Palliative Options of Last Resort

A Comparison of Voluntarily Stopping Eating and Drinking,

Terminal Sedation, Physician-Assisted Suicide,
and Voluntary Active Euthanasia
Timothy E. Quill, MD; Bernard Lo, MD; Dan W. Brock, PhD

Palliative care is generally agreed to be the standard of request help hastening death. Between 10% and 50% of patients
care for the dying, but there remain some patients for in programs devoted to palliative care still report significant
whom intolerable suffering persists. In the face of ethical pain 1 week before death.1·57 Furthermore, patients request a
and legal controversy about the acceptability of physician- hastened death not simply because of unrelieved pain, but be¬
assisted suicide and voluntary active euthanasia, volun- cause of a wide variety of unrelieved physical symptoms in

tarily stopping eating and drinking and terminal sedation combination with loss of meaning, dignity, and independence.8·9
have been proposed as ethically superior responses of last How should physicians respond when competent, termi¬
resort that do not require changes in professional stan- nally ill patients whose suffering is not relieved by palliative
dards or the law. The clinical and ethical differences and care request help in hastening death? If the patient is receiv¬
similarities between these 4 practices are critically com- ing life-prolonging interventions, the physician should discon¬
pared in light of the doctrine of double effect, the active/ tinue them, in accordance with the patient's wishes. Some
passive distinction, patient voluntariness, proportionality patients may voluntarily stop eating and drinking (VSED). If
between risks and benefits, and the physician's potential the patient has unrelieved pain or other symptoms and accepts
conflict of duties. Terminal sedation and voluntarily stop- sedation, the physician may legally administer terminal seda¬
ping eating and drinking would allow clinicians to remain tion (TS). However, it is generally legally impermissible for
responsive to a wide range of patient suffering, but they are physicians to participate in physician-assisted suicide (PAS)
ethically and clinically more complex and closer to or voluntary active euthanasia (VAE) in response to such pa¬
physician-assisted suicide and voluntary active eutha- tient requests. The recent Supreme Court decisions that de¬
nasia than is ordinarily acknowledged. Safeguards are termined that there is no constitutional right to PAS placed
presented for any medical action that may hasten death, great emphasis on the importance of relieving pain and suf¬
including determining that palliative care is ineffective, fering near the end of life.10·11 The Court acknowledged the
obtaining informed consent, ensuring diagnostic and legal acceptability of providing pain relief, even to the point of
prognostic clarity, obtaining an independent second opin- hastening death if necessary, and left open the possibility that
ion, and implementing reporting and monitoring pro- states might choose to legalize PAS under some circumstances.
cesses. Explicit public policy about which of these prac- In this article, we compare VSED, TS, PAS, and VAE as po¬
tices are permissible would reassure the many patients tential interventions of last resort for competent, terminally ill
who fear a bad death in their future and allow for a predict- patients who are sufferingintolerably in spite ofintensive efforts
able response for the few whose suffering becomes intol- to palliate and who desire a hastened death. Some clinicians and
erable in spite of optimal palliative care. patients may find some of the differences between these prac¬
JAMA. 1997;278:2099-2104 tices to be ethically and psychologically critical, whereas others

PALLIATIVE CARE is the standard ofcare when terminally

perceive the differences as inconsequential. We will define and
compare the practices,examineunderlying ethicaljustifications,
ill patients find that the burdens of continued life-prolonging and consider appropriate categories ofsafeguards forwhichever
treatment outweigh the benefits.1"4 To better relieve suffering
practices our society eventually condones.
near the end of life, physicians need to improve their skills in pal¬
liative care and to routinely discuss it earlier in the course of DEFINITIONS AND CLINICAL COMPARISONS
terminal illness. In addition, access to palliative care needs to be With VSED, a patient who is otherwise physically capable
improved, particularly for those Americans who lack health in¬ of taking nourishment makes an active decision to discontinue
surance. However, even the highest-quality palliative care fails
all oral intake and then is gradually "allowed to die," primarily
or becomes unacceptable for some patients, some of whom
of dehydration or some intervening complication.12"14 Depend¬
From the Program for Biopsychosocial Studies, University of Rochester School of
ing on the patient's preexisting condition, the process will
Medicine and Dentistry, Rochester, NY (Dr Quill); the Program in Medical Ethics, Uni- usually take 1 to 3 weeks or longer if the patient continues to
versity of California, San Francisco (Dr Lo); and the Center for Biomedical Ethics,
Brown University, Providence, RI (Dr Brock).
Dr Quill's views do not necessarily reflect those of the University of Rochester Health Law and Ethics section editors: Lawrence O. Gostin, JD, the
School of Medicine or its Department of Medicine. Georgetown/Johns Hopkins University Program on Law and Public Health,
Reprints: Timothy E. Quill, MD, Department of Medicine, The Genesee Hospital, Washington, DC, and Baltimore, Md; Helene M. Cole, MD, Contributing Editor,
224 Alexander St, Rochester, NY 14607 (e-mail: [email protected]). JAMA.

Downloaded From: http://jama.jamanetwork.com/ by a University of Pittsburgh User on 06/19/2015

 take some fluids. Voluntarily stopping eating and drinking has to use TS in patients who lack decision-making capacity but
several advantages. Many patients lose their appetites and appear to be suffering intolerably, provided that the patient's
stop eating and drinking in the final stages of many illnesses. suffering is extreme and otherwise unrelievable, and the sur¬
Ethically and legally, the right of competent, informed pa¬ rogate or family agrees.
tients to refuse life-prolonging interventions, including arti¬ Nonetheless, TS remains controversial2123 and has many of
ficial hydration and nutrition, is firmly established, and vol¬ the same risks associated with VAE and PAS. Like VAE, the
untary cessation of "natural" eating and drinking could be final actors are the clinicians, not the patient. Terminal sedation
considered an extension ofthat right. Because VSED requires could therefore be carried out without explicit discussions with
considerable patient resolve, the voluntary nature of the ac¬ alert patients who appear to be suffering intolerably or even
tion should be clear. Voluntarily stopping eating and drinking against their wishes. Some competent, terminally ill patients
also protects patient privacy and independence, so much so reject TS. They believe that their dignity would be violated if
that it potentially requires no participation by a physician. they had to be unconscious for a prolonged time before they die,
The main disadvantages of VSED are that it may last for or that their families would suffer unnecessarily while waiting
weeks and may initially increase suffering because the patient for them to die. Patients who wish to die in their own homes may
may experience thirst and hunger. Subtle coercion to proceed not be able to arrange TS because it probably requires admis¬
with the process may occur if patients are not regularly of¬ sion to a health care facility. There is some controversy in the
fered the opportunity to eat and drink, yet such offers may be anesthesia literature about whether heavily sedated persons
viewed as undermining the patient's resolve. Some patients, are actually free of suffering or simply unable to report or re¬
family members, physicians, or nurses may find the notion of member it.2426 In some clinical situations, TS cannot relieve the
"dehydrating" or "starving" a patient to death to be morally patient's symptoms, as when a patient is bleeding uncontrolla¬
repugnant. For patients whose current suffering is severe and bly from an eroding lesion or a refractory coagulation disorder,
unrelievable, the process would be unacceptable without se¬ cannot swallow secretions because of widespread oropharyn-
dation and analgesia. If physicians are not involved, palliation geal cancer, or has refractory diarrhea from the acquired im¬
of symptoms may be inadequate, the decision to forgo eating munodeficiency syndrome (AIDS). Although such patients are
and drinking may not be informed, and cases of treatable de¬ probably not conscious of their condition once sedated, their
pression may be missed. Patients are likely to lose mental death is unlikely to be dignified or remembered as peaceful by
clarity toward the end of this process, which may undermine their families. Finally, and perhaps most critically, there may be
their sense of personal integrity or raise questions about confusion about the physician's ethical responsibility for con¬
whether the action remains voluntary. tributing to the patient's death.21·22
Although several articles,12·13 including a moving personal With PAS, the physician provides the means, usually a pre¬
narrative,14 have proposed VSED as an alternative to other scription of a large dose of barbiturates, by which a patient can
forms of hastened death, there are no data about how fre¬ end his or her life.1·3,27 Although the physician is morally re¬
quently such decisions are made or how acceptable they are to sponsible for this assistance, the patient has to carry out the
patients, families, physicians, or nurses. final act. Physician-assisted suicide has several advantages.
With TS, the suffering patient is sedated to unconsciousness, For some patients, access to a lethal dose of medication may
usually through ongoing administration of barbiturates or ben- give them the freedom and reassurance to continue living,
zodiazepines. The patient then dies of dehydration, starvation, knowing they can escape if and when they choose.28·29 Because
or some other intervening complication, as all life-sustaining patients have to ingest the drugby their own hand, their action
interventions are withheld.15"18 Although death is inevitable, it is likely to be voluntary. Physicians report being more com¬
usually does not take place for days or even weeks, depending on fortable with PAS than VAE,3032 presumably because their
clinical circumstances. Because patients are deeply sedated dur¬ participation is indirect.
ing this terminal period, they are believed to be free of suffering. Opponents of PAS believe that it violates traditional moral
It can be argued that death with TS is "foreseen" but not and professional prohibitions against intentionally contributing
"intended" and that the sedation itself is not causing death.1518 to a patient's death. Physician-assisted suicide also has several
The sedation is intended to relieve suffering, a long-standing and practical disadvantages. Self-administration does not guarantee
uncontroversial aim of medicine, and the subsequent withhold¬ competence or voluntariness. The patient may have impaired
ing of life-sustaining therapy has wide legal and ethical accep¬ judgment at the time of the request or the act or may be influ¬
tance. Thus, TS probably requires no change in the law. The enced by external pressures. Physician-assisted suicide is lim¬
recent Supreme Court decision gave strong support to TS, say¬ ited to patients who are physically capable of taking the medica¬
ing that pain in terminally ill patients should be treated, even to tion themselves. It is not always effective,33·34 so families may be
the point of rendering the patient unconscious or hastening faced with a patient who is vomiting, aspirating, or cognitively
death.10·11 Terminal sedation is already openly practiced by some impaired, but not dying. Patients brought to the emergency de¬
palliative care and hospice groups in cases of unrelieved suffer¬ partment after ineffective attempts are likely to receive un¬
ing, with a reported frequency from 0% to 44% of cases.1·6·7,15"20 wanted life-prolonging treatment. Requiring physicians to be
Terminal sedation has other practical advantages. It can be present when patients ingest the medication could coerce an am¬
carried out in patients with severe physical limitations. The bivalent patient to proceed, yet their absence may leave families
time delay between initiation of TS and death permits second- to respond to medical complications alone.
guessing and reassessment by the health care team and the Physician-assisted suicide is illegal in most states, but no
family. Because the health care team must administer medi¬ physicians have ever been successfully prosecuted for their
cations and monitor effects, physicians can ensure that the participation.3 Several studies have documented a secret prac¬
patient's decision is informed and voluntary before beginning tice of PAS in the United States. In Washington State, 12% of
TS. In addition, many proponents believe that it is appropriate physicians responding to a survey had received genuine re-

Downloaded From: http://jama.jamanetwork.com/ by a University of Pittsburgh User on 06/19/2015

 quests for PAS within the year studied.8 Twenty-four percent may hasten death (forgoing life-sustaining treatment and
of requests were acceded to, and over half of those patients high-dose pain medications) and those that are impermissible
died as a result. An Oregon study showed similar results.35 (PAS and VAE).1·2·4·46·47 Both TS and VSED have been argued
Physician-assisted suicide is usually conducted covertly, with¬ to be ethically preferable alternatives to PAS and VAE on the
out consultation, guidelines, or documentation. Public contro¬ basis of similar arguments.12,13'16·19 In this section, we will criti¬
versy about legalizing PAS continues in the United States. cally examine these analyses. We also discuss the issues of
After narrow defeats of referenda in the states of Washington voluntariness, proportionality, and conflict of duties, which
and California, an Oregon referendum was passed in 1994 that may ultimately be more central to the ethical evaluation of
legalized PAS, subject to certain safeguards.36 After a series of these options. We suggest that there are more problems with
legal challenges, the Oregon legislature required that the ref¬ the doctrine of double effect and the active/passive distinction
erendum be resubmitted to the electorate this November be¬ than are ordinarily acknowledged and that TS and VSED are
fore implementation, and it was repassed this November by a more complex and less easily distinguished ethically from PAS

margin of 60% to 40%. The US Supreme Court ruled that laws and VAE than proponents seem to realize. Our discussion in
in the states of Washington and New York prohibiting PAS this section will be restricted to the potential ethical permis¬
were not unconstitutional, but the Court simultaneously en¬ sibility of these actions and not the public policy implications.
couraged public discussion and state experimentation through
the legislative and referendum processes.10·11·37·38 Doctrine of Double Effect
With VAE, the physician not only provides the means, but is When evaluating an action, the doctrine of double effect dis¬
the final actor by administering a lethal injection at the patient's tinguishes between effects that a person intends (both the end
request.1·3·25 As practiced in the Netherlands, the patient is se¬ sought and the means taken to the end) and consequences that
dated to unconsciousness and then given a lethal injection of a are foreseen but unintended.21,22'48,49 As long as the physician's
muscle-paralyzing agent like curare. For patients who are pre¬ intentions are good, it is permissible to perform actions with
pared to die because their suffering is intolerable, VAE has the foreseeable consequences that it would be wrong to intend. In
advantages of being quick and effective. Patients need not have this view, intentionally causing death is morally impermissible,
manual dexterity, the ability to swallow, or an intact gastroin¬ even if desired by a competent patient whose suffering could
testinal system. Voluntary active euthanasia also requires ac¬ not otherwise be relieved. But if death comes unintentionally as
tive and direct physician participation. Physicians can ensure the consequence of an otherwise well-intentioned intervention,
the patient's competence and voluntariness at the time of the act, even if foreseen with a high probability, the physician's action

support the family, and respond to complications. The directness can be morally acceptable. The unintended but foreseen bad
of the act makes the physician's moral responsibility clear. effect must also be proportional to the intended good effects.
On the other hand, VAE explicitly and directly conflicts with The doctrine of double effect has been important in justifying
traditional medical prohibitions against intentionally causing the use ofsufficient pain medications to relieve sufferingnear the
death.39 Although intended to relieve suffering, VAE achieves end of life.1,2,4,46,47 When high-dose opioids are used to treat pain,
this goal by causing death. Furthermore, VAE could be con¬ neither the patient nor the physician intends to accelerate death,
ducted without explicit patient consent.40·41 If abused, VAE could but they accept the risk of unintentionally hastening death in or¬
then be used on patients who appear to be suffering severely or der to relieve the pain. The doctrine ofdouble effect has also been
posing extreme burdens to physician, family, or society, but have used to distinguish TS from PAS and VAE.15,16,18,19 Relief of suf¬
lost the mental capacity to make informed decisions. fering is intended in all 3 options, but death is argued to be in¬
The Netherlands is the only country where VAE and PAS are tended with PAS and VAE but is merely foreseen with TS. Yet
openly practiced, regulated, and studied, although the practices to us it seems implausible to claim that death is unintended when
remain technically illegal. According to the Remmelink re¬ a patient who wants to die is sedated to the point of coma, and in¬
ports,9·42·43 VAE accounts for 1.8% to 2.4% of all deaths, and travenous fluids and artificial nutrition are withheld, making
PAS, another 0.2% to 0.4%. In 0.7% to 0.8% of deaths, active death certain.21,22,50 Although the overarching intention ofthe se¬
euthanasia was performed on patients who had lost the capacity dation is to relieve the patient's suffering, the additional step of
to consent, raising concern about whether guidelines restricting withholding fluids and nutrition is not needed to relieve pain, but
VAE to competent patients can be enforced in practice.44 is typically taken to hasten the patient's wished-for death. In
United States laws prohibiting VAE, however, are stricter contrast, when patients are similarly sedated to treat conditions
than those governing PAS and more likely to be prosecuted. like status epilepticus, therapies such as fluids and mechanical
Physicians are also more reluctant to participate in VAE even ventilation are continued with the goal of prolonging life.
if it were legalized.30·31 Even less is known about the secret According to the doctrine of double effect, intentionally tak¬
practice of VAE than of PAS in the United States. The recent ing life is always morally impermissible, whereas doing so
Washington State study showed that 4% of physicians had foreseeably but unintentionally can be permissible when it
received a genuine request for VAE within the year studied, produces a proportionate good. As applied to end-of-life medi¬
and 24% of those responded by administering a lethal injec¬ cal decision making, the intentions of the physician are given
tion.8 Voluntary active euthanasia was recently legalized in a more moral weight than the wishes and circumstances of the
province of Australia, but this legalization was subsequently patient. An alternative view is that it is morally wrong to take
reversed by the legislature.45 the life of a person who wants to live, whether doing so inten¬

ETHICAL COMPARISONS BETWEEN THE PRACTICES

tionally or foreseeably. In this view, what can make TS mor¬
ally permissible is that the patient gives informed consent to
Many normative ethical analyses use the doctrine of double it, not that the physician only foresees but does not intend the
effect and the distinction between active and passive assis¬ patient's inevitable death.
tance to distinguish between currently permissible acts that The issue of intention is particularly complicated because

Downloaded From: http://jama.jamanetwork.com/ by a University of Pittsburgh User on 06/19/2015

 the determination of what is intended by the patient or phy¬ The application and the moral importance ofboth the active/
sician is often difficult to verify and because practices that are passive distinction and the doctrine of double effect are noto¬
universally accepted may invoi ve the intention to hasten death riously controversial and should not serve as the primary basis
in some cases.21,51 Death is not always intended or sought when of determining the morality of these practices.
competent patients forgo life support; sometimes patients sim¬ Voluntariness
ply do not want to continue a particular treatment, but hope
nevertheless that they can live without it. But some patients We suggest that the patient's wishes and competent con¬
find their circumstances intolerable, even with the best of sent are more ethically important than whether the acts are
care, and refuse further life support with the intent of bringing categorized as active or passive or whether death is intended
about their death. There is broad agreement that physicians or unintended by the physician.54"56 With competent patients,
must respect such refusals, even when the patient's intention none of these acts would be morally permissible without the
is to die.1"4·46·47·51 However, such practices are highly problem¬ patient's voluntary and informed consent. Any ofthese actions
atic when analyzed according to the doctrine of double effect. would violate a competent patient's autonomy and would be
both immoral and illegal if the patient did not understand that
The Active/Passive Distinction death was the inevitable consequence of the action or if the
According to many normative ethical analyses, active mea¬ decision was coerced or contrary to the patient's wishes. The
sures that hasten death are unacceptable, whereas passive or ethical principle of autonomy focuses on patients' rights to
indirect measures that achieve the same ends would be per¬ make important decisions about their lives, including what
mitted.1,2,4,46,47,52 However, how the active/passive distinction happens to their bodies, and may support genuinely autono¬
applies to these 4 practices remains controversial.21-27 Volun¬ mous forms of these acts.27,52

tary active euthanasia is active assistance in dying, because However, because most ofthese acts require cooperation from
the physician's actions directly cause the patient's death. Stop¬ physicians and, in the case of TS, the health care team, the au¬
ping life-sustaining therapies is typically considered passive tonomy of participating medical professionals also warrants con¬
assistance in dying, and the patient is said to die of the under¬ sideration. Because TS, VSED, PAS, and VAE are not part of
lying disease no matter how proximate the physician's action usual medical practice and they all result in a hastened death,
and the patient's death. Physicians, however, sometimes ex¬ clinicians should have the right to determine the nature and
perience stopping life-sustaining interventions as very ac¬ extent of their own participation. All physicians should respect
tive.53 For example, there is nothing psychologically or physi¬ patients' decisions to forgo life-sustaining treatment, including
cally passive about taking someone offa mechanical ventilator artificial hydration and nutrition, and provide standard pallia¬
who is incapable of breathing on his or her own. Voluntarily tive care, including skillful pain and symptom management. If
stopping eating and drinking is argued to be a variant of stop¬ society permits some or all of these practices (currently TS and
ping life-sustaining therapy, and the patient is said to die ofthe VSED are openly tolerated), physicians who choose not to par¬
underlying disease.12,13 However, the notion that VSED is pas¬ ticipate because of personal moral considerations should at a
sively "letting nature take its course" is unpersuasive, be¬ minimum discuss all available alternatives in the spirit of in¬
cause patients with no underlying disease would also die if formed consent and respect for patient autonomy. Physicians
they stopped eating and drinking. Death is more a result ofthe are free to express their own obj ections to these practices as part

patient's will and resolve than an inevitable consequence of his ofthe informing process, to propose alternative approaches, and
disease. Furthermore, even ifthe physician's role in hastening to transfer care to another physician if the patient continues to
death is generally passive or indirect, most would argue that request actions to hasten death that they find unacceptable.
it is desirable to have physicians involved to ensure the patient
is fully informed and to actively palliate symptoms. Proportionality
Both PAS and TS are challenging to evaluate according to the The principles ofbeneficence and nonmaleficence obligate the
active/passive distinction. Physician-assisted suicide is active in physician to act in the patient's best interests and to avoid caus¬
that the physician provides the means whereby the patient may ing net harm.52 The concept of proportionality requires that the
take his or her life and thereby contributes toa new and different risk of causing harm must bear a direct relationship to the danger
cause of death than the patient's disease. However, the physi¬ and immediacy ofthe patient's clinical situation and the expected
cian's role in PAS is passive or indirect because the patient ad¬ benefit of the intervention.52,57 The greater the patient's suffer¬
ministers the lethal medication. The psychological and temporal ing, the greater risk the physician can take of potentially contrib¬
distance between the prescribing and the act may also make PAS uting to the patient's death, so long as the patient understands
seem indirect and thereby more acceptable to physicians than and accepts that risk. For a patient with lung cancer who is anx¬
VAE.3"32 These ambiguities may allow the physician to charac¬ ious and short of breath, the risk of small doses of morphine or
terize his or her actions as passive or indirect.21,50 anxiolytics is warranted. At a later time, if the patient is near
Terminal sedation is passive because the administration of se¬ death and gasping for air, more aggressive sedation is war¬
dation does not directly cause the patient's death and because the ranted, even in doses that may well cause respiratory depres¬
withholding of artificial feedings and fluids is commonly consid¬ sion. Although proportionality is an important element of the
ered passively allowing the patient to die.15,16,19 However, some doctrine ofdouble effect, proportionality can be applied indepen¬
physicians and nurses may consider it very active to sedate to un¬ dently ofthis doctrine. Sometimes a patient's suffering cannot be
consciousness someone who is seeking death and then to with¬ relieved despite optimal palliative care, and continuing to live
hold life-prolonging interventions. Furthermore, the notion that causes torment that can end only with death.58 Such extreme cir¬
TS is merely "letting nature take its course" is problematic, be¬ cumstances sometimes warrant extraordinary medical actions,
cause often the patient dies of dehydration from the withholding and the forms of hastening death under consideration in this ar¬
of fluids, not of the underlying disease. ticle may satisfy the requirement of proportionality. The re-

Downloaded From: http://jama.jamanetwork.com/ by a University of Pittsburgh User on 06/19/2015

 quirement of proportionality, which all health care interventions 2. Informed consent: Patients must be fully informed about
should meet, does not support any principled ethical distinction and capable of understanding their condition and treatment
between these 4 options. alternatives (and their risks and benefits). Requests for a has¬
tened death must be patient initiated, free of undue influence,
Conflict of Duties and enduring. Waiting periods must be flexible, depending on
Unrelievable, intolerable suffering by patients at the end of the nearness of inevitable death and the severity of immediate
life may create for physicians an explicit conflict between their suffering.
ethical and professional duty to relieve suffering and their un¬ 3. Diagnostic and prognostic clarity: Patients must have a
derstanding of their ethical and professional duty not to use at clearly diagnosed disease with known lethality. The prognosis
least some means of deliberately hastening death.57,59 Physicians must be understood, including the degree of uncertainty about
who believe they should respond to such suffering by acceding to outcomes (ie, how long the patient might live).
the patient's request for a hastened death may find themselves 4. Independent second opinion: A consultant with expertise
caught between their duty to the patient as a caregiver and their in palliative care should review the case. Specialists should also
duty to obey the law as a citizen.58 Solutions often can be found review any questions about the patient's diagnosis or prognosis.
in the intensive application of palliative care, or within the cur¬ A psychiatrist should consult if there is uncertainty about treat¬
rently legitimized options of forgoing life supports, VSED, or able depression or about the patient's mental capacity.
TS. Situations in which VSED or TS may not be adequate in¬ 5. Documentation and review: Explicit processes for docu¬
clude terminally ill patients with uncontrolled bleeding, obstruc¬ mentation, reporting, and review should be in place to ensure
tion from nasopharyngeal cancer, and refractory AIDS diarrhea accountability.
or patients who believe that spending their last days iatrogeni- The restriction of any of these methods to the terminally ill
cally sedated would be meaningless, frightening, or degrading. involves a trade-off. Some patients who suffer greatly from in¬
Clearly the physician has a moral obligation not to abandon pa¬ curable, but not terminal, illnesses and who are unresponsive to
tients with refractory suffering60; hence, those physicians who palliative measures will be denied access to a hastened death and
could not provide some or all of these options because of moral or forced to continue suffering against their will. Other patients
legal reservations should be required to search assiduously with whose request for a hastened death is denied will avoid a pre¬
the patient for mutually acceptable solutions. mature death because their suffering can subsequently be re¬
lieved with more intensive palliative care. Some methods (eg,
SAFEGUARDS
PAS, VAE, TS) might be restricted to the terminally ill because
In the United States, health care is undergoing radical reform of current inequities of access, concerns about errors and abuse,
driven more by market forces than by commitments to quality of and lack of experience with the process. Others (eg, VSED)
care,61·62 and 42 million persons are currently uninsured. Capi¬ might be allowed for those who are incurably ill, but not immi¬
tated reimbursement could provide financial incentives to en¬ nently dying, if they meet all other criteria, because of the in¬
courage terminally ill patients to hasten their deaths. Physi¬ herent waiting period, the great resolve that they require, and
cians' participation in hastening death by any of these methods the opportunity for reconsideration. If any methods are ex¬
can be justified only as a last resort when standard palliative tended to the incurably, but not terminally, ill, safeguards should
measures are ineffective or unacceptable to the patient. be more stringent, including substantial waiting periods and
Safeguards to protect vulnerable patients from the risk of mandatory assessment by psychiatrists and specialists, because
error, abuse, or coercion must be constructed for any of these the risk and consequences of error are increased.
practices that are ultimately accepted. These risks, which have We believe that clinical, ethical, and policy differences and
been extensively cited in the debates about PAS and VAE,3941 similarities amongthese 4 practices need to be debated openly,
also exist for TS and VSED. Both TS and VSED could be both publicly and within the medical profession. Some may
carried out without ensuring that optimal palliative care has worry that a discussion of the similarities between VSED and
been provided. This risk may be particularly great if VSED is TS on the one hand and PAS and VAE on the other may
carried out without physician involvement. In TS, physicians undermine the desired goal of optimal relief of suffering at the
who unreflectively believe that death is unintended or that it end of life.40,41 Others may worry that a critical analysis of the
is not their explicit purpose may fail to acknowledge the in¬ principle of double effect or the active/passive distinction as
evitable consequences of their action or their responsibility. applied to VSED and TS may undermine efforts to improve
The typical safeguards proposed for regulating VAE and pain relief or to ensure that patient's or surrogate's decisions
PAS63"66 are intended to allow physicians to respond to unre¬ to forgo unwanted life-sustaining therapy are respected.67
lieved suffering, while ensuring that adequate palliative mea¬ However, hidden, ambiguous practices, inconsistent justifica¬
sures have been attempted and that patient decisions are tions, and failure to acknowledge the risks of accepted prac¬
autonomous. These safeguards need to balance respect for tices may also undermine the quality of terminal care and put
patient privacy with the need to adequately oversee these patients at unwarranted risk.
interventions. Similar professional safeguards should be con¬ Allowing a hastened death only in the context of access to
sidered for TS and VSED, even if these practices are already good palliative care puts it in its proper perspective as a small
sanctioned by the law. The challenge of safeguards is to be but important facet of comprehensive care for all dying pa¬
flexible enough to be responsive to individual patient dilem¬ tients.1"4 Currently, TS and VSED are probably legal and are
mas and rigorous enough to protect vulnerable persons. widely accepted by hospice and palliative care physicians.
Categories of safeguards include the following: However, they may not be readily available because some
1. Palliative care ineffective: Excellent palliative care must physicians may continue to have moral objections and legal
be available, yet insufficient to relieve intolerable suffering for fears about these options. Physician-assisted suicide is illegal
a particular patient. in most states, but may be difficult, if not impossible, to sue-

Downloaded From: http://jama.jamanetwork.com/ by a University of Pittsburgh User on 06/19/2015

 cessfully prosecute if it is carried out at the request of an 30. Cohen JS, Fihn SD, Boyko EJ, et al. Attitudes toward assisted suicide and
euthanasia among physicians in Washington State. N Engl JMed. 1994;331:89-94.
informed patient. Voluntary active euthanasia is illegal and 31. Bachman JG, Alchser KH, Koukas DJ, et al. Attitudes of Michigan physi-
more likely to be aggressively prosecuted if uncovered. In the cians and the public toward legalizing physician-assisted suicide and voluntary
United States, there is an underground, erratically available euthanasia. N Engl J Med. 1996;334:303-309.
32. Duberstein PR, Conwell Y, Cox C, et al. Attitudes toward self-determined
practice of PAS and even VAE that is quietly condoned. death. J Am Geriatr Soc. 1995;43:395-400.
Explicit public policies about which of these 4 practices are 33. Preston TA, Mero R. Observations concerning terminally-ill patients who
choose suicide. J Pharm Care Pain Symptom Control. 1996;1:183-192.
permissible and under what circumstances could have impor¬ 34. Admiraal PV. Toepassing van euthanatica (the use of euthanatics). Ned
tant benefits. Those who fear a bad death would face the end of Tijdschr Geneeskd. 1995;139:265-268.
life knowing that their physicians could respond openly if their 35. Lee MA, Nelson HD, Tilden VP, Ganzini L, Schmidt TA, Tolle SW. Legalizing
worst fears materialize. For most, reassurance will be all that is assisted suicide: views of physicians in Oregon. N Engl J Med. 1996;334:310-315.
36. Alpers A, Lo B. Physician-assisted suicide in Oregon: a bold experiment.
needed, because good palliative care is generally effective. Ex¬ JAMA. 1995;274:483-487.
plicit guidelines for the practices that are deemed permissible 37. Compassion in Dying v Washington, No. 94-35534,1966 WL 94848 (9th Cir,
Mar 6,1996).
can also encourage clinicians to explore why a patient requests
38. Quill v Vacco, No. 95-7028 (2d Cir, April 9,1996).
hastening of death, to search for palliative care alternatives, and 39. Gaylin W, Kass LR, Pellegrino ED, Siegler M. Doctors must not kill. JAMA.
to respond to those whose suffering is greatest.58,60,68-70 1988;259:2139-2140.
40. Teno J, Lynn J. Voluntary active euthanasia: the individual case and public
Dr Lo is supported by National Institute of Mental Health Center grant policy. J Am Geriatr Soc. 1991;39:827-830.
MH42459 and by the Robert Wood Johnson Foundation. 41. Kamisar Y. Against assisted suicde\p=m-\ev n a very limited form. Univ De-

We want to thank Diane Meier, MD, and Frank Miller, PhD, for their work on troit Mercy Law Rev. 1995;72:735-769.
42. vanderMaas PJ, vanderWal G, Haverkate I, et al. Euthanasia, physician-
early drafts of the manuscript and an anonymous reviewer who gave persistent assisted suicide and other medical practices involving the end of life in the
clarifying feedback.
Netherlands, 1990-1995. N Engl J Med. 1996;335:1699-1705.
References 43. van der Wal G, van der Maas PJ, Bosma JM, et al. Evaluation of the noti-
fication procedure for physician-assisted death in the Netherlands. N Engl J
1. Foley KM. Pain, physician-assisted suicide, and euthanasia. Pain Forum. Med. 1996;335:1706-1711.
1995;4:163-178. 44. Hendin H. Seduced by death. Issues Law Med. 1994;10:123-168.
2. Council on Scientific Affairs, American Medical Association. Good care of the
45. Ryan CJ, Kaye M. Euthanasia in Australia: the Northern Territory rights
dying patient. JAMA. 1996;275:474-478. of the terminally ill act. N Engl J Med. 1996;334:326-328.
3. Quill TE. Death and Dignity: Making Choices and Taking Charge. New 46. President's Commission for the Study of Ethical Problems in Medicine and
York, NY: WW Norton & Co; 1993:1-255. Biomedical and Behavioral Research. Deciding to Forego Life-Sustaining Treat-
4. American Board of Internal Medicine End of Life Patient Care Project Com-
ment: Ethical, Medical and Legal Issues in Treatment Decisions. Washington,
mittee. Caringfor the Dying: Identification and Promotion ofPhysician Com- DC: US Government Printing Office; 1982.
petency. Philadelphia, Pa: American Board of Internal Medicine; 1996. 47. The Hastings Center Report. Guidelines on the Termination of Life-Sus-
5. Kasting GA. The nonnecessity of euthanasia. In: Humber JD, Almeder RF,
Kasting GA, eds. Physician-Assisted Death. Totowa, NJ: Humana Press; 1993: taining Treatment and the Care of the Dying. Briarcliff Manor, NY: Hastings
25-43.
Center; 1987.
48. Marquis DB. Four versions of thedouble effect.J Med Philos. 1991;16:515-544.
6. Coyle N, Adelhardt J, Foley KM, Portenoy RK. Character of terminal illness
49. Kamm F. The doctrine of double effect. J Med Philos. 1991;16:571-585.
in the advanced cancer patient. J Pain Symptom Manage. 1990;5:83-93.
7. Ingham J, Portenoy R. Symptom assessment. Hematol Oncol Clin North 50. Quill TE. The ambiguity of clinical intentions. N Engl J Med. 1993;329:1039\x=req-\
Am. 1996;10:21-39. 1040.
51. Alpers A, Lo B. Does it make clinical sense to equate terminally ill patients
8. Back AL, Wallace JI, Starks HE, Pearlman RA. Physician-assisted suicide
and euthanasia in Washington State. JAMA. 1996;275:919-925. who require life-sustaining interventions with those who do not? JAMA. 1997;
9. vanderMaas PJ, van Delden JJM, Pijnenborg L. Euthanasia and Other 277:1705-1708.
Medical Decisions Concerning the End of Life, Amsterdam, the Netherlands: 52. Beauchamp TL, Childress JF. Principles ofBiomedical Ethics. 3rd ed. New
Elsevier; 1992. York, NY: Oxford University Press; 1994.
53. Edwards MJ, Tolle SW. Disconnecting a ventilator at the request of a pa-
10. Vacco v Quill, 117 SCt 2293 (1997).
11. Washington v Glucksberg, 117 SCt 2258 (1997). tient who knows he will die. Ann Intern Med. 1992;117:254-256.
12. Bernat JL, Gert B, Mogielnicki RP. Patient refusal of hydration and nutri- 54. Orentlicher D. The legalization of physician-assisted suicide. N Engl J Med.
tion. Arch Intern Med. 1993;153:2723-2727. 1996;335:663-667.
13. Printz LA. Terminal dehydration, a compassionate treatment. Arch Intern 55. Drickamer MA, Lee MA, Ganzini L. Practical issues in physician-assisted
Med. 1992;152:697-700. suicide. Ann Intern Med. 1997;126:146-151.
14. Eddy DM. A conversation with my mother. JAMA. 1994;272:179-181. 56. Angell M. The Supreme Court and physician-assisted suicide: the ultimate
15. Cherney NI, Portenoy RK. Sedation in the management of refractory symp- right. N Engl J Med. 1997;336:50-53.
toms: guidelines for evaluation and treatment. J Palliat Care. 1994;10:31-38. 57. de Wachter MAM. Active euthanasia in the Netherlands. JAMA. 1989;262:
16. Troug RD, Berde DB, Mitchell C, Grier HE. Barbiturates in the care of the 3316-3319.
terminally ill. N Engl J Med. 1991;327:1678-1681. 58. Quill TE, Brody RV. 'You promised me I wouldn't die like this': a bad death
as a medical emergency. Arch Intern Med. 1995;155:1250-1254.
17. Enck RE. The Medical Care of Terminally Ill Patients. Baltimore, Md:
Johns Hopkins University Press; 1994. 59. Welie JVM. The medical exception: physicians, euthanasia and the Dutch
18. Saunders C, Sykes N. The Management of Terminal Malignant Disease. criminal law. J Med Philos. 1992;17:419-437.
3rd ed. London, England: Hodder Headline Group; 1993:1-305. 60. Quill TE, Cassel CK. Nonabandonment: a central obligation for physicians.
19. Byock IR. Consciously walking the fine line: thoughts on ahospice response Ann Intern Med. 1995;122:368-374.
to assisted suicide and euthanasia. J Palliat Care. 1993;9:25-28. 61. Emanuel EJ, Brett AS. Managed competition and the patient-physician
20. Ventafridda B, Ripamonti C, DeConno F, et al. Symptom prevalence and relationship. N Engl J Med. 1993;329:879-882.
control during cancer patients' last days of life. J Palliat Care. 1990;6:7-11. 62. Morrison RS, Meier DE. Managed care at the end oflife. Trends Health Care
21. Brody H. Causing, intending, and assisting death. J Clin Ethics. 1993;4: Law Ethics. 1995;10:91-96.
112-117. 63. Quill TE, Cassel CK, Meier DE. Care of the hopelessly ill: proposed criteria
22. BillingsJA. Slow euthanasia. J Palliat Care. 1996;12:21-30. for physician-assisted suicide. N Engl J Med. 1992;327:1380-1384.
23. Orentlicher D. The Supreme Court andphysician-assisted suicide: rejecting 64. Brody H. Assisted death. N Engl J Med. 1992;327:1384-1388.
assisted suicide but embracing euthanasia. N J Med. 1997;337:1236-1239.
Engl 65. Miller FG, Quill TE, Brody H, et al. Regulating physician-assisted death.
24. Moerman N, Bonke B, Oosting J. Awareness and reduring
call general N Engl J Med. 1994;331:119-123.
anesthesia: facts and feelings. Anesthesiology. 1993;79:454-464. 66. Baron CH, Bergstresser C, Brock DW, et al. Statute: a model state act to
25. Utting JE. Awareness: clinical aspects; consciousness, awareness, and pain. authorize and regulate physician-assisted suicide. Harvard J Legislation. 1996;
In: RosenM,Linn JN. General Anesthesia. London, England: Butterworths; 33:1-34.
1987:171-179. 67. Mount B, Flanders EM. Morphine drips, terminal sedation, and slow eu-
26. ans Patient's experience of awareness during general anesthesia;
EvJM. thanasia: definitions and facts, not anecdotes. J Palliat Care. 1996;12:31-37.
consciousness, awareness pain. In:
and M, Linn JN. General Anesthesia.
Rosen 68. Lee MA, Tolle SW. Oregon's assisted-suicide vote: the silver lining. Ann
London, England: Butterworths; 1987:184-192. Intern Med. 1996;124:267-269.
27. Brock DW. Voluntary active euthanasia. Hastings Cent Rep. 1992;22:10-22. 69. Block SD, Billings A. Patient requests to hasten death: evaluation and man-
28. Quill TE. Death and dignity. N Engl J Med.1991;324:691-694. agement in terminal care. Arch Intern Med. 1994;154:2039-2047.
29. Rollin B. Last Wish. New York, NY: Warner Books; 1985. 70. Quill TE. Doctor, I want to die: will you help me? JAMA. 1993;270:870-873.

Downloaded From: http://jama.jamanetwork.com/ by a University of Pittsburgh User on 06/19/2015