-worth of the patient as well, at a time when their declining capabilities may make them question their continued purpose in life.[12] Moreover, assuming such a role may do irreparable harm to the profession itself. Patients may justly be concerned if the physician pledged to safeguard their life and health is the same one offering to usher them out of it. If support for PAS and euthanasia grows, and with it the expectation that this will become the responsibility of the medical profession, it may lead to one of the most crucial ethical and moral dilemmas facing the profession in the near future.