Him ORIGINAL CONTRIBUTION Neonatal End-of-Life Decision Making Physicians’ Attitudes and Relationship With Self-reported Practices in 10 European Countries io de Vo Monique Kaminski, MSe Louis AA. Koliée, MD. Audranas Kueinskas, MD Margaret Reid, PhD Rodolfo Saracei, MD ior the BURONIC dy Gi VANCES IN PERINATAL MEDI- cine have dramatically im- proved neonatal survival in ev ery industrialized country.” However, results concerning long-term. residual morbidity have not been equally satisfactory, particularly among the most premature newborns.°" As the number fof very preterm multiple births in- creases because of assisted reproduc- tive techniques,’ the indiscriminate ap- plication of invasive life-sustaining treatment to every infant irrespective of long-term prognosts is increasingly ques: tioned." However, no consensus exists on which patients might be candidates for palliative care rather than for intensive care of on the criteria on which such choices might he based. AL one ex- (©2000 American Medical Association. AI rights reserved. (Reprinted) AMA, November 15,2000 Context The ethical issues surrounding end-of-life decision making for infants with adverse prognoses are controversial. Little empirical evidence is available on the atitudes and values that underie such decisions in different counties and cul- tures. Objective To explore the variability of neonatal physicians’ attitudes among 10 Eu- ropean countries and the relationship between such atitudes and self-reported prac- tice of end-of life decisions. Design and Setting Survey conducted during 1996-1997 in 10 European coun- tries (France, Germany, Italy, the Netherlands, Spain, Sweden, the United Kingdom, Estonia, Hungary, and Lithuania) Participants. total of 1391 physicians (response rate, 89%) regularly employed in 1142 neonatal intensive cae units (NICUS). ‘Main Outcome Measures Scores on an attitude scale, which measured views re- garding absolute value of ite (score of 0) vs value of quality of fe (score of 10); self- Feport of having ever set limits to intensive neonatal interventions in cases of poor neurological prognosis. Results Physicians more likely to agree with statements consistent with preserving life at any cost were from Hungary (mean atitude scores, 5.2 [95% confidence inter= val (Ci), 49-5.5), Estonia (4.9 [95% Cl, 43-5.5), Lithuania (55 [95% Cl, 48-6.1), and Italy (5.7 {95% Cl, 5.3-6.0)), while physicians more likely to agree with the idea that quality of lfe must be taken into account were from the United Kingdom (ati- tude scores, 7.4 {95% Cl, 7.1-7,7), the Netherlands (7.3 [95% Cl, 7.1-7.5)), and Sweden (6.8 [95% Cl, 64-7 3). Other factors associated with having a pro-quality- oflife view were being female, having had no children, being Protestant or having no religious background, considering religion as not important, and working in an NICU with a high number of very low-birth-weight newborns. Physicians with scores reflecting a more quality-of-life view were more likely to report that in their practice, they had set limits to intensive interventions in cases of poor neurological prognosis, with an adjusted odds ratio of 1.5 (95% Cl, 1.3-1.7) per unit change in attitude Conclusions In our study, physicians’ ikelhood of reporting setting limits to inten sive neonatal interventions in cases of poor neurological prognosis is related to their attitudes. After adjusting for potential confounders, country remained the most im- portant predictor of physicians’ attitudes and practices. ‘AMA, 20002862451-2459 vw amacom treme, vitalists support theidea anal Alto fon and Menkes ofthe EURON listssupport the idea ofanab- Seren are sted atte end of ths ace solute intrinsic value of human life (the Gyafaponfing huthor and Reps: Rebas- so-called sanctity-of- life position) and re- Isto, MD, Deparment of ube Heh, Miguel Het a Tite, tater Uivesty, Campus San an Cha Vans Jectany form of discontinuation of (dm 87), Aleant, Span 03850 (e-mall-rebagi sustaining treatment except for cases of Guth). No.9 2454 ‘Downloaded From: http:/jamanetwork.com/pdfaccess.ashx?url=/data/journals/jama/4760 by a Universita Torino User on 03/13/2017NEONATAL END-OF-LIFE DECISION MAKING Imminent death In contrast, others be- lieve that the value of life is related to certain present or future capacities (such 4s, ata minimum having sel-conscious- ness, the ability to establish a relation ship with other human beings, and the capacity to derive some pleastre from existence), which define its quality, and thence a physician's duty to sustain i" A numberof intermediate positions that may be identified between these 2 ex tremes have been the source of ongo- ing discussions among ethicists, legal experts, and policymakers."* However, litle empirical evidence is available on whether and how relevant these con- cepts are to those actually involved in making decisions: the parents and, e pecially, the physicians. The extent 10 ‘which physicians’ personal values and attitudes are assoctated with their prac- lice of end-oF life decision making re- mains unknown, A recent European study, EURONIC (European Project on Parents’ Informa- lion and Ethical Decision Making in Nes natal Intensive Care Units: Stall Ate tudes and Opinions) carried out on a large representative sample of neonatal Intensive care units (NICUs)" in se\ tral countries has shown that both the Frequency of physicians involvement in cthical decision making and the type of choices made varied across countries. Practices such as continuation of cur rent treatment without intensifying tad withholding of emergency interven- lions appear widespread, In contrast, the Frequency of physicians reporting with- drawal of mechanical ventilation was [highest in the Northern European coun- triesand lowest inthe South Mediterr such as Haly and Spain. This study, based on thesame sample of respondents in the EURONIC project."* explores physicians'attitudes and values underlying their medical behaviors. The specific objectives were 2) to describe physicians’ auitudes toward value of life and life with dis- ability, appropriate use of medical tech- nology, and relevance of factorssuch as family burden, economic costs, and legal constraints to decision making; (2) to estimate the influence of physicians 2452 JAMA. November 15,2000 ol254, No 19 Reprint) country oforiginand other personal and professional characteristics on their atli- tudes; and (3) to assess the relationship between attitudesand self-reported prac- Lice of end-of life decisions. METHODS: Sample Theseope and design ofthe project have been described elsewhere.” The sample includes the 7 Western European coun- tries of Prance, Germany, Italy, the Neth- erlands, Spain, Sweden, and the United Kingdom, and the 3 Central and East- em European countries of Estonia, Hun- gary, and Lithuania. We also surveyed Luxembourg, but because it only had 1 eligible NICU, anonymity in reporting was precluded and these results are not presented herein. In each country, we fdeniifiedall the NICUssatistying 4 pre- defined eligibility eriteria: routine are of very low-birth-weight neonates (1500 g), with at least 20 admissions per year; possibility of prolonged me- chanical ventilation; pediatrician or neo- natologist (in Sweden, a nurse neona- tologist also would qualify) on duty in the hospital on a 24-hour basis; and no transfer of patients to other units for medical reasons, In the Netherlands, Sweden, Hun- gary, Estonia, and Lithuania, all the eli- sible units were asked to participate. In France, Spain, aly, Germany, and the United Kingdom, a random sample of NICUs was selected after stratification by geographical area. In Italy NICUs also were stratified by the number of Intensive care beds (<5, =5), and in Germany by university affiliation." At the time of the study, all part- lime and full-time physicians regu larly paid by the hospital to earry out clinical activities (“employed”) in the selected units were asked to partic spate Questionnaires Data collection was performed during 1999-1997. Structured questionnaires were used to record data on the NICU's organization and policies and to survey the physicians views, attitudes, and prac- ices in end-of-life decision making. The physicians’ questionnaire was anonymous and sell-administered to protect confidentiality. It included 57 questions asking participants about Q) their personal views on and atti- tudes about the limitation of intensive care for ethical reasons, the identifica- tion and role of decision makers, and the relevance of such factors as costs of health care, foreseen advances of medi- cal knowledge, and legal constraints; (2) theirself-reported practices through- ‘out professional lifeand inthe lat prob- lematic clinical case that they had et countered; and (3) their professional and demographic characteristics The questionnaire was originally pr pared in English and subsequently translated into each country’s respec live national language. The transla- lion accuracy was checked by transla. ing the questionnaire back into English and by simultaneous review of the na- tional versions by a panel of transla tors to ensure identical semantic con- tent in each language Physicians’ attitudes and beliefs and their relationship withsell-xeported prac lice are the focus of this article. Atti tudes were explored by asking respon- dents’ agreement, on a 5-point Likert- type response scale (from “strongly agree" to ‘strongly disagree”), with alist ff 12 statements dealing with different aspects of ethical decision making, The statements were developed after review fof the medical literature and after in- depth, qualitative, personal inteviews conducted among.asmall sample of phy- siclans and nurses in Haly and France"; the statements were examined for con- tent validity by the international mul- Lidiseiplinary group of experts associ- ated with the EURONIC study Statistical Analysis and Construction of an Attitude Score All questionnaire coding and computer storage were completed at the coordi- nating center in Trieste, Italy. Com- ments by respondents and answers to the few open questions were integrally ran- scribed and translated. Data analysis was performed using STATA statistical soft- ‘ware.* Weights were used toaccount for (©2000 American Medical Association, All rights reserved. ‘Downloaded From: http:/jamanetwork.com/pdfaccess.ashx?url=/data/journals/jama/4760 by a Universita Torino User on 03/13/2017the different sampling fractions applied in the various countries and stra, Results are presented as weighted proportions and 95% confidence inter vals (CIs). In the calculation of the lat ter, SEs were adjusted to account for the cluster sampling study design, ie, the nonindependence of observations ‘within the same unit Factor analysis was used to identily the chief underlyingdimension of theset ‘of 12attitudinal statements, The firs fac- toraccounted for 80.8% ofthe variance and within i7 statements (statements: 1, 2,3,4,6, 7, and 8) were selected ac- cording to their higher actor loadings. Thesestatements were found tobe highly Intercorrelated (reliability Cronbach a=-71)and represented agreement with theidea thatanewbornshilehastobesus- lained irrespective of outcome, in con- trast with the position that quality of ile hastobe considered oo, Anatutudescore ‘was therefore derivedasasum of the an- swers to these 7 selected statements ‘weighted by their factor loadings," and the scale was normed to vary between, ‘NEONATAL END-OF-LIFE DECISION MAKING indicating total agreement with the idea ofan absolute value oflife (the prolife ap- proach), and 10, corresponding tomaxi- ‘mal disagreement with this position (the quality-of-life approach). A multivariate linear regression analy sis was used to identify the variables as- sociated with a physician's atitude score, with the score as dependent variable. In- dependent variables included country personal characteristics of the respon- dents (sex; age; having had children; re- ligious background, coded as Catholic, Protestant, other, ornone; and religious ness, defined on the basis of how impor- tanta respondent considered religion in hisher life); professional charact tics (position, length of experience in NICU, type of clinical work, involve ment in follow-up of infants alter dis- charge, and in research); selected unit characteristics (numberof intensivecare beds and of very low-birth-weight neo- nates admitted in 1996; level; working ata teaching hospital; existence and type of a hospital ethics committee, and of a written unit poliey about ethical dect- sion making). The variables retained in the final model were correlated with the attitude score at P<.10 and remained sig- nificant at that probability level once en- tered in the multivariate model, Vari- ables not significantly related to the attitude score were removed from the model after determining thatthe effect estimates for country and the other vari ables associated with the score were not substantially modified.” To assess the relationship between physicians’ attitude score and thetr prac lices, adjusted odds ratios (ORs) wer estimated through a multivariate logis- lic regression model using as outcome variable the physicians’ selF-report of having ever decided (by themselves or together with others) to set limits on i lensive interventions because of poor neurological prognosis. The atuitude score was included in the model as a continuous variable, jointly with coun- try and other personal, professional, and uunit-related factors found significa associated with the self-reported prac- lice in the univariate analysis, tr "Tr sae, Response Rate and characterstcr of Sucy Population The United Variables Italy Spain France Germany Netherlands Kingdom Sweden Hungary Estonia_Lithuania ‘Sample Sze (Response Rate, %e) No. oLNICUS recited 22 (109 181951784) 29/88) 10/100) 94x) 16,9 _16(@0)_2(100)_9,100) io. crespondeg pysciney TTBS) — 206 4) 206 PIB EN TSE BUTE —BaaT| TO Ga) THT) PHA 4a (6411264 108,69) Physician Sociodemographic Characteristics srs) 4865) 07H ey Buy 9.0) ato) SA a) BAT TOG 40168) 2425) 79,60) TO (6a) T9168) cee 2) A625) DAN) RHA 1H) — HB-B) BTA) BEeH oF GH OSG) Tar 670A T_T aaa] STITH) Tere oo) ae SAS) a1 Tee) aoe Say PIT 7g 0 erg mare 1) ateyeH IY _5/1B) Pa oe] Pos TO Taa es) eS AT oT ee) BS] Brea 0 Pa amt Toe) Tosi are) 40) I wi TO 7 oa 2) 3) ew a se4(6 10461) eta) 1116 97H 98K) sot) 311g) 98) otimperant BST) TOO) THe) 106 S07) er TOea) Tawa TO Biperencein NCU. y “6 Toms) e045) 160.61) _ 15069 see) se saa Se 15/54) ae Tao 6) aoa) OSE aA, 2aeN sae 3186) THT) 5 Trea) 62,31) 85 17) ABT) Ta 2 wea) (©2000 American Medical Association, All rights reserved, (Reprinted) JAMA, November 15, 2000-Vol 204, No. 19. 2482 ‘Downloaded From: http:/jamanetwork.com/pdfaccess.ashx?url=/data/journals/jama/4760 by a Universita Torino User on 03/13/2017NEONATAL END-OF-LIFE DECISION MAKING RESULTS. Response Rate and Characteristics of the Responding Physicians and Prance was Catholic, while in substantial proportion of physi- Sweden, Germany, and the United cians in Spain, Germany, Hungary, Kingdom itwas Protestant Religion was Lithuania, France, and Italy indicated, considered extremely or fairly impor- that may con- Othe 166 cligible NICUsasked to par- tant especially in Italy, Germany, and _ stitute a “slippery slope” likely to lead. mitation of trealme ticipate, 142 accepted, giving an over- Spain to abuses. In every country, however, all response rate of 86% (TABLE 1). A similar or even higher proportions total of 1391 completed question- Physicians’ Attitudes agreed that the “slippery slope” argu- naires were returned from the 1550 eli and Attitude Score ment also may work in the opposite di- sible physicians (overall response rate, TABLE 2 shows the proportion of phy- rection and ht intensive care sell my 8096: range, 69% 100%) scianstesponcingagrecorstonglyagree represent a source of abuses through The distribution of soctodemo- with the 12 statements exploring al-ovettcatinent graphic and professional characteris- tudes. Thirty-three percent of physi-— Inall countries the majority of physi- tics of the physicians varied across coun-_ciansin aly, 25% in Lithuania.and24% clans agreed that burden for the family tries (Table 1).In Sweden and Germany in Hungary agreed with the statement: is relevant when making end-of-life de most physicians weremen, whilethe op- “Because human life is sacred, every- cisions fora child. More than half of the posite was rue in the Netherlands, Hun- thing possible should bedone to ensure sample in the Baltic countries thought gary, and the Baltic countries. In Swe- _aneonate’ssurvival, however severe the that legal constraints preclude the pos- den and Italy physicians tended to be prognosis.” Most physicians in every sibility of decisions to limit treatment, colder than in the other countries. country qualified severe mental disabil- while a very small percentage did so in When asked, “In what religiousback-ityasanoutcome equal to or worse than Sweden (3%) and France (5%). A sub- ground wereyou brought up?" the most death, whileconsensus was lowerwhen stantial percentage of physicians in frequentanswer in Spain, Italy, Lihua- severe physical disability was at stake. Lithuania (549%) and in Italy (20%) sup- ‘Table 2 Physicians: Agreement With 12 Statements Related to Attitudes About Value of Li Tay Spain France Gomany oo! See! Soa! ese! No. 9% 95% Ci) ‘No. _% (95% Ci) No. _% (95% Cf) (No. _% (95% Cl T Gocamonmanionmced aayingposbedoid a8 e325) a Tolloah 1 5eH — 6 Se be done to ensure anecrate’s sural, however severe the prognosis Z Eien wih sor piyecal Gosaily sone Ne sawas 5 AT USH BOAT BL ASRORH 1S a ROT, batter than no ea & Bien wih seers menial daabiiy, cone Ween 61 236) Ba 9 awa a TSO, batter than noe a a © Lanting ensve care, isn only hawiamay solcted 17 20235 6ST EES eo ESA ERTS stuatons sa “sippery stope” thal wil ato abuses 5. lrsive cares a ‘sippy slope” tel to eadto BT ae ES aT ET therapeutic agressive The borden hat a casted chia wi represent for he 7 Seay See 2 Sie wo AT fami snot o relevant when making ethical Seton for that neonate 7 There sno oom er efical ecsors wren ela doos OT GS ETSH__aS Tw Osea aa SATA natal any tars of Vetment & Brey recrate shouldbe geen he maimananontt Te _ PO ESH ae TONSA SEI STAT ‘lene car naspectng of eulcome, bacause he ical ‘xparance acquired wl ens ther abet nthe ure The incrasing cons of Peal careforprslarnneabons G0 TO 2 OTTO ST ABTA Te OTH {nd deabied chien not slow us ore! each patent regress of euteore 10, Foman aca port of wav. have eno aorarce B Sra sae e a7 aT ea baton withhotdng and withdraw of ens core TT Foman aca pont of vw, tars aference ee otvenn witha of ners of ergs wth he purpose af nding TZ. Withoting niensve care wihou srutimenusy ing 100 SSIS aT aayTay 12 SS wTeH SE TEED, ‘Sve measures fo end tna ncrateee dangerous Bocas t makasit mers kl fo the neonate tobe Several deabed hehe ues rand acrnsraon 2454 JAMA. November 15,2000 vol 206, No. 19 (Reptnes). (©2000 American Medical Association, AI rights reserved ‘Downloaded From: http:/jamanetwork.com/pdfaccess.ashx?url=/data/journals/jama/4760 by a Universita Torino User on 03/13/2017ported treatment for every neonate ire spective of outcome, "because the elini- cal experience acquired will benefit others in the future.” Although most phy- sicians in each country did not believe that costs of health care should affect nontreatment decisions, one fourth or more agreed with this statement in France, the United Kingdom, and the Baltic countries, ‘Most physicians in every country but Lithuania appeared to make an ethical distinction between withholding inten- sive care from the very beginning and withdrawing it afterward. About one third of physicians in France, the Neth- crlands, and Estonia found no ethical difference between treatment with- drawal and the administration of drugs ‘with the purpose of ending a patient's Iie, ie, active euthanasia; in Prance and the Baltic counties more than half of the respondents agreed that “withhold ‘NEONATAL END-OF-LIFE DECISION MAKING ing intensive care without simulta ncously taking active measures to end life" may increase the likelihood of se- vere future disability Physicians’ answers to the 7 stat ments selected through factor analysis (statements 1, 2,3, 4,6, 7, and 8 in Table 2) were used to construct the at- ude score, the distribution of which fs presented in the FIGURE. The score varied across countries with median val- tues ranging from 5.5 in Hungary and Estonia to 7.8 in the United Kingdom, The Baltic countries, Hungary, and ltaly ranked lowest, tending toward a more prolife auitude, while the United King- dom, the Netherlands, and Sweden had the highest scores. Spain, Germany, and France ranged in the middle. Italy Spain, Germany, and Hungary showed larger variability around their median, values, indicating a less homogeneous position of respondents. ing. Physicians’ Attitudes A multiple linear regression analysis was conducted to identify the physi- cians’ personal and professional char- acteristics and the units structural conditions, which might explain the variability of the attitude score within and among countries, The conditions that were signil cantly related toa higher attitude seore, ie, having more quality-of-life beliefs, were being female, having had no children, being Protestant or having no religious background, considering ligion not important, having an inte mediate length of professional experi- cence (6-15 years), and working in units ‘witha higher number of very ow-birth- ‘weight admissions (TABLE 3) In every country nonreligious re- spondents had a higher mean auuivude score, ie, a stronger quality-of-life “The Netherlands United Kingdom Sweden Fangary Estonia Tiwana ho. % (05% 09! Ne. % (5% cH! he 605%! he. % (05% c9! fo 2% 05%) Ne. 960554 0)! T4105 3 ser 78a 2 aaqiea) 3 i712 7 2516-57) Tsao ony ECCS CSCIC 2 ese 8 Z oe ea TO SEWER OBSTET Bape ea wae 6 5 Baye a array Sa 7a re aE We wr way a ay 5 = a 7 @ Bem eee aaa a 7ST ea (©2000 American Medical Association. AI rights reserved. (Reprinted) JAMA, November 15, 2000-Vol 204, No. 19. 2485 ‘Downloaded From: http:/jamanetwork.com/pdfaccess.ashx?url=/data/journals/jama/4760 by a Universita Torino User on 03/13/2017NEONATAL END-OF-LIFE DECISION MAKING titude, than religious respondents, but this difference varied significantly (est for interaction, P=,02) between coun- tries. As shown in TABLE 4, the differ- ences of adjusted mean scores between Figure, Deibuton ofthe Attude Score by County religious and nonreligious respondents . 1 ineach country varied between 0.3 in Es- = tonia to 1.6 in Lithuania. Among physt- ‘ cians who considered religion impor- tant, those from Italy, Hungary, and the 8 Baltic countries showed scores signili- cantly lower than in the other coun- 0h == tries, and therefore demonstrated amore eas nga ee prolife auitude, However, when reli- “Then teeing the mide of ath box the medan he box ends rom the 25th pecntietwihe | SOR Was Teported as not important, Ital 75th prcntle nterqurte range [GRD, The err bts extend fo the upper nd mer adecent vue The an physicians did not significantly dif= Upper ajcent values defied asthe gest ala pit es than or equal fo (50h percenle+15>i0R) fer from their Spanish, French, and Shuthe lower adjacent vale defined asthe smallest gata pont grater than or equal fo 25% (Se il hile thece Dercente=15>IGR) Alowerscore representa prlfesttuge shghscre amorequaty-o-ifexttuge, German colleagues, while those from Hungary and Estonia continued to pre- sent the lowest scores, ‘Table 3, Rests of Mulovarlate Linear Regression Analysis Used to loentty Predictors of the ‘Ate Seore Vales" td ae Seas Attitudes and Self-reported Vibe atone “NSN” PVaket Practices coy xe When asked, “nthe course of your pro- fessional life, have you evr decided, by Spa ie " = yourself or together with others, to'set rey limits om intensive interventions” be- Tra Naa am cas of very severe neurologieal prog ‘United Kingcom 001 nosis, most physicians in every coun- Suede try but Haly and Hungary answered Farge allrmatvely (era proportions across Estonia countries ranged from 46% in Italy 10 im = 90% in Sweden). ‘en e160.89 The likelihood of having reported this = aAe2ER 004 ype of nontreatment decision (TABLE 5) increased among those with higher at- ong _titude scores (univariate OR per unit change of score, 1.6; 95% Cl, 141.7), and the estimate did not change sub- Finnghad rar ¥ 9.62 gt Frovesiond eparireey 6 ros fc flo 6.115962 a aspsam oo? stantially when potential confounders 15 6215985) were included in the logistic model Fagus background Other factors significantly associated with Cathal 6.1(6.9.62) the outcome variable were length of ex Poi 5516558) fo rience in an years), a sas we pe NICU (Zo years), nlor professional position, and work- ing in units with # higher number of 65 (64.68) intensive care beds. Ta Baie Te) prance ofaigar ‘Notimpertant o inperan 5s O°! In contrast, the effect of religious io of veiw br warp paoraes background and religiousness was no tla veen ® NCU per year longer statistically significant once the “Coury median 60/6862) oor 08 ally sig County mean 84/6566) attitude score was entered into the Wine nae nous came A can eT Tritrce ries neorenwa model. Women physicians did not sig- sta hes ahs sore reed byte rods, Lower tes conerpore to mer rote atmuce; ngerDificantly differ from their male col- "rou amore aoe ame leagues after adjusting for profes- Vatertar sheoricasiasetancethesnam tiem actenvatticanthestinanas COB 2456 JAMA, November 15,2000 vol 204, No. 19 (Reptnes) (©2000 American Medical Association, AI rights reserved ‘Downloaded From: http:/jamanetwork.com/pdfaccess.ashx?url=/data/journals/jama/4760 by a Universita Torino User on 03/13/2017NEONATAL END-OF-LIFE DECISION MAKING ‘Table 4, Dstroution of Adjstes Means of Attude Score by County af by Prysioans Sell report Relgjousnes: Physicians Rating Religion as Not important (a= 631) Physicians Rating Religion as Important (= 685) No. of ‘Atitude Seore® No.of ‘Attude Seore™ Country Observations (etc Valet Observations (@5% ch, PYauet my 33 SGI 17 BouTeT Sea ao BAS o-69) a 3 BT 6262 a Franca = aaeE72 a Te Sie T8, 007 Germany oo e215 65) a5 Tot BoG562 001 The ethane oy TST, 007 2 Baar 001 Unites kgs a6 ToUsaa, 001 aT aos 79) 001 ‘Swecen 3 726870) 0 73 SS 8579) oo Hungary 6 S56158) O07 = aeazeay Ei Estonia 7 Br aee5) 001 a ae (68 a Tirana 1 ereeray EN = a5 G685) = ‘Table 5. Predictors of Self-reported Decisions to Limit Intensive Care Ever Made Because of Poor Neurological Prognosis Univariate Analy Marivarate Arava Outcome Variable, Unadjusted on [Adjusted OFG Pradictive Variables Nos (estat, pValuet (@5%ch) pValuet Taunt a +176) 10 10 Sea THe Teno ao, 2ag ea Francs ) Read Rees ‘Germany 754 60), PETEEE reo. oe 8) a0 e5-64) 0 degor09 cot Ba 74) Ba. 70) aust Sveden 2 oe aaeoza5, Fangay "ary "ss 750735 TO) 208 5(0a57 iran Te) 130720 750829 iiudeccorebaNESTTTT TBTETT Tor Tae TT Tor Bex ‘Men 507 (6) 10 10 = sata cress | aewsra | " Feige bacroane Gatheie 503 (6) 10 10 7a OTST , SESE . aT 5) 52 a oao2-19 6 THO TE asez47 FrpararcnTao — vo vo ieee sie or J osm] a Frotessonal earns y 6 373.67 10 ee a EaTERD, ] cos ] cas 5 00,61 Boze Frsessara postin Constant 37167 10 eg 0a 78 Tw ] cos i cas House ofeer 08 63) 060A oaz0n Tie olrtnsve cae bade — , - se sme] J % (©2000 American Medical Assocation. Al rights reserved. (Reprinted) JAMA, November 15, 2000 24, No. 19.2457 ‘Downloaded From: http:/jamanetwork.com/pdfaccess.ashx?url=/data/journals/jama/4760 by a Universita Torino User on 03/13/2017NEONATAL END-OF-LIFE DECISION MAKING Country, as indicated by the higher values of the adjusted ORs, remained the strongest single explanatory factor of differences in practice after adjust- ing for the effect of the other determi nants, No significant interactions wes found between atitude score and the other variables included in the model ‘COMMENT The results ofthis study support 3 main conclusions, Pst, we found that the at- Litudes of European neonatologists to- ‘ward sanctity of life vs quality of life war- ied both within and across countries, British and Dutch physicians scored hhighest in the quality-oF life position, ‘while those from Hungary and the Bal- liccountzies appeared to have the stron gest prolife attitude, followed by ltaly, Spain, and Germany. Within-country variability tended to be larger in coun- tries, such as ltaly and Spain, where le islation is more restrictive” and/or the influence of religion is greater. Pres- cence of mental rather than physical disability, burden tothe family, and per- ceived ethical difference between with- hholdingand withdrawing treatment, and active euthanasia appeared to be rel- evant to the majority of physicians in most countries. However, consider ations related to legal constraints, costs of health care, and acquisition of clini- cal experience were rated as less impor- lant to decision making, Second, in every country a physi- an'sattitide score was significantly as- sociated with his/her likelihood of re- porting having ever decided to set lirats fon intensive care interventions be- cause of poor neurological prognosis, ie, on quality-of-life grounds, Third, even alter controlling for poten- Uial confounders in multivariate anal sis, country remained a strong signill- cant predictor of both physicians’ attitudes and practice, suggestingan effect ‘ofculturaland social factors beyond those particular to physicians and their units Several articles have described physi- ans autitudes*™* and, less frequent practices! regarding nontreatment cstons forboth adults and neonates. To cour knowledge, however, only 1 other 2458 JAMA. November 15,2000 ol254, No 19 Reprint) study explored their relationship by us. ing attitudes as a predictor of self- reported practice." In that study, Penn sylvania internists were asked their willingness to withdraw life support in response to hypothetieal cases: those with a greater willingness to withdraw treat- ment were also more likely to report hav- ing done so during the preceding year The authors concluded that clinicians differ considerably in their atitudes to- \ward life value and support and that these attitudes are reflected in their practices. They also suggest that physicians should communicate such personal prefe ences to their patients, ‘Our study confirms the importance of attitudes in influencing a physician's practice. In addition to country and at- titudes, other factors associated with practice in mullivatiate analysis were professional position, length of experi- ence, and working in larger units. In contrast, the effects of religious back- ground and religiousness were no longer statistically significant once the atti- tude score was included in the model, indicating that the attitudes might rep- resent an intermediate step mediating the elfect of physicians’ religious be- liefs on thetr practices, Given the rapid advances in medical knowledge and technology in neonatal care, physicians are increasingly likely to be confronted with difficult decisions about starting or continuing invasive ife- sustaining treatments in the face of very poor long-term prognosis. Because of their predominantly technical raining, however, they may feel poorly equipped to deal with these highly value-laden choices, especially insocieties that do not engage in open ethical, legal, and pro- fessional debate on such issues. Accord ing to a recent review of legislation and oficial guidelines across Europe regard- ing limiting treatmentin neonates most countriesallow nontreatment decisions for newborns who will inevitably die regardless medical interventions, but limitation of intensive care on the basis offuture quality oflifeismuch morecon- troversial. In some counties, such as lualy, the law is strongly protective of human life, especially when minors are Involved, to the extent that resuscita- lion ofa preterm neonate is mandatory and practiced, even when the birth results Irom induced late abortion.”® But in general, physicians are operating in a Tegal vacuum, since most countries lackspecific rues either in their national law oF in their codes of professional medical ethies. Only in the United King- dom and the Netherlands have several ‘court eases tested the limits of what is permissible. In the United States, the American Academy of Pediatrics does support for- ‘going life-sustaining treatment when phy sicians consider the treatment notin the newborn's best interest.” However, un dertheso-called “Baby Doe" regulations and subsequent child abuse amend ments," nontreatment decisions for new- horns and infants are permitted in only 2 fairly restitivecircumstances:(1) when the infant is irreversibly comatose; (2) ‘when treatment would purely prolong the process ofdying;and (3) when it would, Je futile and inhumane. In every other cease, withholding treatment would be cconsidereda form of medical neglectlead- ing to loss of elighility for federal and ing. Asaconsequence USneonatologists hhave often been reported to follow the ‘wait until certainty strategy."*"* which prescribes maximal reatmentof every in Fant until virtual certainty of death or reversiblecomaisreached. AUSsUrv found thatphysictans frequently felt pres- sured to overireat infantsbecause of fed- eral regulations, fear of legal action, and technological developments. They be- lieved thatthe regulations did notallow adequate consideration of infants sulle ingand interfered with parents right 10 determine which course ofactionisin the best interest oftheir child. More recently, in discussing the legacy of Baby Doe, & group of US pediatricians noted that “while adults were gaining. right odie thatincluded theright to forego fluid and nutrition, the parents of neonates wer losing the right to forego ll but the most inelficactous treatments. Some limitations ofthis study should, benoted. The attitude score wasderived through factorial analysis by selecting, froma wider pool, thosestatements that (©2000 American Medical Association, All rights reserved. ‘Downloaded From: http:/jamanetwork.com/pdfaccess.ashx?url=/data/journals/jama/4760 by a Universita Torino User on 03/13/2017showed a high correlation and internal consistency, as well as content validity ‘with attitudes toward life valueand sup- port. Although this wastrue for the whole sample, some of these statements showed ower correlation in some countries: therefore, the consistency at country level might be not as good as the global one Underreporting of end-of-life practices cannot beexcluded although efforts made to protect confidentiality, as wellascon- sistency of results with other stud- ies,” make it less likely. Finally, the numbers of physicians participating in Estonia and Lithuania were small, pos- sibly resulting in unstable estimates, but the physicians who partiipated repre- sent all oF nearly all eligible physicians in those countries. To our knowledge study exploring the neonatologists at- titudes toward quality of life and the re- lationship with their own practices in a representative sample of NICUs froma large number of European countries. In- ternational comparisons of this nature are important for several reasons. Learn- ing about the effeets of cultural factors ‘on treatment choices*may help broaden the perspective of those who frame these decisions as purely medical. Compara tive studies could also be useful forseek- ing a better understanding of the ef- feets of the law and societal and cultural values on elinieal practice.” Ourinvestigation revealed that difer- cnt culturesand legaland religious con- texts influence both physicians’ atti- tudes and end-of-life practices within Europe. This sll the more relevant in light ofthe important finding that phy- sicians' personal attitudes do relate to their practice ofend-of-lifedecisions. To the extent that this relationship reflects dlrect causality (with aitudes influene- ing practices, rather than the reverse), international cross-cultural studies such as this one may facilitate a critical se analysis ofphysicians behaviors, and ead toa decision-making process grounded fon a more sound ethical bass. this is the first Author Afliatons: Deparment of Public Heath [Miguel emander Univer, Alcart, Span (Dr Re: bgt) Unis of Eiemelogy ap Neonatal inter ‘Se Cae Buta GatlaCten Hopital Drs Ci (©2000 American Medical Association, All rights reserved, ‘NEONATAL END-OF-LIFE DECISION MAKING ide Vordenit 2nd ogi) Teste, and Dison ‘of Epidemiology, FC, National Researen Coun (Dr Sarat, Pity Deparment of Obstet and Gynaszlogy, Vasa Kole Hostal, Tet, Hue {21 (Dr Bebo; Deparment of Pedates. Marth {ther Unversty ile, Germany (Dr Hansel ‘emisogea Resch Unton Petal and Women. Heath 0149 NSERM (Or ars Vil and Unt ‘of Reseach on Repreduction INSERM CI 89-08 {DrLenain, Touuse. France: Department of Neons {ology Unies Hepa ot Mimegen meget Nethetangs fr Kole), Neonatal Cini Un ‘erst, Uthuana (Dr Kuan) Newbam and Pre Imatuté Chen's Department, Talon Mesptl to fia (O Lev Cer for Medea Teton Asses ime, Lnkopng Unwesty. Swesen (Dr Perse {nd DepartmentotPube ett Unvestyolegon, Scotian (Dr es) ‘ther Member he EURONI Study Group ance: ‘A DuguetandIt Cae! Gemary-G-Be-Vot Correa po and Lena The United Kingda) Har ‘woodapndl€ eae a BencolV Cast, S'Nocdo, i ze ard Sisat; Lnentour F Mousty and Savod he NeterandL Hanes Dieta aeLeeue Randag and Sauer Spam 1a and: Pe Pers: and Seder 8. Werner, Funding Support: Th sty was funded y con ‘eat A-C198-1283 fom the European Commie Sion projet cordnato, Dr Cultin, and EU oe, Braadous. ‘Adnowedgmet: Wethan Mil for adie ont ‘teal ue and Naa forherworcon esi data nasi We hankthelocaleaoriatosneadsoftne pariopting unts, and allcaleagues wna tok tine ‘utof ther busy westoancverourgoestonnare We “Ecinowedge te contibution of he RTF Ist, party A DelfAngla in data management REN te ent iy ate BOs Ener eee cc ibs bers eet peerectentetarend eae Shenae Seat aaes LSmaoerhemacind Secor eee hal Adolesc Mea 3a 152-025-135, ER orn re toe Se aah beethermssiisae PA cede bt ge SOS Soe rar ts soiemaenca: en BBs HEN ee es aattne see aan oe SNe pale estan ates sits it RTs Ces Fe rd Meta te eagle toni aa Bsr ses scene SR ti le re Sa Eee \isting communication and patiptoin etic de ‘ion a comparton ofeontalunpaes née. ‘on re Cha Fetal Neonatal Ea 13588 FO4-3t 45, Gare M, Cosme Saguret Kans, Cat I uespres de deauors nauesen antraton neo aa Frere. rch Pelt 1997, 2662-670. 46. Sata Sats Sotware(eorpute roar. Re lease 60. Calege tation, Tella Cop, 1998. {7 Lemeshow ,LeteoeurL Darigues IF Lafont SOrgoqozo IM, ComamengesD. lustaten oa Si tating to account complex survey corset Sons Afn/ Enema. 1998 148298 306 18 Oppentam AN Questonnare Desig lnteview ing snd stage Measurement. London, England Pier Pubes it 1992. 49. Maldonad G, Greenland S. Semdstion ty of ‘ofoundeslecbonsateges. AJ Epidel. 193 ‘36973 936 20. Mette HE Cut M, Bl VotG, etal Wt foing/utbaraning eatment Yom neonates Jed cy 195825 40.486 21. Taares ID, Kane D, Howel MC, Shannon OC Beatie litte afecingcetion making nce fective newbome Pediatnce 1977/60 197-20 BE: Canace Sechan Th, Cook usegSO8 Bar {en Mt Maral reasoning and desir in ler of ‘eon eae Ped Rex 182;1686 $50, 23. Kopelman L rons, Kopeman & Neortoo- ‘islude tne “Baby Doe regan ng! ‘seu 318077 683, 24. Sanders aR Donahue PK Obert MA Rosen ante TS, Alen MC. Perceptions ofthe st of buy. Pratl, 189515498502 35. Noup iM. Tresinent oferty premature ne Boms’ 2 survey of attudes among Dansh phy Gans, Acta Paediatr 199587 896902, 36. herp Me Treatment of severely dceased ne Bins survey of atudes among Dansh phys Gan, Acie Pediatr 1999/8848 27° bu RS, Campbell AGM. Mal ad tia fermen te sec are nuey Meng! Me. 1973; daoaon.8bs 2a. Ryan CA yneP, Kuhn 5, Tyeblhan No re ‘ction and wire of the'apy In neat Enda peitre tense care unt in Canad. Po Sh isaras3e-a3e 39. Wat SH, Patnidge IC Death inthe intersivecare husey sian pace of witaraving and wt hotng te suppor Pediat 1397.9868-70. 30. van der Hecde A, van ger Maas, vader Wal, fal Megs end-of fe decison neonates sn {tin the Nefernds ance 997 990951235, 3 Chr NA, Asch DA Physica case fits aocated wth decors to withaaw ie sup ott Am Public Heath, 1995 85:367-372. 532" sopra alfaarto, na made outa Gn Aaian! Comer ela er Marzo 16, 1999. 533. Amencan Academy of Pecatics Commitee on Bloeths Culelnesonogongifesustarang med. (a venient Paiaties, 19493532-536. S34. realy Hi. Baby Doe ana beyond: the past and ‘tue of Goverment regalton a the United States, In: Golgwortn Siveran W, Stevenson DK Young We ther and Pernatoogy: New Yar, WY: Ox ford Univer Pres 1995256306, 35, Roden NK. Treating Baby Doe the ethics of un tay Hastings Cent Rep. 1805 34 56. Lantos, Tyson, Alen A, el a. Withling hd wihdrving if ustaning etme neon ‘intesve cae eae rte 1990 Ach isch, soe reesierr22 37. Nise HE, Fowse PW. Lif Death and Dec Sone: Doctors and Nurse eft on Neonatal Pra ce Chose England Hochand& Hochland: 1996, 538. Paper eaine and Cult: Varebesa Tea iment in the United Sates, England West Germany, ara New York NY rey Hl & Conc 198, 538. WaldeLevm& riematond ponpectves on tea: ‘ant chen peonatalntesive cre un. oe Se Ms. 19803001 512, (Reprinted) JAMA, November 15, 2000Vol 204, No. 19 2459, ‘Downloaded From: http:/jamanetwork.com/pdfaccess.ashx?url=/data/journals/jama/4760 by a Universita Torino User on 03/13/2017