CLINICALETHICSIdealsofpatientautonomyinclinicaldecisionmakingastudyon - PDF document

CLINICALETHICSIdealsofpatientautonomyinclinicaldecisionmaking:astudyonthedevelopmentofascaletoassesspatientsandphysiciansviewsAMStiggelbout,ACMolewijk,WOtten,DRMTimmermans,JHvanBockel,JKievitSeeendofarticleforauthorsaffiliationsCorrespondenceto:DrAMStiggelbout,DepartmentofMedicalDecisionMaking,LeidenUniversityMedicalCenter,POBox9600,2300RCLeiden;a.m.stiggelbout@Received13January2003Revised21May2003Accepted2June2003 samepatientmayprefertoleavethedecisiontothephysician,becauseultimatelysheorhedoesnothavethecouragetodecide(Molewijk,unpublishedobservations).Astheprevailinginterpretationofautonomyinthemedicalliteratureappearstobetheliberalindividualistoneand,asinpractice,patientpreferencesareatoddswiththisinterpreta-tion,wewishedtoassesswhetherpatientidealsaredifferentfromtheirpreferencesorwhetherotherviewsthantheliberalindividualistprevailinpractice.Studyingpatientideals,andsocialpracticesingeneral,isimportantfromtheperspectiveofintegratedempiricalethics.1011Integratedempiricalethicsresearchreferstostudiesinwhichethicistsandsocialscientistscooperateincombiningmoraltheoryandsocialscientificresearchmethods.Suchstudiesresultfromtheviewthat,intheend,thereisnofundamentaldistinctionbetweenfactandvalue,andthattheempiricalandthenormativeareinterconnected.Ontheonehand,factsproducedbydescriptivesciencesareinterwovenwithdisciplinespecificepistemicvalues„forexample,theinfor-mationpresentedtopatientsmaypartlyshapeacertainkindofpatientautonomy.Ontheotherhand,everymoraltheoryisbasedonempiricalpremises.Inthecontextofourresearch,differentmoraltheoriesonpatientautonomyassumedifferentideaswithrespecttotheidentityandrationalityofhumans.Onecannotconstructanormativetheoryonpatientautonomywithoutreferringtoaninterpretationofwhathumanbeingsactuallyare.Ifempiricalresearchshowsthatacertainempiricalpremiseisnotfeasible,thenitmightchallengethevalidityoftheassociatedmoraltheory.Peopleshouldnotbeheldtoastandardthattheyareunlikelytosatisfy.Forexample,ifsocialscientificresearchshowsthatmanypatientsdonothavethecognitivecapacitytounderstandmedicalinformation,onemayquestiontheliberalindividualistnotionofpatientautonomy.Integratedempiricalethicsthusdoesnotbelieveintheprimafaciemoralauthorityofasocialpracticenorinthatofamoraltheory.Itisconcernedwiththecontextualrelevanceofmoraltheoryforthemoralityofspecificsocialpracticesandviceversa(thatis,moralityderivesitsmeaningfromthecontext).Agoalofusingempiricaldatamaybetheimprovementofexistingmoraltheories.Forexample,onemightwishtomodifythetheoryofpatientautonomytowardsthesocialpracticeofsurgicaldecisionmakinginelderlymenonthebasisofempiricaldata.Toassessthemoralbeliefswithinthepracticeofsurgicaldecisionmaking,wewishedtodevelopascalethatassessespatientsandcliniciansidealsofautonomy,nottheirpre-ferences,andthatlinkstheseidealstoelementsfromthemostimportantprevailingethicaltheories(tobedescribedinthenextsection).Asomewhatsimilarattemptatelicitingideals„limitedtosurgeons,notpatients„hasbeendescribedbyFalkumandFørde,whorestrictedtheirassessmenttothecentraldimensionsofthreeofthefourwellknownmodelsofthephysician-patientrelationshipofEmanuelandEmanuel(thepaternalistic,theinformative,andthedeliberative).AscalethatonfirstexaminationseemstohaveasomewhatsimilarnormativeobjectiveisthePatient-PractitionerOrientationScaledevelopedbyKrupatetalInthisscale,theitemsareframedinanormativesense(thedoctorshouldƒ),butneverthelessitisaimedatassessingpatientcentredversusdiseasecentredpreferences,andsolelyfromapatientperspective.Weexplicitlywishedtoassesstheseviewsbothincliniciansandintheirpatients,becauseitisintheirinteractionthatdecisionsaremade.Inanyphysician-patientrelation,whetherpaternalisticorother,bothpartiesplayaroleandtheviewsofbotharethereforerelevant.Thestudypresentedherewasperformedinthecontextofalargeresearchprojectondecisionsupportandpatientautonomyinsurgicaldecisionmaking.Ourstudyhasanexplorativecharacter.Wedidnotstartfromonenormativepointofview,norwasourultimategoaltoquantifytheprevalenceofthesupportforthevarioustheories.Wewishedtoassesswhetheraspectsofethicaltheoriesotherthanliberalindividualismcanbedistinguishedinmedicalpractice.Aspreferencesforinformationhavebeenfoundtobethesecondmajordimensionofpatientautonomyinthedescriptiveliterature,wewereinterestedintheroleofriskinformationindifferentconceptsofpatientautonomy.Informationmaybeperceivedasnecessaryforautonomousdecisionmaking,butatthesametimeitssometimescomplexandthreateningnaturemayobstructautonomousdecisionmakingifpatientsfeeloverwhelmedbytheinformation.Thispaperdescribesourattemptatdevelopingandtestingaquestionnairetoassesspatientsandsurgeonsidealsofautonomy.Itdescribestheitemselection,thefactorstructure,andinternalconsistencyofthescale.Itpresentsthevaliditychecksweperformed:assessingcorrelationsamongthesubscalesoftheinstrument,andcomparingscoresontheinstrumentwithaquestiononpatientspreferencesforparticipationintreatmentdecisionsthatisoftenusedindecisionmak

ingresearch.Assomeoftheitemscoveredinourquestionnairepertaintothedecisionmakingrole,weexpectedourquestionnairetocorrelatewiththisquestion,butthecorrelationwasnotexpectedtobeveryhigh,giventhemorenormativeperspectivethatweheld.TheconstructionoftheIdealPatientAutonomyScaleThestartingpointfortheconstructionofthequestionnairewasthefollowingsixmoralconceptsofpatientautonomyfromtheethicsliterature(MolewijkAC,unpublisheddata).Theliberallegalconceptofrespectforautonomy,basedonBerlinsnegativeconceptoffreedom,whichstressesfreedomfrominterferencebyothers.Itprotectsvulner-ablepatientsfromunwantedinterferencebyphysiciansbydemandingrespectforapatientsintegrity.Itisthusdirectedtowardsthosewhotreatorcareforthepatient,nottothepatienthimself.TheliberalindividualistconceptofautonomyasdefinedbyBeauchampandChildress,whodefineautonomouspatientsaschooserswhoactintentionally,withunder-standing,andwithoutcontrollinginfluencesthatdeter-minetheiractions.Itthusassumesthatpatientsarerationalandreasonableagents.Proponentsofthisconceptofautonomybelievethataphysiciansobligationtorespectautonomyoutweighs,toacertaindegree,profes-sionalobligationsofbeneficence.Autonomyascriticalreflection,withinthetheoryofproceduralindependence,isnormativeinitscriteriaforofautonomousdecisionmaking.Ifthecriteriaofproceduralindependencearesatisfied,thisapproachoffersroomforconscioussubmissiontosomeformofexternalauthority(forexample,aphysician,areligion,aleader,etc).Inthisconcept,thepaternalisticphysiciancanplayarole.Centraltothisconceptisthecriticalreflectionondecisions.Aftercriticalreflectionontheirpreferences,patientsmaychoosetoletthephysiciandecidewhichtreatmentisbest.Actualautonomyasidentification,whichreferstotheprocessofidentificationwithonesactions.Itoriginatedinphenomenologicalapproachestohealthcare,inwhichidentificationreferstojudgingexperiencesinthelightofonesvalueorientationtowardsactions.Itisimportantinthisinterpretationforapatienttoidentifyhimorherself *Theauthorsviewedthefourthmodel„theinterpretivemodel„tohavelimitedrelevanceformedicalconsultationsascompared(forexample)withpsychotherapy.Aninstrumenttoassesspatientsandphysiciansidealsofpatientautonomy269www.jmedethics.com withthedecisionmade,notsomuchtomakethedecisionhimorherself.Identityrefersnotsomuchtoidentityasanindependentbeing,buttoidentityasasocialbeinginfluencedbyemotionsandbytieswithothers.Dependenceisnotproblematicifonecanidentifywiththesourceuponwhichorwhomoneisdependent.Socraticautonomy,whichstressestheimportanceofcaring,andthuslinkstheconceptofautonomywithexistentialfragilityandpatientsvulnerability.Thecoreisnotanethicsofpersonalresponsibilityandselfdetermi-nation,butanethicsofcareandofinterdependence.1718Autonomyasnegotiatedconsentreferstointerpersonalandsocialcommunication,insteadofindividualandrationalconsiderations.Communicationshouldbeunder-stoodasaprocessofnegotiationinwhichpeopletrytodefinethemeaningofthesituationinwhichtheyareinvolved.Patientsandphysiciansshouldhaveanintentiontomutualunderstandingandtoexplicitandjustcommu-nication.Perhapsthistheorybestfitsmanyoftheideasonshareddecisionmakingthathaveevolvedinrecentyears.Usingthetheories,wedeveloped55initialstatementstobeansweredona5pointscale(rangingfromfullydisagreetofullyagree).Asstatedintheintroduction,wewerealsointerestedinthespecificroleofriskinformation,becauseinformationisconsistentlyassociatedwithautonomyintheliterature.Further,riskinformationformsanimportantcomponentofourdecisionsupport.Insomeofthemoralconceptsitisnotevidentwhattheroleofinformationwouldbeintheidealofautonomy.Thereforetheitemsthatpertaintoriskinformationarenotclearlylinkedtotheautonomyconcepts.Thestatementswerepilotedinthreesamples:instudentsofhighlevelvocationaltraininginmedicine;ingeneralpractitionersduringacourseincontinuingmedicaleducation;andinaneurysmpatients.Inthispilot,wehadtriedtorepresenteachtheorywithasimilarnumberofitems.Becauseoflackinvarianceofsomeoftheseitems,wedecidedtolimitthequestionnairetoitemsshowingsufficientvariance.Further,theoriginalsetofstatementsincludeditemsthatdidnotdiscriminatewellbetweenthetheories,asthetheoriesinsomerespectsshowoverlap.Followingthepilotwethereforedecidedtousethedifferentelementsfromthetheoriesmerelytoobtainawiderangeofpossibleviewsoftheidealofautonomythatpatientsanddoctorsmayhold.Thus,ouraimwasnottodevelopsubscalesthatexactlyrepresentmoraltheories,butsubscalesthatmayresultinanormativemapofpossiblepatient-physicianrelationships,basedonthesetheories.Theproposedfutureuseforthequestionnaireistocomparepatientsandsurgeonspositionsonsuchanormativemap(Molewijk,unpublishedobserva-tions),andthereforethequestionnairewasdevelopedforandevaluatedinbothgroups.Onthebasisoftheresponsestothisquestionnaireandpsychometricanalysesthereof,afinallistof22statementswascreated(seeAppendix1).Twoversionsoftheinstructionswerecreated:apatientversionandaphysicianversion.Inbothversions,therespondentswereaskedwhattheythoughtconstit

utedagoodpatient-physicianrelationship,basedontheirnormsandvalues.Itwasexplicitlystatedthattheyshouldnotindicatewhathadbeentheirexperiencesofar,orwhattheythoughtwouldbefeasibleormostcommonlyencountered.Thesurgeonswereaddressedasphysicians„theywerenotaskedtoimaginebeingpatients.Includedasavaliditycheck(asthefinalitemofthequestionnaire)wasaquestiononthepatientspreferenceforparticipationintreatmentdecisionmaking,whichwasdevelopedbySutherland.Thisitemasksforthepatientspreferreddecisionmakingroleona5pointscale.Thequestionnairewastestedintwosamples:asampleofaneurysmpatientsandasampleofsurgeonsandsurgicalresidents.Thepatientswereparticipantsinarandomisedtrialontheeffectofindividualisedevidencebaseddecisionsupportonthedecisionmakingaboutelectiveaneurysmsurgery.PatientswererecruitedfromtheLeidenUniversityMedicalCenter,andfromtheWesteindeandtheLeyenburgHospitalsinTheHague,allintheWestoftheNetherlands.Patientsreceivedthequestionnairebymailaroundthetimeoftheirfirstappointmentwiththevascularsurgeon.Aprepaidenvelopewassentforreturningthefilledoutquestionnaire.Thestudywasapprovedbytheresearchethicscommitteesofallthreehospitals.Thesurgeonswerethoseparticipatinginthelargerresearchproject,aswellassurgicalresidentsfromthesameregion(betweenyears1and6oftheirtraining).Thesurgeonsandresidentswerealsosentthequestionnairewithaprepaidreturnenvelope.DataanalysisAglobaloverviewofthedataanalysisispresentedhere.Fortechnicaldetails,thereaderisreferredtoAppendix2.Thefirststepinitemreductionwasbasedonextremenessofscores.Itemsthatmostrespondentsagreewitharenotveryinformativeandwerethereforeomitted.Next,afactoranalysiswascarriedoutonthereducedquestionnairetoseewhethersubscales(factors)couldbeformedthatwererelatedinsomewaytotheautonomyconcepts.Thefactorsthusfoundweresubmittedtoasecondroundofitemreduction.Weeliminateditemswithineachfactor(subscale)separately,inastepwisefashion,basedonCronbachsmeasureofinternalconsistency.Inthisprocessofdeletingredundantitemstheitemreductionwasnotbasedpurelyonstatisticalgrounds„thecontentoftheitemswasalsotakenintoaccount.Inthiswaywetriedtocreateashortquestion-nairethatcoveredtherelevantdomains.Weanalyseddataforpatientsandclinicianstogether,butwealsoassessedfactorstructureandreliabilityforthetwogroupsseparatelytoseewhetherdifferenceswereobserved.Indecidingtodeleteanitemornotwealsotooktheinternalconsistencyinthetwosubgroupsintoaccount.Summaryscalescoreswerecalculatedbysummingtheitemscores(withoutweighting)andtransformingthesetoa0…100scale.AssociationwiththeSutherlandquestionwasassessedbycorrelationsandone-wayanalysisofvariance.Atotalof160patientswereeligiblefortherandomisedtrial:17(11%)ofwhomrefusedtoparticipate(inthetrial).Ofthe143patientsthusavailableforthissubstudy,23(16%)didnotreturntheirquestionnairewhereas18(15%)ofthereturnedquestionnaireswereunusablebecausetheauton-omyquestionnairehadnotbeenfilledin.Oftheremaining102patients(71%),onepatientwhohadmissingvaluesformorethan25%ofthe22itemswasomittedfromtheanalysis.Fivepatientswhoagreedfully(score=5)to20(90%)ormoreoftheitemswereomittedfromtheanalyses.Thisleftuswith96patientsavailableforanalysis.Meanagewas72(SD8)years,90%weremale,and44%hadlower,40%intermediate,and16%highereducation.Fiftyeightofthe71surgeonsandresidentsapproachedreplied(82%). Inourpilotstudy,wehadusedthetermyouridealofthepatient-physicianrelation,butthisturnedouttobetooabstractforthepatients.Afterthoroughdiscussionwedecidedtousethetermgoodrelationinstead,withthequalificationsasgivenabove(MolewijkAC,unpublisheddata).270Stiggelbout,Molewijk,Otten,etalwww.jmedethics.com ItemreductionThefirstcriterionforitemreductionappliedwasextremityofmeanscore.Wedeletedfouritems(2,6,9,17)becausetheyhadameanofgreaterthan4.5andstandarddeviationoflessthan1.0.Inthesubsequentfactoranalysis,fourfactorswerefound(seetable1;seeAppendix2fordetailsofthestatisticalanalysis).Itemreductionbasedoninternalconsistencyproduced4scales(seetable2).Thefirst,andmostreliablefactor(=0.83),wasfoundtobeafactordescribingtheopinionthatthephysicianshoulddecideandreflectedarelationshipbasedontrust.Wewillrefertothisfactorasthedoctorknowsbestfactor.Item1(Thedoctormustchoosethetreatmentwiththeleasthealthrisk,evenwhenthepatientwantsatreatmentwithmorehealthrisks)alsoloadedonthisfactor,buteliminatingthisitemdidnotreducereliabilityoverall,andimprovedthereliabilityforsurgeons.Item8(Everypatientisobligedtoactivelythinkabouttheappropriatetreatment)loadedmoststronglyonthisfactor,butthefactorloadingwasonlymoderate(0.41)andnotmuchhigherthantheitemsloadingontheotherthreefactors.Further,inclusionofthisitemreducedinternalconsistencyandwasthereforedecidedagainst.Thesecondfactor=0.62)reflectedtheviewthatthepatientshoulddecide.Thethirdfactor(=0.66)describedtheviewthatapatientisentitledtothewishnottoparticipate,andtothewishnottoreceivethreateninginformation.Thefourthfactorconsistedoftwoitems(items5and14)thatreflectrequirementsforriskinformation=0.63).Thefourf

actorsexplained19%,13%,11%,and11%,respectively,ofthevarianceafterrotation.Anitemthatwewouldhavepreferredtoretainisitem7:Ifdoctorandpatientproperlyconsultwitheachother,itdoesnotmatterwhomakesthefinaldecision.Itdidnotclearlyfitwithanyofthescales,butconcernedanimportantconceptfromtheethicsliterature.Wehadfeltittoreflectboththetheoryofnegotiatedconsentandthatofactualidentification.Inthefactoranalysisitwasoriginallygroupedwithitem10(whichwehopedwouldreflectproceduralindependence/criticalreflectionandSocraticautonomy),andwhenforcingfourfactorsitwasgroupedwithitems5and14(themorelegalaspectsofinformationprovision).Forbothofthesecombinations,reliabilitywasinadequatewhentheitemwasincluded.Otheritemsthatwereexpectedtocorrelatewiththisitem(thatweremeanttoreflectnegotiatedconsentoridentification„seeAppendix1)wereitemsthathadbeenendorsedbyalmosteveryone,andhadbeendeletedinthefirstroundofitemreduction„forexample,items2and17).Inquestionnairedevelopmentitisnotrecommendedtokeepasingleitem(forreasonsofreliability)andthereforeitwasdeleted.Aswewereconcernedaboutthepositiveansweringtendencyamongpatients,wetestedtheimpactofagreement.Weassessedhowmanyparticipantsfullyagreedwithfourpairsofitemsthatwefeltwereunlikelytoelicitfullagreementinbothinstancesinapair(items1and21,items8and11,items8and15,anditems10and13).Thoughstrictlyspeakingtheseitemsarenotoppositesofascale,wefeltthatagreeingtobothitemsinapairindicatedacquiescencebias.Fivesurgeonshadfullyagreedwithbothitemsinonepaironly,and20,17,andninepatientshadfullyagreedwithbothitemsinone,two,andthreepairs,respectively.Afactoranalysiswithoutthe26patientswhohadagreedtotwoormoreoftheabovepairsresultedinasimilarstructureandsimilarreliabilities,exceptforareducedreliabilityinpatientsforthenon-participationfactor,from0.59to0.52.CorrelationamongscalesApositivecorrelationwasseenbetweenthedoctorknowsbestscaleontheonehand,andboththerighttonon-participationscaleandtheobligatoryriskinformation Table1Factorstructureafterexploratoryfactoranalysis(n=144)IIIIIIIV20.Thedoctorcanpresumethatthepatientknowsthatpeoplecandieduringseriousoperations.16.Ifdoctorandpatientcannotagreeonwhichtreatmentisbest,thedoctorshouldmakethefinaldecision.18.Thepatientshould,withoutmuchinformationontheriskinvolved,confidentlyundergoanoperation.12.Duringtheconversation,thepatientmustsubmithimselfwithconfidencetotheexpertiseofthedoctor.0.650.434.Itisbetterthatthedoctorratherthanthepatientdecideswhichthebesttreatmentis.1.Thedoctormustchoosethetreatmentwiththeleasthealthrisk,evenwhenthepatientwantsatreatmentwithmorehealthrisks.8.Everypatientisobligedtoactivelythinkabouttheappropriate0.410.3522.Asitconcernsthebodyandlifeofthepatient,thepatientshoulddecide.13.Thepatienthimselfmustchoosebetweenthevarioustreatments.0.6719.Itgoestoofarwhenthedoctordecideswhichtreatmentisbestforthepatient.21.Ifapatientchoosesatreatmentwithmorehealthrisks,thedoctorshouldrespectthistreatmentdecision.10.Itgoestoofarwhenpatientsthemselveshavetodecidewhichtreatmentisbestforthem.11.Patientsshouldhavetherightnottobeinvolvedinthedecisiononthetreatment.15.Patientswhobecomeafraidwhenthinkingaboutthetreatmentdecisionshouldbeleftinpeacebythedoctor.3.Ifthepatientdoesnotwanttoreceiveinformationaboutrisksthedoctorshouldrespectthis.14.Beforeapatientconsentstoatreatmentheshouldreceiveallinformationontherisksinvolved.5.ThepatienthastobeinformedonalltherisksinvolvedinanOnlyfactorloadingsgreaterthan0.3aredisplayed.Aninstrumenttoassesspatientsandphysiciansidealsofpatientautonomy271www.jmedethics.com scaleontheother(seetable3).Apositivecorrelationwasalsoseenbetweenthepatientshoulddecidescaleandtherighttonon-participationscale.Thecorrelationbetweentheobligatoryriskinformationscaleandtherighttonon-participationscalewasonlymoderate.AssociationwiththeSutherlandquestionAsthedistributionoftheSutherlandquestionwasskewed(seetable4),weassessedtheassociationwiththescalesnotonlybycorrelationsbutalsobymeansofonewayanalysisofvarianceusingthetestforlineartrend.Theassociationswiththefourscaleswereasexpected(seetables3and4).Therighttonon-participationscaleandtheobligatoryriskinformationscalewerenotlinearlyassociatedwiththeitem.Thepatientshoulddecidescalemostclearlyshowedtheexpectedassociation(=0.45,p0.01),butthedoctorknowsbestscalealsoshoweda(negative)association0.22,p0.05).Indeed,intheAnalysisofVariance(ANOVA)alineartrendwasseenforthesetwoscales.DISCUSSIONWedevelopedtheIdealPatientAutonomyScaletoassessidealsofpatientautonomyfromabroaderperspectivethanthatofliberalindividualismalone.Basedonfactoranalysisandreliabilityanalyses,theoriginalquestionnaireof22itemswasreducedto14itemsthatformedfourscaleswithmoderatetogoodreliability.Byincorporatingaspectsfromvariousethicaltheoriesasubtlerpictureofthephysician-patientroleemergesthanisgenerallysuggestedbythemedicalliterature.Twoofthefourfactorsreflectthewellknowndistinctionbetweenphysician-centredandpatient-centreddec

isionmaking.Thethirdisasubtle,butimportantaspectofthepatient-physicianrelationship,namelytherightofthepatienttodecidewhethertodelegatethedecisionbacktothephysician.Thefourthreflectstherightandthedutyofpatientstobemadeawareofrelevantrisks.Whenwecomparethefactorstructurewithouroriginalclassificationoftheitemsaccordingtoethicaltheories,aremarkshouldbemade.Inthepilotstudywehadabandonedtheideathatwecouldcoveralltheorieswithequalnumbersofstatements.Theoriginalsetofstatementsincludedmanyitemsthatdidnotdiscriminatebetweentheories,andseveralitemsshowednovariationbetweenrespondents.Thus,wedonotpretendthatouroriginalsetofitemsreflectedthetheoriesinabalancedway.Alimitationofourstudyistheselectivesampleofpatients.Themeanageofourpopulationwashighandthelevelofeducationwaslow.Patientswereselfselectedinthattheywerewillingtoparticipateinourtrial,butasonly11%refusedparticipationthiswillnothaveintroducedalargebias.Justover30%,however,didnotreturnthequestion-naireordidnotfillouttheIdealPatientAutonomyScale.Furthertestinginotherpatientpopulationsisthusneeded.Afactorthatcorrespondedcloselytoourpreconceivedideasaboutalternativeautonomyconceptswastherighttonon-participationfactor.Allthreeitemsreflectedtosomeextenttheconceptofproceduralindependence,andtwocouldalsobereferredtoastoSocraticautonomy(seeAppendix1).Thisfactormaybeseenasthecounterpartoftheidealpatientintheliberalindividualistsense:theviewthatapatientisentitledtonotparticipateandtonotreceiveinformationthatcausesfear.ThedoctorknowsbestfactorencompassessomeoftheitemsthatwehadclassifiedbeforehandasSocraticautonomyorasproceduralindepen-dence.Theytherebystronglyincorporatetheconceptoftrust„animportantvalueinmedicine.Thepatientshoulddecidefactorcorrespondedquitewellwithourconcepts:allfouritemshadbeenclassifiedbeforehandas(tosomeextent)liberalindividualist,eventhoughoneitem(number22)couldalsobeseenasidentification.Thedoctorknowsbestscaleandthepatientshoulddecidescaleemergedastwodistinctfactorsthatseemtorepresenttheoppositesofacontinuum.Forcingthetwofactorsintoonedidreducethereliabilityto0.71,andsubsequentstepwiseeliminationofitemsledtotheoriginaldoctorknowsbestscale(with0.83).Thus,thepatientshoulddecidescalereflectsadistinctconcept,whichisinaccordancewithourexpecta-tionsasdiscussedabove.Theobligatoryriskinformationfactor,finally,consistedoftwoitemsthatwehadbothclassifiedaprioriasliberaljuridical.Fouritemsweredeletedbecausealmostallparticipantsagreedwiththem;theystatedtheobvious.Threeitems(1,8,and10)couldbedeletedfromthescalestoimprovereliabilityastheircontentwassufficientlycoveredbyotheritemsinthequestionnaire.Anitemwedeletedfromtheanalysisforreasonsofreliabilityisitem7:Ifdoctorandpatientproperlyconsultwitheachother,itdoesnotmatterwhomakesthefinaldecision.Wehadfeltthatitreflectedboththetheoryofnegotiatedconsentandthatofactualidentification.Otherswho Table3Correlationamongscales(Pearsons),n=144DKBPatDnonPRIDoctorknowsbest(DKB)Patientshoulddecide(PatD)Righttonon-participation(nonP)0.43Obligatoryriskinformation(RI)0.310.120.20Sutherlanditem0.030.020.01(twosided);0.05(twosided). Table2ThefoursubscalesoftheIdealPatientAutonomyScale(IPAS)ScaleIDoctorknowsbest:=0.83(surgeons:=0.59;patients:=0.70)16.Ifdoctorandpatientcannotagreeonwhichtreatmentisbest,thedoctorshouldmakethetreatmentdecision.4.Itisbetterthatthedoctorratherthanthepatientdecideswhichisthebesttreatment.12.Duringtheconversation,thepatientmustsubmithimselfwithconfidencetotheexpertiseofthedoctor.20.Thedoctorcanpresumethatthepatientknowsthatpeoplecandieduringseriousoperations.18.Thepatientshould,withoutmuchinformationontheriskinvolved,confidentlyundergoanoperation.ScaleIIPatientshoulddecide:=0.62(surgeons:=0.62;patients:=0.63)13.Thepatienthimselfmustchoosebetweenthevarioustreatments.21.Ifapatientchoosesatreatmentwithmorehealthrisks,thedoctorshouldrespectthistreatmentdecision.19.Itgoestoofarwhenthedoctordecideswhichtreatmentisbestforthe22.Asitconcernsthebodyandlifeofthepatient,thepatientshouldScaleIIIRighttonon-participation:=0.66(surgeons:=0.57;patients:=0.59)3.Ifthepatientdoesnotwanttoreceiveinformationaboutrisks,thedoctorshouldrespectthis.15.Patientswhobecomeafraidwhenthinkingaboutthetreatmentdecisionshouldbeleftinpeacebythedoctor.11.PatientsshouldhavetherightnottobeinvolvedinthedecisionontheScaleIVObligatoryriskinformation:=0.63(surgeons:=0.54;patients:=0.54)5.Thepatienthastobeinformedonalltherisksinvolvedinanoperation.14.Beforeapatientconsentstoatreatmentheshouldreceiveallinformationontherisksinvolved.272Stiggelbout,Molewijk,Otten,etalwww.jmedethics.com wouldliketousethequestionnaire,andwhovaluethementionedconcepts(MolewijkAC,unpublisheddata),shouldconsiderkeepingthisitemandaddingsimilaritemstoformareliablescale.Thepositivecorrelationbetweenthedoctorknowsbestscaleandtherighttonon-participationscalewasintheexpecteddirection.Thepositivecorrelationbetweenthephysiciandecidesscaleandtheobligator

yriskinforma-tionscaleconfirmsthefindingsintheliteraturethatinformationprovisionanddecisionmakingaredistinctconcepts,andthatthephysicianwhodecidesforthepatientsshouldstillfulfilallrequirementsregardingriskdisclosure.Thepositivecorrelationbetweenthepatientshoulddecidescaleandtherighttonon-participationscalemaybeexplainedbythefactthatbothscalesencompassitemsstatingthatthepatientswordislaw.Thiscorrelationmayalsoindicatethatthosewhobelievethatapatientshoulddecidealsobelievethatincertainsituationsthepatientmaydecidetodelegatehisorherdecisiontothephysician„aviewthatisreflectedamongothersinthetheoryofproceduralindependence.ThesehypothesesdeservefurtherOurfinalquestionnaireconsistsof14items,whichformthescalespresentedintable2.Asstatedabove,item7couldwellberetainedbutsimilaritemsshouldthenpreferablybeaddedtocreateareliablescale.Threeofthescaleshaveanoflessthan0.70,whichislessthanideal.Butgiventheearlystagesofourvalidationresearch,thehighlyabstractnatureoftheconceptsthatwewishtoassess,andthefactthatourpatientshadgenerallyloweducationlevels(andthereforemayhavehaddifficultyunderstandingtheitems,whichmayhavereducedtheinternalconsistencyoftheresponses),webelieveittobeacceptable.Itwillbeworththetimeandefforttoincreasethenumberofitemsandreducemeasurementerrorinotherways.Further,reversescoringofsomeoftheitemsshouldreduceacquiescence.Mostcorrelationsamongscales,andbetweenthescalesandtheSutherlandquestion,supportedthevalidityofthescale.Wedonotknowwhetherthefactorstructurethatwefoundcanbereproducedinothercultures,becauseidealsofautonomydonotonlydifferbyculture,butperceptionsofaspectsasnormativelyinherenttoautonomymayalsobedifferent.Wethereforeurgeotherstoadaptandusethisscaleandtopresentpsychometricanalysestoobtainmoreinsightintothiscomplicatedbuthighlyinterestingconcept.Thescalecanthenbeusedtocomparepopulationswithrespecttotheirviewsonautonomy,andtoassesswhatpatientorphysiciancharacteristicsarecorrelatedwithcertainidealtypes.Asstatedintheintroduction,webelievethatsocialpracticeshouldbeusedasamoralsourceasmoralityderivesitsmeaningfromthecontext.Useofourquestionnairemayenrichthecurrentdebatesontheethicalaspectsofevidencebasedpatientchoiceandshareddecisionmaking.2024Inourstudy,supportwasclearlyfoundforotherconceptsofautonomythantheliberalindividualist,suchasthatofproceduralindependence.Thisconceptdoesnotfitverywelltheparadigmofevidencebasedpatientchoice.Thus,moreinsightintotheviewsonautonomythatarefoundinpracticemayontheonehandhelpsharpenethicaltheory.Ontheotherhand,itmaypointtobeliefsandattitudesinpracticethatarelessideal,fromanormativeviewpoint,andneedmodification.AuthorsaffiliationsAMStiggelbout,ACMolewijk,WOtten,JKievit,DepartmentofMedicalDecisionMaking,oftheLeidenUniversityMedicalCenter,Leiden,TheNetherlandsJHvanBockel,DepartmentofSurgeryoftheLeidenUniversityMedicalCenter,Leiden,TheNetherlandsDRMTimmermans,DepartmentofSocialMedicineoftheFreeUniversityMedicalCenter,Amsterdam,TheNetherlandsSupportedbyagrantoftheNetherlandsCouncilforScientificResearchAPPENDIX1ORIGINALPATIENTAUTONOMYSeenextpage.APPENDIX2DETAILSOFTHESTATISTICALIntheinitialexploratoryfactoranalysis,fivefactorswerefoundaftervarimaxrotation,witheigenvaluesgreaterthan1.0,ofwhichthelasttwoeachconsistedoftwoitemsonly(items5and14,anditems7and10,respectively).Becausethelatterfactorhadverypoorreliability(=0.27),andasthefactors5to18laymoreorlessonastraightlineonthescreeplot(whichseparatesimportantearlyfactorsfromrandomerrorinthelaterfactors,andusuallyresultsinfewerfactorsthanasolutionbasedoneigenvalues),weforcedtheproceduretoextractfourfactors(seetable1).Weperformedanadditionalfactoranalysistoassesstheimpactofthedecisiontodeletethefouritemswithextremescores.Thesamefactorstructureemerged,andinasubsequentreliabilityanalysisthefouritemsweredeletedanyway. Table4Mean(standarddeviation)scalescoresbyleveloftheSutherlandquestion(n=144)N(%)DoctorknowsPatientshoulddecideRighttonon-participationObligatoryriskPhysicianshoulddecide,basedonallthatisknown1384.6(17.4)20.2(16.8)74.4(24.2)87.5(16.1)Physicianshoulddecide,stronglytakingthepatientsopinionintoaccount5555.9(30.2)43.4(24.1)51.5(28.1)79.8(23.0)Physicianandpatientshoulddecidetogether,onbasisof5063.3(29.9)58.8(19.6)62.7(29.3)82.8(26.1)Patientshoulddecide,stronglytakingthephysiciansopinionintoaccount1946.8(30.4)62.8(21.3)61.0(29.9)88.2(16.4)Patientshoulddecide,basedonallthatheorsheknowsorhearsaboutthetreatments743.6(31.2)60.7(20.0)47.6(26.2)78.6(22.5)Significantdifferencebetweenlevels(p=0.003)andlineartrend(p=0.008).Significantdifferencebetweenlevels(p0.001)andlineartrend(p0.001).Significantdifferencebetweenlevels(p=0.05).Aninstrumenttoassesspatientsandphysiciansidealsofpatientautonomy273www.jmedethics.com Theresultingsummaryscalescoresareobtainedbysum-mingtheitemscores(withoutweighting)andtransformingthesetoa0…100scaleasfollows:summaryscore=(100(scalescoreminimumscore)/(maximumscoreminimumscore)).BeauchampTL,ChildressJF.Principlesofbi

omedicalethics.NewYork,Oxford:OxfordUniversityPress,1994.EndeJ,KazisL,AshA,etal.Measuringpatientsdesireforautonomy:decisionmakingandinformation-seekingpreferencesamongmedicalJGenInternMed:23…30.BenbassatJ,PilpelD,TidharM.Patientspreferencesforparticipationinclinicaldecision-making:areviewofpublishedsurveys.BehavMedNeaseRFJr,BrooksWB.Patientdesireforinformationanddecision-makinginhealthcaredecisions:theAutonomyPreferenceIndexandtheHealthOpinionSurvey.JGenInternMedEmanuelEJ,EmanuelLL.Fourmodelsofthephysician-patientrelationship.DworkinG.Thetheoryandpracticeofautonomy.NewYork:CambridgeUniversityPress,1988.AgichGJ.Reassessingautonomyinlong-termcare.HastingsCenterReportDegnerLF,SloanJA,VenkateshP.TheControlPreferencesScale.CanJNursEntwistleVA,SheldonTA,SowdenA,etal.Evidence-informedpatientchoice.PracticalissuesofinvolvingpatientsindecisionsabouthealthcareIntJTechnolAssessHealthCareMolewijkAC,StiggelboutAM,OttenW,etal.Implicitnormativityinevidence-basedmedicine.Apleaforintegratedempiricalethicsresearch.HealthCareAnalysisMolewijkAC,StiggelboutAM,OttenW,etal.Empiricaldataandmoraltheory.Apleaforintegratedempiricalethics.Medicine,HealthCareandPhilosophy.AEuropeanJournal2004;(inpress).WeaverGR,TrevinoLK.Normativeandempiricalbusinessethics:separation,marriageofconvenience,ormarriageofnecessityBusinessEthicsQuarterly1994;:129…43.FalkumE,FørdeR.Paternalism,patientautonomy,andmoraldeliberationinthephysician-patientrelationship.AttitudesamongNorwegianphysicians.SocSciMed:239…48.KrupatE,YeagerCM,PutnamS.Patientroleorientations,doctor-patientfit,andvisitsatisfaction.PsychologyandHealth707…19.SutherlandHJ,Llewellyn-ThomasHA,LockwoodGA,etal.Cancerpatients:theirdesireforinformationandparticipationintreatmentdecisions.JRSoc:260…3.BerlinI.Fouressaysonliberty.NewYork:OxfordUniversityPress,1969.FoucaultM.Dezorgvoorzichzelf.(LesoucideSoi)Geschiedenisvandesexualiteit.Deel3.Nijmegen:SUN,1985.ManschotH.Kwetsbareautonomie:Overafhankelijkheidenonafhankelijkheidindeethiekvandezorg.[Vulnerableautonomy:Ondependenceandindependenceintheethicsofcare].In:ManschotH,VerkerkM,eds.Ethiekvandezorg:Eendiscussie[Ethicsofcare:adiscussion].Amsterdam:Boom,1994:97…118.MoodyHR.Ethicsinanagingsociety.Baltimore/London:TheJohnsHopkinsUniversityPress,1992.CharlesC,GafniA,WhelanT.Decision-makinginthephysician-patientencounter:revisitingthesharedtreatmentdecision-makingmodel.SocSci:651…61.BlandJM.AltmanDG.Cronbachsalpha.KrosnickJA.Surveyresearch.AnnuRevPsychol:537…67.SmithR.Medicinescorevalues.309:1247…8.CoulterA.Partnershipswithpatients:theprosandconsofsharedclinicaldecision-making.JHealthServResPolicy Appendix1OriginalpatientautonomyquestionnaireNormativestatementRetainedinfinal1.Thedoctormustchoosethetreatmentwiththeleasthealthrisk,evenwhenthepatientwantsatreatmentwithmorehealthrisks.LJNo2.Itisgoodifthepatientsasksthedoctorcriticalquestions.LI,NCNo3.Ifthepatientdoesnotwanttoreceiveinformationaboutrisks,thedoctorshouldrespectthis.LI,NCYes4.Itisbetterthatthedoctorratherthanthepatientdecideswhichthebesttreatmentis.LJ,SA,PIYes5.Thepatienthastobeinformedonalltherisksinvolvedinanoperation.LJ,LIYes6.Thedoctorshouldnotonlypayattentiontothemedicalaspects;heorsheshouldalsopayattentiontothepersonalexperiencesofthepatient.I,SANo7.Ifdoctorandpatientproperlyconsultwitheachother,itdoesnotmatterwhotakesthefinaldecision.NC,I,SA,PISeeDiscussion8.Everypatientisobligedtoactivelythinkabouttheappropriatetreatment.LI,NC,INo9.Ifpatientsdonotunderstandsomethingtheyshouldtelltheirdoctorthis.NC,PINo10.Itgoestoofarwhenpatientsthemselveshavetodecidewhichtreatmentisbestforthem.PI,SANo11.PatientsshouldhavetherightnottobeinvolvedinthedecisiononthePI,SAYes12.Duringtheconversation,thepatientmustsubmithimselfwithconfidencetotheexpertiseofthedoctor.PI,SAYes13.Thepatienthimselfmustchoosebetweenthevarioustreatments.LIYes14.Beforeapatientconsentstoatreatmentheshouldreceiveallinformationontherisksinvolved.LJYes15.Patientswhobecomeafraidwhenthinkingaboutthetreatmentdecisionshouldbeleftinpeacebythedoctor.SA,PIYes16.Ifdoctorandpatientcannotagreeonwhichtreatmentisbest,thedoctorshouldtakethefinaldecision.LJ,LIYes17.Beforethedecisionontheappropriatetreatmentistaken,thegoalsandwishesofthepatientmustbeclearlyknown.NC,SA,INo18.Thepatientshould,withoutmuchinformationontheriskinvolved,confidentlyundergoanoperation.PI,SAYes19.Itgoestoofarwhenthedoctordecideswhichtreatmentisbestforthepatient.LI,LJYes20.Thedoctorcanpresumethatthepatientknowsthatpeoplecandieduringseriousoperations.LJ,PIYes21.Ifapatientchoosesatreatmentwithmorehealthrisks,thedoctorshouldrespectthistreatmentdecision.LI,LJYes22.Asitconcernsthebodyandlifeofthepatient,thepatientshoulddecide.LI,IYesI,actualidentification;LJ,legalliberal;LI,liberalindividualism;PI,proceduralindependence(criticalreflection);NC,negotiatedconsent;SA,Socraticautonomy(existentialfragility);asclassifiedindependentlybytworesearchers(AMS,ACM).274Stiggelbout,Molewijk,Otten,etalwww.jmedethics.