The principle that a competent person has a constitutionally protected liberty interest in refusing unwanted medical treatment may be inferred from our prior decisions.
Cruzan v. Director, Missouri Department of Health, 497 US 261, 278 (1990)
A seriously ill or dying patient whose wishes are not honored may feel a captive of the machinery required for life-sustaining measures or other medical interventions.
Justice Sandra Day O'Connor, concurring opinion in Cruzan, 497 US at 28
The roles of judges, legislators, and administrative officials in influencing care at the end of life vary from the dramatic to the commonplace. On the dramatic end of the continuum are the court cases about the legality of physician-assisted suicide, which were argued before the U.S. Supreme Court as this report was being drafted. In contrast, the right of people to refuse unwanted life-sustaining and other treatments—once the subject of highly charged court cases—is now commonly accepted and enforced (if not always perfectly).
Documenting the impact of statutes, regulations, case law, and administrative actions on clinicians, patients, families, and others can be difficult. In addition, the applicability of various statutes and judicial precedents to specific patient circumstances is quite often a matter of dispute and speculation rather than straightforward matching of law to facts. Nonetheless, in the committee's view, the legal issues discussed here raise concerns either about their possible effects on compassionate and effective care for those approaching death or about the unrealistic expectations they may create or both.
This chapter considers laws relating to prescription of opioids, informed consent and advance directives, and assisted suicide. Among those with clinical, administrative, or similar involvement in end-of-life care, much of the debate about issues such as prescription regulation or informed consent is practical. For example, how can prescription laws be modified so that they do not discourage effective pain management but still respond to legitimate concerns about misuse of controlled substances? For some issues, most notably assisted suicide and euthanasia, ethical concerns may dominate legal discussions, but practical issues also arise as described later in this chapter. The focus here is primarily on how laws may affect the quality of care for dying patients.
Although the impact of malpractice litigation on medical practice is a complex and disputed question, it is discussed only briefly because the committee did not view the prospect of malpractice litigation as likely to have a significant impact on end-of-life care specifically. The committee, however, recognized concerns that physicians may engage in defensive medicine (e.g., ordering extra tests, prescribing unnecessary medications, performing hopeless CPR) because they fear being sued for a bad outcome that plaintiffs might attempt to attribute to lack of a test or procedure. Similarly, decisions might sometimes be influenced by the fear of being sued for not following a family's wishes, even if those wishes were contrary to the doctor's clinical judgment and the patient's own wishes. The committee did not find evidence that physicians were concerned about liability for failure to intervene to relieve pain or other symptoms.
In any case, many of the steps proposed in this report would tackle problems of undertreatment, overtreatment, or mistreatment of dying patients in ways that should reduce the potential for litigation and physician uncertainties and fears about being sued. At the practitioner level, these steps include changing clinicians' attitudes, knowledge, and practices so that they communicate more effectively with patients and families, engage patients and families in a process of goal setting and decisionmaking that increases trust and minimizes misunderstanding, and properly assess and treat pain and other symptoms. At the system level, they include strategies for measuring, monitoring, and improving care that seek to identify and respond to the preferences, experiences, and feelings of patients and families. If, however, these strategies fail to correct the deficits identified in Chapter 3 and if patients come to understand that the standards of care (e.g., practice guidelines) call for efforts to relieve symptoms, then litigation stemming from inattention to symptom management might become more likely—but not necessarily productive. The primary injured plaintiff would, in the case of a dying patient, likely have died, and although a family could claim injury and testify about the decedent's suffering, damages would be hard to establish. In addition, the status of practice guidelines in the courts
is still evolving. Overall, the committee was doubtful that malpractice litigation could be relied upon as an instrument to improve care at the end of life.
Prescription Laws and Barriers to Pain Relief
All patients who suffer pain—not just the dying—deserve relief through treatments that are known to be effective for most pain. Indeed, early treatment of pain as a part of a continuum of good care for those who are seriously ill may be the best approach to minimizing pain at the end of life. Other parts of this report document deficiencies in pain management and gaps in scientific knowledge. This section examines how effective pain management may be compromised by prescription drug laws that are intended to minimize drug addiction and diversion of drugs from legal to illegal sources. (Relief of dyspnea may also be affected by these laws, although this has not been the subject of much attention.) Because these laws both arise from and interact with the misperceptions and attitudes of physicians, medical boards, lawmakers, patients, and the public, reform needs to go beyond revisions in written policies to affect knowledge and values.
The Problem of Diversion and Regulatory Responses
Diversion occurs when persons with legal access to controlled substances distribute them or use them for illegal purposes or when people fraudulently obtain drugs from legal sources (Cooper et al., 1992; Cooper et al., 1993).1 Pain relief medication, for example, might be prescribed to phony patients and then sold on the streets. Alternatively, people might forge prescriptions or misrepresent their symptoms to secure prescriptions. Newspaper articles and television news reports periodically expose the problems of diverted opioids and clinician addiction. No reliable studies document the extent of opioid diversion specifically or compare it to other illegal sources (e.g., illegal imports and domestic production). A 1990 household survey estimated that 4 percent of the population over the age of 12 had used prescription analgesics, stimulants, tranquilizers, or sedatives at least once for nonmedical reasons in the preceding year, and almost 1.5 percent were currently using them (NIDA, 1991). A California estimate puts the
dollar value of diverted controlled substances during the mid-1980s at somewhere between $500 million and $1 billion (Marcus, 1996).
Legal and regulatory policies intended to prevent diversion include triplicate prescriptions and limits on the number of medication dosages that may be prescribed at any one time. These policies are burdensome and appear to deter legitimate prescribing of opioids (see, e.g., Cooper et al., 1992; IOM, 1995a, 1996d; Joranson, 1995a). Triplicate prescription programs require the prescribing physician to complete detailed, multiple-copy prescription forms. The forms themselves are often difficult to obtain and, if incorrectly filled out, must be completed again by the physician. The triplicate forms also become available to the state medical board, which may choose to pursue disciplinary measures on the basis of such information. Electronic forms and monitoring systems would ease the burden on physicians as well as allow easier monitoring but such systems have not been widely adopted or rigorously evaluated nor have appropriate norms to guide such oversight been developed and tested.
Some states have laws limiting the dosages a physician may prescribe to one patient at any given time. These laws force patients who suffer pain that requires frequent medication to request and renew prescriptions repeatedly. This not only inconveniences both patients and physicians but may subject patients to possible interruptions in pain management if something disrupts the timely requests and responses. Such problems are a special concern for patients who are not in a medical facility but are at home or in a care facility without an on-site physician.
The committee recognizes the problems created by illegal drug use and drug diversion and the need for law enforcement responses. It, however, knows of no evidence, anecdote, or other reason to believe that the prescription of opioids in the care of dying patients contributes in any meaningful way to drug diversion problems.
Effects on Care at the End of Life
The effect of anti-diversion policies on their intended targets is unclear. They do, however, appear to affect the rate of prescriptions and perhaps increase the use of less effective or even harmful medications (Cooper et al., 1993; Joranson and Gilson, 1994a, b; IOM, 1995a, 1996d). One study reported that when Texas introduced a multiple-copy prescription program, prescriptions for opioids to control pain were halved (Sigler et al., 1984). It is not known whether this dramatic drop resulted from declines in inappropriate prescribing and diversion or whether physicians and pharmacists became reluctant to prescribe appropriate medications. Nonetheless, the magnitude of the change makes it reasonable to expect that the regulation had some impact on patient care (Von Roenn et al., 1993; Wastila and
Bishop, 1996). Surveys of physicians—discussed further below—suggest that anti-diversion and anti-addiction policies combined with social antipathy toward real or imagined addiction discourages effective, appropriate, and legal pain prevention and management.
Options for Improvement
How can laws be constructed and interpreted in ways that minimize drug diversion without obstructing effective medical management of pain? Options include (1) replacing triplicate forms with electronic reporting of prescriptions and (2) allowing standing prescriptions for outpatients (to be monitored by home health care professionals or pharmacists). In addition to reducing regulatory barriers to effective pain prescribing practices, states could require that pain experts or palliative care specialists be represented on state medical boards to help inform board policies and interpretations. Information collected from triplicate or electronic prescriptions might also be analyzed to identify questionable prescribing practices, which could be used to guide education of physicians and pharmacists about effective and appropriate use of opioids. Another IOM committee has already recommended additional research on the effects of controlled substance regulations on patient care and scientific research (IOM, 1996d).
The creation of new addictions is a separate issue from the diversion of drugs to the black market. A collection of social forces joins with legal restrictions to create a general antipathy toward drug use that flows into the area of medical practice and undermines effective pain management. Even the terminology muddies the waters when chronic use of opioids, which produces physical dependence, is sometimes equated with addiction. For example, California law defines addicts as "habitual users," which might include patients with chronic pain who regularly and appropriately take opioids necessary to manage their pain (Marcus, 1996).
States have addressed the perceived problem of medically induced drug addiction through varied combinations of laws, regulations, and medical board disciplinary policies. Because the committee concluded that policies often reflect inadequate understanding of the mechanisms of pain and addiction, these mechanisms will be described before the policies are considered.
Mechanisms of Pain and Addiction
Efforts to devise reasonable anti-addiction policies are complicated by
ignorance and confusion about the biological and psychological mechanisms of pain management and addiction (Bruera et al., 1987; WHO, 1990; Nestler et al., 1993; Von Roenn et al., 1993; Portenoy et al., 1994; Buchan and Tolle, 1995; Joranson, 1995a; Portenoy, 1996). Research indicates that addiction in patients appropriately receiving opioids for pain is very small, ranging from roughly 1 in 1,000 to less than 1 in 10,000 (Porter and Jick, 1980; Angell, 1982; Jaffe, 1985; Rinaldi et al., 1988; Portenoy and Payne, 1992; Portenoy, 1996).
The committee concluded that drug tolerance and physical dependence should be more uniformly and clearly distinguished from addiction. Tolerance occurs when a constant dose of a drug produces declining effects or when a higher dose is needed to maintain an effect. Physical dependence on opioids is characterized by a withdrawal effect following discontinuation of a drug. Such dependence is a common effect in chronic pain management, but it is not restricted to opioids. Other agents such as beta-blockers, caffeine, and corticosteroids also produce physical dependence. Further, clinical evidence suggests that patients receiving opioids can be easily withdrawn from them in favor of an alternative, effective pain control mechanism if that is clinically indicated. Typical practice is to reduce the dose by fractions, stopping administration of opioids altogether after a week or so (Doyle et al., 1993). This practice may not be relevant, however, for dying patients.
Neither physical dependence nor tolerance should be equated with addiction or substance abuse. Portenoy and Kanner (1996) proposed that "addiction is a psychological and behavioral syndrome characterized by (1) the loss of control over drug use, (2) compulsive drug use, and (3) continued use despite harm" (p. 257). This is consistent with a definition proposed by the American Medical Association: "the compulsive use of a substance resulting in physical, psychological, or social harm to the user and continued use despite that harm" (Rinaldi et al., 1988, p. 556). The federal Controlled Substances Act defines an addict as someone who habitually uses an opioid in ways that endanger public health or safety (AHCPR, 1994a).
Unfortunately, the general term substance dependence is often used as a synonym for addiction, perhaps because the latter is more stigmatizing. For example, the American Psychiatric Association sets out criteria for dependence rather than addiction in its Diagnostic and Statistical Manual of Mental Disorders (4th ed., 1995). Despite a disclaimer that the scheme focuses on "maladaptive" substance use, the discussion of substance dependence may nonetheless mislead (p. 181). A later disclaimer about distinguishing legitimate medical purposes from opioid dependence is not specific, given that, as described below, many seem to be confused about what is legitimate. The committee is particularly concerned about misinterpreta-
tion of criteria related to tolerance, withdrawal, and overuse. Tolerance and withdrawal are, in general, clinically acceptable (although not necessarily invariable or desirable) consequences of effective use of opioids to manage pain, and "overuse" as defined above may be difficult to distinguish from increasing use due to uncontrolled pain, which may result from increasing pathology, tolerance, or other sources (Weissman and Haddox, 1989). Similarly, some behaviors suggestive of addiction may be confused with those resulting from inadequately managed pain or anxiety about the reliability of pain management.
Responses to the problem of addiction take several forms including some of those already identified in the discussion of drug diversion. Federal and state laws and regulations attempt to control the prescribing behavior of physicians, nurses, and pharmacists by criminalizing certain activities. In addition to legislatures and courts, state medical boards set policies that, although not having the official force of law, may be just as powerful in their effect. These policies dictate the standards by which physicians may be professionally disciplined. Laws and medical board policies are also intertwined, in that legislatures may place legal limitations on the extent of a medical board's powers.
Medical Board Policies. State medical boards may establish guidelines on pain-prescribing practices that constitute official statements of board policy. Such guidelines describe acceptable medical practice and notify health care practitioners of professional boundaries. Violating them may lead to disciplinary action. The sometimes restrictive perspective of state boards could interfere with the treatment of pain. In 1987, for instance, the Washington State Medical Disciplinary Board stated that it did "not recognize repeated prescribing of controlled drugs as appropriate therapy for chronic pain" (cited in Joranson, 1995a, pp. 2–3).
Several state medical boards have issued guidelines that deal with the use of opioids to treat intractable pain.2 In California, the nursing and pharmacy boards have also created guidelines addressing the same issue (Joranson, 1995b). These guidelines are intended not only to instruct phy-
sicians and other caregivers on the proper use of opioids in pain management but also to reduce physicians' fear of attracting board discipline for such use. Another way for state medical boards to improve pain control might be for the boards to educate the physicians within their states about how to comply with laws, regulations, and board-set standards. Information collected from triplicate prescription forms could be used in this educational effort.
Some state boards, however, continue to require that physicians avoid the potential for addiction and that they justify the continued prescribing of opioids (Joranson, 1995b). A survey of state medical board members conducted in 1991 showed that most would discourage the use of opioids to relieve chronic, noncancer pain; a third of them said they would investigate such a prescription as a potential violation of the law (Joranson et al., 1992). There is still, it seems, an inappropriate sense of distrust on the part of the medical boards, which this committee believes has developed, in part, on the basis of misperceptions discussed above about the nature and consequences of dependence and addiction.
Laws and Regulations. In 1974, the federal government, through the Federal Intractable Pain Regulation, clarified the federal law that prohibits physicians from prescribing opioids to detoxify or maintain an opioid addiction (Code of Federal Regulations, Title 21 Part 1300). The regulation states that the prohibitive regulations are "not intended to impose any limitation on a physician … to administer or dispense narcotic drugs to persons with intractable pain in which no relief or cure is possible or none has been found after reasonable effort" (21 Code of Federal Regulations, Title 21 Sec. 1306.7[c]). The policies of the Drug Enforcement Administration are similarly explicit.
Even when antiaddiction laws exempt those with intractable pain, the protections generally do not extend to those already addicted (Joranson, 1995a). When these people become patients suffering intractable pain, physicians are not free to prescribe opioids to relieve their suffering. This problem becomes especially acute in the AIDS wards of many urban hospitals.
At the state level, a number of prescribing laws include provisions that could interfere with effective medical use of opioids. For example, in New Jersey, regulations call on physicians "periodically to either cease the medication or taper down the dosage … to reduce the addiction propensity for the patient" (Joranson, 1995a).
In 1988, the Commonwealth of Virginia passed the first state law addressing the need to treat pain in terminally ill cancer patients (Joranson, 1995a). The legislation—despite its positive provisions—also illustrated the misperceptions surrounding the treatment of pain. It allowed physicians to
prescribe heroin to their terminally ill patients even though heroin is not legally available under federal law and has no significant advantages over other available opioids (Joranson, 1995a).
Texas was the first state to pass an Intractable Pain Treatment Act, in 1989. California followed suit in 1990 and Florida in 1994 (Joranson, 1995a).3
Some state pain treatment laws, (e.g., Colorado and Washington) recognize the benefits of pain control and allow physicians to prescribe controlled substances but do not address concerns about inappropriate discipline by medical boards. The Texas and California acts do address this problem by prohibiting medical board discipline of physicians who follow the provisions within the laws. Both acts also define intractable pain (following the model of the Federal Intractable Pain Regulation 4), authorize physicians to prescribe controlled substances to treat intractable pain, and prohibit health care facilities within the states from limiting such prescriptions. California's act requires an evaluation of the patient by a specialist.
Effects on Care at the End of Life
Surveys suggest that physician apprehension about addiction and anti-addiction regulations is widespread (Cleeland et al., 1986; Portenoy, 1990; Weissman, Joranson et al., 1991; Hill, 1993; Von Roenn et al., 1993). Such apprehension is not limited to physicians within the United States. In a survey of all the governments in the world conducted by the International Narcotics Control Board (within the United Nations International Drug Control Program), 47 percent of responding governments cited health care provider reluctance due to concerns about legal sanctions as an impediment to medical use of opioids (Joranson and Colleau, 1996).
The frequency of punitive action against physicians for apparently legitimate prescribing practices is unknown, but the committee heard many
anecdotes about threatening statements by medical disciplinary boards and about physicians who find the scrutiny and requirements sufficiently burdensome that they choose not to prescribe medications needed to manage pain effectively. In addition, the earlier discussion of regulations to limit drug diversion indicate that these policies may discourage the appropriate medical use of opioids and may discourage research to develop better medications.
Options for Improvement
More states could pass carefully drawn pain treatment laws. The American Medical Association (AMA) recently adopted a resolution to create a model state law, based on the Texas and California acts (AMA, 1996a). By protecting physicians from disciplinary actions, the AMA hopes to "provide patients with the security and knowledge that intractable pain resulting from terminal illness need not persist in a chronic, unrelieved manner" (AMA, 1996a, p. 4).
Although such laws constitute an important step to promote effective pain management for patients, they may not go far enough or may imply clinical clarity that does not exist. By making positive statements about the benefit of opioid use in the control of pain, legislators hope to reduce the fear of arbitrary medical board discipline. Yet they do not, in all cases, mark a clear area of medical practice in which physicians feel free to manage their patients' pain. The more specific laws, for example those that set out detailed prescription practices, may actually afford physicians less leeway in the practice of medicine. Additionally, by carving out an area of pain treatment that is immune from medical board discipline, there may be an implication that other forms of pain treatment should be subject to disciplinary review.
Even the strongest intractable pain law is still limited by the term intractable. Many cases are ambiguous, and physicians may believe that they must delay opioid treatment until pain is far enough along to be called intractable. An additional problem arises when state laws define addiction without regard to pain management. As noted earlier, California defines addicts as "habitual users," which might include patients taking opioids for chronic pain. Such confusing definitions once again expose physicians to the threat of medical board discipline.
Finally, the legal affirmations in these laws of the importance of pain control do not, in themselves, correct practice patterns or improve physician training. Laws could, however, encourage patients to expect diligence in pain relief, including use of generally effective medications. Medical boards could consider disciplining physicians who fail to apply proven methods of pain control.
Overall, the committee is encouraged by recent actions to revise drug prescribing. It urges continued review of restrictive state laws, revision of provisions that deter effective pain relief, and evaluation of the effect of regulatory changes on state medical board policies, physician attitudes and practices, patients, and illegal or harmful drug use.
Informed Consent and Advance Care Planning
A series of legal decisions over the past three decades has affirmed the right of people to refuse unwanted medical treatments (President's Commission, 1982; Faden et al., 1986; Appelbaum et al., 1987). As stated in an important 1960 California Supreme Court case, "Anglo-American law starts with the premise of thoroughgoing self-determination," which includes the right of individuals to refuse medical treatments (Natanson v. Kline, 1960). This legal reasoning reinforced a shift in emphasis in medical ethics from a dominant paternalism (i.e., action in the best interest of patients as judged by physicians) toward autonomy (i.e., patients' right to choose the course they prefer) (Childress, 1982).
One means for recognizing patient autonomy in decisionmaking is informed consent, which means that patients voluntarily accept (or refuse) a medical intervention after disclosure of its expected benefits and risks and discussion of the alternatives. For dying patients who are unconscious or in such distress that they cannot reasonably communicate their wishes when a treatment decision needs to be made, the legal concept of informed consent may have limited application.
In response, the concept of advance care planning was devised to allow people (whether or not they are "active patients") to specify how they want to be treated should serious illness or injury leave them without the capacity to make decisions or communicate (see, e.g., President's Commission, 1982; AARP, 1986; Emanuel and Emanuel, 1989; Annas, 1991; Burt, 1994). Documents used in advance care planning, called advance directives, take several forms, including surrogate decisionmaking arrangements and what are popularly called "living wills." For purposes of this report, advance directives refer particularly to statements intended to be legally binding.5
As discussed in Chapter 3, advance care planning is a broader, less legally focused concept than that of advance directives. It encompasses not
only preparation of legal documents but also discussions with family members and physicians about what the future may hold for people with serious illnesses, how patients and families want their beliefs and preferences to guide decisions (including decisions should sudden and unexpected critical medical problems arise), and what steps could alleviate concerns related to finances, family matters, spiritual questions, and other issues that trouble seriously ill or dying patients and their families. Impediments to advance planning and the implementation of written directives may be less a matter of law than of ordinary inertia or unwillingness to consider unpleasant matters. The rest of this section discusses resuscitation orders, living wills, designation of surrogates, and the Patient Self-Determination Act of 1991.
Do Not Resuscitate Orders
Do not resuscitate orders or DNRs are orders placed by a physician with a patient's or surrogate's consent into the patient's treatment chart. As discussed in Chapter 2, it is not unusual for severely ill patients, who may be dying from any of a variety of diseases, to suffer cardiac or respiratory arrests. The normal action when this occurs is called a "code."6 DNRs, or "no-codes," inform hospital staff or other caregivers that, in the event of such an episode, no attempts at revival should be made. Even when attempted, success rates of cardiopulmonary resuscitation are often low, especially for elderly patients (Murphy et al., 1989). For that reason, DNRs are sometimes called DNARs or "Do Not Attempt to Resuscitate" orders.
Because DNRs are physicians' orders, they come out of the clinical rather than the legal tradition. They thus have more in common with orders for medication or lab tests than they do with such legal documents as living wills or durable powers of attorney. Additionally, many hospitals had DNR options in place before they were required to do so by law. DNRs might, however, have some legal significance, if courts take them into account when determining whether a patient's preferences have been followed. Also, because the decision by the physician to place the DNR in the chart should be made in consultation with the patient and should reflect a patient's decision to forego certain forms of life-prolonging treatment, DNRs share with living wills and durable powers of attorney a role in the process of advance care planning.
As of 1990, 40 states and the District of Columbia allowed adults to create what is popularly called a "living will" (Strauss et al., 1990). These statutes vary in their particulars, but they generally envision that individuals may make legally binding arrangements to the effect that they shall not be sustained by medical treatment that artificially prolongs the dying process if they are in a terminal condition and can no longer make decisions.
The statutes include several safeguards against abuse. Most include a requirement that the two witnesses to the signing of the document be neither related to the patient nor involved in his or her treatment or financial support. Also, the determinations that the patient fits the statutory definition of terminal and is unable to make decisions sometimes must be made by at least two physicians. A mentally competent individual is always entitled to revoke his or her advance directive. The statutes vary on whether nutrition and hydration are considered "artificially life-sustaining" treatments. Some statutes explicitly exempt nutrition and hydration from the care a patient may choose to refuse, others give the signer the option to explicitly include them, while a third group is silent on the matter (Strauss et al., 1990).
Skeptics of living wills argue that these documents, which may be standard forms approved by the legislature of some states, provide little practical guidance in real life clinical situations, which often involve many more factors or contingencies than anticipated by standards forms (see, e.g., Brett, 1991; Lynn, 1991). Indeed, by leading patients to believe that the signing of a living will means that their preferences for an end-of-life treatment plan have been made clear, these documents could even discourage active and ongoing discussions among patients, their families, and health care professionals. In contrast, a document designating a surrogate decisionmaker could encourage such communication.
Designation of Surrogate Decisionmakers
Another legal option for advance care planning involves the designation of a surrogate to act on one's behalf in the event one becomes incompetent to make decisions about medical care. State statutes (or, in some cases, sections within the living will statutes) vary in the amount of authority a person can assign to a surrogate. For example, in California, the patient's agent, who is assigned durable power of attorney7 for health care,
may make all health care decisions that the patient could make for himself or herself, had he or she the capacity (California Civil Code Sec. 2500). The attorney-in-fact's duty is to follow the wishes of the power's grantor, but specific instructions need not be included in the document. In contrast, Nevada and Rhode Island that require statutory forms be used (Nevada Chapter 449 Secs. 2–8; Rhode Island Sec. 23-4.10-1). Grantors of the power of attorney choose options on the form, instructing their agents when to consent to or refuse life-sustaining treatments.
In one sense, although the statutes that provide for standard forms and checked options seem more specific, they may still lead to ambiguities of definition and decision. For example, when an agent is instructed to refuse treatment when that treatment's burdens outweigh the expected benefits, it remains up to the agent (with the help of the health caregivers and others involved) to make the determination. In fact, under the broader powers available under California's statute, the grantor and the agent may be more likely to sit down together and discuss the grantor's wishes, rather than have the grantor check a box and leave it at that.
Other states place even more limits on the powers of the agent. In New York, power of attorney may not be used to delegate medical decisionmaking authority, only to communicate the wishes of the grantor (Strauss et al., 1990). This inflexible provision restricts people's ability to plan ahead and may prevent humane care at the end of life.8
Decisions regarding dying children involve special considerations (Lantos and Miles, 1989; Strain, 1994; AAP, 1995; Fleischman, 1996). Although specific state laws vary, those below a certain age are legally unable to agree to or refuse medical treatment, and so others must make decisions for them. Even so, the best interests of these patients often oblige caregivers to discuss the situation with the children in ways appropriate for their developmental level and physical condition. This discussion may go beyond the sharing of information to ask children what they want for themselves (see discussion in Chapter 3). Problems arise when those with the power to consent to treatment for children disagree with each other or with clinicians. For health care providers, parental decisionmaking may also be complicated by spousal disagreement or evidence of child abuse.
Parents' decisionmaking discretion is not absolute, and pediatricians view themselves as having a professional obligation to look after the best interest of their patients (AAP, 1995). In some cases, their conclusions may conflict with those of the patients' families. Some of the most difficult cases arise from parents' demands for what clinicians regard as "futile" or "inhumane" care. The possibilities for resolving conflict include sensitive conversations between the child's physician and the parents; involvement by social workers, ethicists, pastoral counselors, or others trained in working with grief-stricken families; mediation by a hospital ethics committee; or recourse to the legal system. The latter is widely viewed as a last resort because of the burden it places on families, the stress it creates for clinicians, and the potential for negative publicity for families and institutions.
The Patient Self-Determination Act
The Patient Self-Determination Act (PSDA) was enacted by Congress in 1990 and went into effect in December 1991 (White and Fletcher, 1991; GAO, 1995b). The PSDA requires health care institutions that receive Medicare or Medicaid funds to provide written information to adult patients about state laws regarding advance directives. It also requires those institutions, among other things, to note any advance directive in a patient's file, not to discriminate between patients on the basis of whether they have an advance directive, and to educate staff and community about the availability of advance directives.
The purpose of the PSDA was to encourage greater awareness and use of advance directives so that situations of ambiguity, as illustrated by the Nancy Cruzan case, might be avoided. In that 1990 case, the United States Supreme Court recognized a competent patient's right to refuse life-sustaining treatment, but left it to the lower courts to determine whether testimony of Nancy Cruzan's previously expressed oral wishes was persuasive. In Cruzan v. Director, Missouri Health Department, Justice Sandra Day O'Connor suggested in a concurring opinion that written advance directives could dispel such ambiguity. That year, Congress passed the PSDA.
The law, however, appears to have had modest effects (Teno, Lynn et al., 1994; Morrison et al., 1994; Emanuel, 1995a; see also Chapter 3). There are no national studies on the rates of persons completing advance directives, but studies of discrete populations (e.g., nursing home residents or hospital patients) conducted both before and after passage of the PSDA show rates between 5 percent and 29 percent (GAO, 1995b; Yates and Glick, 1995). The SUPPORT investigators found a small increase of seriously ill patients having an advance directive since the PSDA went into effect (from one in five to one in four), but this increase did not translate
into higher rates of documented resuscitation discussions or DNR orders for patients who seemed to want them (Teno, Lynn, Wenger et al., 1997).
Although it requires health care organizations to provide information, the PSDA does not specify the content of that information. Often, patients are informed of their rights regarding advance directives during admission to a hospital or long-term care facility. The information is provided on a piece of paper, one of many that crosses the table during this usually stressful time. Other problems with implementation of the PSDA exist. One study found problems in the accessibility of previously completed advance directives during subsequent hospitalizations (Morrison et al., 1995). Another study found that, of the patient charts that indicated the existence of an advance directive, only 57.5 percent actually contained a copy of the directive (Yates and Glick, 1995). The study also revealed that a mere 32 percent of medical institutions covered by the law had done any community education on advance directives. The lack of involvement of physicians, especially primary doctors, also contributes to the tendency of patients to overlook the information offered.
The committee, while recognizing the value of advance directives, questions the urgency of intensive efforts to universalize their use. In this area of decisionmaking at the end of life, the law's favorite product—the legally binding document—may sometimes stand in the way of, rather than ease, the process, especially if these documents are naively viewed as ultimate solutions to the difficulties of decisionmaking. Rather, the documents known as advance directives should be seen as a set of tools useful in the ongoing process of advance care planning. Methods must be developed for encouraging continuing conversation among patients, their families, and the health professionals involved in their care. Less legalistic ways to approach planning and decisionmaking at the end of life were discussed in Chapter 3.
"Physician-assisted suicide" refers to a practice by which physicians provide, but do not directly administer, the means for a patient voluntarily to hasten his or her own death. This typically is done by prescribing lethal doses of medication that the patient then ingests. "Euthanasia," in contrast, is a practice by which the means of hastening death are administered directly by the physician, for example, when a doctor injects a patient with a lethal medication.
Controversies about assisted suicide have received recent widespread attention as a result of two lawsuits challenging the constitutionality of New York and Washington laws that prohibit physician-assisted suicide. The U.S. Supreme Court heard arguments on the cases (Vacco v. Quill, No.
95-1858 and State of Washington v. Glucksberg, No. 96–110) in early 1997.9 The litigation followed popular referenda in California and Washington in 1990 and 1991 in which proposals to legalize physician-assisted suicide were defeated. In 1992, however, Oregon voters approved a similar proposal. Oregon thus became the first jurisdiction in the United States to provide formal legal recognition of the practice of physician-assisted suicide, although court challenges delayed implementation of the law and legislative reconsideration was being discussed as this report was completed.
The committee agreed that it would not take a position on the legality or morality of assisted suicide, but it did examine some of the issues that might arise if the Supreme Court ultimately ruled either that a terminally ill person who is mentally competent and voluntarily chooses suicide has a constitutional right to self-administer lethal drugs received with the assistance of a physician or that it was constitutionally permissible for individual states, such as Oregon to permit the practice. Many of these issues were explored in friend-of-the-court briefs filed with the Court.10
Although proposals to legalize physician-assisted suicide typically include various safeguards or restrictions to protect patients and physicians, these provisions involve a number of ambiguities that might make them impossible or impractical to implement. For example, as noted earlier in this report, the status of being "terminally ill" has not been satisfactorily defined conceptually or in application because no boundary prognosis correlates precisely with an important clinical change and none can reliably be supported by data (Lynn et al., 1996). Subjective definitions of illness can be criticized as being so variable as to seem capricious. Already, several hospices have been challenged over terminal illness identifications, prognoses of survival, and small percentages of patients who survive for more than six months (see Chapter 6). In the case of care that is widely viewed as beneficial, the acceptance of some prognostic errors for a large population of patients is reasonable. It is harder to be so sanguine about such errors when the issue is assistance in suicide.
On June 30, 1997 (after the initial release of this report), the Supreme Court ruled that there is no general constitutional right to physician assistance in suicide. Some of the justices, however, wrote statements that suggested that a narrowly defined right might be upheld in specific circumstances.
See, for example, the briefs filed by the American Geriatrics Society, the American Medical Association, the American Nurses Association, the American Psychiatric Association et al., the Project on Death in America, and Ronald Dworkin, Thomas Nagel, Robert Nozick, John Rawls, Thomas Scanlon and Judith Jarvis Thomson. The latter was reprinted in The New York Review of Books, March 27, 1997, pp. 41–47. Several briefs are available at www.soros.org/death/brieftxt.html.
The criterion of voluntariness also presents problems in determining patient status and articulating boundaries (e.g., what constitutes undue influence by another party). Further, the question can be raised whether serious socioeconomic disadvantage nullifies voluntariness. If a desirable treatment would bankrupt a patient's family and, therefore, a patient chooses suicide, should a physician be authorized to assist? The dilemma between complicity with societal inequalities (by allowing assisted suicides) and magnification of them (by refusing assistance in suicides) is not readily resolvable.
Similarly, requiring that patients be mentally competent raises questions about what standards will be used, what threshold will be set, how fluctuating capacities will be handled, and what will be done about directions in advance. If competence requires very good mental functioning, then few persons known to be near death may qualify. If, however, one cannot direct suicide in advance of becoming incompetent, then people may consider preemptive suicide far in advance of death.
Proposals typically require that self-administrated prescription drugs be authorized by a physician. If many physicians consider themselves ethically or otherwise precluded from doing so, pressure for more involvement of nonphysicians is likely to arise and, perhaps, to require new safeguards.
In sum, the proposed restrictions and intended safeguards in initiatives to legalize physician-assisted suicide are problematic: difficult to define, uncertain in implementation, or vulnerable to unanticipated and unwanted consequences for those they propose to protect. Resolving uncertainties would likely be a difficult process for clinicians, and the courts almost certainly would be involved in further challenges to the implementation of assisted-suicide laws.
Other questions can be posed concerning autonomy—an individual's right to exercise free choice regarding his or her life. This is the core principle that is advanced in favor of physician-assisted suicide. The committee agrees that this principle is a centrally important value. It also believes that the current serious deficiencies in the provision of care to dying people—deficiencies highlighted throughout this report—themselves compromise the autonomy principle by depriving individuals of many choices that should, and realistically can, be made available to them. As discussed in Chapter 5, substantial numbers of dying people today suffer from avoidable pain and other symptoms, and many of the arguments for physician-assisted suicide reflect fear of pain. Offering these patients just two options—either physician assistance for hastened death or continued life with untreated pain—is a highly constricted choice that undermines the principle of autonomy. Truly autonomous choice would allow for adequate relief of pain and other distressing symptoms, adequate psychological support from properly trained health care professionals, and adequate financial and per-
sonal service support for home care in preference to impersonal hospital or nursing home settings.
If, one way or another, Oregon proceeds, the committee believes that its implementation of legal physician-assisted suicide should be carefully and intensively monitored. One key objective would be to learn whether legal safeguards are truly effective. A second objective would be to determine whether general deficiencies in the care of dying people influence individual choices for physician-assisted suicide and whether legalization stimulates correction of deficiencies. If the Oregon law is implemented, advantage should be taken of the opportunity to develop a more adequate factual basis for evaluating the competing claims for and against legal recognition of physician-assisted suicide.
Individual committee members had varied views about the morality, legality, and administrability of assisted suicide. The group fully agreed, however, that the current deficiencies in the provision of care for dying people are so extensive that they may provide inappropriate incentives for people to choose hastened death if that option is made available to them without accompanying remedial measures to improve their care. The nation should not need the prod of assisted suicide to drive it to act in behalf of the dying, although this committee, realistically, believes that media coverage of the assisted suicide cases has put the issues before the public and the professions in a very attention-getting fashion.
Reliable, excellent care at the end of life is an objective that should be supported, not impeded, by public policy. Unfortunately, some laws, regulations, and policies of public/private regulatory bodies may obstruct good care, either by their specific provisions or by the fear and misunderstanding they create. Drug-prescribing laws stand out in this regard and, in the view of the committee, warrant revisions to minimize discouragement of effective pain management. Other laws and regulations reflect an overly optimistic view of the effectiveness of laws and legal documents in clarifying how people wish to be treated when dying. Legal documents have a role to play but should not deflect attention from the more significant and complex process of advance care planning as considered in Chapter 3.
Deficiencies in care of the dying were recognized well before recent assisted suicide referenda, legislative activities, and court challenges. Nonetheless, much of the recent attention to deficiencies in end-of-life care arose only when the issue of assisted suicide came before the Supreme Court. Even if assisted suicide becomes legal, both society and the professions should feel confident that no one who chooses suicide does so because care systems are deficient in meeting their needs.