Journal ofMedical Ethics 1997; 23: 361-367

Marginally effective medical care: ethicalanalysis of issues in cardiopulmonaryresuscitation (CPR)Mark Hilberman, Jean Kutner, Debra Parsons, and Donald J Murphy The Carbondale Clinic, Carbondale,Colorado, University of Colorado Health Sciences Center, St Joseph's Hospital, and the Colorado Collective forMedical Decisions, Denver, Colorado, USA

AbstractOutcomes from cardiopulmonary resuscitation (CPR)remain distressingly poor. Overuse of CPR isattributable to unrealistic expectations, unintendedconsequences of existing policies andfailure to honourpatient refusal ofCPR. We analyzed the CPR outcomesliterature using the bioethical principles of beneficence,non-maleficence, autonomy andjustice and developed aproposalfor selective use ofCPR.

Beneficence supports use ofCPR when most effective.Non-maleficence argues against performing CPR whenthe outcomes are harmful or usage inappropriate.Additionally, policies which usurp good clinicaljudgment and moral responsibility, thereby contributingto inappropriate CPR usage, should be consideredmaleficent. Autonomy restricts CPR use when refusedbut cannot create a right to CPR.

J7ustice requires that we define which medicalinterventions contribute sufficiently to health andhappiness that they should be made universallyavailable. This ordering is necessary whether onebelieves in the utilitarian standard or wishes medicalcare to be universally available on fairness grounds.Low-yield CPR fails justice criteria.

Cardiopulmonary resuscitation should be performedwhen justified by the extensive outcomes literature; notperformed when not desired by the patient or notindicated; and performed infrequently when relativelycontraindicated.

IntroductionCardiopulmonary resuscitation (CPR) can be extra-ordinarily successful. During heart surgery andelectrophysiological studies the heart is stopped andstarted, with uniform success. Otherwise, recoveryfrom cardiac arrest remains unlikely. Cardio-pulmonary resuscitation represents the opportunityfor life when cardiac arrest occurs. Yet it remainssimply an intervention which is neither intrinsically

Key wordsEthics; bioethics; cardiopulmonary resuscitation; CPR;clinical decision-making; death and dying.

effective nor benign. Use of this intervention canrestore good health and wellbeing to some survivorswhile the overall benefit to others is low and someare left significantly impaired.

Existing guidelines promote CPR to restore lifewhen cardiac arrest occurs from cardiac causes.'Statutes which give an implied consent to emergencytreatment are used to endorse CPR as the defaultresponse to cardiac arrest. It has become commonpolicy to require CPR unless CPR is explicitlyrefused2 3 or futile.4 Yet, cardiac arrest normallyaccompanies death and an extensive medical litera-ture supports selective use of CPR.5 6 Cardio-pulmonary resuscitation is not an appropriateresponse to death which occurs as a consequence ofadvanced age or illness.7 8Two anecdotes illustrate limitations in the current

Do Not Resuscitate (DNR) approach:

1. A university hospital cardiac resuscitation team,composed of house physicians, nurses and respira-tory therapists, arrived at the bedside of a ninety-year-old with advanced cancer whose heart hadstopped. A DNR order had not been written, andhospital policy required the resuscitation team toproceed with CPR despite their grave medical andmoral reservations. An attending physician relievedthem of this burdensome policy requirement.2. A robust 60-year-old was admitted to hospitalwith vague chest pains for diagnosis and treatment.After evaluation he was questioned about his CPRpreferences, a new conversation promoted byAmerican laws intended to promote refusal of bur-densome life-prolonging interventions. He indicatedthat he did not wish CPR and a DNR order waswritten. Within the hour he developed ventricularfibrillation (a treatable, lethal cardiac rhythm distur-bance.) Simple electrical defibrillation would havebeen life-saving. However, the staff felt compellednot to intervene due to the DNR order and wereprepared to let him die. Fortunately, his heart spon-taneously reverted to a normal rhythm. When askeda second time, the patient changed his mind.

When people are under stress it is unrealistic to

copyright. on S

eptember 4, 2020 by guest. P

rotected byhttp://jm

e.bmj.com

/J M

ed Ethics: first published as 10.1136/jm

e.23.6.361 on 1 Decem

ber 1997. Dow

nloaded from

362 Marginally effective medical care: ethical analysis of issues in cardiopulmonary resuscitation (CPR)

expect consistently sound decision-making. Patientswho will not benefit from CPR usually benefit fromother medical and nursing care. However, a writtenDNR order may represent "giving up" to patients,families and providers. For example, 92% of neurol-ogists surveyed indicated that a DNR order for astroke patient implied withdrawal of other treatmentand nutrition. ' Neurologists make the same linkagefor patients in a vegetative state.'0A CPR paradigm must recognize the complexity

of the clinical environment and the importance ofprovider judgment and must respect patientautonomy. The heart can stop in patients at any ageand in conjunction with many underlying illnessesand the rules should facilitate prompt, effective andappropriate CPR. In this paper we present ouranalysis ofCPR issues using the four bioethical prin-ciples of beneficence, non-maleficence, autonomyand justice. "

BeneficenceMoral agents should take positive steps to help others.In medicine this normally involves the restoration ofhealth and function and the relief of pain and suffer-ing. These goals were clearly accomplished in theearly steps towards effective resuscitation: in the late1 940s and early 50s respiratory intensive careincreased survival in bulbar poliomyelitis from about15% to over 50%.12 13 A decade later, 14 of 20patients (70%) treated by closed chest cardiacmassage survived intact.'4 However, investigatorswho followed Kouwenhoven et al at Johns Hopkinssubsequently reported hospital discharge rates of 14%in 1985,'5 and below 10% in 1994.16 The success rateof 70% was never duplicated.5 6The greatest benefit from CPR, with survival rates

over 20%, was reported when cardiac arrest occuredduring anaesthesia, from drug overdose, and withcoronary disease or a primary ventricular arrhyth-mia.5 6 A 1995 hospital discharge rate of only 17%followed CPR in coronary care unit patients, whoare closely monitored and by skilled staff.'7 Thetreatment goal is to prevent cardiac arrest, whichfrequently represents therapeutic failure and difficultdisease.

Patients survive CPR infrequently when non-cardiac major illness or organ dysfunction precedecardiac arrest. Cardiopulmonary resuscitationsurvival is extremely poor (<5%) when cardiacarrest occurs in patients with renal failure, cancer(unless disease is minimal), or AIDS; and in theabsence of irreversible underlying illness, followingtrauma, haemorrhage, sustained hypotension orpneumonia.5 6 18 19 Limiting use of CPR improvedresults in cancer patients who had a 10-5% post-CPR survival when 7-10% of those who arrestedreceived CPR.20

Field CPR promptly initiated in the streets ofSeattle resulted in a 36% hospital survival rate, the

best in the recent literature.2' In other systems theresults are so poor that the whole endeavour shouldbe reconsidered. Specifically, fewer than 2%survive field CPR to hospital discharge in Chicagoor New York, due to traffic-delayed initiation ofresuscitation.2223

Non-maleficenceWe wish to examine the admonition to "do no(deliberate) harm" in terms of outcomes, policies,and appropriateness.

OUTCOMESThe incidence of brain injury following CPR variesfrom 10-83%.24 28 In one study, 55 of 60 childrendied following prolonged field resuscitation; all fivesurvivors were in persistent coma or a vegetativestate at the time of hospital discharge.27 Manypatients regard severe disability following significantbrain injury as worse than death.29 Cardiopulmonaryresuscitation becomes maleficent when the risk ofbrain injury is high.

POLICIESAmerican DNR policies were established to protectpatients from unilateral physician DNR decisionsand generally require CPR unless explicitlyrefused.2 3 Rigid field CPR rules resulted in at leastone dramatic headline: Paramedics rush dead peopleto hospitals, costing millions.30 Since even a relativelybrief interruption of blood flow to the brain or heartresults in severe injury, resuscitation can onlysucceed if applied promptly. Thus, Swedish inves-tigators reported that survival exceeded 80% withbystander CPR and ambulance arrival in less thantwo minutes, but was less than 6% with ambulancearrival time over six minutes or no bystanderCPR.25 Nevertheless, in 1993 it was noted thatChicago paramedics were required to resuscitateunless the victim was decapitated, in rigor mortis, ordecomposing.23 In some states, emergency crewsare bound to proceed with CPR despite evidenceat the scene that CPR is not wished. In one suchstate, 7% of out-of-hospital resuscitations wereunwanted.2'A generation ago moral responsibility was placed

squarely upon the individual to act appropriatelyregardless of orders given. Nevertheless, Americanhealth professionals have yielded to policies and lawswhich require that they provide CPR even whenjudgment and conscience protest. Actions andpolicies which violate judgment and conscienceshould be considered maleficent.

APPROPRIATENESSCardiopulmonary resuscitation is a rough, somewould say abusive, intervention. When life issnatched from death, this is inconsequential.However, cardiac arrest normally precedes death

copyright. on S

eptember 4, 2020 by guest. P

rotected byhttp://jm

e.bmj.com

/J M

ed Ethics: first published as 10.1136/jm

e.23.6.361 on 1 Decem

ber 1997. Dow

nloaded from

Mark Hilberman, Jean Kutner, Debra Parsons and DonaldJ Murphy 363

and providers are appropriately disturbed when theyperform CPR on people afflicted by advancedillness, the debilities of old age, or dementia.

Age per se is not a contraindication to CPR.17 31Nevertheless, aging is associated with the accumula-tion of infirmities and illnesses7 32 33 which, withlong-term nursing home residence and diminishedfunction, remain predictors of poor CPR out-come.20 34-36 Survival to hospital discharge followingCPR in public, home, and nursing home settings was36%, 13% and 3% respectively, where "severechronic disease" was present in 6%, 33% and 68%respectively.2'One description of nursing home CPR survivors

presents the concern better than dry survival per-centages:

One of the two nursing-home residents whosurvived was an 87-year-old woman who spent 30days in the hospital and died eight months afterreturning to the nursing home, demented, cachec-tic, with a large sacral pressure sore. The other wasan 81-year-old man who, after a 60-day hospitaliza-tion, returned to the nursing home and died there14 days later.33

We should not treat our elders this way.Inappropriate CPR is maleficent.

Net benefit, futility and the decisionalproblemWhat about interventions of little benefit? Whodecides to use them? When a person's life is charac-terized by advanced illness, dependency or dementiathe net benefit of CPR - few survivors with a poorquality of life - is inadequate. Cardiopulmonaryresuscitation is not maleficent when net benefit islow, yet such poor outcomes are an inadequate goalfor a medical intervention. Giving the patient thedecision does not resolve the issue. The scientificmedical enterprise depends upon evaluating thera-peutic efficacy against defined goals. While issuesinvolved in death and dying are difficult, the need fordefined goals remains essential.

Physicians face similar problems when recom-mending any treatment. Patients commonly presentproblems which are more complex than the precisestudies in the literature. The provider cannot beabsolved of exercising necessary professionaljudgment both in selecting appropriate therapies anddeciding how best to present this information to theindividual. Cardiac arrest may force rapid decisionsand preclude discussion. However, advance direc-tives can play only a limited role in reaching adecision, as the circumstances of the arrest alsodetermine if intervention is appropriate.

"Futile ... 1. serving no useful purpose: ineffec-tive, fruitless".37 Tomlinson and Brody recognizedthat to declare a medical intervention futile involved

balancing complex uncertainties and decisionalresponsibilities.38 Schneiderman and Jecker exploredfutility and developed a quantitative definition offutility which requires that an intervention has failedthe last 100 times used.39 This implies unreasonablecertainty. It is better to recognize that certainty willelude us and that these decisions must involve a sub-jective balancing of burdens and benefits.The argument that professional decisional respon-

sibility is necessary is most effectively developed inThe Death of Common Sense, a critique of theAmerican desire to eliminate judgment in humanaffairs by developing evermore specific regulations.40Individualized complex discussion with patientand/or family is part of the evaluation of benefits andburdens which remains essential.4' One importantlesson of the twentieth century is that when the statedecides these issues, it often puts its citizens in peril.Insurors (and America's "health maintenanceorganizations") have inherent conflicts of interest indeciding medical care, and lack both the moralcapacity and the close involvement with the patientand family to make such decisions.

Autonomy (liberty)There is widespread support for the right to refuseunwanted treatments. However, autonomy impliesmore and may not best characterize this principle.For Kant and Rawls an autonomous moral decisionwas a rational and dispassionate decision. Rawls setthis forth clearly - he required that his justice rule-makers, his imaginary choosers, be ignorant of theirfuture status in the just society.42 From autonomysome have drawn the inference that the patientshould be able to demand treatment. This extensionrequires that we re-examine this principle andresolve two issues. First, it is clear that the patient isalways concerned with the outcome of clinicaldecisions and is therefore not strictly autonomouseven when judgment is not clouded by pain, fear orsuffering. Second, it is the principle of justice whichcreates a right to receive something, not autonomy.43

In their most recent formulation Beauchamp andChildress more accurately present the principle as"Respect for Autonomy""; however, we believe thereare two formulations which may be superior. JohnStuart Mill's On Liberty remains the most eloquent:

"The object of this essay . . . is to assert one verysimple principle, . . . that the sole end for whichmankind are warranted, individually or collectively,in interfering with the liberty of action of any of theirnumber, is self protection. That the only purpose forwhich power can be rightfully exercised over anymember of a civilized community, against his will, isto prevent harm to others. His own good, eitherphysical or moral . . . [provides] good reasons forremonstrating with him, or reasoning with him, orpersuading him, or entreating him, but not for

copyright. on S

eptember 4, 2020 by guest. P

rotected byhttp://jm

e.bmj.com

/J M

ed Ethics: first published as 10.1136/jm

e.23.6.361 on 1 Decem

ber 1997. Dow

nloaded from

364 Marginally effective medical care: ethical analysis of issues in cardiopulmonary resuscitation (CPR)

compelling him or visiting any evil in case he dootherwise".43

This is wonderful and blends nicely with a profes-sional sense of obligation to assist (educate, remon-strate, reason, persuade or entreat) the patient tomake the best decision. Coercion is precluded.

In The Foundations of Bioethics, H TristramEngelhardt Jr proposes two principles of biomedicalethics: I. The Principle of Permission and II. ThePrinciple of Beneficence. If less impassioned, thisdiscussion is up to date and shifts an impossibleautonomy to necessary (and accepted) permission."His views are also concordant with the analysis ofjustice which follows.The importance of obtaining fully informed per-

mission has been taught in medical schools foryears. When we surveyed 463 nurses, physiciansand other professionals in the Denver area aboutCPR, 96% respected patient refusal of CPR(unpublished observations.) Documenting permis-sion and recognizing the right to revoke permissionensure that this critical principle is honoured.Neither respect for autonomy, respect for thepatient's liberty interests, nor the principle of per-mission create a right to receive treatment as hasbeen implied when autonomy per se has been takenas a trump prinicple.

Advance directives permit an individual somecontrol over end-of-life care; however, we cannotrequire that the paperwork be in order to delivergood emergency medical care. As to the inappro-priate use of technology, professionals are generallymore willing to forgo non-beneficial care than are

3 4546patients. Death occurs quite infrequently inhospitalized patients - in a cancer referral centeronly 1 5% of patients sustained a cardiac arrest.20This low incidence may help explain the avoidanceof routine DNR discussions.

Finally, false hope undermines the ability of thepatient to make an informed judgment. LimitingCPR availability to those situations where its goalsmay be reasonably obtained enhances respect forliberty, autonomy and the permission process.

JusticeMoral justice considerations involve the creation ofrights to receive something, the resolution of compet-ing individual demands and the balancing of socialgoals. The issue is whether there should be a moraljustice requirement to provide medical care to those inneed on utilitarian grounds,43 or because justice asfairness requires that we mitigate the inequalitieswhich randomly occur in life.42 The libertarianargument, that medical care should be treated likeother desirable goods and left to the individual, retainsvalidity but is less widely supported. Physicians areaccustomed to sufficient societal resources to care forthose in need - resources generally provided as private

or governmental insurance or direct institutionalsupport. None the less, to determine whether there isa moral justice requirement for a decent minimum ofmedical care for all"1 we must examine several import-ant issues.

MEDICAL CARE AND HEALTH"The medical art is proved to be good by its conduc-ing to health; . . . ".43 Many medical interventionsmay not contribute to health. Americans spend hugeamounts on medical care, yet live no longer thancitizens of other developed countries,47 nor muchlonger than American Christian Scientists, whoeschew medical care.48 Primary health determinantsinclude: food, clothing, shelter, clean water, immu-nizations, family, education, personal wealth,societal wealth, a safe work environment, non-violent surroundings, personal behaviours, basicmedical care and clean air.49 52

Given those determinants, policies whichmaximize health would concentrate on economicprosperity, safety, responsible personal behaviourand basic medical care. That leads us to a hypo-thetical question: under what circumstances shouldwe consider CPR basic, optional, experimental orcontraindicated medical care?

CAN WE AFFORD TO MAKE CPR AND OTHEREXPENSIVE MEDICAL INTERVENTIONS UNIVERSALLYAVAILABLE?The risk-sharing argument - if we each contribute asmall amount to the kitty, then expensive interven-tions can be provided to those few in need - beginsto fail as more and more expensive treatments aredeveloped. Examples of the cost/benefit of variousinterventions is provided in Table 1. As a reference,we have inserted in that table the US GrossDomestic Product (presently approximately$25,000/person/year) as it estimates availableresources and is an important determinant of theamount of medical care consumed.47Some treatments are intrinsically costly, some are

marginally effective and some must be repeatedmany times. Regulation and litigation have drivencosts upwards. All cost issues need to be addressed.While aggregate expense data are needed, it isevident that it is not possible to deliver all medicalinterventions regardless of cost or effectiveness.Therefore, further definitions are needed before wecreate a right to medical care.

SOME DEFINITIONSIt is necessary to define basic medical care in terms ofoutcomes or purpose. The traditional labels ofmedicalnecessity,53 standard and experimental therapy54 arestrained by the demands of medical progress. Todevelop a preliminary definition we considered thegoodness of appendectomy, penicillin for uncompli-cated pneumonia, immunizations, fracture care andhospice care. Moreover, adult cardiac operations

copyright. on S

eptember 4, 2020 by guest. P

rotected byhttp://jm

e.bmj.com

/J M

ed Ethics: first published as 10.1136/jm

e.23.6.361 on 1 Decem

ber 1997. Dow

nloaded from

Mark Hilberman, Jean Kutner, Debra Parsons and DonaldJ Murphy 365

Table 1 Some examples of the cost/benefit of medical care

$ Cost/year ofTreatment life saved Source

Cervical cancer screening in elderly (1,400)* Mandelblat & Fahs, 198855Aortic valve replacement56 2,800 Hilberman**Left main coronary bypass 6,400 Weinstein and Stason, 198257CPR in 40% of patients with high survival <10,OOOt Theimann et al, 199416Mild-moderate hypertension, beta-blocker 10,900 Edelson et al, 199018US GDP/person/year -25,000 [The amount we produce each year to provide all our needs]End stage renal disease program 30,000 Levinsky 199360Cholesterol reduction with colestipol 63,900 Kinosian and Eisenberg, 198861Mild-moderate hypertension, ACE inhibitor 72,100 Edelson et al, 199058Triple vessel bypass, good heart muscle 95,000 Weinstein and Stason, 198557ICU care, AIDS, PCP and ventilator: 1981-91 175,000 Wachter et al, 199562ICU care for cancer: hematologic group 189,300 Schapira et al, 199363CCU care, low risk of MI 208,000 Fineberg et al, 198464CPR in 60% of patients with low survivaltt -400,000 Theimann et al, 199416, Murphy and Matchar 19906"

Dollar values are as published, the assumptions of each author have not been reexamined.*A cost savings was achieved when future treatment was considered.**Using cited data source for the difference in survival between medical and surgical treatment and a cost per cardiac operation of$20,000.t$20,500/hospital survivor, and more than half lived two years.ttThis is the cost per hospital survivor- 0/5 lived two years in the Theimann study.- = approximately.

generally achieve a one-year survival >90%/o. Bycontrast some treatments for advanced breast canceraccelerate death in two of three treated. Therefore wepropose that basic medical care should:

1. Prevent, cure, palliate, or yield a one-year survivalgreater than 75 per cent;2. Produce little toxicity or long term disability;3. Be affordable; and4. Be distinctly more beneficial than burdensome.

Other therapy should be considered either optionalor experimental.

Despite its preliminary nature, we believe it usefulto apply this definition to CPR. As burdensomeoutcomes seem roughly correlated with survival, weuse survival:

* CPR, in patient groups with anticipated survival of20-50% is experimental, generally beneficial care inneed of further evaluation and refinement.* CPR with anticipated survival of 5-20% ismarginal experimental care, in need of further evalu-ation and refinement.* CPR with expected survival below 5% or withdelayed initiation has proven an unsuccessful exper-iment and is not to be performed.

Summary of ethical argumentSelective CPR is supported by consideration of thebalance between burdens and benefits and of themoral requirement that known harmful actions beavoided. Decisional authority to use or withholdCPR must reside in providers who can use theirtraining, skills and knowledge to provide the bestavailable care. Respect for individual refusal ofunwanted treatment is widely accepted. However, ademand for CPR based solely upon the patient'swishes is difficult to support. Beneficence, non-

maleficence and securing permission remain theprimary responsibility of providers. Justice consider-ations seem especially worthwhile for therapeuticevaluation and societal consideration. Justice con-siderations support limited use of CPR and indicatethe need for better definition of medical care and therelationship of medicine to health.

CPR recommendationCardiac arrest must occur for CPR to be a relevantintervention, but cardiac arrest is not a sufficientindication for CPR. Since the decision not toperform CPR is irreversible, it is appropriate forthere to be a bias toward its initiation. However, theextensive outcomes literature and ethical analysisjustify a more limited application of CPR than dopresent DNR policies. Therefore, we recommend:

* CPR should be performed when it is indicated.* CPR should not be performed when it has beenrefused or is not indicated.* CPR should be performed infrequently when theintervention is relatively contraindicated.

Other medical care should be provided to patientsindependently of resuscitation status. Providersmust specifically exclude race or socioeconomicfactors from consideration.

Definition: CPR means full resuscitation forcardiac or respiratory arrest, including chest com-pressions, ventilation, and pharmacological orelectromechanical support.CPR is indicated: 1. For witnessed arrests. 2. For

a cardiac rhythm of ventricular fibrillation or tachy-cardia; 3. During operations and procedures; and, 4.As part of well justified experimental protocols.CPR is not obligatory.CPR is not indicated: 1. If the patient does not

want CPR. 2. If the arrest is unwitnessed, unless

copyright. on S

eptember 4, 2020 by guest. P

rotected byhttp://jm

e.bmj.com

/J M

ed Ethics: first published as 10.1136/jm

e.23.6.361 on 1 Decem

ber 1997. Dow

nloaded from

366 Marginally effective medical care: ethical analysis of issues in cardiopulmonary resuscitation (CPR)

some sign of life persists; 3. If CPR is not startedwithin six minutes of the arrest, or has continuedmore than 30 minutes (except hypothermia); 4. Forpatients in a Persistent Vegetative State, in coma, orwith severe heart or lung failure, advanced cancer, orother end-stage illness.CPR is relatively contraindicated: 1. If it is

known that the patient had significant physicaldeterioration prior to the cardiac arrest. 2. Forpersons who have severe dementia, and possibly forthose with moderate dementia (CPR is intended toprevent premature death and is not appropriate ina person who has advanced and debilitatingsymptoms of aging.) 3. For patients with cancer(who rarely survive CPR according to the medicalliterature.) However, some patients have minimalcancer and deserve CPR. 4. For victims of the AIDSepidemic for whom cardiac arrest is a late complica-tion. Exciting advances in treatment recentlyreported seem most likely to delay the occurrence ofcardiac arrest but not alter subsequent outcome.

It is intended that indications will be revised as newknowledge emerges. In any scenario we expect thatproviders will need to make individual judgments.

AcknowledgementsThe Colorado Trust partially supported this projectthrough its funding of the Colorado Collective ForMedical Decisions, an open organization of medicalcare professionals and citizens concerned with inap-propriate or ineffective medical care. Members ofthe Adult Intensive Care Subcommittee shared theirinvaluable expertise, experience and insight with us,and helped formulate the recommendation. Ourthanks to those who commented on the manuscript:Father Bill Pounds, Susan Fox Buchanan, BarbaraLundy, Ranaan Gillon and the Jfournal of MedicalEthics reviewers. The analysis in this paper is theresponsibility of the authors.

Mark Hilberman, MD, is Director of the CarbondaleClinic in Carbondale, Colorado. Jean Kutner, MD,MSPH, is Assistant Plrofessor of Medicine in theUniversity of Colorado Health Sciences Center, Denver,Colorado. Debra Parsons, MD, FACP, is Director ofClinical Inpatient Services, Internal Medicine ResidencyProgram at St Jfoseph Hospital, Denver, Colorado andDonald J Murphy MD, FACP, is a Geriatrician andDirector of the Colorado Collective for Medical Decisions.

References1 Emergency Cardiac Care Committee andSubcommittees AHA. Guidelines for emergencycardiac care, I-IX. Jrournal of the American MedicalAssociation 1992; 268: 2172-295.

2 Rothman DJ. Strangers at the bedside. New York: BasicBooks, 1991.

3 Zussman R. Intensive care, medical ethics and the medicalprofession. Chicago: University of Chicago Press, 1992.

4 Emergency Cardiac Care Committee and Sub-committees AHA. Guidelines for emergency cardiaccare, VIII: ethical considerations in resuscitation. Journalof the American Medical Association 1992; 268: 2282-8.

5 Cohn EB, Lefevre F, Yamold PR, Arron MJ, MartinGJ. Predicting survival from in-hospital CPR: meta-analysis and validation of prediction model. Journal ofGeneral Internal Medicine 1993; 8: 347-53.

6 Ebell MH. Prearrest predictors of survival following in-hospital cardiopulmonary resuscitation: a meta-analysis. Journal ofFamily Practice 1992; 34: 551-8.

7 Nuland SB. How we die. New York: Knopf, 1993.8 Golub ES. The limits of medicine. New York: TimesBooks/Random House, 1994.

9 Alexandrov AV, Pullicino PM, Meslin EM, et al.Agreement on disease-specific criteria for do-not-resus-citate orders in acute stroke. Stroke 1996; 27: 232-7.

10 Quality Standards Subcommittee of the AmericanAcademy of Neurology. Practice parameters: assess-ment and management of patients in the persistentvegetative state (summary statement). Neurology 1995;45: 1015-18.

11 Beauchamp TL, Childress JF. Principles of biomedicalethics [4th ed]. New York: Oxford University Press,1994.

12 Bower AG, Bennett VR, Dillon JB. Investigation of thecare and treatment of poliomyelitis patients. Annals ofWestern Medicine and Surgery 1950; 4: 561-82, 687-716.

13 Lassen HCA. Preliminary report on the 1962 epidemicof poliomyelitis in Copenhagen. Lancet 1953; 1: 37.

14 Kouwenhoven WB, Jude JR, Knickerbocker GG.Closed-chest cardiac massage. J7ournal of the AmericanMedical Association 1960; 173: 1064-7.

15 Dans PE, Nevin KL, Seidman CE, MacArthur JC,Kariya ST. Inhospital CPR 25 years later: why hassurvival decreased? Southern Medical Journal 1985; 78:1174-8.

16 Theimann DR, Bass EB, Powe NR, Halperin HR,Steinberg EP. Pre-arrest comorbidity, outcome andcost in inpatient CPR (abstract). Circulation 1994; 90,4, 2: supp 2: I-288.

17 Brymer C, Gangbar E, O'Rourke K, Naglie G. Age as adeterminant of cardiopulmonary resuscitation outcomein the coronary care unit. J7ournal of the AmericanGeriatrics Society 1995; 43: 634-7.

18 Ballew KA, Philbrick JT, Caben DE, Schorling JB.Predictors of survival following in-hospital cardio-pulmonary resuscitation: a moving target. Archives ofInternal Medicine 1994; 154: 2426-32.

19 Herlitz J, Ekstrom L, Wennerblom B, Axelsson A,Bang A, Holmberg S. Prognosis among survivors ofprehospital cardiac arrest. Annals ofEmergency Medicine1995; 25: 58-63.

20 Vitelli CE, Cooper K, Rogatko A, Brennan MF.Cardiopulmonary resuscitation and the patient withcancer. J7ournal of Clinical Oncology 1991; 9: 111-15.

21 Dull SM, Graves JR, Larsen MP, Cummins RO.Expected death and unwanted resuscitation in the pre-hospital setting. Annals ofEmergency Medicine 1994; 23:997-1002.

22 Lombardi G, Gallagher EJ, Gennis P. Outcome of out-of-hospital cardiac arrest in New York City. Jrournal ofthe American Medical Association 1994; 271: 678-83.

23 Becker LB, Han BH, Meyer PM, Wright FA, RhodesKV, Smith DW, et al. Racial differences in the inci-dence of cardiac arrest and subsequent survival. NewEngland Journal ofMedicine 1993; 329: 600-6.

copyright. on S

eptember 4, 2020 by guest. P

rotected byhttp://jm

e.bmj.com

/J M

ed Ethics: first published as 10.1136/jm

e.23.6.361 on 1 Decem

ber 1997. Dow

nloaded from

Mark Hilberman, Jfean Kutner, Debra Parsons and Donald _J Muirphy 367

24 Bedell SE, Delbanco TL, Cook EF, Epstein FH. Survivalafter cardiopulmonary resuscitation in the hospital. NewEngland Journal ofMedicine 1983; 309: 569-76.

25 Herlitz J, Ekstrom L, Wennerblom B, Axelsson A,Bang A, Holmberg S. Effect of bystander initiatedcardiopulmonary resuscitation on ventricular fibrilla-tion and survival after witnessed cardiac arrest outsidehospital. British Heart Journal 1994; 72: 408-12.

26 Berek K, Lechleitner P, Luef G, et al. Early determina-tion of neurological outcome after prehospital cardio-pulmonary resuscitation. Stroke 1995; 26: 543-9.

27 Ronco R, King W, Donley DK, Tilden SJ. Outcomeand cost at a children's hospital following resuscitationfor out-of-hospital cardiopulmonary arrest. Archives ofPediatric Adolescent Medicine 1995; 149: 210-14.

28 Rogove HJ, Safar P, Sutton-Tyrrell KS, et al. Old agedoes not negate good cerebral outcome after cardio-pulmonary resuscitation: analyses from the brain resus-citation clinical trials. Critical Care Medicinie 1995; 23:18-25.

29 Solomon NA, Glick HA, Russo CJ, Lee J, SchulmanKA. Patient preferences for stroke outcomes. Stroke1994; 25: 1721-5.

30 Herhold S. Fruitless rescues: paramedics rush deadpeople to hospitals, costing millions. Mercury News1993 Aug 15: 1A,ff.

31 Tresch D, Heudebert G, Kutty K, Ohlert J, VanBeek K,Masi A. Cardiopulmonary resuscitation in elderly patientshospitalized in the 1990s: a favorable outcome.3ournal ofthe Amnerican Geriatrics Society 1994; 42: 137-41.

32 Gordon M, Cheung M. Poor outcome of on-site CPRin a multi-level geriatric facility: three and a half yearsexperience at the Baycrest Centre for Geriatric Care.Jtournal of the Americani Geriatrics Society 1993; 41:163-6.

33 Applebaum GE, King JE, Finucane TE. The outcomesof CPR inititated in nursing homes. Yournral of theAmerican Geriatrics Society 1990; 38: 197-200.

34 Bialecki L, Woodward RS. Predicting death afterCPR: experience at a non-teaching communityhospital with a full-time critical care staff. Chest 1995;108: 1009-17.

35 Kentsch M, Stendel M, Berkel H, Bueller-Esch G.Early prediction of prognosis in out-of-hospital cardiacarrest. Intensive Care Medicine 1990; 16: 378-83.

36 Martens PR, Mullie A, Buylaert W, Calle P, vanHoeyweghen R, and the Belgian Cerebral ResuscitationGroup. Early prediction of non-survival for patientssuffering cardiac arrest - a word of caution. In1ten1siveCare Medicine 1992; 18: 11-14.

37 Gove PB and the Merriam-Webster editorial staff.Webster's third new international dictionary. Springfield,MA: G and C Merriam Co, 1976.

38 Tomlinson T, Brody H. Futility and the ethics of resus-citation. Journal of the Anerican Medical Association1990; 264: 1276-80.

39 Schneiderman LJ, Jecker NS. Wrong miedicinle: doctors,patients and futile treatmienzt. Baltimore: The JohnsHopkins University Press, 1995.

40 Howard PK. The death of comnmoni senise: how law is suffo-catingAmerica. New York: Random House, 1994.

41 McCormick RA. To save or let die. Journal of theAnmerican Medical Association 1974; 229: 172-6.

42 Rawls J. A theory of justice. Cambridge, MA: TheBelknap Press of Harvard University Press, 1971.

43 Mill JS. Utilitarianism. Oni libertv anid utilitarianiismiz.New York: Bantam Books, 1993.

44 Engelhardt Jr HT. The founidationis of bioethics [2d ed].New York: Oxford, 1996.

45 Morgan R, King D, Prajapati C, Rowe J. Views of elderlypatients and their relatives on cardiopulmonary resuscita-tion. British Medical Journal 1994; 308: 1677-8.

46 Meystre CJN, Ahmedzai S, Burley NMJ. Terminally illpatients may want to live. British Medical Journal 1994;309: 409.

47 Peet J. Survey of health care. The Econonmist 1991July 6.

48 Simpson WF. Comparative longevity in a collegecohort of Christian Scientists. Journal of the AnmericanMedical Association1 1989; 262: 1657-8.

49 Drucker PF. A century of social transformation.AManaginlg in a tinie ofgreat chanige. New York: TrumanTalley Books/Dutton, 1995: ch 21.

50 McGinnis JM, Foege WH. Actual causes of death inthe United States. Yournal of the Anmericani MedicalAssociation 1993; 270: 2207-12.

51 Hurowitz JC. Toward a social policy for health. NewEngland Journal of Medicinle 1993; 329: 130-3.

52 Kiple KF, ed. The Canmbridge world history of hunmandisease. Cambridge: Cambridge University Press, 1993.

53 Bergthold LA. Medical necessity: do we need it? HealthAffairs 1995; 14: 180-90.

54 Reiser SJ. Criteria for standard versus experimentaltherapy. Health Affairs 1994; 13: 127-36.

55 Mandelblatt JS, Fahs MC. The cost-effectiveness ofcervical cancer screening for low-income elderlywomen. Journial of the Amiericani Medical Association1988; 259: 2409-13.

56 Horstkotte D. Quality of life after heart valve replace-ment. In: Walter PJ, ed. Quality of life after open heartsurgery. Dordrecht: Kluwer, 1992:101-4.

57 Weinstein M, Stason W. Cost-effectiveness of interven-tions to prevent or treat coronary heart disease. AnnpiualReview of Public Health 1985; 6: 41-3.

58 Edelson JT, Weinstein MC, Tosteson ANA, et al. Long-term cost-effectiveness of various initial monotherapiesfor mild to moderate hypertension. Journial of theAnerican Medical Associationi 1990; 263: 408-13.

59 Califf RM, Harrell FE, Lee KL, et al. The evolution ofmedical and surgical therapy for coronary arterydisease, a 15-year perspective. Journal of the AmiericanMedicalASsociation 1989; 261: 2077-86.

60 Levinsky NG. The organization of medical care:lessons from the Medicare End Stage Renal DiseaseProgram. New Eniglanid Jolunal of Medicine 1993; 329:1395-9.

61 Kinosian BP, Eisenberg JM. Cutting into cholesterol.J7olunal of the Americani Medical Associationz 1988; 259:2249-54.

62 Wachter RM, Luce JM, Safrin S, et al. Costs andoutcome of intensive care for patients with AIDS,pneumocystis carinii pneumonia, and severe respira-tory failure. Yournial of the Anerican Medical Association1995; 273: 230-5.

63 Schapira DV, Studnicki J, Bradham DD, et al.Intensive care, survival and expense of treating criticallyill cancer patients. Jouirnial of the American MedicalAssociationi 1993; 269: 783-6.

64 Fineberg HV, Scadden D, Goldman L. Care of patientswith a low probability of acute myocardial infarction.New England Journal of Medicine 1984; 310: 1301-7.

65 Murphy DJ, Matchar DB. Life-sustaining therapy: amodel for appropriate use. Jouirn1al of the AmericanMedical Associationi 1990; 264: 2103-8.

copyright. on S

eptember 4, 2020 by guest. P

rotected byhttp://jm

e.bmj.com

/J M

ed Ethics: first published as 10.1136/jm

e.23.6.361 on 1 Decem

ber 1997. Dow

nloaded from