!"#$%$&'()(#*+(,+-"./0($-(12-+./32%4(56+/#&+%# !"#*$67.89(:6";+(1<(=200+6 ,+32+>+?(>$6@7.89 A$"6;+9(5*+(B/.#2%4.(C+%#+6(,+D$6#E(F$0<(GGE(H$<(I(7!"4<E(GJKG8E(DD<(LLMLK N"O02.*+?(O'9(The Hastings Center A#/O0+(P,19(http://www.jstor.org/stable/3561339 . !;;+..+?

9(QRSQGSLQGL(QT9TU
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

The Hastings Center is collaborating with JSTOR to digitize, preserve and extend access to The Hastings Center Report.

http://www.jstor.org

IllllllIMl

IIIIIllllllll

IIII

FOURCASES, FOURSENSES OF AUTONOMY

Autonomy &

the

Refusal

of

Lifesaving Treatment

by BRUCEL. MILLER
the patient recovered, he asked that if he had a further cardiovascular collapse no steps should be taken to prolong his life, for the pain of his cancer was more than he would needlessly bear. He wrote a note to this effect in his case records and the hospital staff knew of his feelings.3 hospital with injuriesand internalbleeding caused when a treefell on him. He needed whole bloodfor a transfusion but refused to give the necessary consent. His wife also refused. Both were Jehovah's Witnesses,holding religious beliefs thatforbid the infusion of whole blood. The hospital lawyer brought a petition to the home of a judge. The patient's wife, brother, and grandfather were present to express his strong religious convictions. The grandfather wants to live in the Bible's promised new world where life will never end. A few hours here would nowhere compare to everlasting life." Thejudge was concerned with the patient's capacity to make such a decision in light of his serious condition. She recognized the possibility that the use of drugs might have impaired his judgment. The hospital lawyer replied that the patient was receiving fluid intravenously but no drugs that could impair his judgment. He was conscious, knew what the doctor was saying, was aware of the consequences of his decision, and had with full understandingexecuted a statementrefusing the recommendedtransfusionand releasing the hospitalfrom liability. The judge went to the patient's bedside. She asked him whetherhe believed that he would be deprived of the for "everlasting life" if transfusion were oropportunity dered by the court. His response was, "Yes. In other words, it is between me and Jehovah; not the courts .... I'm willing to take my chances. My faith is that strong .... I wish to live, but not with blood transfusions.Now get that straight." The patient had two young children. There was a family business and money to providefor the children, and a large family willing to care for them.4 piratory infection suddenly developed severe headache, stiff neck, and highfever. He went to an emergencyroom for help. The diagnosis was pneumococcal meningitis, a bacterial meningitis almost always fatal if not treated. If treatmentis delayed, permanent neurological damage is likely. A physician told the patient that urgent treatment was needed to save his life and forestall brain damage. The patient refused to consent to treatmentsaying that he wanted to be allowed to die.S
The Hastings Center Report, August 1981

applied-has reacted against utilitarianismbecause of its tendencyto regardthe individualas little more thana recipient of good and evil. To avoid the perniciouseffect of this notion, many philosophershave insisted thatthe conceptof a person as an autonomousagent must have a centraland role in ethicaltheory.' From this position there independent is firm ground to resist coercion and its less forceful, but more pervasive, cousins: manipulationand undue influfor ence. It also provides a warrant treatinga person's own choices, plans, and conception of self as generally dominant over what anotherbelieves to be in that person's best interest. In biomedicalethics, the conceptof a personas an autonomous agent places an obligation on physicians and other healthprofessionalsto respectthe values of patientsand not to let their own values influencedecisions abouttreatment. The conflictof patientvalues andphysicianvalues becomes most troublesomewhen a patientrefuses treatmentneeded to sustain life and a physician believes that the patient should be treated.The conflict can be resolved by taking a firm line on autonomy:any autonomousdecision of a patient must be respected. On the otherhand, the physician's obligationto preservelife can be placed above the patient's right to autonomy and refusals of treatmentcan then be overriddenwhen they conflict with "medicaljudgment."2 The notion of medicaljudgmentused here is not clear, and it may only be a gloss for "whatdoctor thinks best." Neither extremeposition is tenable;both are insensitive to the complexities of such cases, and the second removes the right to autonomyaltogether.But, the conflict between autonomy and medical judgmentis not as sharpas it seems.

and normative ethics-both theoretical ontemporary,

CASE2. A forty-three-year-oldman was admittedto the

said that the patient "wants to live very much . . . He

Four Cases
Considerthe following cases. CASE1. A doctor, sixty-eight years of age, had been retiredfor five years after severe myocardial infarction. He was admitted to a hospital after a barium meal had shown a large and advanced carcinoma of the stomach. Ten days after palliative gastrectomy was performed, the patient collapsed with a massivepulmonaryembolismand an emergencyembolectomywas done on the ward. When StateUniversity. manities Program,Michigan
22 is L. BRUCE MILLER professor of philosophy, Medical Hu-

man with mild upperresCASE3. A thirty-eight-year-old

I _

_ I

_I ___ _

__

CASE 4. A fifty-two-year-oldmarriedman was admittedto a medical intensive care unit (MICU) after a suicide attempt. He had retired two years earlier because of progressive physical disability related to multiple sclerosis (MS) during the fifteen years before admission. He had successfullyadapted to his physical limitations, remaining actively involved in family matters with his wife and two teenage sons. However, during the three months before admission, he had become morose and withdrawn.On the evening of admission, while alone, he had ingested an unknownquantityof diazepam. Whenhis family returnedsix hours later, theyfound the patient semiconscious. He had left a suicide note. On admission to the MICU, physician examination showed several neurologic deficits, but no more severe than in recent examinations. Thepatient was alert and fully conversant. He expressed to the house officers his strong belief in a patient's right to die with dignity. He stressed the "meaningless" aspects of his life related to his loss of function, insisting that he did not want vigorous medical intervention should serious complications develop. This position appeared logically coherent to the MICU staff. However, a consultation with membersof the psychiatric liaison service was requested. During the intitial consultationthe patient showed that the onset of his withdrawaland depression coincided with a diagnosis of inoperable cancer in his mother-in-law, who lived in another city. His wife had spent more and more time satisfying her mother'sneeds. Infact, on the night of his suicide attempt,thepatient's wife and two sons had left him alone for the first time to visit his mother-in-law.6

In the firsttwo cases, the most compelling intuitionis to The patientsare competent, respectthe refusalof treatment. they exercisingtheir rightof autonomyto refuse treatments believed not in theirinterest.The patientin Case 1 believed furtherresuscitationwas needless, for it would only briefly prolonga life of greatsuffering.His concernwas for life on would earth.The patientin Case 2 believed the transfusions deprive him of salvation. His concern for life hereafter made whateverlife on earthhe could get from the transfusions insignificant. In Case 2 the SuperiorCourt and the Court of Appeals recognized the patient's right to refuse transfusion and none was given. Though the patient's chargedfrom the hospital. In Case 1, two weeks after the embolectomythe patientsuffered acute myocardialinfarction; his heartwas restartedfive times in one night. He reheartstopped,plans were being made to put him on a respirator. The Jehovah's Witness was fortunate,retaininghis life on earthwithoutriskingthe loss of life everlasting.The physicianwas not so lucky;his rightto an autonomousdecision concerningthe mannerof his own death fell victim to the technological imperative-"If you can do it, you should do it."
Center The Hastings covered to linger for three weeks in a coma. On the day his chances were thought very slim, he recovered and was dis-

Cases 3 and 4, however, incline to the view that patient autonomymay be overridenby medicaljudgment.In Case 3 there is no apparentreason to justify the death of this otherwisehealthyvictim of meningitis. His medical condition is not hopeless, as was the condition of the doctor in Case 1, nor does he have a religious objectionto treatment like the Jehovah's Witness in Case 2. Our intuitionis to treathim againsthis will. In Case 4 the patient'sdisability may give us pause; it does prevent a full life, yet he had became ill and the family manageduntil his mother-in-law to her needs. We might expect thatfamily began attending discussion of the problem could lead to a resolution that would restorethe patient'sdesire to live. At firstglance the positionthatalthoughthereis a rightto autonomyfrom which patients can refuse lifesaving treatment, the rightis not absoluteand sometimesmedicaljudgment can overrideit is a tenable one; for there is nothing surprisingabout a right that is not absolute.7However, acknowledging the limits of rights does not mean that rights can be overridenwhen their exercise conflicts with others' judgments. If medical judgment can override the right to refusalof treatment,then all four patientsshouldhave been treatedagainst their will, for in each case a physician believed thatthe patientshould be treated.If this is implausible, given our intuitionson Cases 1 and 2, then we have to say that autonomyis supremeand the refusalsof lifesaving treatmentshould have been respectedin all four cases. One way aroundthis impasseis to develop a list of conditions thatmustbe takeninto accountto determinewhethera should be respected,8for example, age refusal of treatment of the patient, life expectancywith and without treatment, the level of incapacitywith and withouttreatment,the degree of pain and suffering, the effect of the time and circumstancesof deathon family and friends, the views of the family on whetherthe patientshould be treated,the views of the physician and other medical staff, and the costs of with it the refusalof This is a plausibleapproach; treatment. for meningitiscan be justifiablyoverridenand the treatment refusal of treatmentfor the doctor suffering from cancer justifiablyrespected. The meningitis patient is young and will recover without residualdefect to lead a full life; the cancerpatientwill die soon in any case, is sufferinggreatly, and even thoughresuscitatedis not likely to survive with a capacity for conscious awareness. The problemwith this approachis twofold. First, the list is of characteristics so vague, and hence subjectto alternathat tive interpretations, the right to autonomy, and with it the right to refuse lifesaving treatment,can again be overruled. In practiceit might turnout thatrefusalsof treatment would be respectedonly if there were few negative consequences and everyone agreed with the decision. Second, this view shifts the focus from the patient's refusal to the patient's condition. Appealing to a list of diagnostic and prognosticfeatures and to the consequences for others of treatment versus nontreatmentmakes the decision one 23

II

about the patient ratherthan one by the patient. The patient'srefusalbecomes simply one of many factorsto weigh in arrivingat a decision. But the thrustof placing the patient'srightto autonomyin the forefrontof medicalethics is to counteract thattendencyto securethose decisionsfor just theirs. An approachthatprepatientsthat are appropriately serves this prioritymust be developed. Four Senses of Autonomy If the concept of autonomy is clarified, we will have a of more rigorousunderstanding what the rightto autonomy to respectthatright, thus illuminating is and what it means the problemsregardingrefusals of lifesaving treatment.At the firstlevel of analysisit is enoughto say thatautonomyis that self-determination, the rightto autonomyis the rightto make one's own choices, and that respect for autonomyis the obligationnot to interferewith the choice of anotherand to treatanotheras a being capableof choosing. This is helpful, but the concept has more than one meaning. There are at least four senses of the concept as it is used in medical ethics: autonomyas free action, autonomyas authenticity, autonomyas effective deliberation,and autonomyas moral
reflection.9

Autonomyasfree action. Autonomyas free actionmeans an actionthatis voluntaryand intentional.An actionis voluntaryif it is not the result of coercion, duress, or undue influence. An action is intentionalif it is the conscious object of the actor. To submit oneself, or refuse to submit oneself, to medical treatment is an action. If a patient wishes to be treatedand submitsto treatment,thataction is intentional.If a patientwishes not to be treatedand refuses that may treatment, too is an intentionalaction. A treatment be a free actionby the physicianand yet the patient'saction is not free. If the meningitisvictim is restrainedand medication administered against his wishes, the patienthas not submittedto treatment.If the patient agrees to voluntarily pain relief medication,but is given an antibioticwithouthis knowledge, the patient voluntarilysubmittedto treatment, but it was not a free action because he did not intend to receive an antibiotic. The doctrine of consent, as it was before the law gave us the doctrine of informed consent, requiredthatpermissionbe obtainedfrom a patientand that the patient be told what treatmentwould be given; this maintainsthe right to autonomyas free action. Permission to treat makes the treatmentvoluntaryand knowledge of what treatmentwill be given makes it intentional. Autonomy as authenticity. Autonomy as authenticity means that an action is consistent with the person's attitudes, values, dispositions, and life plans. Roughly, the Ourinchoatenotionof authenpersonis actingin character. ticity is revealed in comments like, "He's not himself today" or "She's not the JaneSmithI know." For an actionto it be labeled "inauthentic" has to be unusualor unexpected, in itself or its consequences, and have relativelyimportant
24

no apparentor profferedexplanation.An action is unusual for a given actorif it is differentfrom what the actoralmost always (or always) does in the circumstances,as in, "He always flies to Chicago, but this time he took the train."If an action is not of the sort that a person eitherusually does or does not do, for example, something more like getting marriedthan drinkingcoffee, it can still be a surpriseto those who know the person. "What!George got married?" A person'sdispositions,values, and plans can be known, and particular actions can then be seen as not in conformity with them. If the action is not of serious import, concern To about its authenticityis inappropriate. ask of a person who customarilydrinks beer, "Are you sure you want to drinkwine?" is to make much of very little. If an explanation for the unusualor unexpectedbehavioris apparent,or given by the actor, that usually cuts off concern. If no explanationappearson the face of thingsor if one is given that to is unconvincing, then it is appropriate wonder if the acthe person wants to take. Often we tion is really one that will look for disturbancesin the person's life that might account for the inauthenticity. It will not always be possible to label an action authentic or inauthentic,even where much is known abouta person's attitudes,values, and life plans. On the otherhand, a given dispositionmay not be sufficientlyspecific to judge that it action. A generouspersonneed would motivatea particular not contributeto every cause to merit that attribute.If a person's financialgenerosity is known to extend to a wide range of liberalpolitical causes, not making a contribution to a given liberalcandidatefor political office may be inauthentic. On the other hand, most people have dispositions that conflict in some situations;an interestin and commitment to scientificresearchwill conflictwith fear of invasive procedureswhen such an individualconsidersbeing a subject in medicalresearch.Many questionsaboutthis sense of autonomy cannot be explored here, for example, whether there can be authenticconversionsin a person'svalues and life plans. Autonomyas effective deliberation. Autonomy as effective deliberationmeans action taken where a person believed thathe or she was in a situationcalling for a decision, was aware of the alternativesand the consequencesof the alternatives,evaluatedboth, and chose an action based on is that evaluation.Effective deliberation of course a matter of degree; one can be more or less aware and take more or less care in makingdecisions. Effective deliberationis distinct from authenticityand free action. A person's action can be voluntaryand intentionaland not result from effective deliberation,as when one acts impulsively. Further,a person who has a rigid pattern of life acts authentically when he or she does the things we have all come to expect, but withouteffective deliberation.In medicine, there is no if effective deliberation a patientbelieves thatthe physician makes all the decisions. The doctrineof informedconsent, which requiresthat the patientbe informedof the risks and
The Hastings Center Report,August 1981

benefits of the proposedtreatment its alternatives,proand tects the right to autonomywhen autonomyis conceived as effective deliberation. GeraldDworkinhas shown that an effective deliberation must be more than an apparentlycoherent thought process.10 A person who does not wear automobile seat belts may not know that wearing seat belts significantlyreduces the chances of death and serious injury. Deliberationwithout this knowledge can be logically coherentand lead to a decision not to wear seat belts. Alternatively,a personmay know the dangers of not wearing seat belts, but maintain that the inconvenience of wearing them outweighs the reduced risk of seriousinjuryor death. Both deliberations are noneffective:the firstbecause it proceedson ignoranceof a crucialpiece of information; second because it assigns a the nonrationalweighting to alternatives. It is not always possible to separatethe factual and evaluative errorsin a noneffective deliberation.A patient may refuse treatment because of its pain and inconvenience,for example, kidney dialysis, and choose to run the risk of serious illness and death. To say that such a patienthas the relevantknowledge, if all alternativesand their likely conassequenceshave been explained, but made a nonrational of priorities,is much too simple. A more accurate signment characterization may be that the patient fails to appreciate certainaspects of the alternatives.The patientmay be cognitively aware of the pain and inconvenience of the treatment, but becausehe or she has not experiencedthem, may believe that they will be worse than they really are. If the patient has begun dialysis, assessment of the pain and inconvenience may not take into account the possibilities of adaptingto them or reducing them by adjustmentsin the treatment. In order to avoid conflating effective deliberationwith reachinga decision acceptableto the physician, the following must be kept in mind: first, the knowledge a patient is needs to decide whetherto accept or refuse treatment not equivalentto a physician's knowledge of alternativetreatments and their consequences; second, what makes a is weightingnonrational not thatit is differentfrom the physician's weighting, but eitherthatthe weightingis inconsistent with other values that the patientholds or that there is good evidence thatthe patientwill not persistin the weightof ing; third,lack of appreciation aspectsof the alternatives is most likely when the patient has not fully experienced them. In some situationsthere will be overlapbetween deof and terminations authenticity effective deliberation.This does not undercutthe distinctions between the senses of autonomy;ratherit shows the complexity of the concept. Autonomyas moral reflection. Autonomy as moral reflectionmeansacceptanceof the moralvalues one acts on." The values can be those one was dealt in the socialization process, or they can differ in small or large measure.In any case, one has reflected on these values and now accepts them as one's own. This sense of autonomyis deepest and
The Hastings Center

most demandingwhen it is conceived as reflectionon one's complete set of values, attitudes,and life plans. It requires rigorous self-analysis, awarenessof alternativesets of values, commitmentto a method for assessing them, and an abilityto put them in place. Occasional,or piecemealmoral reflectionis less demandingand more common. It can be broughtabout by a particularmoral problem and only requires reflection on the values and plans relevant to the problem. Autonomy as moral reflection is distinguished from effective deliberation,for one can do the latterwithout questioningthe values on which one bases the ohoice in a deliberation.Reflectionon one's values may be occasioned by deliberationon a particular problem, so in some cases it may be difficult to sort out reflection on one's values and plans from deliberation using one's values andplans. Moral reflection can be related to authenticityby regardingthe what sortof personone will be andin formeras determining comparisonto which one's actionscan be judged as authentic or inauthentic.

ResolvingApparentConflicts
The distinctionof four senses of autonomycan be used to resolve the apparentconflict between autonomyand medical judgmentthatthe four cases generate.The action of the Jehovah'sWitnessin Case 2 is autonomousin at least three of the senses. It was a free action because it was voluntary and intentional. The patient was not being coerced and knew whathe was doing. It was an authenticactionbecause it was demanded by a strongly held religious belief. A Jehovah'sWitness who acceptedtransfusionunderthe circumstances would be regarded as one who lacked the strengthof commitmentto resist earthly temptations;this would not be cause for blame, for it is understandable and, if you are not a Jehovah'sWitness, commendable.The acbecause the pation was the result of effective deliberation he had a choice, was aware of the alternatives tient knew and their consequences, evaluatedthem on his values, and made a choice. The situationwas so clear and his belief so probablydid not take much time strongthatthe deliberation effective deliberationis long and painstaking and thought; only when the matterfor decision is perceived to be difficult. Whetherthe patientengaged in moralreflectionis difficult to determine.The case is not sufficientlydetailed to know whetherthe patientever carefullyreflectedon his religious beliefs. One can have strongbeliefs withoutever having thoughtcarefullyaboutthem. Further,since no position has been taken on just what the standardsfor adequate moral reflectionare, it is not possible to make a determination even if all the facts were there. Whetherone can, or should, choose a life plan or a religious belief by reasoned at inquiry(effective deliberation the most generallevel) is a matterof controversyin philosophy and theology. In Case 1, the physicianwith cancer, the refusalof treatment was a free action, authentic,and the resultof effective 25

deliberation.The decision to treat the patient after he had in refusedresuscitation the event of cardiovascular collapse was clearly a violation of his autonomy.He did not volunhe tarily submit to treatment; was treated against his will even though no force or threatof force had to be used. He did not intend to submit to treatment,his conscious desire was not to be treated. The authenticityof the refusal of treatmentis less a matterof identifying a particular strong belief and showing that the action is in accord with it, than it is a matterof the patientannouncingthatfurther resuscitation would incur needless suffering. Because this patientis a physicianwho has seen such sufferingand is now undergoing it, it is more likely that the assertionis coming from the patient's values, and not as something that is not an authenticexpressionof himself. The refusal also appearsto the be the result of effective deliberation: patientknew the alternatives and their consequences, his assessment and weighting of them cannot be regardedas nonrationalor a lack of appreciation.Again we do not know whether and values that how this patienthas reflectedon the fundamental determinehis judgment,but to requirethat he subjectthem to some sort of reflectionbefore his wishes have to be reof spectedwould be to set the standards autonomytoo high. In Case 4, the man with MS who attemptedsuicide, the action of the patientis a free action, that is, voluntaryand intentional,and it is the result of effective deliberation,but it is not authentic.This was the outcome of the case: The patient had too much pride to complain to his wife about his feelings of abandonment.He was able to recognize that his suicide attempt and his insistence on death with dignity were attemptsto draw thefamily's attentionto his needs. Discussion with allfour family membersled to improvedcommunicationand acknowledgmentof the patient's special emotional needs. After these conversations, the patient explicitly -retracted both his suicidal threats and his demand that no supportivemedical efforts be undertaken.12

It is tempting to say that the actions of the patient, the suicide attemptand the refusal of treatment,were not free actionsbecause they were neithervoluntarynor intentional. Even though the patient was not coerced directly by aninto the actionsby his conditionand other,he was pressured the circumstancesof his mother-in-law'sterminalillness. Further,it was not an intentionalaction because he did not really want to die; he wanted attentionand support. This position is not defensible. First, the claim that the actions were not voluntaryrests on the fact that the pressure of as circumstances a motivatingfactor can be as strongas the directthreatof anotherperson. Indeed, it is easy to imagine to cases whereit would be stronger.It is important preserve of voluntariness,and treatsimia clear and distinctconcept lar but distinguishablesituationsunder a differentrubric. Second, the claim thathis action was not intentional,that is, not his conscious object, is wrong for two reasons. It 26

fails to distinguishthe action of taking the overdose of diazepam from his saying that he wanted no treatmentand that he wanted to die with dignity. It was his conscious desire to take the diazepamandto refuse treatment; whether it was his conscious desire to die with dignity is a separate matter.Its answerrequiresadducingconsiderations bethat long to the notion of autonomyas authenticity.The belief thatthe patientdid not want to die dependson knowing that he had gotten on very well for many years, thathis change of view was coincidentto the illness of his mother-in-law and the family's attentionto her, that a desire to die is not consistentwith the values revealedby the past severalyears of the patient'slife, and that there is no apparentor proffered explanation a changein values but insteadan explaof nation of the alleged desire for death with dignity as a way of assertinghis demandson his family. His suicide attempt, the taking of diazepam, and refusal of treatmentwere free actions, but they were not authentic. The claim that the patient'staking of diazepam was the result of effective deliberationis more difficult to defend. The hospital staff regardedthe explanationas logically coof herent,but the appearance logical coherenceis not suffiIf cient for effective deliberation. the patientlackedrelevant knowledge, made a nonrationalassignment of weights to or one or alternatives failed to appreciate of the alternatives its consequences,then the decision to take an overdose and requestthatlifesaving measuresnot be startedwould not be the result of effective deliberation. The case description lacks the detailrequiredto reacha definiteconclusionon all of these. The most difficultis whetherthe patientoverestimatedthe difficultyof continuinghis life as a victim of MS; it is hard to imagine that he lacked knowledge, and even though we might regardhis weighting of deathversus continued life in his condition as mistaken, the severity of his conditionand the difficultyof coping with it do not readily supporta claim that it is not a rationalweighting. Further, he is the personwith MS andhe is the one who has suffered it for fifteen years; for anotherperson to believe that the the patientfails to appreciate severityof his conditionseems on its face to discountthe most relevantexperience. On the other hand, his appreciationat the time of the attempted suicide might be said to have been alteredby the perceived is threatto his care and comfort. More information required as to decide this;even if his actionis ultimatelyregarded the result of effective deliberation,it is still not authentic.One might even say that the lack of authenticityinfluences the effectivenessof the deliberation.More strongly,if an action is not authentic,whetherit is the resultof effective deliberation becomes somewhat irrelevant;it is ratherlike asking whethera personeffectively decided a matterbased on values or plans thatwere not his own. As in the othercases, we have no clear evidence that the patient ever engaged in moralreflection,or if he did thatthe values andlife planshe reflectedabouthad directbearingon the issues presentedby his attemptedsuicide. It does seem that a person who has
The Hastings Center Report, August 1981

had to manage his life with a seriously debilitatingillness must have given some thoughtto what is important him to and what sort of life plan is suited to him. In the case of the patient in the emergency room with meningitis, his refusal of treatmentis autonomousin the sense of free action. But is it authenticand the result of This is difficult to determineunless effective deliberation? someone in the emergencyroom knew the patientwell or is able to get to know the patient well. Presumablyno one knows him and thereis not enoughtime to get to know him; if treatmentis delayed there is risk of brain damage and death. Assuming thatthe patienthas the capacityfor autonto omy in all four senses, treatinghim would be contrary his in the sense of free action;whetherit would be a autonomy violation of his autonomy in the senses of authenticity, effective deliberation,and moralreflectioncannotbe determined. Treatinghim would makepossible his furtherdeliberationon whetherhe wished to live. On balance, it is more respectfulof autonomy, given all four senses, to treathim againsthis will. On the otherhand, it might be arguedthat meningitis has made the patient incompetent.Though the patienthas voluntarilyand intentionallyrefused treatment, his disease has removed his capacityto act authenticallyor to effectively deliberate.If this is so, then the patient'srefusal is not an autonomousaction, and the obligation to respect the autonomyof patientswould not be abridgedby treatingthe patient against his wishes.

A Bridge between Paternalismand Autonomy

This discussion shows thatthere is no single sense of autonomy and that whether to respect a refusal of treatment of requiresa determination what sense of autonomyis satisfied by a patient'srefusal. It also shows that thereneed not be a sharp conflict between autonomy and medical judgwith ment. Jacksonand Youngnerarguethatpreoccupation patient autonomy and the right to die with dignity pose a "threatto sound decision making and the total (medical, social and ethical) basis for the 'optional' decision."13 Sound decision making need not run counterto patientautonomy;it can involve a judgmentthat the patient'srefusal sense. of treatmentis not autonomousin the appropriate What sense of autonomyis requiredto respect a particular refusal of treatmentis a complex question. is If a refusalof lifesaving treatment not a free action, that is coercedor not intentional,thentherecan be no obligais, to tion to respectan autonomousrefusal.It is important note that if the action is not a free action then it makes no sense to assertor deny that the action was autonomousin any of the other senses. A coerced action cannot be one that was chosen in accord with the person's characterand life plan, nor one thatwas chosen aftereffective deliberation,nor one that thatwas chosen in accordwith moralstandards the person has reflectedupon. The point is the same if the actionis not intentional.When a refusal of treatmentis not autonoThe Hastings Center

mous in the sense of free action, the physicianis obliged to see that the coercion is removed or that the person understands what he or she is doing. Is it possible that coercion cannotbe removedor thatthe action cannotbe made intentional?This could be the case with an incompetentpatient, not externallycoerced, but subject to an internalcompulhis sion, or who lacked the capacity to understand or her situation.For incompetentpatientsthe questionof honoring refusals of treatmentdoes not arise; it is replaced by the issue of who should make decisions for incompetentpatients, an issue beyond the scope of this article. is If a refusalof treatment a free actionbut thereis reason or to believe thatit is not authentic not the resultof effective deliberation,then the physician is obliged to assist the patient to effectively deliberateand reach an authenticdecithat sion. This is whathappenedin Case 4. It is not required everyone bringabout, make possible, or encourageanother to act authenticallyand/oras a result of effective deliberation. Whethersuch an obligationexists depends on at least two factors:the natureof the relationshipbetween the two persons and how serious or significantthe action is for the actor and others. Comparethe relationshipsof strangers, mere acquaintances,and buyer and seller on the one hand, with those of close friends, spouses, parentand child, physician and patient, or lawyer and client. To borrow, and somewhatextend, a legal term, the latterare fiduciaryrelationships;a close friend, parent,spouse, physician, or lawat yer cannottreatthe otherpersonin the relationship arms' but has an obligation to protect and advance the length, interestsof the other. For example, we have no obligation to advise a mere acquaintanceagainst making an extravagant and unnecessarypurchase, though it is an option we have so long as we do not go so far as to interferein someone else's business. The situation is different for a good friend, a close relative, or an attorneywho is retainedto give financialadvice. The otherfactor, the seriousnessof the action, is relevant to medical and nonmedicalcontexts. If, inspiredby the lure of a "macho" image, my brotherimpulsively decides to buy yet anotherexpensive automobile,how I respondwill depend on how it will affect him and his dependents.If a that patientrefuses a treatment is elective in the sense thatit benefithim if done but will not have adverseconsemight quences if not done, a physician can accept such a refusal even though it is believed not to be the result of effective deliberation.On the other hand, if the refusal of treatment has serious consequencesfor the patient, the physicianhas the obligationto at least attemptto get the patientto make a decision thatis authenticandthe resultof effective deliberation. For the patientwith meningitiswho refuses treatment (Case 3) the consequencesof the refusalare indeed serious, to but there is no opportunity determinewhetherthe deciand, and the resultof effective deliberation sion is authentic to encourage and make possible an authenticand if not, effectively deliberateddecision.

27

---- --

A crucialissue is whethera refusalof lifesaving treatment in thatis autonomous all four senses can be justifiablyoverridden by medical judgment. It will help here to compare the Jehovah'sWitness case with a somewhat fanciful expansion of the meningitiscase. The former'srefusal is autonomous in three of the four senses and could be judged autonomousin the sense of moral reflection if we knew more about the patient's acceptanceof his faith and had a clear idea of the criteriafor moral reflection. Though the belief of Jehovah's Witnesses are not widely shared, and many regard as absurd the belief that accepting a blood transfusionis prohibitedby biblical injuncton, their faith has a fair degree of social acceptance. Witnesses are not regarded as lunatics. This is an importantfactor in the recognitionof theirrightto refuse transfusion.Supposethat the meningitisvictim had a personalset of beliefs that forbid the use of drugs, thatafteryears of reflectionhe came to the view that it was wrong to corruptthe purityof the body with foreign substances.Suppose thathe acts on this belief consistentlyin his diet and medical care, that he has carefully thoughtaboutthe fact thatrefusalin this circumstance may well lead to death, but he is willing to run that risk because his belief is strong. This case is parallel to the Jehovah'sWitnesscase; the principaldifferenceis thatthere is no large, organizedgroup of individualswho share the belief and have promulgatedand maintainedit over time. One reactionis to regardthe patient as mentally incompetent, with the centralevidence being the patient's solitary stance on a belief thatrequiresan easily avoided death. An is alternative approach to not regardthe patientas incompebut to see treatmentas justified paternalism.Finally, tent, the position could be that a refusal of lifesaving treatment that is fully autonomous,thatis, in all four senses, must be respectedeven though the belief on which it is founded is to eccentricand not socially accepted. Which approach take would requirean analysis of incompetence, a definitionof paternalism,and an examinationof when it is justified.14 Defining paternalismas an interferencewith autonomyin one or more of the four senses might be an illuminating approach. The conflict between the rightof the patientto autonomy and the physician'smedicaljudgmentcan be bridgedif the conceptof autonomyis given a morethoroughanalysisthan it is usually accordedin discussions of the problem of refusal of lifesaving treatment.In some cases where medical judgmentappearsto overrideautonomy,the four senses of autonomyhave not been taken into account.
REFERENCES 'Alan Donagan, The Theory of Morality (Chicago: The University of Chicago Press, 1977); Ronald Dworkin, Taking Rights Seriously UniversityPress, 1977); John Rawls,A Theoryof (Cambridge:Harvard Justice (Cambridge:HarvardUniversity Press, 1971). 2David L. Jackson and Stuart Youngner, "Patient Autonomy and 'Death with Dignity,"' The New England Journal of Medicine 301 (1979), 404. 3Thiscase is drawnfrom W. St. C. Symmers, Sr., "Not Allowed to

Die," British Medical Journal 1 (1968), 442; it is reprintedin Tom L. Beauchamp and James F. Childress, Principles of Biomedical Ethics (New York: Oxford University Press, 1979), p. 263. 4This case is drawn from In Re Osborne, 294 A.2d 372 (1972). SThis case is drawn from Eric J. Cassell, "The Function of Medicine," Hastings Center Report 6 (1976), 16. 6This case is drawn from Jackson and Youngner, p. 406. 7Ibid., p. 408. 8MarkSiegler, "CriticalIllness: The Limits of Autonomy" Hastings Center Report 8(1977), 12-15. 9Beauchampand Childress, pp. 56-62; Donagan, p. 35; Gerald Dworkin, "Autonomyand Behavior Control,"Hastings CenterReport 6(1976), 23; and "Moral Autonomy" in H. TristramEngelhardtand Daniel Callahan, Morals, Science and Society, (Hastings Center, 1978), p. 156; HarryG. Frankfurt,"Freedomof the Will and the Concept of a Person," TheJournal of Philosophy 68(1971), 5; BernardGert and Timothy J. Duggan, "Free Will as the Ability to Will" Nous 13 (1979), 197; Charles Taylor, "Responsibility for Self" in Amelie Rorty, ed. The Identitiesof Persons (Berkeley: University of California Press, 1976). l?GeraldDworkin, "Paternalism,"in Richard A. Wasserstrom,ed. Morality and the Law (Belmont: WadsworthPublishing Co., 1971). "This brief account draws on Dworkin, "Moral Autonomy," and Taylor. '2Jacksonand Youngner. 31Ibid., 405. p. "Paternalism";Bernard Gert and Charles M. Culver, "4Dworkin, "PaternalisticBehavior," Philosophy and Public Affairs 6(1976), 45; and "The Justification of Paternalism," in Wade L. Robison and Michael S. Pritchard, eds. Medical Responsibility: Paternalism, Informed Consent, and Euthanasia (Clifton, N.J.: HumanaPress, 1979), pp. 1-14.

Visiting Senior Scholar Program at The Hastings Center 1982-1983

for is Center now accepting The Hastings applications its to commence 1982-1983 VisitingSeniorScholar Program, 1, September 1982. The Undera grantfromthe HenryR. LuceFoundation, senior HastingsCenteris able to appointan established scholarfor the academicyear 1982-1983 to a visiting staff fellowship. The senior scholar will be residential half to spendapproximately of his or her time at expected and the Center projects will partin ongoingresearch taking to pursuepersonalresearch be free for the other half will of interests applicants be exinterests.The research to overlapwith ongoing work or projectsat the pected and musthavea distinguished Center. scholarly Applicants to recordin a field or fields directlypertinent publication the workof the Center.The VisitingSeniorScholarwill be expectedto spendten full monthsin residenceat the Center.The stipendwill rangefrom$35,000-$40,000 for fundswill be provided the ten-month periodandadditional in to to assist in costs incidental relocating Hastings. The HastingsCenteris an equal opportunity employer. write to Bonnie or For applications furtherinformation, The Baya, PersonnelCoordinator, HastingsCenter,360 N.Y. Broadway, Hastings-on-Hudson, 10706.

28

The Hastings Center Report, August 1981