Although the American Medical Association [4], the American Nurses' Association [5], and the American Geriatrics Society [6] oppose the legalization of physician-assisted suicide, the medical profession is increasingly acknowledging that physician-assisted suicide may have a role in terminal care. Surveys of the attitudes of physicians from several regions of the United States indicate that many physicians would consider helping a terminally ill patient commit suicide in selected cases [2, 7, 8]. The medical profession is developing greater expertise in managing terminal illness [9] but will need to develop similar expertise in responding to requests for physician-assisted suicide, if it is legalized.

To date, the debate about physician-assisted suicide has centered on the moral and legal considerations that surround its legalization. Exploring the concrete effects that this legalization would have on practice provides an important perspective that has not been adequately considered. In this paper, we reflect on information in the literature, offer our perspective on some of the practical issues that will need to be addressed by individual practitioners, and suggest areas of research that should be explored if physician-assisted suicide is legalized.

Patient Motivation

The American Medical Association and the American Geriatrics Society recommend that when a patient requests assisted suicide, the physician should seek to understand the sources of the patient's suffering and the reasons why the patient is making the request at that particular time [6, 10]. Limited information about why patients request assisted suicide is currently provided by four sources: public opinion polls, anecdotal reports by physicians, surveys of physicians' experiences with assisted suicide and euthanasia, and interviews with terminally ill patients. In a national telephone survey done in 1991, 52% of 1004 persons in the United States said that they might consider assisted suicide or euthanasia. Among the reasons cited were not wanting to be a burden, not wanting to live in pain, not wanting to be dependent on machines, and not wanting to be dependent on others for personal care [3].

Quill [11] reports that a request for physician-assisted suicide may be prompted by inadequate relief of physical symptoms, depression, problems in personal relationships, or spiritual issues. Block and Billings [12] add that untreated psychological symptoms, such as anxiety, grief, and sleep problems, may also contribute to a patient's desire to hasten death.

Several surveys have documented physicians' thoughts about what motivates patients to request physician-assisted suicide. Investigators in the Netherlands have reported that physicians who are experienced with assisted suicide and euthanasia perceive that severe physical and emotional suffering were the usual basis for requests for assistance with dying. According to these physicians, the most common reasons for these requests were wanting to die in a dignified way, being in pain, being dependent on others, and being tired of life [13, 14]. Physical symptoms other than pain that frequently underpinned such requests included dyspnea, anorexia, nausea, constipation, insomnia, and anxiety. Differences in law, culture, and health care systems, however, limit our ability to generalize these findings to persons in the United States [15].

In a recent survey [8], 93% of physicians in Oregon thought that a patient might request physician-assisted suicide because he or she feared being a burden to others; 83% thought that the request might result from financial pressures on the patient. In a survey done in Washington State [16], physicians who had received requests for physician-assisted suicide cited fearing future loss of control, being a burden, being dependent on others for personal care, and losing dignity as the primary reasons for the requests. In summary, physicians perceive that issues of dignity, control, and independence motivate requests for physician-assisted suicide more often than do unrelieved pain or other symptoms.

Recent studies based on interviews with terminally ill patients have shed new light on the reasons why patients seek physician-assisted suicide. One survey [17] found that 8.5% of 200 terminally ill patients with cancer expressed a pervasive desire to die that was associated with depression, pain, and poor social support. In another study [18], 27% of terminally ill patients with cancer had thought seriously about physician-assisted suicide or euthanasia, but only 1.9% had discussed these options with their physicians. Patients who were depressed, who were not religious, who were more physically dependent, and who were more affluent were more likely than others to have discussed these options with their physicians. In a study of 378 patients with human immunodeficiency virus (HIV) infection [19], those patients who would consider physician-assisted suicide were more likely than patients who would not consider physician-assisted suicide to be white, to have symptoms of depression, to have little social support, or to have had experience with a friend or family member who had a terminal illness.

Many questions about other factors that may motivate a request for physician-assisted suicide have remained unanswered by research to date. For example, to what extent is a request for physician-assisted suicide an attempt to battle perceived loss of control or to receive reassurance that the final decision remains with the patient? Do terminally ill patients enhance their sense of control by requesting a lethal dose of medication to have on hand in case their suffering should become intolerable in the future? To what extent do financial, emotional, or care burdens imposed on others by the patient's illness influence the patient's decision? As managed care formulates its response to the needs of terminally ill patients, the availability of palliative care options may play a role in influencing patients to consider physician-assisted suicide. The Oregon law [1] states that physician-assisted suicide must be "voluntary" on the part of the patient. Judgments about whether or not the act is voluntary are left to the physician. Although it is not clear how the clinician is supposed to reliably judge the sources of requests, physicians in the Netherlands study reported having a high degree of certainty that patients were acting autonomously when requesting physician-assisted suicide and euthanasia [13, 14].

The Patient's Mental State

Dementia and delirium often rob patients of the cognitive apparatus they need to make informed decisions and consider alternatives. In the early stages of dementia, the patient may be aware of the deterioration process and may express a desire to hasten death before becoming impaired. Cases of persons with dementia who have requested and received a physician's assistance in a suicide have been reported [20]. The point at which the disease process affects a person's ability to give truly informed consent is nebulous and needs more definition. Delirium, found in many patients with cancer and the acquired immunodeficiency syndrome, is underrecognized and underdiagnosed [21, 22]. Even subtle delirium can affect a patient's ability to see options clearly and make informed decisions.

Depression may influence decision making by diminishing the ability to appreciate life's benefits and magnifying life's burdens [23]. Depression, however, does not necessarily impair a patient's decision-making capacity or result in decisions that hasten death [24-26]. As previously noted, many studies show that depression is common among patients considering physician-assisted suicide [17-19]. Depression can be difficult to diagnose in terminally ill patients [27, 28]. Somatic symptoms may be attributable to either depression or the disease itself. When mild, such depressive psychological symptoms as dysphoria, hopelessness, diminished self-esteem, and difficulty experiencing pleasure may be realistic responses to a terminal prognosis and the limitations associated with terminal illness [29].

Research on the efficacy of the treatment of depression in terminally ill elderly patients has been limited. Although patients with severe medical illnesses are routinely excluded from depression trials, it has been suggested that the response to pharmacotherapy in medically ill persons with depression is much less pronounced than that in physically healthy persons [30]. Knowledge of the time to efficacy is especially important; studies from the Netherlands indicate that, in most cases, life is shortened by less than 1 month when euthanasia occurs [13].

Even when clinical depression is absent and decision-making abilities seem to be intact, psychotherapeutic interventions may be helpful. At the time when a patient requests physician-assisted suicide, the patient may be experiencing isolation, guilt, or low self-worth as a result of being a burden to others. Anger or other emotions may limit the patient's acceptance of alternatives. Illness, exhaustion, and various personality traits may interfere with the patients' ability to think flexibly, to consider other alternatives, and to imagine tolerating the burden of even palliative interventions. The patient may perceive that the family and physician are withdrawing, and anger about loss and abandonment may be turned inward. Consultation with a mental health professional may be useful, even if the patient has no mental disorder, to delineate approaches that attempt to reframe alternatives for the patient and counter negative thinking [12].

The Physician-Patient Relationship

If physician-assisted suicide is legalized, some aspects of the physician-patient relationship will change. Physicians and patients will need to negotiate an understanding of roles and expectations with regard to the physician's participation in the patient's suicide. The Oregon law, like many discussions of physician-assisted suicide, assumes that all patients will have an ongoing relationship with a primary care provider whom they know and trust. However, an estimated 35 million persons in the United States reside in areas that have shortages of primary care [31]. For patients who have a primary care provider, the length of time that the relationship has existed and the quality of the interaction between patient and practitioner vary. This is particularly true as the patient's illness grows more complex and specialists (such as cardiologists, oncologists, and neurologists) become involved.

For the physician who has known the patient over a period of time, both while the patient has been in crisis and while the patient has been in relatively good health, judgments about the effects of external pressures, depression, decision-making capacity, and motivation may be relatively easy to make [11]. For the physician who encounters a patient only when that patient is inquiring about the option of suicide, these judgments are more difficult. This may make physician-assisted suicide more accessible to patients who have primary care providers than to those who do not. If patients are to be dependent on their primary care providers for assistance in suicide, physicians will need to decide when to let the patient know whether they would be willing to participate in the patient's suicide. Should this discussion take place during the initial contacts between physician and patient, or should the physician defer the discussion until the patient makes a request? Surveys document that about one third of physicians have moral, ethical, or religious objections to assisting in suicides [8], but these physicians must still empathetically address the suffering of patients who request physician-assisted suicide [10]. As a profession, we need to define whether physicians who have a moral objection to suicide also have an obligation to refer patients who seek physician-assisted suicide to other providers.

The medical profession will also need to set limits on the permissibility of physician-assisted suicide. Patients who are physically unable to administer their own medications may still wish to consider suicide and may request that their physicians accommodate their needs. Some patients, in order to control the mode of their death after they have lost the mental capacity to act, may request "physician-assisted suicide" in an advance directive. To extend the same rights to all patients but to avoid possible abuses of physician-assisted suicide, professional organizations will need to develop clear clinical practice guidelines.

Defining Prognosis

A key issue in most legislation or discussion about physician-assisted suicide is the labeling of the patient as "terminal." Two definitions of this state are "Terminally Ill: means having an incurable or irreversible condition that has a high probability of causing death within a relatively short time with or without treatment (one year)" [32] and "‘Terminal Disease’ means an incurable and irreversible disease that has been medically confirmed and will, within reasonable medical judgment, produce death within six months" [1].

Many problems arise when physicians try to diagnose a disease that will be terminal or try to recognize the terminal phases of an illness. For example, a person who has recently been infected with HIV can be considered to have a condition that will be terminal, yet 10% to 17% of such persons are still without sequelae of immunodeficiency at 20 years [33]. Cardiac disease is the leading cause of death in the United States [34], but persons with atherosclerotic disease are not considered to be terminally ill even though their deaths may occur at any moment.

Patients who are terminally ill may have a single disease process (such as a brain tumor) that will, in and of itself, cause death; they may have a disease (such as leukemia) that weakens them to the point where a second condition (such as pneumonia) may overwhelm and kill them; or they may have a combination of diseases, each of which makes the other incurable (for example, severe lung disease and cardiac disease). The prognosis will alter as the patient makes decisions about treatment of the primary disease or intercurrent illnesses.

Determining prognosis and communicating probabilities is a complex task that sometimes produces misleading results. Although the methods for estimating survival in critically ill patients are increasing in sophistication (the methods used in the Study to Understand Prognoses and Preferences for Outcomes and Risks of Treatments, for example, or the Acute Physiology and Chronic Health Evaluation [APACHE III]) [35, 36], relatively few data are available on predicting survival for a patient who is not yet critically ill but who has a disease that will become terminal. The "6-month" prognosis was originally used to discuss survival with chemotherapeutic regimens and was codified to define eligibility for Medicare Hospice benefits. Nevertheless, half of the physicians who responded to the Oregon survey [8] were not confident that they could predict that a patient requesting physician-assisted suicide had less than 6 months to live.

A patient's interpretation of statistical information can vary according to the ways in which the information is presented; for example, a patient may have a 20% chance of surviving 2 months, but the mean time to death may be 2 weeks [37]. The patient will not be concerned just with the prognosis for death but will want to know when certain events are likely to happen. The patient may want to know when he or she might become dependent on others for basic functions (such as toileting), when he or she is likely to lose mental capacity, or when certain symptoms will limit enjoyment of life. Physicians need to be alert to the possibility that some patients who want to have the option of assisted suicide may feel compelled to make a decision to act before they become too debilitated to commit the act.

The Physician's Obligation to Others

The physician-patient relationship does not occur in isolation, and neither does the act of suicide. The physician interacts with other professionals, and the roles of these other professionals in the patient's suicide will need to be defined. For the patient in the community, these professionals may include visiting nurses, social workers, therapists, hospice personnel, clergy, and dispensing pharmacists. Health care institutions will need to develop policies and procedures that address the right to know or "need to know" for all professionals involved in the patient's care. If other consultants have been involved in the patient's care, their continued involvement with regard to the planned suicide will need to be negotiated. Managed care and other insurance benefits need to include evaluation by the primary care provider and mandated consultations for a mental status assessment, a second medical opinion, or palliative care expertise. Patients will also need to be aware that some life insurance policies have a 1-year to 2-year period of exclusion after the purchase of benefits for suicide.

Similarly, the physician's obligation to discuss the method of death with the patient's relatives and other loved ones, both before and after the act of suicide, must be considered. The Oregon law [1] requires that physicians ask patients to disclose their suicide plans to family members, but patients can refuse to do this. However, if death certificates and medical records accurately record the circumstances of death, this information will be accessible to many persons. Profound and enduring psychological trauma to family members and others may result when a person commits suicide [38]. It is not known whether this trauma extends to the family members of terminally ill persons who choose to commit suicide with a physician's assistance. The effects on families, especially those who are excluded from the decision by the patient or who disagree with the patient's decision, may be significant.

Method of Suicide

One of the arguments for physician-assisted suicide is that physician participation avoids incomplete suicide attempts. However, physicians have not been taught which medications to prescribe for suicide, and there is no research or case series in the professional literature to which they can refer to find prescribing information. Half of the physicians in the Oregon survey [8] were not sure what they would prescribe. It is doubtful that one standard method will be equally effective for all patients because patients have different levels of drug tolerance. In surveys from Washington State, Oregon, and the Netherlands, physicians have reported prescribing barbiturates, narcotics, tricyclic antidepressants, and benzodiazepines, often in combination. The relative efficacy and limitations of these methods are not documented in the scientific literature. Physicians who are willing to participate in physician-assisted suicide will need to gain expertise in this area to prevent incomplete suicides. The Food and Drug Administration may need to re-license drugs for this purpose.

Conclusions

The moral and constitutional issues surrounding physician-assisted suicide have been extensively debated. We suggest that the practical issues involved in implementing laws allowing physician-assisted suicide are substantial and should be part of the ongoing debate. If physician-assisted suicide is to become part of standard medical practice in the United States, we must begin to resolve the difficulties associated with its practical application. Future legislation must be carefully drafted in view of new information about these issues to avoid pitfalls, such as setting unrealistic expectations about the accuracy of prognosis. Professional societies will need to cooperate to develop practice guidelines for negotiating the necessary exchange of information in the physician-patient relationship, evaluating and treating the underlying reasons for requests for physician-assisted suicide, handling referrals and interactions among health care professionals, and addressing confidentiality and reimbursement issues. Individual physicians will need to sort out for themselves what their personal positions will be and how their practice and referrals will be affected. Professional education efforts that address palliative care options and reliable prescribing information will be essential. Key research must be done to address the lack of data in many areas, including the efficacy of palliative care and the treatment of depression in relieving the underlying reasons for requests for physician-assisted suicide; the differences in motivation between those who inquire about physician-assisted suicide, those who request it, and those who actually commit it; the ability of physicians to prognosticate in various situations and settings; and the appropriate methods to be used for physician-assisted suicide.

Dr. Lee: Medical Service, Portland Veterans Affairs Medical Center, 3710 Southwest U.S. Veterans Hospital Road, Portland, OR 97207.

Dr. Ganzini: Mental Health Division, Portland Veterans Affairs Medical Center, 3710 Southwest U.S. Veterans Hospital Road, Portland, OR 97207.