Clinical Ethics in Anesthesiology. A Case-Based Textbook
2. End-of-life issues
15. The principle of double effect in palliative care: euthanasia by another name?
Denise M. Dudzinski
Mrs. Ryan was a 58-year-old woman with widely metastatic breast cancer. She entered a hospice a month prior to this hospital admission. A Do Not Attempt Resuscitation (DNAR) order had been written 2 months ago at her request. Per her living will, she was willing to forgo artificial nutrition and hydration when she was no longer able to take food and water by mouth. Her pain and symptoms were well controlled on an outpatient regimen of celecoxib, amitriptyline, lorazepam, oxycodone hydrochloride, and morphine sulfate. After 2 weeks on this regimen, she was admitted to the hospital due to unbearable pain in her back, pelvis, and shoulders (presumed due to bone metastases). In order to relieve pain and symptoms, intravenous fentanyl and lorazepam were administered and titrated up when Mrs. Ryan exhibited verbal or physical signs of distress or pain. She was receiving 700 mcg/h of fentanyl with 100 mcg boluses every hour for breakthrough pain. In keeping with Mrs. Ryan’s wishes, her physicians achieved their therapeutic goal of relieving her pain and symptoms, but they knew there may be a “double effect.” The medications could suppress respiration and hasten her death.
What is the ethical rationale for allowing physicians to take this risk?
The principle of double effect (DE)
St. Thomas Aquinas first used the term “double effect” to refer to “the duality of results of a single human action” in discussing killing in self-defense.1 If one attempts to defend oneself, and the assailant is killed, it does not mean that death was the defender’s intention. Catholic moral theology distinguishes what someone intends and the side effect (“foreseen but unintended consequences”), suggesting that people remain responsible for, but are less culpable for, such double effects.
A modern formulation of DE is:
(T)he traditional rule of double effect specifies that an action with two possible effects, one good and one bad, is morally permitted if the action: (1) is not in itself immoral; (2) is undertaken only with the intention of achieving the possible good effect, without intending the possible bad effect even though it may be foreseen; (3) does not bring about the possible good effect by means of the possible bad effect; and (4) is undertaken for a proportionately grave reason.2
In the context of caring for a terminally ill patient, DE allows healthcare providers to: (1) provide adequate pharmacological pain and symptom management (a good action) knowing that (2) the foreseen but unintended consequence is that the medications may suppress respirations and hasten death. The provider must intend only good pain management. She may not hope to save the patient from suffering by causing the patient’s death, as this would constitute euthanasia which is illicit in this moral framework. (3) Death (the bad effect) is not a means to alleviate suffering (the good effect). (4) Pain and suffering at the end of life is a “proportionately grave reason” to justify the use of opioids, but the risk of hastening death is only ethically justified if the patient is terminally ill.
Preconditions and components of DE
Certain virtues and prohibitions are taken for granted in the employment of DE. For example, alleviating pain at the end of life is morally virtuous, and intentionally killing or assisting in the death of another person is always prohibited. Physicians need not hold these beliefs. There are a myriad of other ethical frameworks that may inform moral behavior, such as maximizing benefit and minimizing harm (utility) or acting according to duty (deontology). For example, if a competent, terminally ill patient requested euthanasia by means of high doses of opioids, one could argue this is morally permissible because it is consistent with the patient’s wishes, the patient is dying, and euthanasia will mercifully end her suffering in the manner she chooses. This is a legitimate ethical argument, but it is not an argument based in DE. Ultimately, each physician must decide if he finds DE reasoning persuasive, and to do so he should understand both the preconditions necessary for an appeal to DE and its specific components.
Prior to invoking DE, the patient must be dying and there must be evidence that the patient’s pain and suffering cannot be managed using less risky means, which always include non-pharmacological management such as spiritual and social support. The expert management of pharmacological and non-pharmacological treatments is essential to the concept of proportionality. Jansen and Sulmasy describe proportionality: “A physician’s therapeutic response to terminal suffering is justified, even if it imposes a high risk of hastening the patient’s death, if and only if (1) the measures implemented are directly proportionate to the intensity of the patient’s suffering; [and] (2) the measures implemented are appropriate for the type of suffering the patient is experiencing …”3 In addition, we must exhaust all “equally efficacious alternatives with fewer side-effects.”4 For example, if Mrs. Ryan was receiving moderate doses of narcotics and anxiolytics but was still in pain and her doctor chose to suddenly administer an extremely high dose of fentanyl, DE could not legitimately be invoked. There is no medical or ethical justification to leap to such a high dose of fentanyl without first testing accelerated but intermediate doses, and therefore no “proportionately grave reason” (e.g., unremitting pain despite carefully titrated analgesia). An appeal to DE can only be made after the patient is receiving expert pain and symptom management. One cannot rely on DE to justify poor clinical care.
The principle of DE does not address the issue of consent
In cultures where respect for autonomy is a legal and ethical priority, the patient or legal surrogate should understand and accept the risks of pain management. Patients and surrogates consent to all medical interventions, so palliative care should be no exception. However, we respect patient autonomy due to obligations that are independent of DE. DE is silent on consent. Instead, DE revolves around intention.
Doctors who want to appeal to or better understand DE should be able to distinguish between the intention of the act and the intention of the actor. The intention of the act is the goal of an action and should be discernible by looking at the drugs and dosage increments as well as the way pain and symptoms are assessed. For example, lorazepam, oxycodone hydrochloride, morphine sulfate, and fentanyl effectively alleviate pain and anxiety. However, the administration of potassium chloride, under the auspices of analgesia, would be impermissible because, under the principle of DE, therapy must still be governed by the standard of care, and potassium chloride cannot alleviate pain.
In the interest of good palliative care, pain and symptoms should be assessed frequently so that the patient can receive additional medication if she needs it and so that the dosage can be maintained if pain relief has been achieved. When physicians use appropriate analgesics and write orders to titrate medication based on evidence of pain and suffering, such as groaning, agitation, verbal complaints, diaphoresis, hypertension, or unexplained tachycardia, they demonstrate that the intention of the act is geared toward alleviating pain. Unmonitored continuous infusions or orders to titrate up irrespective of signs of pain, anxiety, or dyspnea may suggest poor clinical management or a covert intention to hasten death. Neither is acceptable under DE. Expert clinical management and exhaustive efforts to avoid risks of respiratory depression are preconditions of DE. Pre-emptive dosing in anticipation of signs and symptoms of suffering is not forbidden under DE, provided it meets all DE criteria, is proportional, incremental, and in keeping with expert clinical management and monitoring.
The motives and intentions of the actor (the physician ordering medication, nurse administering the medication, etc.) are more difficult to ascertain. The physician should aim only to alleviate the patient’s suffering, not to hasten death. DE asks physicians to carefully examine their motives and assumes that one’s private moral intentions are morally relevant, even though we can never be certain of another’s intentions. Put simply, we trust the stated goals and motivations of providers, unless there is good reason or evidence to question them.
DE permits extremely high opioid doses
DE places no upper limit on opioid or other medication dosages. The patient’s narcotic tolerance, age, underlying diseases and organ dysfunction, current level of sedation, and previous alcohol or drug use account for pronounced variations in narcotic and benzodiazepine requirements. An experienced anesthesiologist might be astounded by the dosages a patient requires and may continue to titrate up, provided she doses based on ongoing clinical assessments to alleviate targeted distressing symptoms such as dyspnea, agitation, and pain, and provided narcotics are not increased after symptoms are relieved. DE unquestionably supports this clinical approach, because it meets the proportionality condition.
Double effect reasoning, physician aid-in-dying and euthanasia
DE is invoked when physician aid-in-dying (a.k.a. physician-assisted suicide) and euthanasia are illegal and/or deemed immoral. Washington and Oregon Death with Dignity Acts allow competent adults residing in those states who “have been determined by the attending physician and consulting physician to be suffering from a terminal disease, and who … voluntarily expressed their wish to die, may make a written request for medication that the patient may self-administer to end his or her life in a humane and dignified manner.”5 These Acts describe physician aid-in-dying (PAD) as practiced in the US. Even outside the specific provisions of state Death with Dignity Acts, physician aid-in-dying occurs when the physician complies with a competent patient’s request for a prescription for lethal medication and the physician understands that the patient intends to self-administer the medication for the purpose of ending his or her life. It is not PAD when a patient hoards medications prescribed for therapeutic purposes and overdoses, provided the physician was not aware of the patient’s intention.
In contrast, “euthanasia occurs when a third party administers medication or acts directly to end a person’s life.”6 It is the action of another person that causes the patient’s death, and he acts out of mercy to end the patient’s suffering. Physicians may feel they are on the cusp of euthanasia when treating pain and symptoms in dying patients. Granted, in some cases, the line between the two may not be clear, but certainty is not required, even were it possible. DE addresses this concern and helps to distinguish “slow euthanasia” from the DE of effectively treating terminal pain and suffering.
Billings and Block describe one example of slow euthanasia as the practice of “hanging the morphine drip” without adjusting medications for targeted symptoms.7 Because the dying process tends to be slower, those involved might feel that death is simply the DE of aggressive comfort care measures, however the interval between treatment and death is irrelevant in DE. One can intend to hasten death slowly or quickly. For reasons already outlined, DE does not support hanging the morphine drip, failing to monitor the patient’s symptoms, and succumbing to pressures from family or providers to end the anguishing experience of witnessing a longer-than-anticipated dying process. These pressures should be addressed in other ways such as educating the family and providing emotional and spiritual support to the family and providers.
Legal and professional appeals to double effect reasoning
In the interest of good anesthesia practice, physicians widely hold dual commitments to relieve pain at the end of life and to minimize respiratory depression and other undesired side-effects of analgesia. This professional obligation does not rely on the principle of DE. However, when this balancing act is no longer sustainable, both medical societies and courts seem to support DE reasoning. The American Medical Association states that “physicians have an obligation to relieve pain and suffering and to promote the dignity and autonomy of dying patients in their care. This includes providing effective palliative treatment even though it may foreseeably hasten death.”8 Similarly, laws distinguish intended versus unintended action. For example mens rea crimes, crimes of direct intent, result in stricter punishments than crimes of neglect. Intention is relevant to culpability, because “we can refuse to cause harm intentionally, but can’t avoid all harm that occurs as a side effect.”9
In Vacco v. Quill, the United States Supreme Court responded to the argument that ending or refusing lifesaving medical treatment is “nothing more nor less than assisted suicide.”10 Chief Justice William Rehnquist wrote that that the distinction between withdrawing or withholding medical interventions and assisted suicide “comports with fundamental legal principles of causation and intent.” “When a patient refuses life-sustaining medical treatment, he dies from an underlying fatal disease or pathology; but if a patient ingests lethal medication prescribed by a physician, he is killed by that medication.”10 Indeed, it is precisely uncertainty concerning cause of death that raises the ethical dilemma: Am I hastening death by adequately treating pain? DE permits death as a side effect not because we assert, without proof, that the underlying disease was the cause of death, but because the physician is not culpable even if the medications he administered were the cause of death.
The Supreme Court seems to concur. “(I)n some cases, painkilling drugs may hasten a patient’s death, but the physician’s purpose and intent is, or may be, only to ease his patient’s pain. A doctor who assists a suicide, however, “must necessarily and indubitably, intend primarily that the patient be made dead.”11 In Vacco v. Quill, Justice Rehnquist wrote “Just as the State may prohibit assisting suicide while permitting patients to refuse unwanted lifesaving treatment, it may permit palliative care related to that refusal, which may have the foreseen but unintended ‘double effect’ of hastening the patient’s death.”10
Criticisms of DE
DE’s emphasis on intention invites criticism. Timothy Quill, Rebecca Dresser, and Daniel Brock write, “(A)ccording to modern psychology, human intention is multilayered, ambiguous, subjective, and often contradictory. The rule of double effect does not acknowledge this complexity; instead, intention is judged according to the presence or absence of a clear purpose. Clinicians familiar with the requirements of the rule may learn to express their intentions in performing ambiguous acts … in terms of foreseen but unintended consequences …”12 On the contrary, DE does acknowledge the complexity of human intention, but still requires moral honesty and integrity, not perfection, in examining private intentions. Under virtue-based traditions, the fact that human psychology and motivation is complex and sometimes ambiguous – that we often have mixed intentions – does not exempt one from setting morally appropriate goals, examining mixed intentions, and acting in good faith. Certainly, a physician can appeal to DE to justify his intention to hasten death and none may be the wiser. DE does not “let him off the hook” simply because it has no mechanism to objectively test for dishonesty. Rather, his intentions never met criteria for DE despite his claim.
• Until relatively recently, dying patients were routinely undertreated for pain because physicians feared that the treatment would hasten death.
• The principle of double effect permits aggressive treatment of pain when death may be an unintended effect of that treatment.
• DE requires that the therapy in question be within the standard of care for treating the targeted symptoms.
• DE does not provide moral justification for poor clinical care.
1 Summa Theologica of St. Thomas Aquinas, I–II.
2* Sulmasy, D.P. and Pellegrino, E.D. (1999). The rule of double effect: clearing up the double talk. Arch Intern Med, 159(6), 545–50.
3* Jansen, L.A. and Sulmasy, D.P. (2002). Proportionality, terminal suffering and the restorative goals of medicine. Theoret Med Bioeth, 23(4), 321–37.
4* Young, R. (2007). Medically Assisted Death. Cambridge, UK: Cambridge University Press. p. 251.
5 Revised Code of Washington, 70.245.020.
6* Braddock, C. and Tonelli, M.R. (2009). Physician Aid in Dying. Ethics in Medicine Website, University of Washington, Seattle WA. http://depts.washington.edu/bioethx/topics/pas.html
7* Billings, J.A. and Block, S.D. (1996). Slow euthanasia. J Palliative Care, 12(4), 21–30.
8* Council of Ethical and Judicial Affairs. (1992). Decisions Near the End of Life. Chicago, IL: American Medical Association.
9* Hawryluck, L.A. and Harvey, W.R. (2000). Analgesia, virtue, and the principle of double effect. J Palliative Care, 16 Supplement, S24–30.
10 Vacco v. Quill, 521 US793 (1997).
11 Washington v. Glucksverg. 521 U.W. 702 (1997).
12* Quill, T.E., Dresser, R. and Brock, D.W. (1997). The rule of double effect – a critique of its role in end-of-life decision making. N Engl J Med, 337(24), 1768–71.
Battin, M. (2008). Terminal sedation: pulling the sheet over our eyes. Hastings Ctr Rep, 38(5), 27–30.
Boyle, J. (2004). Medical ethics and double effect: the case of terminal sedation. Theoret Med Bioeth. 25(1), 51–60.
Hawryluck, L.A., Harvey, W.R., Lemieux-Charles, L, and Singer, P.A. (2002). Consensus guidelines on analgesia and sedation in dying intensive care unit patients. BMC Med Ethics, 3(3): E3.
Jansen, L.A. (2010). Disambiguating clinical intentions: the ethics of palliative sedation. J Med Philos, 35(1), 19–31.