Journal ofMedical Ethics 1998;24:308-313

The role of futility judgments in improperlylimiting the scope of clinical researchWilliam Harper The University ofAlabama, USA

AbstractIn medical research, the gathering and presenting ofdata can be limited in accordance with the futilityjudgments of the researchers. In that case, researchresults falling below the threshold of what theresearchers deem beneficial would not to be reportedin detail. As a result, the reported information wouldtend to be useful only to those who share thevaluational assumptions of the researchers. Shouldthis practice become entrenched, it would reducepublic confidence in the medical establishment,aggravate factionalism within the researchcommunity, and unduly influence treatment decisions.I suggest alternative frameworks for measuringsurvival outcomes.(7ournal ofMedical Ethics 1998;24:308-313)Keywords: Clinical research; consumer; futility; medicalethics; outcome measure

The notion of medical futility has received a greatdeal of attention in the literature. Usually, the dis-cussion centres around the question of attendingphysicians withholding treatments deemed bythem to be futile (ie, deemed by them not to con-stitute a medical benefit to the patient). It seemsto have gone unnoticed, however, that medicalresearchers are in a position to act on futilityjudgments when deciding whether certain dataare worth recording and reporting. A researchercould reason that there is little advantage in usingvaluable time and space to report in detailtreatment outcomes which do not benefit patients.It would then seem to make more sense to lumpall non-beneficial outcomes into a single reportingcategory, and to focus instead on differentiatingthe various benefits, if any, likely to be afforded bya treatment. Whether any researchers haveactually followed such a line of reasoning is diffi-cult to say, but it can be said that some research-ers in fact do not report in detail informationabout treatment outcomes which they deem to bewithout benefit to the patient. Lantos et al, studiedthe effectiveness of cardiopulmonary resuscitation(CPR) on low birthweight babies. The authorsdetermined that none of the 38 babies receiving

CPR in the first 72 hours after birth survived tohospital discharge.' They judged that CPR was ofno benefit to the subjects of the study, andconcluded that CPR in infants of very low birth-weight is "virtually futile".2 Throughout thepaper, Lantos et al, construe terms such as"success", "benefit", "futile", "effectiveness","validated", and "proved" in terms of survival tohospital discharge. Even terms such as "survival"and "mortality" are employed by the authorsmerely as shorthand for "survival to hospital dis-charge" or "dying before discharge". The abstractsays simply that "none of the babies who receivedCPR in the first three days of life survived".3 Lan-tos et al report virtually nothing of the treatmentresults that fell short of survival to hospitaldischarge. This omission makes sense given theparameters they demarcate for medical futility.

Rubenfeld et al seek to develop guidelines forthe withdrawal of life-saving mechanicalventilation.4 The authors write: "It would be moreclinically useful to define a subgroup of patientswhose survival is so low that reasonable physiciansand patients would agree that intensive care canno longer effectively fulfil the goals oftransplantation".5 They state that life supportshould be withdrawn from such patients.6 Ruben-feld et al make clear that they regard survival shortof hospital discharge as not beneficial to patients,and as unduly burdensome on family, caregiversand society.7 Alternative assessments are brushedaside as not rational, as limited by "arbitrary fac-tors and cognitive biases".6 The authors accord-ingly define "surviving" patients as "those whowere alive 30 days after extubation and who weredischarged from the hospital".8 The only datumreported for survival short of "survival" is the totalnumber of patients failing to "survive". The judg-ment by Rubenfeld et al that a life span short ofhospital discharge is not worth living logicallysupports their evident judgment that informationabout the length of such survival is not worthreporting.

Several recent research reports make the futilityjudgment that outcomes short of "survival" do notbenefit the patient.9'-4 No information is given of

copyright. on January 30, 2022 by guest. P

rotected byhttp://jm

e.bmj.com

/J M

ed Ethics: first published as 10.1136/jm

e.24.5.308 on 1 October 1998. D

ownloaded from

Harper 309

shorter survival intervals. As with Lantos et al andRubenfeld et al such information would be primafacie pointless given the judgment that such resultsare of no medical benefit. It makes more sense toreport only the total of such results, under a head-ing such as "failed to survive".The Utstein Consensus Conference recom-

mends that responsible personnel record thefollowing data: date and time of death in hospital,with exact time for patients who die withintwenty-four hours; the Overall Performance Cat-egory and Cerebral Performance Category of theGlasgow-Pittsburgh Outcome Categories at timeof discharge and at one year, with the best scorethe week before death recorded for those dyingwithin one year, and the best outcome achievedduring hospitalisation and in the year after thearrest.' -'" However, with respect to admittedpatients dying in hospital, the conference recom-mends for researchers only that they "should tabu-late the number of patients who die in the hospi-tal, with special notation of patients who diewithin the first 24 hours of admission".'8 Thiswould make sense if one regarded the moredetailed information on survival short of dis-charge as not important enough to include inresearch reports.

I want to emphasise that none of the authorscited offers any justification for the choice of ameasure of survival rates. My point here is thatfutility judgments potentially play an importantrole in the choice of treatment outcome measuressince no one would choose to measure uselessinformation. The choice of a treatment outcomemeasure is important since the measure filters theinformation to be reported. But, I will argue, i. itis wrong at present to limit the reporting ofsurvival rates in accordance with the researcher'sfutility judgments, and ii. an alternative method ofreporting survival outcomes should be used.The ability to limit the scope of research

according to the researchers' futility judgmentsgives medical researchers the ability to shapeinformation in the image of their personaljudgments about when life is worth living. Treat-ment results falling below the researcher's mini-mum acceptable treatment outcome (MATO) canbe lumped together and reported as a total undersome heading equating to "failure to achievemeaningful results". Patients with treatmentsurvival goals falling below the measures ofsurvivability used by researchers are denied themore specific information they would need tomake an informed judgment, according to theirown values, about whether to attempt thetreatment in question. From the report, theycould know only that in a certain number of cases

the treatment failed to achieve a given result. Theycould not know the extent, if any, to which itachieved lesser results they might deem accept-able.One response might be that such patients are

simply misguided in their choice of treatmentgoals, and that we are saving patients grief byrefusing to give treatments that will not truly ben-efit them. This paper is not the place to try andsettle the medical futility debate. It would seem,though, given the extremely personal nature ofend-of-life decisions, that the burden of proofwould lie on those who would allow health careprofessionals unilaterally to act in accordance withfutility judgments. To my knowledge, no argu-ments have been offered to justify the practice oflimiting the reporting of research results inaccordance with futility judgments. Argumentshave been made in favour of allowing attendingphysicians unilaterally to withdraw treatmentsjudged to be futile, but even these arguments havebeen less than convincing to many authors.'9-27Until a clearly adequate justification of thepractice has been advanced, researchers shouldrefrain from demarcating the limits of reportedinformation in accordance with their own futilityjudgments.

Default positionThe adoption of survival to hospital discharge as aMATO seems to have become a default positionin the futility debate. "In the literature, in-hospitaldeath seems to have become the standardaccount, for the data usually offered to supportclaims of imminent demise futility are data aboutsurvival to discharge".28 It is difficult to assess theextent to which futility judgments have influencedthe reporting of information in recent research.This is because very few authors offer any justifi-cation for their selection of an outcome measure.What we know is that many authors, includingsome researchers, have argued explicitly for thefutility of outcomes short of survival to discharge,and that at present much of the research literatureuses survival to discharge as the primary measureof survivability, often the only measure.

Eisenberg et al in their 1990 review of 74 stud-ies of survival rates, report that 51 of the 74 arti-cles (69%) define "survival" as "discharged fromthe hospital".29 I conducted a search ofMEDLINE using the search strings "survival rate+ CPR", "survival rate + burn", and "survival rate+ ventilation". The searches using "survival rate +CPR" and "survival rate + burn" went back oneyear and generated eleven and fourteen citationsrespectively. The search using "survival rate +ventilation" returned many more entries, so I went

copyright. on January 30, 2022 by guest. P

rotected byhttp://jm

e.bmj.com

/J M

ed Ethics: first published as 10.1136/jm

e.24.5.308 on 1 October 1998. D

ownloaded from

310 The role offutility judgments in improperly limiting the scope of clinical research

back only six months, generating eleven citations,for a total of thirty-six citations. Of the total cita-tions, three did not have survival as a focus, onewas untranslated from the Chinese, and four werenot accessible. The twenty-eight remaining arti-cles break down as follows:

A. Seven explicitly define survival as survival todischarge.30"36 Five of these give no furthersurvival data. Two note whether non-survivorswere admitted alive after receiving CPR fromemergency medical treatment (EMT) units;one of those also tracks survivors at one yearafter discharge.

B. Seven did not explicitly define survival, butclearly are equating it with survival todischarge.9 '1 3741 Five of these give no furthersurvival data. Two track survivors at one yearafter discharge.

C. Seven studies simply offer no clear indicationof the measurement used for survival.'1-14 42-44

D. Six directly report a broad range of detaileddata on the survival times of patients.4550

E. One study adopts a survival measure of 180days, giving no additional information onsurvivability.5"

Thus, of a total of twenty-eight studies, fourteen(50%) equate "survival" with "survival to dis-charge". That the authors in group B would usethe crucial term "survival" without defining it,relying on context and the reader's backgroundknowledge to make clear the assumed parametersof survival, is testimony to the widespread accept-ance of survival to discharge as the propermeasure of survival rates. Of the fourteen papersin groups A and B, thirteen (six from group A andseven from group B) give no information at all ontreatment results less than survival to discharge.Of the total papers, only seven (one from group Aand the six in group D; twenty-nine per cent of thetotal) give clear, systematic information aboutsurvivability short of discharge.The above is not intended to constitute a statis-

tically valid survey of the literature. Rather, itpoints out that survival to hospital discharge is avery commonly used outcome measure forsurvivability, often the sole measure. When thelatter, survival to discharge filters information,intentionally or not, in the same way as would anexplicit MATO. This motivates a discussion as towhether futility judgments can justify the practiceof limiting the reporting of treatment outcomes inaccordance with any particular survival measure,especially survival to discharge.There are several reasons for thinking that futil-

ity judgments should not be used to limit the

reporting of research results. At the social level,research that accommodates only a narrowvaluational attitude alienates all those members ofsociety who do not share that valuational attitude.This will do nothing to instil confidence that themedical profession shares, or at least respects, thepublic's interests. Medical research, like theprofession of medicine itself, is a social enterprise,dependent to some extent on the good will of thepublic. Whether from self interest, or because theresearcher values social harmony and tolerance,the researcher should be willing to make room fora broad range ofviews concerning the value of life.Note that the practice of limiting the reporting ofsurvivability to a single outcome measure such assurvival to discharge also denies relevant infor-mation to those patients who would accept treat-ment only if it meant survival in terms muchlonger than the measure used by the researcher.(Only five papers of the twenty-eight cited above(18% of the total) discuss survival at one year.)Thus, the use by researchers of survival todischarge as the sole measure of survivability willdeny patients with treatment goals on either sideof that valuational point information relevant totheir making informed treatment decisions conso-nant with their own values. This problem is a gen-eral one that applies to any single treatmentoutcome measure that might be chosen by aresearcher. Whatever measure the researcherchooses, patients whose own MATOs fall above orbelow that measure can find themselves lackinginformation crucial to making informed treatmentdecisions.

Cultural warfareThe use of MATOs in reporting survival out-comes threatens further to import cultural warfareinto the research arena. At the administrativelevel, ever-present limitations on funding forceadministrators to choose which researchers re-ceive funding. Why would a responsible adminis-trator of funds choose to "waste" money onresearchers who use measures that do not track"meaningful outcomes", or, on the other hand,who set "unreasonably high standards" of out-comes? Administrators of funds will feel morallyobligated to freeze out researchers whose studiesexemplify the "wrong" values. The divisive effectsof such exclusionism would not be helpful to theprofession, but would be practically unavoidableonce the research community got into the game ofmaking narrow determinations of the value of life.When the information needed for making prac-

tical ethical decisions is shaped according tocertain categories and measurements there is atendency uncritically to adopt those categories

copyright. on January 30, 2022 by guest. P

rotected byhttp://jm

e.bmj.com

/J M

ed Ethics: first published as 10.1136/jm

e.24.5.308 on 1 October 1998. D

ownloaded from

Harper 311

and measurements. Regardless of whether re-searchers intend this result, by omitting infor-mation about the effects of treatment short of sur-vival to hospital discharge, family members aresteered away from considering the worth oftreatments that could lengthen the patient's lifeshort of survival to discharge. Some researchersmight see this as a positive result, but whether suchinfluence is desirable is, I would suggest, a matter tobe publicly aired and debated before such a meansof influence is used by the research community.

Physician decisions, and even hospital policies,concerning specific treatments will be impactedby the practice of using a single measure ofsurvivability. Physicians and hospitals cannot ingood faith offer treatments for which there is noevidence of effectiveness, and survival effectsother than those measured by the researcher willnot be reported. When the data have been filteredby the use of a MATO in accordance with aresearcher's futility judgment, the question oftreatment appropriateness will have been begged.

Two frameworksHow should data be reported? Lacking soundempirical data of the distribution ofMATOs heldby patients, I can give only sketches of two frame-works that seem appropriate in light of theconsiderations discussed in this paper. A goodframework for recording treatment outcomes willbe one that is feasible for researchers, does notunduly draw resources from other valuableprojects, and provides information important tothe consumers of the report. No precise, finalanswer is possible as to the optimum way tobalance these legitimate, competing interests. Aquick solution would be to make records ofsurvival after initiation of treatment at hour one,day one, week one, month one, year one, and yearfive. While somewhat arbitrary, and grounded inconventional divisions of time, I believe that theabove schedule would give most patients outcomeinformation adequate to make treatment deci-sions in light of their personal values. This sched-ule would be modified when better informationwas available of the actual distribution ofMATOsheld by patients.

Better than such a schedule would be the use ofa graph plotting all survival results, or a Kaplan-Meier curve showing the probabilities of survival.The latter is used by Steiner et al who present prob-abilities of survival in terms of days for up to 60days after admission, and in terms ofmonths for upto 12 months after admission.52 Such graphs seemto be a good way to provide accurate, dense, butunderstandable survivability information to pa-tients. An added plus is that the provided

information seems dense enough largely to forgothe need for empirical research into the distributionofMATOs held by the patient population.

It may be objected that researchers cannotanticipate individual patient situations, and in factshould conduct their research without referenceto possible patient idiosyncracies. This objectionis misleading. I am not asking that research not beconducted in terms of statistical generalisations,nor am I asking that research be tailored for eachindividual patient. I am asking that researcherscollect and report a broad range of data pertainingto survivability and not rely on a single outcomemeasure. It should be pointed out that even thelevel of reporting I have suggested may omit thesurvivability information needed by certain pa-tients holding unusual MATOs. A patient holdinga MATO of survival to ten years would requirethat treatment outcome studies last at least tenyears in order to provide the desired information.However, the fact that not every logically possibleMATO can be accommodated in practice is not ajustification for adopting the other extreme oflimiting the reporting of information to a single,discrete measure of survivability.Note that there is no need to limit the freedom

of researchers or journal editors to voice anymanner of value judgment about the appropriate-ness of a treatment. For example, the editor ofResuscitation makes a comment at the end of onepaper, voicing concern over the possible inappro-priateness of a treatment for certain subgroups ofpatients.53 The comment does not in any wayobscure the reported data, and, however influen-tial the comment may be, it does not impede theindependent judgment of patients and other con-sumers of the report. Likewise, valuationalstatements by authors within the body of theirown texts need not interfere with the reporting ofdata nor impede the decision making of patients.What is objectionable is for researchers to act onthose value judgments by suppressing data impor-tant to independent decision making by patientsand other consumers of the reported information.

In the interest ofsimplicity and brevity, I have notto this point discussed quality of life issues. Formany patients, the quality of life associated withsurvival, and not just survival itself, will be animportant factor in the decision of whether toaccept treatment. "Quality of life" has manifolddimensions. Because of this, it admits of muchgreater variation among patients, and is a muchmore complex issue than simple survival. TheUtstein Consensus Conference, cited above,'5 sets agood minimum standard for the recording ofperformance outcomes. However, I recommendthat researchers report this information in tandem

copyright. on January 30, 2022 by guest. P

rotected byhttp://jm

e.bmj.com

/J M

ed Ethics: first published as 10.1136/jm

e.24.5.308 on 1 October 1998. D

ownloaded from

312 The role offutility judgments in improperly limiting the scope of clinical research

with a more adequate reporting of survivabilityresults, so that the survival and performanceoutcomes for all patients are reported. This will giveconsumers ofthe report a better picture of the likelyoutcome of a given treatment option, and shouldhelp each patient to make an informed decision asto how best to face the future, however abbreviatedthat future may be.

William Harper, PhD, is Adjunct Professor of Philos-ophy at the University ofAlabama, where he teachesmedical ethics.

References1 Lantos JD, Miles SH, Silverstein MD, Stocking CB. Survival

after cardiopulmonary resuscitation in babies of very low birth-weight: is cpr futile therapy? The New England 3ournal ofMedi-cine 1988; 318: 93.

2 See reference 1: 94.3 See reference 1: 91.4 Rubenfeld GD, Crawford SW Withdrawing life support from

mechanically ventilated recipients of bone marrow transplants:a case for evidence-based guidelines. Annals ofInternal Medicinle1996; 125: 625-33.

5 See reference 4: 626.6 See reference 4: 632.7 See reference 4: 625.8 See reference 4: 625 (emphasis added).9 Covington DS, Wainwright DJ, Parks DH. Prognostic indica-

tors in the elderly patient with burns. Journal of Burn Care &Rehabilitationi 1996; 17: 222-30.

10 Hollmig KA, Soehngen D, Leschke M, Kobbe G, Schneider P,Klein RM et al. Long-term survival of recipients of allogeneicbone-marrow transplantation after mechanical ventilation.European J7ournal ofMedical Research 1997; 2: 62-66.

11 Brown DL, Archer SB, Greenhalgh DG, Washam MA, JamesLE, Warden GD. Inhalation injury severity scoring system: aquantitative method. 3rournal of Burn Care & Rehabilitation1996; 17: 552-7.

12 Caldwell Jr FT, Wallace BH, Cone JB. Sequential excision andgrafting of the burn injuries of 1507 patients treated between1967 and 1986: end results and the determinants of death.Jfournal ofBurn Care & Rehabilitation 1996; 17: 137-46.

13 Frye KE, Izenberg SD, Williams MD, Luterman A. Simulatedbiologic intelligence used to predict length of stay and survivalof burns. Journal ofBurn Care & Rehabilitationz 1996; 17: 540-6.

14 Grobmyer SR, Maniscalco SP, Purdue GF, Hunt JL. Alcohol,drug intoxication, or both at the time of burn injury as apredictor of complications and mortality in hospitalizedpatients with burns. J7ournal of Burni Care & Rehabilitation1996; 17: 532-9.

15 Cummins RO, Chamberlain DA, Abramson NS, Allen M, Bas-kett PJ, Becker L et al. Recommended guidelines for uniformreporting of data from out-of-hospital cardiac arrest: the Utsteinstyle. A statement for health professionals from a task force of theAmerican Heart Association, the European Resuscitation Coun-cil, the Heart and Stroke Foundation of Canada, and theAustralian Resuscitation Council. Circulation 1991; 84: 970.

16 Jennett B, Bond M. Assessment of outcome after severe braindamage: a practical scale. Lancet 1975; 1: 480-4.

17 Brain Resuscitation Clinical Trial I Study Group. A rand-omized clinical study of cardiopulmonary-cerebralresuscitation: design, methods, and patient characteristics.Anmerican J3ournal of Enmergency Medicine 1986; 4: 72-86.

18 See reference 15: 967.19 Brody BA, Halevy A. Is futility a futile concept? The Journlal of

Medicine and Philosophy 1995; 20: 123-44.20 Griener GG. The physician's authority to withhold futile treat-

ment. The Journal ofMedicine and Philosophy 1995; 20: 207-24.21 Harper W. Judging who should live: Schneiderman and Jecker

on the duty not to treat. J7ournal of Medicine and Philosophy(forthcoming).

22 Kopelman LM. Conceptual and moral disputes about futileand useful treatments. The Jtournal of Medicine and Philosophy1995; 20: 109-21.

23 Lantos JD, Linger PA, Walker RM, Gramelspacher GP,Shapiro GR, Sanchez-Gonzales MA et al. The illusion of futil-ity in clinical practice. The American Journal of Medicine 1989;87: 81-4.

24 Teno JM, Nelson HL, Lynn J. Advance care planning: prioritiesfor ethical and empirical research. Hastings Center Report 1994;24: s32-s36.

25 Tong R. Towards a just, courageous, and honest resolution ofthe futility debate. The Journal ofMedicine and Philosophy 1995;20: 165-89.

26 Truog RD, Brett AS, Frader J. The problem with futility. TheNew England 3ournal ofMedicine 1992; 325: 1560-4.

27 Sawacki BE. The "futility debate" and the management ofgordian knots. The Journal of Clinical Ethics 1995; 6: 112-27.

28 See reference 19: 137.29 Eisenberg MS, Cummins RO, Damon S, Larsen MP, Hearne

TR. Survival rates from out-of-hospital cardiac arrest: recom-mendations for uniform definitions and data to report. AnnalsofEmergency Medicine 1990; 19: 1250.

30 Brigham PA, McLoughlin E. Burn incidence and medical careuse in the United States: estimates, trends, and data sources.Jfournal ofBurn Care & Rehabilitation 1996; 17: 95-107.

31 Chan DKL, Ho LY, Joseph VT. Mortality among infants withhigh-risk congenital diaphragmatic hernia in Singapore.Journal of Pediatric Surgery 1997; 32: 95-8.

32 Herlitz J, Ekstrom L, Wennerblom B, Axelsson A, Bang A,Lindkvist J et al. Lidocaine in out-of-hospital ventricular fibril-lation. Does it improve survival? Resuscitation 1997; 33:199-205.

33 Murphy DJ, Hope PL, Johnson A. Neonatal risk factors forcerebral palsy in very preterm babies: case-control study. Brit-ish Medical Journal 1997; 314: 404-8.

34 Pranikoff T, Hirschl RB, Steimle CN, Anderson III HL,Bartlett RH. Mortality is directly related to the duration ofmechanical ventilation before the initiation of extracorporeallife support for severe respiratory failure. Critical Care Medicine1997; 25: 28-32.

35 Suominen P, Korpela R, Kuisma M, Silfvast T, Olkkola KT.Paediatric cardiac arrest and resuscitation provided byphysician-staffed emergency care units. Acta AnaesthesiologicaScandinavica 1997; 41: 260-5.

36 Varon J, Fromm Jr RE. In-hospital resuscitation among theelderly: substantial survival to hospital discharge. AmericanJournal of Emergency Medicine 1996; 14: 130-2.

37 Callaham M, Madsen CD. Relationship of timeliness ofparamedic advanced life support interventions to outcome inout-of-hospital cardiac arrest treated by first responders withdefibrillators. Annals of Emergency Medicine 1996; 27: 638-48.

38 Fischer M, Fischer NJ, Schulttler J. One-year survival after out-of-hospital cardiac arrest in Bonn city: outcome report accord-ing to the "Utstein style". Resuscitation 1997; 33: 233-43.

39 Herlitz J, Bang A, Holmberg M, Axelsson A, Lindkvist J, Hol-mberg S. Rhythm changes during resuscitation from ventricu-lar fibrillation in relation to delay until defibrillation, number ofshocks delivered and survival. Resuscitation 1997; 34: 17-22.

40 Lee KH, Angus DC, Abramson NS. Cardiopulmonaryresuscitation: what cost to cheat death? Critical Care Medicine1996; 24: 2046-52.

41 Weston CFM, Wilson RJ, Jones SD. Predicting survival fromout-of-hospital cardiac arrest: a multivariate analysis. Resuscita-tion 1997; 34: 27-34.

42 Arensman RM, Statter MB, Bastawrous AL, Madonna MB.Modern treatment modalities for neonatal and pediatric respi-ratory failure. The American Jfournzal of Surgery 1996; 172: 41-7.

43 Dunagan DP, Baker AM, Hurd DD, Haponik EF. Broncho-scopic evaluation of pulmonary infiltrates following bone mar-row transplantation. Chest 1997; 111: 135-41.

44 Nuinez-Gutierrez H, Castro-Muniozledo F, Kuri-Harcuch W.Combined use of allograft and autograft epidermal cultures intherapy of burns. Plastic and Reconstructive Surgery 1996; 98:929-39.

45 Bavaria JE, Kotloff R, Palevsky H, Rosengard B, Robects JR,Wahl PM et al. Bilateral versus single lung transplantation forchronic obstructive pulmonary disease. The Journal of Thoracicand Cardiovascular Surgery 1997; 113: 520-8.

46 Headley AS, Tolley E, Meduri GU. Infections and the inflam-matory response in acute respiratory distress syndrome. Chest1997; 111: 1306-21.

47 Herlitz J, Ekstrom L, Axelsson A, Bang A, Wennerblom B,Waagstein L et al. Continuation of CPR on admission toemergency department after out-of-hospital cardiac arrest.Occurrence, characteristics and outcome. Resuscitation 1997; 33:223-31.

48 Plaisance P, Adnet F, Vicaut E, Hennequin B, Magne P, Prud-homme C et al. Benefit of active compression-decompression

copyright. on January 30, 2022 by guest. P

rotected byhttp://jm

e.bmj.com

/J M

ed Ethics: first published as 10.1136/jm

e.24.5.308 on 1 October 1998. D

ownloaded from

Harper 313

cardiopulmonary resuscitation as a prehospital advancedcardiac life support: a randomized multicenter study. Circula-tion 1997; 95: 955-61.

49 Sarnaik AP, Meert KL, Pappas MD, Simpson PM, Lieh-LaiMW, Heidemann SM. Predicting outcome in children withsevere acute respiratory failure treated with high-frequencyventilation. Critical Care Medicine 1996; 24: 1396-1402.

50 Schiller WR, Bay RC, Garren RL, Parker I, Sagraves SG.Hyperdynamic resuscitation improves survival in patients with

life-threatening burns. Journal of Burn Care & Rehabilitation1997; 18: 10-16.

51 Steiner T, Mendoza G, De Georgia M, Schellinger P, Holle R,Hacke W. Prognosis of stroke patients requiring mechanicalventilation in a neurological critical care unit. Stroke 1997; 28:711-715.

52 See reference 51: 713.53 See reference 47: 231.

Omentum-to-Brain Transplants*

Beyond your simplistic jokes I can feel new blood bounceinto the cingulate, untangle the Alzheimer twos, and swish away the hibernation debris.Lendrum's stain will fail as morenuclei move to the center, reduce the umbras,and return my social security numbers.Glopy meals draw blood once aimed at the gut, butnow up. Whirl past the heart, the jaw, the ear, andplunge through the dura to the thirsty bouton belt,releasing Eleanor Roosevelt.Each pulse attacks my old Wernicke's,"Now where'd I put the car peas?"

*Goldsmith HS, et aL Regional cerebral blood flow after omental transposition. Acta Neurochirurgica1990; 106:45-52.

Edward V Spudis, Neurologist

Editor's noteReaders who wish to have a few additional clues to this poem should turn to page 319.

copyright. on January 30, 2022 by guest. P

rotected byhttp://jm

e.bmj.com

/J M

ed Ethics: first published as 10.1136/jm

e.24.5.308 on 1 October 1998. D

ownloaded from