The President's Council on Bioethics click here to skip navigation



Table of Contents

The President's Council on Bioethics
Washington, D.C.
September 2005

The Limited Wisdom of Advance Directives
Chapter 2

As the American population ages, the dilemmas and obligations of making caregiving decisions for incapacitated patients—including decisions about when to initiate, forgo, or cease potentially life-sustaining treatments—will only become more widespread and more acute. As we described in the previous chapter, more people will experience longer periods of dependence, including years of mental incapacitation. Deciding the best course of medical care during this extended period of debility will typically fall to surrogates—including family members and friends, health care professionals and social workers, and sometimes state guardians called upon to speak for those without proxies or courts called upon to adjudicate cases in which surrogates disagree. If we fail to think ahead about what we want done and what we owe to those who can no longer speak for themselves, we are more likely to make the necessary decisions in a state of excessive confusion and crisis. But if we plan ahead thoughtlessly or unwisely, we may in fact hamper efforts, when the time arrives, to provide the kind of care that we will then need and deserve.

In the United States, our effort to think ahead about caregiving for incapacitated persons has taken shape mainly around the legal instrument of advance directives—both "instruction directives" that aim to dictate how one should be cared for and "proxy directives" that appoint others to make or execute caregiving decisions.i Advance directives came into existence at a particular time in our recent history, when people began to worry that the healer's art on which we all rely to make life better could also be used in ways that seemed ambiguous or even harmful. People worried especially that life-sustaining medical technologies might keep them alive for too long in what they perceived to be an undignified state, unrewarding to themselves and excessively burdensome to their loved ones. More generally, people worried that decisions might be made without sufficient regard to their own wishes and welfare. Advance directives were created as a way to alleviate these concerns.

A series of high-profile court cases, from the Quinlan case of 1976 to the Cruzan case of 1990, gave further credence to the value of leaving explicit instructions in advance about future treatment preferences, to be followed should one become incapacitated, and of formally designating a trusted surrogate to make medical decisions on one's behalf should one no longer be able to decide for oneself. A consensus seemed to emerge, formally ratified in the federal Patient Self-Determination Act of 1990, that widespread completion of advance directives is the best way to ensure that medical care of the incapacitated near the end of life will conform to the wishes and protect the interests of the patient. More recently, the Schiavo case has led to renewed calls for living wills: If only she had made her treatment preferences clear in advance, some argued, everyone might have been spared the wrenching decisions, bitter court battles, and national drama that ensued.

Of course, the Schiavo case—involving sudden injury to a young person, leading to a persistent vegetative state—is hardly paradigmatic of the social and ethical challenge facing our society. Vastly more typical is the patient suffering the gradual, degenerative decline toward incompetence and physical vulnerability associated with Alzheimer's disease and other dementias. It is primarily for such persons that the value of advance directives has been especially urged. And yet, studies indicate that only a small percentage of Americans actually have formal advance directives, and those that exist are often vague or limited. Thus, in most cases, the burden of decision-making for incapacitated elderly patients still falls on caregivers making contemporaneous decisions. This should not be surprising. Indeed, as we will indicate, in most cases, it will be inevitable.

Our goal in this chapter is to evaluate the wisdom and limits of advance directives—both instruction directives and proxy directives—building on the extensive work of numerous social scientists, legal scholars, medical experts, and bioethicists who have studied the subject in recent years.ii We seek not only to gauge the effectiveness of advance directives in practice, but also to reflect on the broader significance of this approach to aging and caregiving.

We state the conclusion in advance: The need to make decisions on behalf of others will only become more complicated as the American population ages; and it is misleading to think that, through wider use of living wills, competent persons will be able to direct their own care simply by leaving detailed instructions in advance. In fact, the evidence suggests that this "solution" to the problem of caregiving in an aging society is not only unrealistic but in several respects undesirable. Despite years of urging, most Americans do not have living wills, either because they would rather not think about their own dependence and death, or because they are wise enough to know that aging and dying sometimes mean placing oneself in the care of others. Not only are living wills unlikely to achieve their own stated goals, but those goals themselves are open to question. Living wills make autonomy and self-determination the primary values at a time of life when one is no longer autonomous or self-determining, and when what one needs is loyal and loving care. This paradox is at the heart of the trouble with this approach to caregiving.

This does not mean that advance directives or advance care planning are useless or unnecessary. Proxy directives serve the wise and helpful purpose of putting one's trust explicitly in the hands of loved ones who rightly bear the burden of providing care and making decisions. And advance care planning—not only about treatment preferences but also about housing arrangements and long-term care options—is a wise way to come to terms with the possibility of one's own future dependence, at a stage of life when one can still participate in such planning. But in the end, no legal instrument can substitute for wise and loving choices, made on the spot, when the precise treatment dilemma is clear and care decisions are needed. Proxy directives can appoint decisionmakers, but only ethical reflection and prudent judgment can guide them at the bedside. And advance care planning can help prevent future decisions from being made in ignorance or in crisis. But such planning should always aim at providing the best care possible for the patient as he or she might be in the future, which means providing care for a person whose precise needs can never fully be known at the time such advance planning occurs.

We acknowledge at the outset that even the most passionate advocates for living wills, the major focus of the analysis that follows, do not see these legal instruments as sufficient in themselves to address the needs of long-term care for those who suffer from age-related debility and dementia. But living wills serve as an example—perhaps a defining example—of how our society tends to approach the question of caregiving for the incapacitated, including and especially those with dementia. Seeing the limits of this legal approach clarifies the need for the kind of ethical approach that we offer in Chapters 3 and 4. Our critical analysis of advance directives is thus a prelude to the positive ethical guidelines that follow, focused not simply on discerning or executing an incompetent person's prior wishes but on providing the best care possible for the person now placed in our care.

I. Defining Key Terms

In order to understand and evaluate advance directives more fully, we first need to define some key terms more precisely:

Advance treatment directives are written or oral declarations by individuals capable of making informed and voluntary medical decisions. These declarations aim at shaping future care decisions if and when the individual loses the capacity for independent choice. Advance treatment directive (or just "advance directive") is an umbrella term that encompasses both "instruction directives" and "proxy directives."

Instruction directives are written or oral statements expressing a person's actual treatment preferences. Instruction directives can be quite specific, offering detailed descriptions of the medical interventions a person would want administered or withheld in different health situations. More often, they contain general statements, reflecting the person's basic values, about the considerations that should guide those at the bedside.

Living wills are written instruction directives. The earliest living wills typically expressed the person's wish not to receive "heroic" or "extraordinary" measures if death was "imminent." Later versions moved away from these vague terms and gave people the opportunity to refuse specific medical interventions, such as resuscitation, respirator care, antibiotics, or medical nutrition and hydration. Later versions were also designed to allow individuals to request, as well as refuse, particular types of treatment. Many living wills use standard legal forms that individuals can easily fill out at the direction of an attorney; others involve extensive narrative statements about one's personal values and treatment preferences.

Because relatively few people go through the formalities of completing a written advance directive, oral statements are the most common instruction directives available to caregivers. Relatives and clinicians sometimes consider a person's past remarks about the kind of treatments or quality of life that would be acceptable as relevant information in making medical decisions on that person's behalf. Like written instruction directives, oral directives vary in specificity and precision. They can also be more difficult to evaluate because the seriousness of the speaker's intent is not always obvious. For example, statements made in response to watching a film or visiting an ailing relative may or may not be well considered, and their significance for future caregiving decisions is often hard to judge.

Proxy directives, frequently called "health care powers of attorney," may also be written or oral. People making a proxy directive designate someone they trust to make medical decisions on their behalf and to act as their representative if they become incapacitated. People who would prefer one relative over another or a friend over a family member as their representative can use proxy directives to give effect to their preferences. Appointing a formal proxy can be especially valuable if a patient has no close family members, if the patient's relatives are dispersed, or if relatives disagree among themselves.

The two types of directives may be combined in a form that both sets forth the individual's instructions about future care and designates a proxy to cope with the actual treatment situations that later arise.iii

II. The Emergence of Advance Directives in Policy and Law

Advance directives cannot be understood in the abstract, separate from the specific context in which they emerged or the legal and public policy environment in which they now operate. Viewed historically, advance directives originated largely as a response to novel clinical circumstances, in which more and more people whose basic mental and physical capacities had been permanently lost due to illness or injury could be kept alive for extended periods by medical intervention. Many people feared living indefinitely on machines in a profoundly diminished condition. They worried about burdening loved ones, existing as mere shells of their former selves, or bankrupting their family with the costs of long-term care. In response, people sought means to restrict the kinds of medical interventions they would accept should they become incapacitated, or to appoint trusted surrogates to make medical decisions armed with the necessary legal authority to forgo or stop unwanted interventions.

Legal thinkers had a major role in promoting advance directives. A lawyer, Luis Kutner, described an early version of the living will in 1969. Expressing concern about medicine's increasing ability to prolong life in what he called "a state of indefinite vegetated animation," he suggested a written document for people seeking to avoid this fate.1 By preparing such a document, individuals could register in advance their consent to or refusal of proposed future treatments if and when they were unable to engage in independent decision-making. Kutner offered the living will as a device that would allow people to express religious or other beliefs relevant to medical care and would protect clinicians from potential liability for withholding or withdrawing life-sustaining interventions.

During the 1970s, state legislatures and courts began to extend formal legal recognition to advance treatment decision-making. In 1976, California became the first of many states to enact a law designed to allow "natural death" or "death with dignity." The California Natural Death Act declared that adult patients had the right to decide about life-sustaining medical procedures, including the right "to make a written directive instructing [their] physician to withhold or withdraw life-sustaining procedures in the event of a terminal condition."2 The directive would take effect if patients became unable to communicate their contemporaneous views on life-sustaining interventions.

The first judicial support for basing such clinical decisions on a patient's past wishes also came in 1976. Like Theresa Marie Schiavo, Karen Ann Quinlan was a young woman who had suffered severe brain damage and been diagnosed in a persistent vegetative state. Her family sought removal of the respirator that was believed to be sustaining her life. Worried about possible legal implications, physicians and hospital officials asked for a court ruling on the matter.

In its opinion resolving the case, the New Jersey Supreme Court first noted that a competent patient would be free to refuse the respirator. It then engaged in the following thought experiment:

We have no doubt, in these unhappy circumstances, that if Karen were herself miraculously lucid for an interval (not altering the existing prognosis of the condition to which she would soon return) and perceptive of her irreversible condition, she could effectively decide upon discontinuance of the life-support apparatus, even if it meant the prospect of natural death.3

The court then determined that the onset of incapacity failed to eliminate the patient's right to refuse treatment. Although she could then no longer assert the right for herself, her father-guardian could do so on her behalf. And though there was no clear evidence indicating how Quinlan would choose, the court said physicians could forgo treatment if this was her family's "best judgment" as to how she would exercise her right to decide.iv

Legal authorities in other states soon joined California and New Jersey in recognizing the competent individual's right to control future treatment. Today, all states have laws authorizing some form of advance directive.4 These laws vary in scope and coverage. For example, some state laws authorize directives for patients with any irreversible, incurable condition expected to cause death in the near future if treatment is forgone; other state laws adopt a narrower definition of eligible patients. Some state laws require an explicit statement from individuals seeking to refuse medical nutrition and hydration, while others include these measures in the general category of life-sustaining interventions that may be refused. Despite these differences, all of the existing statutes endorse the concept of advance treatment decision-making and protect from liability clinicians who act according to the terms of a properly made directive. The laws do not require individuals to make directives, however; nor do they represent the sole legally acceptable basis for forgoing life-sustaining treatment.

Federal officials, too, have offered support for advance directives. In 1990, Congress enacted the Patient Self Determination Act (PSDA).5 The law requires hospitals, nursing facilities, and other health care organizations receiving Medicare and Medicaid funds to give adults under their care information about their rights under state law to make advance directives. If someone has an advance directive, the information must be noted in that person's medical record.

Finally, the United States Supreme Court has indicated that decisions about future care may be part of the individual's constitutionally protected liberty interest. In Cruzan v. Director, Missouri Department of Health,6 the Court considered the case of another young woman diagnosed in a persistent vegetative state. Nancy Cruzan had no living will or other advance directive, and Missouri state officials, as well as the state courts, blocked her family's request to remove the feeding tube that was sustaining her life. The U.S. Supreme Court held that Missouri could prohibit cessation of treatment in the absence of Cruzan's explicit and precise wish to refuse nutrition and hydration in her current circumstances.

Writing for the majority, Chief Justice Rehnquist said that the Constitution permits (but does not require) states to demand clear and convincing evidence of an incapacitated patient's wish to refuse life-sustaining treatment. This evidentiary demand was justified, he asserted, as "a procedural safeguard to assure that the action of the surrogate [decision-maker] conforms as best it may to the wishes expressed by the patient when competent."7The Court's decision thus affirmed state policies that granted a privileged place to a patient's prior wishes in making current treatment decisions, but also affirmed the right of individual states to establish their own criteria for deciding if and when those prior wishes are clear and thus operative in particular cases. v any commentators believe that Cruzan imposes an obligation on states to respect patients' explicit advance treatment decisions, but since the case did not involve or address this question directly, the specific contours of any putative constitutional right are as yet unclear.

The widespread embrace of advance treatment decision-making by many legal officials is traceable to the law governing medical decisions by competent patients. When courts began to encounter cases involving adult patients refusing life-sustaining interventions, they proceeded to examine the strength of the individual's interest in controlling medical care and to balance that interest against competing considerations, such as the state's interests in protecting life and the ethical integrity of the health professionals caring for the patient. Courts assigned significant weight to the patient's interest in controlling care, on grounds that the patient must bear the physical and other burdens associated with life-prolonging measures. They also observed that in a pluralistic nation such as the United States, individuals have different views on the relative importance of survival and quality of life. Accordingly, very strong justification would be required to impose treatment against the competent individual's wishes.8

When patients could no longer choose their preferred treatment alternative, judges thought that the next best option would be to consult any previous instructions offered by the patient about life-sustaining measures. By giving priority to the individual's former wishes, legal authorities could seemingly continue to respect the individual's interest in controlling medical care. The courts also maintained that this approach would protect vulnerable patients from the risks of having their care determined by relatives and clinicians with their own concerns and agendas.9

Nearly three decades after Quinlan and the first Natural Death Act, advance decision-making remains popular among legal officials, at least in the abstract. Courts regard the patient's prior instructions as the preferred basis for treatment decisions, and both state and federal lawmakers have promoted this approach in legislation. Nevertheless, courts and legislatures also recognize the limits of advance treatment decision-making. The dearth of explicit directives, as well as the implementation problems discussed below, have compelled officials to develop and use other approaches to resolving questions about treatment for incapacitated patients.

Besides legislation authorizing health care proxies to make decisions at the bedside, many states now have statutes recognizing family members as appropriate surrogates for incompetent patients lacking formal directives. Moreover, when advance directives are absent or fail to offer clear guidance to surrogates, two other approaches to clinical decision-making may come into play. These two approaches are based on traditional legal standards governing financial and personal decisions for legally incompetent persons (that is, minors and adults judicially determined to lack decisional capacity). Courts, policymakers, and medical ethicists sometimes apply these approaches in the medical setting as well.

The "substituted judgment" standard combines information about a patient's past with an evaluation of that patient's current situation. As with advance directives, the goal is to determine the treatment alternative most consistent with the patient's earlier values and preferences. But the substituted judgment standard is applied when the patient failed to issue a formal advance directive. The aim is to guess, here and now, what kind of treatment the patient would choose, were he (miraculously) to become competent just long enough to say what he would want done. To reach the substituted judgment in practice, caregivers must rely on more general information, such as the patient's religious affiliation, cultural background, or attitudes toward medical care.

Several difficulties can arise in applying the substituted judgment standard. Relatives and friends may report remarks and behavior that point in different directions or that support a range of treatment options. Questions may arise about whether a patient's statement or action was meant as a deliberate and thoughtful indication of personal beliefs and values. Clinicians may wonder whether evidence about the patient's past is being accurately reported. For these reasons, many believe that the substituted judgment standard should also include a close examination of the incapacitated patient's existing condition and an evaluation of the benefits and burdens that would accompany a decision to administer or forgo treatment.10

The "best interests" standard applies when information about the patient's previous views fails to point to a particular treatment choice. Instead of focusing on evidence about the patient's personal, subjective views, the standard adopts "the perspective of a 'reasonable person,' choosing as most people would choose for themselves."11

Relatives and clinicians applying the best interests standard focus on the patient's present circumstances and consider the positive and negative effects of different treatment decisions. They examine matters such as pain, distress, and probability of treatment success. The best interests standard requires others to weigh and balance different dimensions of the patient's situation, such as the burdens that administering or forgoing treatment would impose and the benefits of the extended life that treatment could provide. Because people value quality of life and survival differently, they may reach different conclusions about appropriate treatment under this

Nine years after its Quinlan decision, the New Jersey Supreme Court offered an extensive analysis of the dementia treatment situation, developing both the substituted judgment and best interests standards in its own specific directions. In re Conroy12 involved an elderly nursing home resident with severe dementia and other health problems. She was conscious but significantly impaired and had a life expectancy of about a year. Her nephew, her only biological relative, asked the court to authorize removal of her feeding tube. There was no advance directive in this case, but the nephew maintained that she had rejected medical interventions throughout her life and would never have permitted the tube to be inserted.

In its ruling, the court set forth three standards to apply in cases like Conroy's. First is the "subjective standard," which applies when there is "clear evidence" that the patient would have made a particular treatment choice in her current circumstances. This evidence could be in the form of a living will; it might also be testimony about the patient's verbal comments, conduct, or religious beliefs. The court found that the available information on Conroy's prior preferences was insufficient to satisfy the clear evidence required under the subjective standard.

Acknowledging that it will sometimes be impossible to know for certain what the patient would choose, the court described two standards (its own versions of substituted judgment and best interests) that would permit non-treatment when clear evidence of a patient's prior wishes and intent is lacking. One is the "limited-objective standard," which applies when, as in Conroy's case, there is "some trustworthy evidence that the patient would have refused the treatment." Under this standard, treatment may be forgone if "it is clear that the burdens of the patient's continued life with the treatment outweigh the benefits of that life for him."

Conroy's third alternative is called the "pure-objective standard." This standard applies when there is no trustworthy evidence that the patient would have opposed treatment. This standard allows treatment to be forgone if two conditions are met: first, if "the net burdens of the patient's life clearly and markedly outweigh the benefits that the patient derives from life," and second, if "the recurring, unavoidable and severe pain of the patient's life with the treatment [are] such that the effect of administering life-sustaining treatment would be inhumane."

In making its decision, the court recognized that the evidence of prior preferences might be more or less probative, depending on the specificity of the individual's statements and the "remoteness, consistency, and thoughtfulness of the prior statements or actions." Its version of substituted judgment, the limited-objective standard, restricts the effect that general evidence of the patient's previous preferences may have on later care. When the evidence is less than clear, the court required decisionmakers to focus on the patient's contemporaneous interests.

By now, several other state supreme courts have joined the New Jersey court in limiting the impact of general evidence about a patient's previous beliefs and values. When asked to resolve cases involving conscious dementia or brain-injured patients, these courts refused to authorize cessation of treatment based on evidence of a patient's informal statements about medical care and quality of life.13 In the absence of a formal advance directive or relatively explicit verbal instructions, the courts based their decisions on the patient's current welfare.

Yet discerning a person's true welfare is a complicated and often controversial legal and ethical question. For some, a person's welfare is best served by always following the advance instructions given by the once-competent self, if such directions exist, even if doing so seems to harm the incapacitated person now here. For others, patient welfare hinges entirely on whether "the net burdens of the patient's life clearly and markedly outweigh the benefits that the patient derives from life," as the Conroy decision put it, suggesting that life with certain burdens is sometimes better not extended even if the life-sustaining treatment itself is not very burdensome. And for still others, serving a patient's welfare means always benefiting the life the person still has, however diminished or burdened by debility and disease.

So far, neither the courts nor policymakers have adequately investigated what patient welfare really means—a task that properly begins in ethical reflection, not legal decision-making. And indeed, the vast majority of cases regarding medical treatment of incompetent patients never come to court (or even to hospital-based ethics committees). Those that do reach the courts do so only because of serious disagreements among the caregivers, because no family member or proxy is available to make decisions, or because hospitals and medical professionals are worried about liability. But these cases remain the minority. Most difficult cases will be resolved not in courts by judges but at the bedside by family, friends, and clinicians struggling with the question: What do we owe the person now in our care? Thus, reflection on what it means to benefit (or burden) incompetent and incapacitated patients will remain a grave necessity. Our legal procedures and arrangements surely influence these ethical deliberations, but they can never replace them and in some cases they might even limit or deform them. We shall explore the ethical dimensions of caregiving in the chapters that follow. But first we must evaluate advance directives on their own terms—by looking at the ethical principles that guide this approach to caregiving and the evidence of this legal instrument's success or failure in practice. We focus first and primarily on advance instruction directives and living wills, the instrument that often gains the greatest enthusiasm but also presents the greatest problems.

III. The Principles and Aims of Advance Instruction Directives

To its advocates, the instruction directive (or living will) seems to offer an effective way of resolving the dilemmas of end-of-life care by keeping the patient firmly in charge of his or her own medical treatment. Since decisions about continuing or withdrawing medical care near the end of life are always weighty, often agonizing, and sometimes highly contentious, the living will appeals to many people as a convenient procedural solution that can bypass the endless substantive debates about what would constitute "best care" for a particular incompetent patient. The living will is seen as a vehicle by which the patient, freely exercising his or her right to informed consent, can authorize the preferred course of treatment or non-treatment, thereby relieving other persons and institutions of the burden of deciding what to do in the hardest cases, where medical intervention or non-intervention is often the difference between living longer or dying sooner.

Here, in greater detail, are the various purposes that living wills are thought to serve, both ethical and practical, and the various arguments offered in their favor.

First, living wills serve and preserve self-determination. Since it is the patient who primarily receives the benefits and bears the burdens of any medical treatment (or non-treatment), most individuals wish if possible to remain active participants in important decisions regarding their own medical care. Facing the prospect of diminished mental competence in old age, many of us are understandably concerned that such medical decisions will increasingly be made for us by others without our participation. A proxy directive can help ensure that those decisions will always be made with input from or at the direction of those who care about us. But filling out a living will seems to provide a way whereby we can direct what happens to us. We can make choices now, while still in possession of our faculties, that would determine the kinds of medical care we receive later in life, when those faculties may have diminished. In this way, the benefits and burdens of end-of-life medical care will be, in some measure, freely chosen by us rather than simply imposed from without. Each individual would remain—by this "remote control"—a self-determining agent, even when he can no longer directly or contemporaneously determine his own fate.

Second, and continuing the theme of personal freedom, living wills provide an opportunity for people to express and give effect to their personal ideas about how they would like their lives to end. In his book Life's Dominion, Ronald Dworkin argues that all of us have certain "critical interests," that is, hopes and aims that lend coherence, integrity, and meaning to our lives. According to Dworkin, these critical interests rightly shape how we see the final chapter of our lives and the specific actions we take to ensure the personal ending we want. In particular, many of us hope to avoid dying in circumstances that are out of character with the rest of our lives or in ways we find unworthy or undignified. As Dworkin puts it, most people "want their deaths, if possible, to express and in that way vividly confirm the values they believe most important to their lives."14The choices inscribed in a living will can later safeguard these critical interests at a time when the patient is losing or has lost reflective control of his or her own life.

Third, and more practically, living wills offer protection against maltreatment. Many people are concerned about the possibility of over-treatment or under-treatment at the end of life, when they are no longer in a position personally to accept or decline such care. We may be worried that life-sustaining measures will be pursued too aggressively, or we may be worried that too little will be done on our behalf, once we are unable to speak up for ourselves. The living will would seem to be a way to protect ourselves prospectively against the imposition of unwanted interventions or the premature cessation of all interventions.

Fourth, living wills can relieve anxiety and facilitate choice. Decision-making about medical care near the end of life sometimes involves a painful dilemma: should the patient's ailments be treated aggressively, with a view to maximizing longevity, or should the main focus be on insuring physical comfort, personal dignity, or the possibility of dying peacefully at home with loved ones? Rather than letting such decisions be made haphazardly or under duress, when the patient is no longer able to make his wishes known, the living will offers a way to adjudicate such tensions calmly and rationally, well in advance of the extreme circumstances in which they are likely to become manifest.

Fifth, living wills can help our loved ones. Patients are often concerned that the painful burden of end-of-life medical decision-making will fall too heavily on the shoulders of their family members, friends, or other devoted caregivers. The living will recommends itself as a way for us, in advance, to give permission to our loved ones to continue treatment in some circumstances or to surrender in others, and thereby to ease their distress in the face of our decline and death. The concern addressed here is not that they might decide badly, but that the very responsibility to decide without explicit guidance would unnecessarily add to their anguish. Even if the course of treatment chosen is ultimately the same as what the family would have chosen uninstructed, the guidance provided by a living will can help the patient's family (as well as the physician) make the painful decisions with a degree of confidence and closure.

Sixth, living wills can protect financial resources. Many people approaching the end of life worry that, in the absence of written directives explicitly limiting medical treatments, a disproportionate share of their family's resources could be squandered on costly medical interventions of limited value. They hope to avoid such a gross misallocation of resources, in which nearly all of their assets are spent down in the final months of life, when they would rather see that wealth used on behalf of their spouse, children, or grandchildren. The living will offers a way to prevent (or at least limit) excessive spending in the patient's final phase of life, by instructing surrogates not to undertake certain costly treatments in circumstances where the benefits seem limited or where life itself has become so diminished, painful, or unrewarding that extending it seems undesirable.

Seventh, living wills can decrease the risk of litigation. Physicians tend to welcome them not only as a guide to treatment but as a hedge against legal liability, should things turn out badly or relatives later regret what was decided. And prospective patients sometimes also want to forfend possible legal battles among family members who, when the time for treatment decision arrives, might disagree strongly about the preferred course of action.

Finally—an indirect benefit—living wills can foster necessary communication. The process of preparing and executing a living will can be seen as a way to promote conversation among loved ones and with doctors about one's values and preferences regarding illness, medicine, and dying. Regardless of whether the explicit directives contained in the document eventually guide the patient's medical treatment, the mere exercise of preparing a living will can encourage greater thoughtfulness and communication between the patient, his family, and his doctor as to how he would like to be cared for. Preparing a living will might be, for many people, the first occasion to articulate for themselves and for those who might care for them just how they would like their lives to end or how they would like to be cared for until the end.

These aims and purposes invite two different kinds of questions: First, how successful are instruction directives and living wills in meeting their own goals? Second, are those goals and purposes adequate to the task of caring well for those who cannot care for themselves? In the two sections that follow, we address these questions in some detail—first empirically, then ethically and philosophically.

IV. Living Wills in Practice: Evaluating Success and Failure

Enthusiasm for instruction directives and living wills, as we have noted, has been widespread in many professional quarters. The idea of the living will has been enthusiastically endorsed not only by Congress and the courts, but also by state legislatures, the Veteran's Administration, medical and legal associations, doctors, lawyers, ethicists, and patient advocacy organizations. A much cited 1991 article in the New England Journal of Medicine, for example, concluded that advance directives are desired by the great majority of patients and by the public in general, and that they can be easily completed in a 15 minute office visit with a physician.15

Advance instruction directives and living wills have been available now for some decades, and enough evidence has begun to emerge to permit a preliminary assessment of the success or failure of this particular form of advance decision-making. A number of serious problems have been documented that call into question whether living wills or other instruction directives will ever achieve the goals set out by their advocates, at least for the vast majority of patients.

1. Few people actually complete living wills.

Despite the widespread acclaim for the idea of living wills, and despite more than thirty years of encouragement, studies show that most Americans do not have one. While the rate of completion of living wills did rise in the years right after the Cruzan decision and the Patient Self-Determination Act (PSDA), there remains a large gap between the number of Americans who claim to believe that living wills are a good idea and the number of Americans who actually have them. By 2001, despite more than a decade of efforts under the PSDA to increase the number of people filling out advance directives, the completion rate nationwide remained under 25 percent.16 Even the chronically or terminally ill do not seem to prepare living wills in substantially higher numbers; one recent study suggested that only about a third of dialysis patients had a living will, even though most of them thought living wills "a good idea."17

In their review of the literature, Angela Fagerlin and Carl Schneider point to a number of reasons why people fail to complete living wills: Some people say they don't know enough about them; others find them too difficult to execute; others simply procrastinate or hesitate to discuss living wills with their doctors. Some people doubt they need a living will; others think that living wills are appropriate only for the elderly or infirm; many suspect that living wills do not affect how patients are treated. Some people are content to delegate decisions to family members, and some see living wills as incompatible with their cultural traditions or ethical beliefs, by putting too much emphasis on self-determination rather than solidarity or by implying that disabled persons are better off untreated.18

The problem does not seem to be lack of information. Many studies suggest that programs designed to increase people's awareness of living wills do not appreciably increase the likelihood of their completing them.19 Instead, people—including many who claim to believe that living wills are a good idea—seem to have substantial reasons for not completing a living will, and by and large they cannot be easily persuaded to change their minds. One recent study suggests, in particular, that most patients prefer not to put specific treatment preferences in writing; and even when individuals complete instruction directives, they typically prefer "to allow surrogate decision makers leeway in decision making."20

It should be acknowledged that, under certain favorable conditions, relatively close-knit communities have succeeded in considerably raising the rate of completion of advance directives. The "Respecting Your Choices" initiative established in La Crosse, Wisconsin in 1993 was a large-scale education and implementation effort throughout the community that featured training and education sessions, standardized materials, and coordination between local hospitals and clinics to ensure that directives would be available when needed and would play a role in treatment decisions.21 A subsequent study reported that up to 85 percent of the targeted patients had completed advance directives at the time of their deaths, and of these, most were recent (the average date of completion was a little more than a year before death), and 95 percent were available to doctors when needed. Nearly all the patients involved felt they had benefited from the process. Several other small, locally based initiatives have shown similar success, but the possibility of generalizing these efforts remains unknown.22 In 2004, Joan Teno and her colleagues reported on a 22-state survey of bereaved family members which estimated that 71 percent of the patients dying in these states had completed either a living will or durable power of attorney, many of them doing so soon before death. But the preponderance of evidence suggests that despite outreach and education efforts to increase the writing of living wills, the majority of Americans at present do not have them.23

2. People who complete living wills may not have clear treatment preferences and may not fully comprehend the clinical conditions they might face in the future.

Even when people are prepared to execute living wills, it is doubtful whether they have clear and definite ideas about the treatment they would want if and when they become incapacitated. There are, to begin with, simply too many possible future situations that the patient must try to imagine, each with its unique combination of burdens, benefits, and risks, making the notion of "informed consent" long in advance of treatment a highly questionable one. And those patients who are tempted to reject certain kinds of future medical intervention (on the ground that they "wouldn't want to live like that") may not understand how short-term use of some of the same interventions could restore them to basic or even normal function. In many cases, people end up contradicting themselves in the instructions they leave, or contradicting their instructions when asked concurrently or subsequently what their real preferences are. As Fagerlin and Schneider suggest, most people find it difficult to accurately predict their preferences "for an unspecifiable future confronted with unidentifiable maladies with unpredictable treatments."24

Indeed, most people contemplating filling out a living will know very little about the various illnesses and treatments that might one day affect them, and they often rely for that information on severely limited discussion with doctors. Moreover, studies suggest that the specific preferences recorded in living wills depend a great deal on how the questions are asked; people often change their minds about what treatment they would want in any given scenario when the illnesses and interventions are described differently. Studies also suggest that patients' individual preferences are often not very stable over time, meaning that a snapshot of those preferences frozen in a living will may not accurately reflect a person's real past preferences when a future decision needs to be made on his behalf. A meta-analysis of several studies suggested that, over periods as short as two years, almost a third of preferences for life-sustaining treatment changed.25

Data from the Robert Wood Johnson SUPPORT study suggests that many patients filling out living wills are confused about what they are being asked to decide, and vague or misinformed about the purpose and effectiveness of the medical interventions they are being asked to choose among.26 More deeply, on such difficult life-and-death questions, patients may simply not have clear and delineable preferences that they can state authoritatively in advance. The healthy may incautiously prefer death to disability, but then change their minds when they are actually sick and find themselves on the precipice between life and death. There is in fact an extensive body of research showing how poor we are at predicting our own preferences and desires, especially in regard to choices far off in the future.27 This inability is likely to be acutely present here, since we have no experience deciding how and when to die.

3. The living wills people complete often fail to convey treatment preferences clearly.

Living wills range from short questionnaires that give room for only the most general answers, to elaborate charts that allow the individual to check off which particular interventions he would or would not want in a variety of possible circumstances, to lengthy "values histories" designed to elicit the patient's overarching beliefs and core values, from which others are supposed to infer what medical treatment the patient would want if he ever becomes incapacitated. The shorter living wills offer families and clinicians very little specific guidance as to how to resolve actual treatment dilemmas. It is not very helpful, for example, to be told merely that the patient would prefer to forgo medical treatment "if the burdens of treatment outweigh the expected benefits." The more elaborate documents contain considerably more information about how the patient would like to be treated in specific scenarios. But unless the patient actually finds himself in one of the covered scenarios, that information may not offer adequate guidance as to how he should be treated. And given the complexity and particularity of every real-life clinical case—often involving co-morbidities, chronic conditions, novel treatment options, and complicated personal narratives—it is unlikely that the imagined scenario fits perfectly with the real-life one.28

There is a further complication. Studies show that patients, having completed relatively precise living wills, are often ambivalent about whether their written instructions should be followed.29 It is therefore questionable whether even the detailed living wills people fill out can be considered reliable or decisive guides to their actual treatment preferences.

Fagerlin and Schneider argue that the short questionnaires convey too little specific information to be very useful, the elaborate scenario-based charts overwhelm the patient with specifics that he cannot comprehend, and the values histories, even if a patient could be induced to complete one, would not provide much clear guidance for health care decisions in concrete situations. As they conclude: "The failure to devise workable forms is not a failure of effort or intelligence. It is a consequence of attempting the impossible."30

4. The living wills people complete often do not get transmitted to those making medical decisions.

Even if a person were to have his clear preferences inscribed in a living will, in many cases the document itself or the information contained therein will not actually reach the people responsible for the incapacitated patient's medical care. A living will signed years in advance may be misplaced or forgotten by the time it is needed. Most patients do not give their living will to their physician; and, even if they do, that physician may not be the one treating the patient by the time he has become incapacitated. One study found that, even when patients had completed living wills before being hospitalized, their medical charts contained accurate information about their directives only 26 percent of the time, and only 16 percent of the charts contained the actual form.31

One novel form of instruction directive has shown high rates of compliance and effectiveness: Physician Orders for Life-Sustaining Treatment (POLST). First developed a decade ago in Oregon and now in use in fifteen other states as well,32 a POLST document (unlike a standard living will) takes the form of a signed doctor's order; it is not completed by the patient, but by a doctor or nurse-practitioner after consulting with the patient or his surrogate. The POLST is a concise form containing specific medical instructions that can be acted on immediately by nurses, doctors, or emergency personnel; it may include "do not resuscitate" and "comfort measures only" orders, and it may indicate whether to administer CPR (cardio-pulmonary resuscitation), antibiotics, intravenous fluids, feeding tubes, artificial respiration, and other medical interventions. Unlike the living will, the POLST governs medical issues that are considered very likely to arise in the near term. According to its developers, it is really only suitable for those expecting to die within the year.33 vii

5. Even when they are transmitted to the medical decisionmakers, living wills often have little effect on surrogate decisionmaking and little impact on the care incompetent patients actually receive.

A recent study by Peter Ditto and colleagues set out to assess whether instruction directives or living wills are effective in improving the accuracy of surrogate decision-making.34 An experiment was performed involving competent outpatients aged 65 or older and their preferred surrogate decisionmakers (mostly spouses or children). All the patients completed a questionnaire asking whether they would want any of four life-sustaining medical treatments in nine different illness scenarios. Subjects in two experimental groups filled out scenario-based instruction directives and subjects in two other experimental groups filled out values-based instruction directives. Subjects in the control group filled out no advance directive at all. The surrogates were then divided into five corresponding groups: In the control group, the surrogates were asked to predict the patient's preferences for the life-sustaining medical treatments in each of the illness scenarios without the benefit of an advance instruction directive. In the four experimental groups, surrogates made such predictions after reviewing the patient's scenario-based or value-based written directive. Surrogates in two of the experimental groups also discussed the contents of the directive with the patient. The researchers then measured the accuracy of surrogate judgment in the various groups, by comparing the predicted preference with the preference actually expressed by the patient.

Strikingly, what the researchers found in this study was that, compared to the control group, none of the interventions produced significant improvement in the accuracy of the surrogates' judgment in any illness scenario or for any medical treatment. When spouses or children of elderly patients made surrogate "decisions" about medical treatment based only on their familiarity with the patient, their judgments were just as accurate as that of spouses and children who had read or read and discussed a detailed living will drawn up by the patient. In all five groups, the accuracy of surrogate decision-making was found to be about 70 percent.

In a companion study, some of the same researchers examined the effectiveness of instruction directives in improving the accuracy with which physicians could predict the treatment preferences of their older patients.35 What they found was that (a) family members generally predict patient preferences more accurately than physicians; (b) the accuracy of predictions by the patient's primary care physician (that is, a doctor who knows the patient) was not significantly improved by reading either a values-based or scenario-based living will; but (c) hospital-based physicians (that is, doctors unfamiliar with the patient) could make more accurate predictions in certain scenarios if they had read the patient's scenario-based living will.

These studies call into question whether living wills are likely to have a significant impact on the medical care received by an incompetent patient, at least in cases where surrogate decisions are made either by relatives of the patient or by physicians who know the patient. This conclusion is borne out by several studies cited by Fagerlin and Schneider, such as one completed in 1998 by Martin Goodman and colleagues,36 which concluded both that "few critically ill seniors have advance directives" and that "the level of care delivered to elderly ICU [intensive care unit] patients is not affected by the presence or absence of advance directive statements." Another study suggests that, in roughly three out of four cases, "previously executed advance directives are not accessible when patients are admitted to hospitals for acute illness";37 and yet another study gives evidence that incompetent patients frequently receive care that is inconsistent with their living will.38

In a recent study, Howard Degenholtz and colleagues found that completing a living will was in fact associated with a lower rate of in-hospital deaths, perhaps suggesting that living wills are effective at communicating patients' preferences regarding life-sustaining medical treatments.39 But, as Joan Teno has pointed out, the mere correlation between having a living will and dying outside the hospital setting does not suffice to prove that the use of living wills causes a lower rate of hospitalization. It could simply mean that those who complete living wills have, on average, a stronger preference for dying at home than those who do not. Teno's own research suggests that, the increased prevalence of advance directives notwithstanding, bereaved family members report many problems with the end-of-life institutional care received by their loved ones.40

6. The side-benefits of preparing living wills are uncertain.

Over the years, responses to these general criticisms of living wills have yielded some alternative arguments in their defense. Supporters of living wills, conceding some of the limitations, argue that in particular cases, they remain appropriate and useful. For example, for competent people about to undergo risky medical procedures, living wills may serve as effective treatment guides. Or, for people who lack close friends or relations to whom they can entrust decision-making power, living wills offer an alternative to what might otherwise be impersonal and unsympathetic care. But advocates also argue that, even in cases in which the living will has no impact on treatment decisions, filling one out is a useful way to get people thinking and talking about end-of-life treatment preferences. Such conversations, they contend, may help doctors and proxy decisionmakers later on when they are asked to make decisions on the patient's behalf.

Living wills are also said to provide a measure of comfort to patients and their surrogates, whether or not their instructions are followed. Their existence can bolster the patient's confidence that his wishes will be followed and relieve some of the stress and misery that beset family members compelled to make decisions about withdrawing life-support from a loved one. Even if the living will does not actually ensure that the patient's preferences are accurately carried out, so advocates claim, its mere existence can help make the burdens of end-of-life care more bearable for all concerned.

Needless to say, some of these alleged benefits are hard to measure. We know of little empirical research that either supports or undermines those claims regarding dementia patients.41 Besides, living wills might give much less comfort if people understood how ineffective they generally are in accomplishing their other stated purposes.

V. Conceptual and Moral Limits of Choosing in Advance

In the previous section we considered a growing body of empirical evidence suggesting that living wills are encumbered with a host of practical problems that severely limit their capacity to realize the goals they were designed to achieve. Some of the practical problems could perhaps be mitigated by improved procedures: for example, longer doctor-patient consultations could be mandated, living wills could be updated more frequently, and so forth. But, in light of the many obstacles, there is little evidence that improved procedures of this sort would measurably enhance the effectiveness of the living will program. The practical value of living wills for the general population would seem to remain highly doubtful, even if we accepted the general principle behind them: the notion that individuals can decide long in advance (as able-bodied and competent persons) what is best for them if and when they become disabled and incompetent. But that principle is itself open to question, even apart from the practical problems. There are more fundamental difficulties with living wills—conceptual and ethical—that cannot be overcome by merely procedural changes.

As we have seen, one of the central goals of the living will is to preserve for the incapacitated patient the same measure of autonomy, self-determination, and freedom of choice that he or she enjoyed while still competent. The living will is based on an attractive idea: that the right of privacy in one's own body—a right from which it follows that no patient should be subjected to medical treatment without his consent—ought to be respected even when the patient is no longer in a position to articulate his preferences. The living will aims to extend "informed consent" into the future by allowing the individual to articulate his values and wishes in the present. But there are serious reasons to doubt the wisdom of treating patient autonomy as the crucial guide for making end-of-life care decisions for patients who are, in fact, no longer autonomous but absolutely dependent on others for their care.

A.The Problem of "Informed Consent"

The living will, as distinguished from the durable power of attorney, binds a currently incompetent patient to limitations he imposed on himself in the past, when he was not able to take into account all the particular circumstances of his present situation. In the most detailed instruction directives, the person presumes to understand the personal and clinical details of an unforeseeable future, and in some cases the healthy self aims rather explicitly to decide whether life in a future diminished state is worth sustaining. But it is hard to see how prudent judgment and informed consent can operate so far removed from the real-life case in all its fine human detail and medical complexity.

This problem is especially clear in those cases in which a living will written long ago requests that particular treatments be withheld that could benefit the patient here-and-now. In such cases, to honor the patient's "autonomy" by carrying out his wishes as stated in his living will may actually be harmful rather than beneficial to the present patient. And since the incapacitated patient is unable to revise his preferences in light of vital new information, it is questionable whether carrying out his written instructions is really an exercise of informed consent at all. Informed consent usually means an agreement made in particular cases in light of current knowledge and well-understood facts. But for incompetent persons, informed consent is always impossible: the patient cannot understand the choices available or the knowledge needed to make them. To claim that the living will extends the right of free choice and self-determination to those who can no longer make informed choices seems like an illusory quest, since the written "choices" are imposed on individuals who are no longer free to change their minds.

It would seem reasonable, in such cases, that those making treatment decisions on behalf of the incapacitated patient should be able to revise or supersede the instructions recorded in the living will, in light of new information unavailable at the time the living will was executed. But then the question becomes: By what standard should the surrogate's judgment be governed when he departs from the patient's explicit instructions? Should the surrogate attempt to conjecture what the incapacitated patient would have wanted, had he known what new circumstances would arise? Or should the surrogate give up on such speculations and simply be guided by a sense of what would constitute the best possible care for the patient here-and-now?

These are surely hard questions, requiring substantive ethical judgments about what being a "benefit" to those entrusted in our care really means. The doctrine of "informed consent" is not a sufficient guide to the perplexities and obligations of ethical caregiving, and a person's prior wishes, even when clearly expressed, should not be the only relevant factor in making caregiving decisions. We need to think foremost about the patient's present welfare, and about our own role as moral agents making life-and-death decisions for others.

B.Prior Wishes and Present Welfare

A challenging illustration of the dilemmas involved in caring for incapacitated patients can be found in an account by Andrew Firlik of an Alzheimer's patient he called Margo, whom he described as "pleasantly demented."42 Firlik, a medical student, visited Margo regularly, and she was cheerful and pleased to see him but never knew his name. Margo enjoyed reading but "her place in the book jumped randomly from day to day." She enjoyed music, seemingly unaware that she was listening to the same song over and over. She enjoyed painting, but always painted the same simple pastel shapes day after day. Firlik eventually concluded that, "despite her illness, or maybe somehow because of it, Margo is undeniably one of the happiest people I have ever known":

There is something graceful about the degeneration her mind is undergoing, leaving her carefree, always cheerful. Do her problems, whatever she may perceive them to be, simply fail to make it to the worry centers of her brain? How does Margo maintain her sense of self? When a person can no longer accumulate new memories as the old rapidly fade, what remains? Who is Margo?43

Using this account of life with dementia, both Ronald Dworkin44 and Rebecca Dresser45 ask us to imagine the following scenario: Suppose that Margo executed a living will when she was first diagnosed with Alzheimer's, asking that no treatment be given to her if she contracted another serious life-threatening illness. Now Margo contracts pneumonia, which could probably be ameliorated with antibiotic treatment. Should her living will be honored and the antibiotics withheld? Or does her present contentment make her continued life worthwhile and override her past misgivings about living with dementia?

Cases like this raise the most profound questions about how much it is given to us to control or orchestrate the shape of our lives in advance, and whether we always possess the wisdom or authority to dictate what is best for a future self. Defenders of the principle of autonomy will argue that we have a right, if we wish, to escape dying under circumstances that are inconsistent with the character of the lives we chose to live. They argue, with Dworkin, that people "want their deaths, if possible, to express and in that way vividly to confirm the values most important to their lives." This means allowing individuals to decide, by means of living wills, whether life-sustaining medical interventions should be abandoned when one becomes mentally incapacitated. Dworkin suggests that it would be an "unacceptable form of moral paternalism" to disregard a patient's written instructions on the ground that he or she still derives some benefit from life in a diminished state. In Dworkin's terms, the patient's "critical interests" (her long-term desire to lead a life of "integrity" and "coherent narrative structure") outweigh her "experiential interests" (her apparent enjoyment of life with dementia).

Certainly many of us, contemplating the prospect of disability or dementia from the perspective of vigorous physical and mental health, are tempted to reject with horror the possibility of a lengthy period of mental and physical incapacity at the end of our lives. Yet, as we have seen, patients who complete living wills sometimes do so without full awareness of what life would be like in all the various circumstances that might arise. They do not and cannot know in advance whether the experience of old age with dementia will still seem valuable to a future self, even though it is not the life they would freely choose. Can individuals really know in advance that such a life would be worse than death? And, more fundamentally, do we possess a present right to discriminate against the very life of a future self, or—even more problematic—to order others to do so on our behalf?

Giving absolute priority to a patient's earlier choices in such cases may seem injurious to the patient in his or her current state; and the significant possibility that a living will was executed without full and necessary knowledge means it lacks the moral weight of an autonomous and contemporaneous choice. Moreover, as Rebecca Dresser has put it, "A policy of absolute adherence to advance directives means that we deny people like Margo the freedom we enjoy as competent people to change our decisions that conflict with our subsequent experiential interests."46

A person's prior wishes and instructions surely count in any judgment about providing care. Simply ignoring the patient's written instructions would give too little regard to the person's former beliefs about the shape and character of a good life. But giving those wishes trumping power may force caregivers to forgo doing what is best for the person who is now entrusted to their care; as moral agents themselves, caregivers cannot simply do what they were told but must also try to do what is best. Margo's apparent happiness would seem to make the argument for overriding the living will morally compelling in this particular case.

But we also need to think about what to do for those patients with dementia who are not quite so happy, whose lives seem filled with difficulty and distress. Does a person's claim on us for treatment, even if she has a living will requesting that treatment be forgone, depend entirely or primarily on the person's experiential well-being? Or are there also "interests" that are not critical or experiential but ontological—the interest of a living person in being alive, in being cared for as an equal member of the community, even in a profoundly diminished and often unhappy condition? These larger considerations take us beyond autonomy to consider the meaning of human dependence and, in our case, the meaning of being a caregiver (and care-receiver) for persons with grave disabilities and advanced dementia. It forces us to ask: Who is this person with dementia now before us, and what do we owe her even or especially when her life seems so diminished?

C. Personal Identity and the Obligations of Care

How we care for Margo—or any person with dementia, with or without a living will—will depend in great measure on how we see her identity: Is she the same person she was when competent, with the same rights and interests? Or is she a quite different person, with possibly different rights and interests? Philosophically puzzling as it is, we are probably inclined in different moments to answer "yes" to each of these questions.

One way to think of Margo as still the same person is to focus on the rational will—on the person as one who deliberates and chooses. If Margo is the same person she was, and if twenty years ago she chose to enact certain directives about her future care, then one might argue that those directives should now be honored, even if they no longer seem to further the care she most needs. Yet this way of picturing her as the same person over time seems, actually, to ignore the significance of time. Instead, it focuses on a single moment in Margo's existence—the moment her living will was enacted—and gives it governance over all future moments. As her capacity to choose diminishes, she can no longer change her instructions; she is stuck forever in that timeless point that is her earlier choice.

If we find this puzzling, as well we might, and if we doubt that Margo's instructions from twenty years ago are really the best guide to the care she now needs, we may be tempted to say that she is no longer the same person. It may seem that the only way to get out from under the tyranny of that timeless moment of choice is to say, however paradoxically, that she is now someone different—a new person with different needs, ways, and satisfactions. We can understand the attractiveness of such a position; for it does, at least, free us to care for Margo in a way that genuinely serves what are now her best interests. It takes seriously both the person she now is and the obligations we have to her.

Understandable as this perspective may be, perhaps it is possible to affirm her continuity of identity over time, yet to do so in a way that allows us to take seriously her previously expressed wishes without giving them trumping power. Margo is still the same person she once was because the body—the fact that human beings have and are bodies—is so important to our understanding of personal identity. Throughout life—both in times when our reason and will are fully developed and active, and in times when they are undeveloped or quiescent—we are physically present in embodied form as the unity of body and psyche. This is true even (as in the case of dementia) when our reason and will have been compromised by disease. It is this continuing bodily presence that is fundamental to being human and that is the locus of all personal presence. Those who embraced or held her hand years ago and who do so again today are sure that they are embracing Margo.

From this perspective, therefore, we want to affirm that Margo remains one and the same person, even as she slides ever deeper into dementia. That helps us understand why, of course, we would not want to ignore preferences she had expressed earlier in lifeviii—or, more important, the character she had developed over time and the things she had cared about. We should respect these central aspects of her person and her story, even if they are not always decisive for the care we think best for her when she is afflicted with dementia.

But although she is still the same person, she is also greatly changed; for one of the truths about embodied beings is that they change over time. The body may reach an optimal moment when reason and will are at their height, but it is also characterized by beginnings and decline—by change, development, and decay. What Dworkin terms our "critical" interests, important as they are to all of us, are themselves only part of that story of the person's development. We would do less than justice to the person Margo is—an embodied person who grows, flourishes, and declines in time—were we to designate one moment as the decisive moment for all decisions about her care, as if an instruction directive enacted twenty years earlier could be an adequate expression of her needs and desires here and now.

There is, of course, another way out of this identity conundrum. We could say that Margo was once a person but is no longer. This would not so much solve as dissolve the problem; for it would suggest that no "person" is present any longer in the bodies of those stricken by dementia, that Margo's living body is simply the corpse in which the real Margo once lived. To take this course, as some have done in recent years, would involve a radical transformation in our civilization's understanding of what it means to be human. It would sanctify reason and will (and qualities such as consciousness and self-awareness) as the qualities which, alone, give one membership in the community of care. It is far better, we think, to take seriously the truth that, whatever else they may be, human beings are embodied beings in time. Each person's life is a story marked at first largely by potential of what is yet to come, then by flourishing of the organism's most characteristic capacities, and finally by decline. But none of these moments in the story is the person. On the contrary, the person is simply the one whose story it is.

What began in this chapter as an attempt to trace our society's search for a largely procedural solution to the problem of caring for those who become incapacitated—a solution that respects patients' wishes and prevents mistreatment by others—has led us to realize that procedural solutions cannot free us from hard philosophical and ethical questions. Indeed, we need to cultivate in ourselves a deeper sense of our solidarity in body and in time—a richer sense of what it means to be dependent on others and to care for those who can no longer make choices for themselves.

The depiction of identity that we have given—of personal identity as continuous throughout the trajectory of the body's life, but of a continuity that incorporates countless changes—does not provide concrete answers to any of the difficult treatment decisions that caregivers must make when caring for dwindling patients. A great deal will always depend on the clinical and personal particulars of each case, and much of this cannot be fully anticipated in advance. But this understanding of the human person does suggest two insights that should guide the ethical reflections that follow. First, those who have lost some characteristically human cognitive capacities are still human beings with identities and still worthy of our care. Second, however much we understandably cling to our autonomy and dread our decline, dependency is very often part of the normal course of human life.

In the end, living wills can never relieve us of the responsibility we have to care for one another as best we can, even in difficult circumstances such as those dementia creates. Nor should we want such relief from responsibility; for our aim is not simply to execute instructions given us, but to develop a true ethic of caregiving. Writing a living will requires facing up to the possibility of decline, debility, and death, but it does so by seeking to exert more self-mastery than may be possible at a time of life when accepting limits and trusting others are often the virtues most needed. In the very effort to spare loved ones the excessive burdens of care, we may send a message that we do not trust them enough to put ourselves in their hands.

Of course, trusting others requires the presence of others who are trustworthy—surrogates who are willing to care, able to make wise decisions, and willing to let go when the time comes to do so. The comparative virtue of proxy directives is that they embody such trust, even if they do not, in themselves, provide ethical guidance about what to do in the face of those dilemmas that devoted proxies ultimately face. And the comparative virtue of advance care planning, as distinct from the more narrowly focused living will, is that it invites individuals and families to face up to the dependencies of the future without necessarily dictating every medical decision. At its best, such planning aims instead at being better prepared for the unknown dilemmas ahead. In our final section, we look beyond living wills to examine, briefly, these other ways of thinking in advance, so that we might understand their genuine virtues and inherent limits.

VI. Beyond Living Wills: The Wisdom and Limits of Proxy Directives and Advance Care Planning

The proxy directive does not ignore the significance of our desire to participate (in advance) in shaping treatment decisions made for us at a time in the future when we can no longer participate concurrently. Precisely by naming someone to serve as our proxy, we take that desire seriously. At the same time, however, this approach emphasizes less the importance of self-determination and correspondingly more the importance of solidarity and interdependence. It invites us to move toward our final days and years not in a spirit that isolates our free decisions from the networks of those who love and care for us but, instead, in a spirit that entrusts our dying to those who have supported us in our living. It enlists them to stay by our side, to the very end.

In this way, the proxy directive is more in accord with our ideals of family and community life than is the instruction directive. Indeed, if we try too hard to solve all problems in advance by stipulating directions for our care, we may cut off the family discussion that is needed when difficult caregiving decisions must be made. It is precisely such discussion that forces us all to take seriously the continued presence of a loved one who is no longer able to participate in decision-making. The very activity of seeking the best possible care pushes us all toward deepened understanding and concern. This does mean, of course, that a proxy directive places greater burdens on family members than does an instruction directive (one of whose aims, as we have seen, is precisely to relieve family members of such stressful burdens). But to care about one's family is to accept such burdens, and we may well wonder whether families would really be better off without the trials of fidelity that aging and dying often present, or whether we would want to live in families if families did not need to care.

To be sure, there are some, especially the very elderly, who may have no family or who may outlive all their obvious proxy choices. And there may be others who, because their family has been a source of pain rather than support, may feel unable to entrust themselves to the care of a family member. In some of these cases a living will may seem the better course to take. But such cases should also spur us all to consider more seriously how to expand the boundaries of communal solidarity—recapturing, for instance, the deep significance of friendship, or finding in community or religious institutions others to whom we can entrust decisions about our care. We should not too readily acquiesce in a vision that isolates us in the time of our dependency, or a vision that rests on the false notion that individuals can precisely determine and manage every fact of their lives until the very end.

The limits of controlling one's own future does not mean that individuals and families should not plan for the future together. The shortcomings of living wills should not obscure the real advantages, both to the patient and to his family, of thinking ahead about some of the dilemmas that might arise as one's capacities diminish, through conversation and prudent planning before an illness like dementia takes its course. Such conversation might focus not so much on the specific medical treatments a patient would or would not want as on other aspects of aging and dying that might matter even more to the person: for example, being steadily cared for during the long period of illness, having the company of one's family and friends at the end, making peace with God, having a chance to say good-bye to a particular person, dying in a quiet and dignified setting, sparing one's family additional anguish, and other considerations not strictly medical. Is the patient concerned more about pain at the end of life or about loneliness? About mental deterioration or about physical dependency? What are his deepest fears, what are her fondest hopes? Knowing how the person feels about these matters, at a stage of life when true collaboration is still possible, can give both guidance and comfort to family members who must eventually make wrenching decisions for the patient, including in many cases the decision to stop treatment and accept death.

Such conversations do not make the decisions, of course, but enrich the perspective of the decisionmakers. For in the end, the best laid plans always require devoted and prudent caregivers, who know what it means to benefit the lives of those in their care, and who possess the character to care well even in the darkest times. Ethical caregiving, in all its aspects, is the order of the day. It is therefore also the subject of the remainder of this report.



1. Kutner, L., "Due Process of Euthanasia: The Living Will, a Proposal," 44 Indiana Law Journal 539, 1969.

2. California Health & Safety Code sec. 7186 (West 1976) (repealed 2000).

3. In re Quinlan, 355 A. 2d 647 (N.J. 1976).

4. For detailed information on state laws, see Meisel, A. and Cerminara, K., The Right to Die: The Law of End-of-Life Decisionmaking. (3d ed.) New York: Aspen Publishers, 2004 (Supp. 2005).

5. 42 U.S. Code Service sections 1395i-3, 1395l, 1395cc, 1395bbb (2005).

6. 497 U.S. 261 (1990).

7. Ibid., p. 280.

8. See, for example, McKay v. Bergstedt, 801 P.2d 617 (1990).

9. Dresser, R. and Robertson, J., "Quality of Life and Nontreatment Decisions: A Critique of the Orthodox Approach," 17 Law, Medicine and Health Care 234, 1989.

10. Buchanan, A., and Brock, D., Deciding for Others: The Ethics of Surrogate Decision Making, New York: Cambridge University Press, 1989.

11. New York State Task Force on Life and the Law, When Others Must Choose, New York: New York State Task Force on Life and the Law, 1992, p. 33.

12. 486 A. 2d 1209 (N.J. 1985).

13. Dresser, R., " Schiavo's Legacy: The Need for an Objective Standard," Hastings Center Report 35(3): 20-22, 2005.

14. Dworkin, R., Life's Dominion: An Argument About Abortion, Euthanasia, and Individual Freedom, New York: Knopf, 1993, p. 211.

15. Emanuel, L. L., et al., "Advance Directives for Medical Care—A Case for Greater Use," New England Journal of Medicine 324: 889-895, 1991.

16. Eiser, A. R., and Weiss, M. D., "The Underachieving Advance Directive: Recommendations for Increasing Advance Directive Completion," American Journal of Bioethics 1: W10, 2001.

17. Holley, J. L., et al., "Factors Influencing Dialysis Patients' Completion of Advance Directives," American Journal of Kidney Disease 30: 356-360, 1997.

18. Fagerlin, A., and Schneider, C. E., "Enough: The Failure of the Living Will," Hastings Center Report 34(2): 30-42, 2004.

19. Eiser, A. R., and Weiss M. D., op. cit.

20. Hawkins, N. A., et al., "Micromanaging Death: Process Preferences, Values, and Goals in End-of-Life Medical Decision Making," Gerontologist 45: 107-117, 2005.

21. Hammes, B. J. and Rooney, B. L., "Death and End-of-Life Planning in One Midwestern Community," Archives of Internal Medicine 158: 383-390, 1998.

22. Garas, N., et al., AHRQ Evidence Reports. Chapter 49. Health Services Technology/Assessment Texts. Available online at books/bv.fcgi?rid=hstat1.section.62397 (accessed September 5, 2005).

23. See Teno, J. "Advance Directives: Time to Move On," Annals of Internal Medicine 141: 159-160, 2004; Teno J. et al., "Family Perspectives on End of Life Care," Journal of the American Medical Association 291: 88-93, 2004.

24. Fagerlin, A., and Schneider, C. E., "Enough," p. 33.

25. Ibid., p. 34.

26. Teno, J., et al., "Do Advance Directives Provide Instructions That Direct Care?" Journal of the American Geriatric Society 45: 508-512, 1997. These researchers found directives from 569 patients out of 4804; few of those contained any instructions and some of the ones with instructions showed clear lack of understanding.

28. Brett, A. S., "Limitations of Listing Specific Medical Interventions in Advance Directives," Journal of the American Medical Association 266: 825-828, 1991.

29. Sehgal, A., "How Strictly Do Dialysis Patients Want Their Advance Directives Followed?" Journal of the American Medical Association 267(1): 59-63, 1992.

30. Fagerlin, A., and Schneider, C. E., "Enough," p. 35.

31. Morrison, R. S., et al., "The Inaccessibility of Advance Directives on Transfer from Ambulatory to Acute Care Settings," Journal of the American Medical Association 274: 501-503, 1995.

32. For information about POLST, see, a website sponsored by the Oregon Health Sciences University's Center for Ethics in Healthcare.

33. "Dr. Tolle helped design the POLST form. 'A POLST form is not for every adult. A POLST form is for someone who their doctor wouldn't be surprised if they died in the coming year.'" Quoted in "A Different Kind of Living Will," Ivanhoe Broadcast News, June 27, 2005. Available online at http:// (accessed September 5, 2005).

34. Ditto, P. H., et al., "Advance Directives as Acts of Communication: A Randomized Controlled Trial," Archives of Internal Medicine 161: 421-430, 2001.

35. Coppola, K. M., et al., "Accuracy of Primary Care and Hospital-Based Physicians' Predictions of Elderly Outpatients' Treatment Preferences With and Without Advance Directives," Archives of Internal Medicine 161: 431-440, 2001.

36. Goodman, M. D., et al., "Effect of Advance Directives on the Management of Elderly Critically Ill Patients," Critical Care Medicine 26: 701-704, 1998.

37. Morrison, R. S., et al., op. cit.

38. Danis, M., et al., "A Prospective Study of Advance Directives for Life-Sustaining Care," New England Journal of Medicine 324(13): 882-888, 1991.

39. Degenholtz, H. B., et al., "Brief Communication: The Relationship Between Having a Living Will and Dying in Place," Annals of Internal Medicine 141: 113-117, 2004.

40. Teno, J. et al., "Family Perspectives on End-of-Life-Care at the Last Place of Care" Journal of the American Medical Association 291: 88-93, 2004. The authors conclude that "many people dying in institutions have unmet needs for symptom amelioration, physician communication, emotional support, and being treated with respect."

41. But see Martin, D. K., et al., "A New Model of Advance Care Planning: Observations from People with HIV," Archives of Internal Medicine 159: 86-92, 1999. This study presents some evidence that advance care planning helped people prepare for death.

42. Firlik, A.D., "Margo's Logo" (letter), Journal of the American Medical Association 265: 201, 1991.

43. Ibid.

44. Dworkin, R., Life's Dominion, pp. 221-222.

45. Dresser, R. "Dworkin on Dementia: Elegant Theory, Questionable Policy," Hastings Center Report 25(6): 32-38, 1995.

46. Ibid., p. 35.



i. "Living wills," about which we will speak at length, are a formal kind of advance instruction directive in which treatment preferences and instructions are put in writing.

ii. A list of these works appears in the Thematic Bibliography in the Appendix. We rely especially on two essays: one by Angela Fagerlin and Carl Schneider, "Enough: The Failure of the Living Will," Hastings Center Report 34(2), March-April 2004, pp. 30-42, the other by Council Member Rebecca Dresser, "Precommitment: A Misguided Strategy for Securing Death with Dignity," 81 Texas Law Review 1823, June 2003. We rely also on Prof. Schneider's presentation to the Council on this topic, December 2, 2004 (transcript available online at

iii. Other terms that have been used in court cases regarding end-of-life decisions for incompetent persons, "substituted judgment" and "best interests" standards, will be defined and discussed in the next section.

iv. As it turned out and as some experts predicted in advance, Karen Ann Quinlan lived another 10 years after the "life-sustaining" respirator was removed, breathing on her own.

v. At a later probate court hearing, witnesses offered more evidence on Cruzan's previous statements and the judge ordered the feeding tube removed.

vi. Different courts and experts have introduced their own variations and refinements on these terms—such as the "limited-objective standard" and the "pure-objective standard," used in the Conroy case discussed below. In the end, however, the crucial distinction is between trying to discern what the person would want if the competent self of the past could speak and trying to do what is best for the incapacitated patient here-and-now. In some cases, these two standards of care lead to similar conclusions; in other cases, there is tension and ambiguity, resulting from different interpretations of a patient's real wishes in the past and different judgments about a patient's real interests in the present.

vii. Inasmuch as the POLST form is focused on decisionmaking in the short term, it cannot be regarded (and is not presented by its sponsors) as a living will in the full sense, viz., a set of instructions that a competent patient can fill out well before the time when he might suffer mental incapacitation. Moreover, the animating idea of the POLST program does not seem an appropriate principle for the prudent use of living wills: the POLST document was designed in the belief that, to be most effective, an instruction directive should be conveyed with the patient everywhere he goes in a simple, standardized form, already signed by a doctor, and capable of being implemented at once by any clinician who encounters it. Yet a living will that is so "effective" in this sense might well be too effective, too easy to act on quickly, when the family might wish to make care decisions more deliberately, in light of changing circumstances and new information. See the exchange of letters between Susan E. Hickman, et al., and Angela Fagerlin and Carl Schneider, "A Viable Alternative to Traditional Living Wills: A Response to 'Enough: The Failure of the Living Will,'" Hastings Center Report (Letters to the Editor) 34(5): 4-6, 2004.

viii. If, as some people speaking loosely allege, Margo is an altogether different person, then Margo's earlier living will should be regarded as irrelevant to the care of her "replacement." Worse, on this theory of altered identity, to refuse treatment for the "new person" solely because Margo so instructed us years ago would be culpable negligence.


Next Chapter Right Arrow

  - The President's Council on Bioethics -  
Home Site Map Disclaimers Privacy Notice Accessibility NBAC HHS