LAUREDOMBRECHT

Medicalend-of-lifedecisionsinstillbirths,neonatesandinfants

Apopulation-basedstudyonprevalenceestimates,viewsandexperiencesinFlanders,Belgium.

Thesissubmittedtofulfiltherequirementsfor

AjointPhDDegreeasDoctorin

HealthSciences

FacultyofMedicineandHealthSciences,GhentUniversity,2019-2020

SocialHealthSciences

FacultyofMedicineandPharmacy,VrijeUniversiteitBrussel,2019-2020

2

Promotors: Prof.dr.LucDeliens(UniversiteitGent)

Prof.dr.JoachimCohen(VrijeUniversiteitBrussel)

Co-promotors: Dr.KimBeernaert(UniversiteitGent)

Prof.dr.KennethChambaere(VrijeUniversiteitBrussel)

MembersoftheProjectGroup:

Prof.dr.FilipCools(UniversitairZiekenhuisBrussel)

Dr.LindeGoossens(UniversitairZiekenhuisGent)

Prof.dr.GunnarNaulaers(UniversitairZiekenhuisLeuven)

DoctoralJuryCommittee:

Chair: Prof.dr.ErnstRietzschel(UniversiteitGent)

Jurymembers: Prof.dr.FreddyMortier(UniversiteitGent)

Prof.dr.KristienRoelens(UniversiteitGent)

Prof.dr.KoenraadSmets(UniversiteitGent)

Prof.dr.KoenPardon(VrijeUniversiteitBrussel)

Prof.dr.EduardVerhagen(UniversiteitvanGroningen)

Prof.dr.KatrienBeeckman(UniversitairZiekenhuisBrussel)

3

MembersoftheFlemishNeonatalIntensiveCareUnitConsortium:

Prof.dr.FilipCools(UniversitairZiekenhuisBrussel)

Dr.LindeGoossens(UniversitairZiekenhuisGent)

Prof.dr.GunnarNaulaers(UniversitairZiekenhuisLeuven)

Prof.dr.LucCornette(AZSint-JanBrugge-Oostende)

Dr.SabineLaroche(UniversitairZiekenhuisAntwerpen)

Dr.ClaireTheyskens(ZiekenhuisOost-LimburgGenk)

Dr.ChristineVandeputte(ZiekenhuisGZASint-Augustinus)

Dr.HildeVandeBroek(ZiekenhuisZNAMiddelheim)

This doctoral thesis is supported by a grant from the Research Foundation Flanders (FWO;G041716N)andthespecialresearchfundofGhentUniversity(BOF;01J06915).

4

5

“There'snotragedyinlifelikethedeathofachild.Thingsnevergetbacktothewaytheywere.”

- DwightD.Eisenhower

ThisdissertationisdedicatedtomybeautifuldaughtersNova*andNisa.

6

Content

Dankwoord 8

Part1 Generalintroduction 11

Introduction 12

Part2 Medicalend-of-lifedecisionsinstillbirths,neonatesandinfantsinFlanders 37

Apost-mortempopulationsurveyonfoetal-infantileend-of-lifedecisions:aresearchprotocol 38

End-of-lifedecisionsinneonatesandinfants:arepeatedpopulation-levelmortalityfollow-backstudy 54

Part3 Attitudes,viewsandexperiencesofhealthcareprovidersonneonatalend-of-lifedecision-making 71

Neonatologistsandneonatalnurseshavepositiveattitudestowardsperinatalend-of-lifedecisions,anationwidesurvey 72

Barrierstoandfacilitatorsofend-of-lifedecision-makingbyneonatologistsandneonatalnursesinneonates:aqualitativestudy 92

Psychologicalsupportinend-of-lifedecision-makinginneonatalintensivecareunits:fullpopulationsurveyamongneonatologistsandneonatalnurses 110

Part4 Generaldiscussionandconclusion 121

Generaldiscussionandconclusion 122

Summaries 159

Englishsummary 160

Nederlandstaligesamenvatting 182

CurriculumVitaeandlistofpublicationsofLaureDombrecht 205

Appendix 209

Appendix1 VragenlijstMedischebeslissingenaanhetlevenseinde,nultotéénjarigen(inDutch) 210

Appendix2 VragenlijstMedischebeslissingenbijeendoodgeboorte(inDutch) 216

Appendix3 VragenlijstAttitudesvanartsenoverlevenseindebeslissingen(inDutch) 222

Appendix4 VragenlijstAttitudesvanverpleegkundigenoverlevenseindebeslissingen(inDutch) 228

Appendix5 TopicGuideinterviewstudieartsen(inDutch) 234

Appendix6 TopicGuideinterviewstudieverpleegkundigen(inDutch) 238

7

Chapters2-6arebasedonthefollowingpublications

Chapter2:Dombrecht,L.,Beernaert,K.,Roets,E.,Chambaere,K.,Cools,F.,Goossens,L.,Naulaers,G., De Catte, L., Cohen, J., Deliens, L. and on behalf of the NICU consortium. A posts-mortempopulationsurveyonfoetal-infantileend-of-lifedecisions:aresearchprotocol.BMCPediatrics,2018August.

Chapter3:DombrechtL.,BeernaertK.,ChambaereK.,CoolsF.,GoossensL.,NaulaersG.,CornetteL., Laroche S., Theyskens C., Vandeputte C., Van de Broek H., Cohen J., Deliens L. End-of-lifedecisions in neonates and infants: a repeated population-level mortality follow-back study(Submittedforpublication).

Chapter4:DombrechtL.,DeliensL.,ChambaereK.,BaesS.,CoolsF.,GoossensL.,NaulaersG.,RoetsE.,PietteV.,CohenJ.,BeernaertK.(andtheNICUconsortium).Neonatologistsandneonatalnurseshavepositiveattitudestowardsperinatalend-of-lifedecisions,anationwidesurvey.ActaPaediatrica,2019April.

Chapter5:DombrechtL.,PietteV.,DeliensL.,CoolsF.,ChambaereK.,GoossensL.,NaulaersG.,Cornette L., Beernaert K., Cohen J., and on behalf of the NICU consortium. Barriers to andfacilitatorsofend-of-lifedecision-makingbyneonatologistsandneonatalnursesinneonates:aqualitativestudy.JournalofPainandSymptomManagement,2019October.

Chapter6:DombrechtL.,Cohen J.,CoolsF.,DeliensL.,GoossensL.,NaulaersG.,BeernaertK.,ChambaereK.onbehalfoftheNICUconsortium.Psychologicalsupportinend-of-lifedecision-making in neonatal intensive care units: full population survey among neonatologists andneonatalnurses.PalliativeMedicine,2019November.

8

Dankwoord

Methetschrijvenvanditdankwoordlegikdelaatstehandaaneenprojectdatdeafgelopen5jaareenbelangrijkdeelvanmijnlevenisgeweest.Hetklinktmisschienwatcliché,maareenwerkzoalsditkomternietzonderdesteuneninzetvaneengevarieerdegroepgemotiveerdemensen,dieikgraageensextraindeverfzet.Eerstenvooralwilikallezorgverlenersbedankendiedevoorbijejarendeelgenomenhebbenaanonsonderzoek.Dedrivediejulliedagindaguittonenvoordekinderenenoudersopjulliedienstisbewonderenswaardig.Datjulliedaarnaasttijdvrijkondenmakenomde(somsuitgebreide)vragenlijstenintevulleneninterviewsaftenemenkanikalleenmaaruitvoerigbedanken.

Ditonderzoeksprojectwasernooitgekomenzonderdeeindelozeinzetenmotivatievanmijndagelijksbegeleiderenco-promotor,Kim.Alklinkthetmisschienvreemd,onderzoeknaarhetlevenseindeisjouophetlijfgeschreven.Jebentnietalleeneenongeloofelijkgoedeonderzoeker,maarjehebtookhethartopdejuisteplaats.Ikkondaarombijjouterecht,nietenkelommijnwetenschappelijke competenties uit te bouwen, maar ook om mijn hart te luchten als hetmoeilijkerging.Elkdocumentwerd,terecht,meteenkritischoognagelezen,waarnajemesteedsweesopzakendietochnetwatbeterkonden.Ikhebvanjougeleerdomsteedstereikennaarhetvolgendeniveau.Ikbendaaromongeloofelijktrotsomjeeerstebegeleidingalspostdoctemogenzijn,enikhoopdatwenogvelejarenmogensamenwerken.

Daarnaastwilikgraagmijnandere(co-)promotorenLuc,JoachimenKennethbedanken.Luc,bedanktommijdekanstegevenomditprojectsuccesvolaf tewerkenbinneneensterkeenhechteonderzoeksgroep.Jekritischeinbrengtijdensdeprojectgroepenheeftdeartikelsbinnendit proefschrift zeker naar een hoger niveau getild. Joachim, bedankt ommijn academischekennissteedsverderoptebouwen,ookalmoestikachterafsomsopzoekenwatjeprecieswouzeggen.Zohebiknietalleenopacademischenmethodologischvlakenormveelbijgeleerd,maarweetiknuookwateentehoogcortisolniveauzoalkanteweegbrengentijdensdelaatstemaandvanmijnzwangerschap.Kenneth,alseensoorttweedebegeleiderstondjesteedsklaaromraadtegeven,ensomsookombrandjes teblussen tijdensperiodesvanstress. Jouwnieuweprof-bureauheeftmisschienslechtsrecentelijk‘therapiestoeltjes’,jouwdeurstonddaarvoorookaltijdopenwanneerikhetnodighad.Bedankthiervoor.

Naastdeonderzoekershebikookenormveelbijgeleerdvandeartsendienauwbetrokkenwarenbijditdoctoraatsproject.Filip,LindeenGunnar,zonderjullieinbrengenhardewerkwasditdoctoraatnooittotstandgekomen.DemedischevaktermenwareninhetbeginvoormijdanookChinees,maarnubenikvolledigmeewanneerertijdensdeinterviewsnonchalantvaneenNECwordt gesproken. De klinische bril opzetten was voor ons als onderzoekers soms nietvanzelfsprekend,endaarombenikheelblijdatjulliebeschikbaarwarenomsomsevenaandealarmbeltetrekkenenonzedatateschetsenindedagelijksemedischepraktijk.

Aan alle consortiumleden vande8NICU’s,dr. Filip Cools, dr. LindeGoossens,dr.GunnarNaulaers, dr. Luc Cornette, dr. Sabine Laroche, dr. Claire Theyskens, dr. ChristineVandeputteendr.HildeVandeBroek,bedanktvoordeinteressantedebattenenreflectiestijdensdeconsortiummeetingsenbezoekenaanjullieziekenhuisdienstdoorheendejaren.

9

Dekrijtlijnenvanhetdoctoraatsprojectaandeneonatalekantstondenvanafhetbeginalietwatduidelijkerafgetekend.Maarprenataaltasttenwebijdestartvanditprojectnoginhetduister.Daaromwilikookgraagdebetrokkengynaecologen,dr.EllenRoets,dr.LucDeCatteendr.LeonardoGucciardobedankenvoorhunonmisbaresteunenkennis.Ellen,aanhetbeginvanditprojecthadikbewonderingvoorjeklinischeexpertise,kennisendrive.Ondertussenbenikook blij om je bij mijn vriendenkring te rekenen. Ik kijk er naar uit om jouw doctoraat alscomplementairnaasthetmijneindekastteplaatsen.

Verderwil ik graagde ledenvandedoctoraatsjury bedankenommijnproefschriftmet eenkritischebliktebekijkenentebeoordelen.

Alsferventefanvanon-campuswerken,deeldeikenormveeltijdmetbureaucollega’s.Lenzo,SigridenCharlotte,bedanktommevanafhetbeginoptevangen.Wewetenallemaalwaardepanda’shunoriginehadden,enikvindhetnogsteeds jammerdatonzebureau-jungleernietgekomenis.EenpaarbureauwisselslaterisLenzodeenigeconstante(duidelijkteherkennenaanhetrommeligebureau).Gelukkigdusmaardatsamenwerkenmetjouzofijnis.Isabel,develebabbelsenWest-Vlaamseinbrengfleurdenelkewerkdagsamenop,albeniktochblijdatjenietmetdetractornaarhetwerkkomt.Enmoestjeernogaantwijfelen:jehoortgewoonbijons!Ook bedankt aan de iets vluchtigere bureau-collega’s: Steven, Mariëtte en Sarah. Hetpsychologenbureau,formerlyknownasbureauK1,KimE.,Naomi,MariëtteenGaelle:alwetenjulliealsgeenanderhoeveelstresserbijeendoctoraatkomtkijken,tochheeftelkvanjulliemeopeeneigenmaniergeleerdhoebelangrijkwork-lifebalancekanzijn.Anne-Lore,weroldensameninhetavontuurvaneendoctoraatbijZrLenookalzienonzeprojectenertotaalandersuit, ik konaltijd evenbinnenspringenophet4e voordenodige raad en/of afleiding.Veerle,bedanktvoorallehulpendevlottesamenwerkingbijhetafnemenvandeinterviews,goteam-kinderlijn!Aanalleanderepanda’svanZrLGent-divisie:bedanktvoordeafter-workdrinks,delunchgesprekken en de hallway-encounters. ZrL-Jette collega’s (special shout-out aan allejuniors), al zien we elkaar minder, de werkbesprekingen, seminaries, uitjes en st(r)afdagenzoudenniethetzelfdezijnzonderjullie.InhetbijzonderookbedanktaanGeertjeenNadineomonzesteedsgroterwordendeonderzoeksgroepaltijdingoeiebanenteleiden.

Jolien,ikkanmegeenbeterebestevriendinwensen.MijnsteunentoeverlaatsindsvervlogenTWA-tijden.NovaenNisaboffenmaarmetzoeenfantastischemeter.Aanmijnlievevriendenvanhetjeugdorkest,wehebbensamenzoveelwoeligewaterendoorzwommen.Ikhebhetgelukomdeeluittemakenvanzoeengevarieerdevriendengroep,waardooriederopzijnmanierkonhelpenomsomswatdrukvandeketeltehalen.Ikkanmevoorstellendatmijndoctoraatsprojectvoorsommigenergabstractoverkwam,maartochstondenjulliesteedsklaarmeteenluisterendoor. Ookmijnboekenclub-vriendjes kon ik onmogelijk vergeten. Onze gemeenschappelijkeliefdevoorboekenbrachtonssamen,maarondertussenzijnwegeëvolueerdnaareenoprechtclubjevriendinnenwaariknietsandersdandankbaarvoorkanzijn.AandeLosers(youknowwhoyouare),bedanktommezosneloptenemeninjullievriendengroep,enomeengrappigenoottekoppelenaanallestressmomenten.

10

Ikwilinhetbijzonderookmijnoudersbedanken.Mamaenpapa,vanjulliehebikgeleerdomsteedsdoor tezettenenhetondersteuitdekan tehalen. Ikkreegvan julliedekansommijndromentevolgen,ookalsdiemenaarhetverreColoradobrachten.Ikovervieljulliesomsmetgrootseideeënenprojecten,ensteedskonikrekenenopjullieonvoorwaardelijkesteun.Ikstavandaagwaarikstadoorjullietoewijding,liefdeenopvoeding,enikhoopdatikjulliehiermeetrotskanmaken.OokdankjewelaanLars,mijngrotekleinebroertje,dieelkeirritatieenspanningmetdenodigerelativiteitkanbekijken.

Mijn lieve Chiara, ik weet niet goed waar ik moet beginnen. Mijn dank aan jou is dan ookeindeloos.Nietalleenluisterjenaarelkespraakwatervaldieuitmijnmondkomt,engeloofmedat zijn er veel, maar je probeert ook nog alles te begrijpen en te helpen waar je kan. Devormgevingvanditboekje,mijnsignature-roze,wasdanooknooitzoperfectgeweestzonderjouwhulp.Tussendeschoonmaak-crisesendestressvaneendoctoraatdoorslaagjijerinomelkedagbetertemakendandevorige.Ikbeloofdaaromookomjenietmeertelatenverhuizendannodig,enomjesteedsallehondenfilmpjesdoorheenmijnsocial-mediafeedonderdeneusteduwen.Geenenkelenummerplaatisveiligwanneerwesamenergensheengaan,jijmaaktmijngezinnetjecompleet.

Dit doctoraat draag ik met veel trots op aan mijn twee prachtige dochters,Nova* en Nisa.Wanneer ik begon aan een doctoraat rond levenseindebeslissingen in de pre- en postnataleperiodehadniemandkunnenvoorspellendatikdriejaarlaterzelfmijnlievemeisjetijdensdezwangerschapzoumoetenafgeven. Jewaszogeliefdengewenst,energaatgeendagvoorbijzonderdatikaanjedenk.Nova,jegafnietalleeneennieuwebetekenisaanmijnleven,maarookaanmijndoctoraatsproject.Mijnmooistesterretjeaandehemel,ikhouvanjou.AllerliefsteNisa,myreasontolive,deredenwaaromikelkedagopstameteenglimlach.Ikhoopdatikjouvanafhetbeginkanlerenomtereikennaardesterren,omhetlevenmeteenrozebriltebekijken,enomsterkinjeschoenentestaan.Erisnietswatjijnietkan.Mamahoudtvanje.

11

Part1Generalintroduction

Chapter1:Introduction

12

Introduction

13

Whentalkingaboutdeath,dyingandend-of-lifecare,peopleoftenautomaticallyenvisionadyingadultwhileinfantmortalityisoftenoverlooked.However,currently,inabout7,9in1000birthsinFlanderstheinfantdiesshortlybeforebirthorbeforetheyevenreachtheageofoneyear1.Despitethefactthatmedicaltreatmentoptionsforillneonateshavebothexpandedandimprovedduringthelastdecades,causingadeclineininfantdeathduringthelasttrimesterofpregnancyandthefirstyearoflife2,therehasbeenlittleimprovementintheprevalenceofpretermbirthandcongenitalmalformations3.Duetoimprovementsinprenataldiagnostictoolssuchasgenetictechniques and prenatal imaging techniques, an increasing number of these congenitalmalformationscannowbediagnosedprenatallyinsteadofafterbirth4,5.Thediagnosisofthesesevereorlethalanomaliesbeforeorafterbirthisoftenthestartofanextremelydifficultdecision-makingprocessregardingtheprognosisandpossibletreatmentoptionsforthechild.

TheleadingcausesoffoetalandinfantilemortalityinBelgiumhardlychangedoverthelasttenyears1. Among all foetal deaths in Flanders, the main causes of death are congenitalmalformationssuchaschromosomalabnormalitiesorspinabifida(30%)andfactorsrelatingtoeitheramaternalmedicalconditionand/orcomplicationsduringpregnancyandchildbirthsuchas problems with the placenta or infections (28,7%)6. In live-born infants, congenitalmalformationsandfactorsrelatingtoamaternalmedicalconditionand/oracomplicationduringpregnancy and childbirth are also among the main causes of death with 24,6% and 23,4%respectively6.Prenatalandneonatalmorbidityandmortalityintroducesthedebateonwhetherornotallavailablemeansoftreatmentshouldbeusedinallcircumstances.Themostcommonneonatalsituationswhereprovidingtreatmenttoprolonglifecouldpossiblybeseenasfutileare:extremelypreterminfantswhoarebornatthelimitofviability,neonateswithlifethreateningorlifelimitingcongenitalanomalies,andacutelydeterioratingillnewbornsadmittedtoneonatalintensivecareunits(NICUs)7.Insomecases,providingtreatmentmightevencausesufferingforthealreadydyingchild,andthedecisiontoforgotreatmentiseasier8.However,withinthegreyzonewherefutilityoftreatmentissuspectedbutnotalwaysclear,resuscitationandintensivecare canbeprovided, or treatment canbewithheldorwithdrawnandpalliative care canbediscussedandprovided8.

Previous research indicated that the death of a neonate is often preceded by apossibly life-shortening end-of-life decision9–11. Aside from these medical decisions after birth, due toimprovementsinthedetectionofcongenitalanomaliesbeforebirth,end-of-lifedecisionscan,andare,takenprenatally4,5,12.Theethicaldilemmainsomeofthesecasesbetweensavingthelifeofthe foetus or neonate, and not knowingwhat the burden of sufferingwill be later on needsthoughtfulandprofessionaldeliberationof theparentsand involvedhealthcareprofessionals.Eventhoughtheseethicaldilemmasneedtobeevaluatedonacase-by-caselevel,consideringthespecific characteristics andmedical situation of the child, populationdata onwhat occurs insimilarsituationscouldbevaluablefortheinvolvedhealthcareprovidersincasesofuncertaintyordisagreementbetweeninvolvedactors.Currently,availableresearchbothwithintheBelgiancontext and abroad is either incomplete or outdated, and thus not helpful as a guide to aidhealthcareprovidersincurrentdailypractice.Withinthestudiesincludedinthisdoctoralthesis,wethereforefocusedonkeycharacteristicsofend-of-lifedecisionsinavulnerablepopulationofchildrenfromaviabletermofpregnancyupuntiltheyreachtheageofoneyear.Theaimofthisdissertationwas twofold: 1) toprovide an accountofwhathappensonapopulation level bymeansofprovidingprevalenceratesonend-of-lifedecisionsandtheirclinicalanddemographic

14

characteristics;and2)togodeeperintowhatitmeanstomakethesedecisionsindailypracticebymapping out attitudes, views and experiences on neonatal end-of-life decision-making ofinvolved healthcare providers, namely neonatologists and neonatal nurses, in order toadequatelyframethesenumbersindailypractice.

Thischapterincludesanintroductionofmydissertationonend-of-lifedecision-makingbeforeandafterbirth.Firstly,theconceptualizationusedinthisdissertationwillbediscussed,includingadefinitionofusedconcepts,andanoverviewoftheconceptualframeworkweusedtoclassifyprenatalandneonatal end-of-lifedecisions.Secondly, ashort overviewof the legal context ofthese decisions in Belgium and in other European countries will be given. Thirdly, someimportantconsiderationsontheseprenatalandneonatalend-of-lifedecisionswillbediscussed.Fourth, I will give an overview of the currently available evidence on all aspects regardingprenatalandneonatal end-of-lifedecisionsdiscussed in thisdissertation, followedby a shortreflection on roles of different actors in neonatal end-of-life decision-making, specificallyfocussing on physicians andnurses as theirperspectivewas studiedwithin thisdissertation.Lastly,theresearchquestions,thestudydesignandmethodologiesusedinthisdissertationaredescribed,andtheoutlineofthisdissertationisspecified.

1.1 Conceptualisationusedinthisdissertation

1.1.1 Definitionofconcepts

Beforewegointodetailonwhatisalreadyknownonthetopicofend-of-lifedecisionsbeforeandshortlyafterbirth,somekeyconceptsusedinthisdissertationneedtobedefined.Whenusedhereafterinthisdissertation,thesedefinitionswillbeimplied.

FoetaldeathEverystillbirthofachildwithabirthweightofmorethan500grams1.Thisoftenconcurswithagestationalageof22weeks,andisinternationallyconsideredtobethelimitoffoetalviability13–15.

NeonataldeathEverydeathofalive-borninfantwithabirthweightofmorethan500gramsupuntiltheageof28daysafterbirth1.

InfantdeathEverydeathofalive-borninfantwithabirthweightofmorethan500gramsupuntiltheageofoneyear1.

PerinataldeathThesumoffoetaldeath(stillbirthfrom22weeksofgestationorabirthweight of500gramormore) andearlyneonataldeathupuntil the live-born infantreachesthethresholdofsevendays1.

Foetal-infantiledeathThesumof foetaldeathand infantdeath, thereforespanningstillbirthfrom22weeksofgestationorabirthweightof500gramormoreordeathofaninfantupuntilhe/shereachestheageofoneyear1.

15

End-of-life decisionsWithin this dissertation, end-of-life decisions are defined asdecisions regarding medical practices with a (potentially) life-shortening effectperformedbyaphysicianorateam.Anoverviewofthepossibleend-of-lifedecisionsbothprenatallyandneonatally, isgivenunder the following ‘conceptual framework’section.

Late termination of pregnancy We define late termination of pregnancy asterminationsofapregnancyfromaviableageofthefoetus(22weeksofgestation)andonwards.

Figure1.1Schematicoverviewofconceptsusedtoindicatefoetalorinfantdeath.

1.1.2 Conceptualframework

A comprehensive framework on which end-of-life decisions are possible within the foetal-infantileperiodwasonlypartiallyavailableatthestartofthisdissertation,focussingsolelyonneonatal end-of-life decisions. We therefore adjusted this previously existing and validatedframeworkof end-of-lifedecisions inneonates16 to accommodatebothprenatalandneonataldecisions,seechapter2ofthisdissertationforadetaileddescriptionofthedevelopmentofthisframework. A comprehensive framework was needed to make valid comparisons possibleprenatallyandneonatally.Twodimensionsofthepre-existingframeworkwereincluded,namelythedimensionofthemedicalact,andthedimensionoftheintentionofthephysician.

Themedicalactconcernsthetypeofmedicaldecisionthatismade.Thisdimensiondistinguishesnon-treatmentdecisionssuchaswithholdingorwithdrawingtreatment,andadministratingdrugsorperformingamedicalintervention17,18.

The intention of the physician on the other hand, distinguishes what the life-shorteningintentionofthemedicalactwas.Thiscouldeitherbenointentiontoshortenlife but the potentially life-shortening effect was taken into account, a co-intention

16

where the potentially life-shortening effectwas not themain goal but itwaspartlyintended,oranexplicitintentiontoshortenthelifeofthefoetusorinfant17,19.

To cover all decisions that could possibly influence the death of a foetus or infant, bothdimensions should be taken into consideration. The resulting framework, including clinicalexamples,canbefoundbelow.

Medical-technicalclassification

Life-shorteningIntention

End-of-lifepracticespriortostillbirth

End-of-lifepracticesinneonatesandinfants

Non-treatmentdecisions

Nointention e.g.notocolysisinpretermlabourat24weeks’gestation

e.g.noantibiotics

Co-intention e.g.nocardiacsurgeryExplicitintention e.g.nointrauterine

transfusionforananaemicfoetus

e.g.withdrawingventilation

Drugadministrationormedicalintervention

Nointention e.g.start-upofanticonvulsivetherapy

Co-intention e.g.administeringmorphine

Explicitintention e.g.terminationofpregnancywithfeticide

e.g.administeringmusclerelaxant

Table1.2Acomprehensiveframeworkofend-of-lifepracticesinthefoetal-infantileperiod.

Asisthecaseneonatally,wedistinguishtwotypesofprenatalmedicaldecisions.Afirstoptionincludes non-aggressive obstetricmanagement or abstinence,where the decision ismade toforgo medical interventions before and after birth such as foetal monitoring, ultrasoundsurveillance,caesareandeliveryorlifesupportinalevelIIIunit20.Thesecondprenatalmedicalactincludesterminatingthepregnancy,herebyactivelyendingfoetallife4,5bypreterminductionoflaboureitherwithorwithoutfeticide(administeringmedicationtointentionallyendthelifeofthefoetusbeforebirth)priortothetermination13.

Somesimilaritiesbetweenend-of-lifepracticespriortostillbirths,andinneonatesandinfantscanbenoted.Firstofall,non-obstetricmanagementcanbeseenasanon-treatmentdecision,morespecificallywithholdinglife-sustainingtreatment,sincetreatmentsthatarenecessarytosustain the lifeof the foetusarenot implemented.Secondly, terminationofpregnancycanbecomparedtodeliberatelyendinglifesincethedeathofthefoetusisactivelyhastened.However,animportantdifferencebetweenbothresearchpopulationsseemstobetheabsenceofa‘greyarea’ in the life-shortening intention prenatally. Where in neonates, decisions exist that arebeneficialforthechildwithanadditional,partiallyintendedlife-shorteningeffectthatwasnotconsideredthemaingoal,thelife-shorteningintentionprenatallyisalwayseithernon-existentor explicitly intended. Additionally, end-of-life practices where drugs or interventions areprovidedwithoutexplicitintentiontohastenthedeathofthefoetus,suchasalleviationofpainor symptoms bymeans ofmedication that couldpossibly shorten the life of a child, are notpossiblebeforebirth.

17

1.2 Legalframework

1.2.1 TheBelgianlegalframework

Whensevereorlethalfoetalanomaliesareidentifiedduringpregnancy,healthcareprofessionalsand parents can decide on end-of-life decisions such as abstinence of treatment4,5,12 orterminationofpregnancy4,5,21.IntheBelgianlegalframework,terminationofpregnancyislegalbefore12weeksofgestation.However,terminationsafter12weeksofgestationarealsolegallypossible,onlywhencompletingthepregnancypresentsaseriousthreattothewomen’shealthorwhenit isestablishedthat,whenborn,thechildwillsufferfromaparticularlysevereailment,acknowledgedtobeincurableatthetimeofdiagnosis22–25.Therefore,decidingonterminationofpregnancyfromtheviableageof22weeksandonwards,asisthescopeofthisdissertation,iscurrentlylegalonlywhentheseclinicalcriteriaaremet.However,fromaviablegestationalageonwards,thediagnosisofsevereorlethalfoetalanomaliescanalsoleadtothedecisiontoforgoaggressivemanagement20.Non-aggressivemanagementorabstinenceoftreatmentprenatallyismadeincaseswheretheinfantisexpectedtodieonitsown,eitherbeforeorshortlyafterbirth20.Inthesecases,thetreatmentoptionshavebeendeemedfutilebytheprevailingmedicalstandardsatthattime,andphysiciansarelegallypermittedtorefusepointlesstreatmentinconsultationwiththeparents.

Afterbirth,theBelgianlegalframeworkbecomesmorecomplicated.Life-sustainingtreatmentscanlegallybewithheldorwithdrawnwhentheyaredeemedfutile.Asidefromwithholdingorwithdrawing treatment, medication to relieve pain and/or symptomswith a potentially life-shorteningeffectcanbegiven.Ethically,somephysiciansstandbehindthedoctrineofdoubleeffect26,wherethewishofdoingsomethingmorallygoodsuchasrelievingsufferingjustifiestheaccompaniedyetundesiredeffectofdoingsomethingmorallybadbyhasteningdeath,aslongasthe‘side-effect’wasn’tintended.Inthiscase,respiratoryfailuremaybeforeseenorexpected,butitisimportantthatthiswasnottheintendedeffect26.Thedoctrineofdoubleeffectfollowstheassumptionthathasteningdeathisalwaysconsideredasundesirableandmorallybad,whichisoftendebated.Furthermore,thisdoctrinedeterminestherightnessorwrongnessofanactbylookingattheintentionofthephysician.Thisisfollowedbymostlegalsystemsasintentionofapersonisoftenvitalfordecidingwhetherornotacrimewascommitted,indicatingthatthesepracticeswouldbelegallyjustifiediftheintentionistoalleviatepainorsufferingfromsymptoms,andifthepossiblelife-shorteningeffectismerelyforeseenbutnotintended27.Wecandebatewhetherornotthisdistinctionbetweentheexplicitorimplicitintentionofthephysiciantohastendeath should even be made, as the consequence of the infant dying was always foreseen.Following this, many ethicists even reject the theory that allowing for an infant to die bywithholdingorwithdrawingtreatment,andadministeringmedicationtoexplicitlyhastendeatharemorallydifferent,astheendresultisthesame.

WithintheBelgianlegalframework,activeterminationoflifeonrequestisonlylegallyallowedunder strict circumstances in adults and recently for capable minors without formal agerestriction28. Inneonatalcare, the legalityofadministrationofmedicationwithapotentialorexplicitlife-shorteningeffectcanbedebated.Thereisastrongpresumptionthatthedeliberateadministration of lethal drugs is illegal, since neonates are not capable of determining theirwishesregardingactiveterminationoflife.However, itremainsanopenquestionunderwhat

18

circumstancesallowinganeonatetodiewouldcountornotcountasillegal.Thelegalhazinessconcerningneonatalend-of-lifedecisionscreatessomewhatofapeculiarcontextforend-of-lifedecisionmakinginperinatalcare,asprenatallyterminationofpregnancyispossiblebutfromthemomenttheinfantisborn,legaloptionsmightbeconsideredaslimitedduetothelackofaclearframework.

1.2.2 ShortoverviewoftheEuropeancontext

Laws regarding late termination of pregnancy differ across countries. In some countries,terminationofpregnancyforfoetalanomaliesisillegalregardlessofthegestationalage,asisthecaseinforexampleMaltaandIreland29.Othercountriesallowterminationofpregnancyincaseoffoetalanomaliesonlyupuntilviabilityisreachedat22weeksofgestation,suchasforexamplein Denmark, Estonia, Lithuania, Norway, Slovakia, Spain and Sweden30. After the viabilitythreshold is reached, some countries allow terminations up until 23 (e.g. Iceland), 24 (e.g.Finland,Latvia,theNetherlands)or25(e.g.Greece)weeksofgestation30.Lastly,somecountriesdonotposegestationalagelimitsonwhetherornotlateterminationofpregnancyislegalincaseof foetalanomalies,as is the case in for exampleFrance,Germany, Sloveniaor Switzerland30.However,whenthefoetalanomaliesaredeemedlethal,othercountriessuchastheNetherlands,Norway, Portugal and Denmark follow this trend and revoke their gestational limit29.Terminationofpregnancytosavethelifeofthemotherontheotherhand,waslegallypermittedin98%ofallworldcountriesattheendofthetwentiethcentury31.

In neonates, most European countries are very consistent with the current Belgian legalframework. Life-supporting treatments can legally be withheld or withdrawnwhen deemedfutile,whichincludesaninterpretationofthephysicianonwhetherornotfutilityisindicated.InEurope,decidingtowithholdorwithdrawlife-prolongingtreatmentininfantswithabsolutelynochanceofsurvivalisconsideredgoodpractice,sinceamajorityofEuropeanphysiciansindicatethattheprimaryobligationincarefortheseinfantsislettingthemdiewithaminimalamountofsuffering10,32.Whentheinfantwouldsurvivewithaconsiderablypoorqualityoflife,themajorityof European neonatologists consider forgoing or not initiating life-sustaining treatmentacceptableifboththemedicalteamandtheparentsagreethattreatmentisnotinthebestinterestofthechild10.Whenaneonateisnolongerdependentonmedicaltreatmentsbutsufferingcannotberelieved,notasinglecountrylegallypermitsrequestsofparentstoexplicitlyendthelifeofanewbornwithlethal(dosesof)medication10.Theonlycountrythatcurrentlylegallycondonesactively ending the life of a neonate under strict conditions is the Netherlands, where theGroningenprotocolimposesrequirementsthatmustbefulfilledbeforedecidingonintentionallyshorteningthelifeofnewbornswithlethal(dosesof)medication10.Threedistinctmedicalcasesofunbearablesufferingweredescribed,namely1)physiologicalfutilityoftreatmentinnewbornswithabsolutelyno chanceof survival;2) infantswhomay survive after aperiodof intensivetreatment,buttheiractualorforeseensufferinginthenearfutureissevereandunbearable;and3)infantswithanextremelypoorprognosiswhodonotdependontechnologyforphysiologicalstabilitybutwhosesufferingissevereandcannotbealleviated33.Inallcases,thediagnosisandprognosis must be certain, and hopeless and unbearable suffering must be present10.Furthermore,thediagnosis,prognosisandunbearablesufferingmustbeconfirmedbyatleastone independent doctor, and both parents must give informed consent10,33. Despite theNetherlandsbeingtheonlycountrywithaclearlyformulatedlegislationregardingadministering

19

drugswithanexplicitlife-shorteningintention,thispracticehappensoutsideoftheexistenceofalegalframeworkacrossEurope34.

1.3 Importantconsiderationsonfoetal-infantileend-of-lifedecisions

Decidingwhetherornottoresorttotheseend-of-lifedecisionsbringswithitextremelydifficultmedical,ethicalandmoraldiscussionsforparentsandinvolvedhealthcareproviders.InFlanders,allneonatalintensivecareunitsissuedaformalconsensustowithholdaprioriintensivecarewhenaninfantisbornbefore24weeksofgestation,includingbothtreatmentsbeforeandafterbirth35. Furthermore, a previous study showed that perinatal healthcare professionalswouldrecommendlimitinglife-supportingtreatmentbothincaseswheredeathinthenearfuturewasverylikely,andincaseswherepatientscouldpossiblyliveformonthstoyearswithcontinuedaggressivetreatment,indicatingthatqualityoflifeisthusalsoanimportantfactorinthedecision-makingprocess36.Wecanthereforestatethatthedilemmaofwithholdingorwithdrawinglife-supportingtreatment,consideredfutileeitherbecausedeathisimminentorbecausethefuturequalityoflifewouldbeextremelypoor,canbeseenaspartofregularclinicalperinatalpractice.Theethicaldoctrineofdoingandallowingclearlydemarcatesbetweennon-treatmentdecisionswhere the child is ‘allowed to die’ by stopping or not starting futile treatment, and activelyintending for the child to die by means of medication7,37. Terminating a pregnancy whencongenitalanomaliesarediagnosed38oradministeringmedicationtointentionallyshortenthelifeofaseverelyill live-bornchild19whenfollowingthisdoctrinecanbeseenasintentionallyinterveningwhilstnon-treatmentdecisionsaremoreinterpretedasa(passive)acceptancethatdeathofthechildisimminent7.However,theconceptualdistinctionbetweendoingandallowingasgroundformoralityhasoftenbeenchallenged39,asitisdifficulttobelievethatthewayinwhichanoutcomeisachievedcouldbemorallymoreimportantthantheactualoutcomeitself40.Asaconclusion,wecanstatethatend-of-lifedecisions,whicharepartofdailyneonatalpractice,arecauseforalotofethicalandlegaldebate.Dealingwithdeathandend-of-lifecare,especiallyinneonatal and prenatal care, is thus by no means easy for parents and involved healthcareprofessionals.

End-of-lifedecisionsshouldbe,andaregenerally,madeinthebestinterestofthechild,meaningthatthebenefitsandburdensofaparticularmedicalinterventionarebalanced,togetherwiththefuture quality of life andwhether or not death is imminent38. However, newborns or viablefoetusescannotdefinetheirownbestinterest,norcantheyparticipateinthedecision-makingprocess, indicating thatasurrogatedecision-makerneedstomake thesedecisions forthem41.Currently,thesedecisionsaremostcommonlysharedbetweenparentsandinvolvedhealthcareprofessionals,withbothpartieshavinganactiveroleinend-of-lifedecision-making42.Consensusshould always be sought43, but when disagreements between healthcare professionals andparents arise, the extent to which the wishes of the parents or the medical opinion of thehealthcareprofessionalsdominatesdependsontheindividualsituation.

20

1.4 Current evidence on aspects of foetal-infantile end-of-life decisionsdiscussedinthisdissertation

1.4.1 Theprevalenceoffoetal-infantileend-of-lifedecisions

Animportantstartingpointinexaminingend-of-lifepracticesbeforeandafterbirthistoprovideinsight into how often end-of-life decisions occur. In Belgium, stillborn babies are officiallyregisteredbymeansofadeathcertificatefrom180daysofgestation(about26weeks)onwardsbytheFlemishAgencyofCareandHealth,andregistrationfrom22weeksonwardisencouragedbutnotmandatory,causingunderreportingofstillbornsbymeansofdeathcertificates in thisgroup.Additionally,allbirths(livebornandstillborn)withabirthweightof500gramsormoreareregisteredthroughtheStudyCentreforPerinatalEpidemiology(SPE).However,bothoftheseregistrationsonlyencompassinformationregardingthepresenceofcongenitalmalformations,withoutdetailedinformationonthereasonforstillbirthsuchaswhetherornotthepregnancywasactivelyterminated.Theycanthereforenotbeusedtoprovideinsightsintotheprevalenceof end-of-life decisions in the prenatal period. Additional to the registrations of stillbirths, aregistration of the practice of abortion in Belgium exists, which is reported on biennially.However,thelastreportdatesbacktoabortionpracticesin2010and2011andsincethen,noreportshavebeenpublished.Forreliableprevalencerates,wethushavetolooktowardsreliable,population-basedstudieswhereallstillbornsareconsideredirrespectiveofthediagnosisofthechild, thedecisionsmade,or thesettinginwhich thestillbirthoccurred. InBelgium,onlytwopreviouspopulation-basedstudiesonend-of-lifepracticesinlatetermpregnancyexist,namelytheMOSAIC study performed in200322 and the European register-based study of Garne andcolleagues in 2000-200524. The first studied registered pregnancy terminations and itsproportion in stillbirths in a specific region in Flanders22. The second study is based on theEUROCAT register (European network of population-based registries for the epidemiologicsurveillance of congenital anomalies), which registers all liveborn, stillborn and terminatedcongenital anomalies24,44. Based on the information gathered in these sources, the estimatednumberoflateterminationsofpregnancyshouldbeatleastfourtimeshigherthanthatreportedbytheevaluationcommitteeofabortionpractices.

In livebornneonates,most studieson end-of-lifedecision-making are either limited to singlecentrestudiesorfocusedsolelyonnon-treatmentdecisions.AvailabledatawasmostlybasedonreviewsofmedicalrecordsinNICUsataspecifichospital.Thesestudiesshowthatbetween40%and93%ofdeathsintheNICUfollowwithdrawaloflife-sustainingtreatments11,45–47,varyingbyregionandphysicianattitudes9,48.TheEURONICstudywasoneoftheonlylargerscalestudies,albeitnotpopulation-based,onphysicians’self-reportedpracticesandattitudes.Thisstudywasperformed across a sample of 143 European NICUs (Belgium not included) in the 1990s49.Population-basedstudies,wherealldeathcasesareconsidered,areidealtostudyincidencesofend-of-lifedecisions,withoutabiasonwhetherornotthechildand/ormotherwereadmittedtoalevelthreehospitalunitsuchasaNICU.However,theonlypopulation-basedstudiesinneonatesarefromtheNetherlands50andonefromBelgiumthatdatesbackto200019.IntheNetherlandsin2000,63%ofalldeathsundertheageofoneyearwasprecededbyanend-of-lifedecision51.Mostofthesedecisionscouldbeclassifiedundernon-treatmentdecisions,andonly1%ofthedecisionsincludedtheadministrationofmedicationtodeliberatelyhastendeath.InBelgiumwesaw that almost 60% of deaths before the age of one year were preceded by an end-of-life

21

decision19. Similar to theNetherlands study,most of these decisionswere classified as non-treatment decisions such as withholding or withdrawing treatment, while 7% consisted ofadministratingmedicationtodeliberatelyhastendeath19.

Thisoverviewonwhatiscurrentlyknownonprenatalandneonatalend-of-lifedecisionsrevealssomemajorgaps.Firstly,noBelgianincidenceratesonend-of-lifedecisionsinviablefoetusesexist,making it verydifficult toprovideanoverviewof currenthealthcarepractice. Previousstudies22,24onlylookedattheprevalenceofterminationofpregnancy(deliberatelyendinglifewithanexplicitlife-shorteningintention).Therefore,notmuchisknownaboutthefullscopeofend-of-life practices, including non-treatment decisions, and their decision-making process.Moreover,bothexistingstudiesonlyincludeddatafromtwoseparateprovincesinBelgium,onlyone situated in Flanders. Secondly, foetuses and neonates where a life-limiting disorder isdiagnosedareinessencethesamepatient.Theonlydifferenceistheoccurrenceofbirthandnotnecessarilyadifferenceindisordersorcongenitalanomalies21,whiletheimpactonparentsandinvolvedcaregiversisverysimilar.Wherepreviousresearchfailedtoconsidertheentirefoetal-infantileperiodbylookingsolelyatend-of-lifedecisionsprenatallyorneonatally,wewillaimatdevelopingasolidresearchmethodologytoexaminebothsimultaneously.Hereby,wewillbeabletostudyshiftsinend-of-lifedecisionsintheentirefoetal-infantileperiod,thatwouldotherwisebemissed. The advanced technologies inprenatal screening and consequent rise inprenataldiagnosesofcongenitaldisorders4,5couldforexamplecauseanincreaseinlateterminationsofpregnancy and a decrease in some end-of-life decisions neonatally. Thirdly, prior to thedevelopmentofthestudiesinthisdissertation,bothprenatalandneonatalFlemishhealthcareprofessionalsstatedtheneedformorerecent,population-baseddataontheprevalenceofend-of-life decisions. Both the studiesmentioned in the paragraph regarding prenatal end-of-lifedecisionsandlateterminationofpregnancy,andthestudiesintheparagraphonneonatalend-of-lifedecision-making,collecteddataintheearly2000s19,22,24.Inlightofeverchangingsocietal,legalandclinicalinfluences,wethusbaseimportantclinicaldecisionsandrecommendationsindaily practice on outdatedpopulation-data. Important societal changes took place that couldpossiblyimpactend-of-lifepractice,includingintheunbornandnewbornpopulation.Therewastheimplementationoflawsonpatientrights,palliativecareandeuthanasiainadultsin2002,andthelawoneuthanasiaforchildrenwithdecisionalcapacityin201452.Neonatesdonotfallunderthis euthanasia law,which is limited to adults and capableminors, yet a possible impact onprenatalandneonatalpracticecannotbeexcluded.Internationally,theGroningenprotocolintheneighboringNetherlandscouldpossiblyhaveanimpactonBelgianprevalencerates.Asidefromlegalchanges,theriseinmedicaltreatmentoptionsforextremelyillneonates5,53couldpossiblyhave changedmedical practice. Therefore, a need for current and reliable incidence rates ofFlemish end-of-life decisions is indicated, not only by researchers but also by Flemishrepresentatives from all eightneonatal intensive care units.Within thisdissertation,wewillthereforeaimtoexaminetheseincidencesonapopulationlevel,ininfantswhodiedbeforetheageofoneyear.

1.4.2 Attitudes of healthcare providers concerning foetal-infantile end-of-lifedecisions

Previousresearchshowedthat,eveninnewbornswiththesamepathology,variabilitybetweentypes of end-of-life decisions can be noted54,55. This is because end-of-life decisions can be

22

influenced by a number of contextual variables such as available hospital resources and theparents’ and clinicians’ social, cultural and religious beliefs54. Aside from these contextualvariables,attitudesofcaregiversplayacrucial role inend-of-lifedecision-making19.Andevenwithinacareteam,importantdifferencesbetweenphysicians’andnurses’attitudestowardsend-of-lifedecisionshavebeen found56. Personal characteristics of healthcareprovidersmaythusplayacrucialroleinend-of-lifedecision-makinginneonates9,19,57,58.

Anattitudesurveystudyin10Europeancountriesin2000foundthatthelikelihoodoflimitinglife-supporting treatments in neonates is dependent on the country of residence, reportedreligionofthephysician,theirgender,whetherornotthephysicianhaschildren,andtheamountof very low-birth-weight infants that areadmitted to theirNICU48. Furthermore, a self-reportquestionnairecombinedwithretrospectivemedicalchartreviewrevealedthatanunintentionallife-shortening effect of administering opioids is considered acceptable for most NICU andpaediatricintensivecare(PICU)nurses59.Thesestudiesarehoweverlimited,sincetheyfailtoinclude attitudes towards decisions that could have been made before the baby was born.Becauseattitudesanddecisionsbeforeorafterbirthcouldpossibly influenceeachother,andneonatologistsareoftenconsultedinprenatalend-of-lifedecisions60,attitudestowardsprenatalandneonatalend-of-lifedecisionsshouldthusbeincludedintooneoverarchingstudytomakevalidcomparisonspossible.Becauseoftheirrelevanceforclinicalpractice,aseparatepartofthisdissertationwillbedevotedtotheexaminationofattitudesregardingfoetal-infantileend-of-lifedecisions of themost involved healthcare providers in neonatal end-of-life decision-making,namelyneonatologistsandneonatalnurses.

1.4.3 Barriers to and facilitators of the neonatal end-of-life decision-makingprocessforhealthcareproviders

Despitethesevereimpactofend-of-lifedecision-makingonNICUstaffmembers61, fewstudieshave focused onwhat the involvedneonatologists andneonatal nurses find either helpful ordifficult inmaking these end-of-lifedecisions.Qualitative studieswithNICU staffmembers inNorwayondecidingwhetherornottocontinuelife-sustainingtreatmentshowthatthelackofcertaintyintheprognosisofthechildandwhattheirsufferingwillbelateron62canbeseenasanimportant barrier in decision-making. Furthermore, these Norwegian studies show that theambivalencebetweenwantingtoincludeparentsandwantingtosparethemsomeofthepain,cancauseindecisionregardingwhether,whenandhowcertaininformationabouttheprognosisneedstobegivenbythehealthcareproviderstotheparents62,63.

Previousstudieson thesebarriersand facilitatorsinneonatalend-of-lifedecision-makingarelimited in that they mainly focus on specific end-of-life practices such as withholding andwithdrawing of treatment9,19,50,51,64 rather than focusing on the entire spectrum of end-of-lifedecisions, or that they mainly focus on the experiences of parents42,64–67, hereby excludinghealthcareprovidersasanimportantco-actorinthedecision-makingprocess.Aseparatechapterinthisdissertationwillthereforefocusonexaminingwhichfactorsneonatologistsandneonatalnursesexperienceaseitherhelpfulordifficultinneonatalend-of-lifedecision-makinginaNICU.Knowledge onwhich factors could either benefit or hinder the neonatal end-of-life decision-makingprocessfromtheviewpointofthemostinvolvedhealthcareproviderscouldbeacrucialstartingpointinformulatingrecommendationstoaidfuturepractice.

23

1.4.4 Psychological support in end-of-life decision-making for healthcareproviders

Neonatologistsandneonatalnurseswhoworkinaneonatalintensivecareunitoftenexperiencestressorsandmoraldistressduetothehighdemandsoftheiroccupation68,69.Especiallyintimeswhenaninfantintheircarecannolongerbenefitfromaggressiveorevenfutiletreatmentandanend-of-lifedecisionneedstobemade68,70.Similarlytopaediatricintensivecareunitstaff,theyexperiencesadness,helplessnessandfrustrationwhentheyareunabletodomorewhenachilddies71.Theycanevenbecalleddisenfranchisedgrievers,sincetheyaregenerallynotrecognizedas a bereaved person by society or their work environment72. Because of this distress,neonatologistsandneonatalnursesarepronetodevelopingcompassionfatigueorburnoutwhentheemotionalpriceofcaringbecomestoohighforthemtocope73–75.PsychosocialsupportforNICUstaffmembersiscurrentlyincludedinrecommendationsforNICUpractices76–78,howevermostrecommendationsandguidelinesconcerningthispsychosocialsupportfocusonprovidingneonatologist andneonatalnurseswith concrete tools tooptimally attend toparents in theirdecision-makingprocessandgrief79–81.Furthermore,researchonhowsupportedtheyactuallyfeelislacking.

To our knowledge, only one study included specific recommendations solely focusing on thebenefittoNICUstaffmembersinaneonatalend-of-lifepalliativecareprotocol70.Catlin&Carter70recommendedformalmeetingsorcounsellingsessionsaspartofregularworkhours,insteadofon a voluntary basis or during unpaid time. Furthermore, they recommended that bothneonatologists and nurses should be able to opt out of end-of-life care by taking on otherassignments.Alastpartofthisdissertationthereforefocussesontheexperiencedpsychologicalsupportofhealthcareprovidersworkinginaneonatalintensivecareunitasanimportantaspectofthefoetal-infantileend-of-lifedecision-makingcontext.Caringfortheonesresponsibleforthecareof critically ill infants couldbe a crucial step towardsprovidingbetter support forbothpatientsandgrievingparentsinaneonatalintensivecareunit74,76.

1.5 The role of neonatologists and nurses in foetal-infantile end-of-lifedecisionmaking

Since foetuses andnewborns areunable tomakedecisions for themselves, it is important toconsidertheirsurrogatedecision-makers.Theparents’righttodecidefortheirchildreninlessserious clinical situations is generallywell accepted82. They aremotivatedby the child’s bestinterest,andaboveall,parentsaregenerallyconsideredashavingthemainauthorityovertheirchild82.However,whentalkingaboutend-of-lifedecisionsthatarelikelytoresultinthedeathoftheinfant,theroleofparentsislessclear.Healthcareprovidersmightwanttoprotectparentsfrom possible negative psychological consequences of deciding on if andwhen their child isallowedtodie83.Furthermore,whilesomeparentswanttobetheultimatedecision-maker84orwanttoatleastbeactivelyinvolved84–86,othersindicatedthattheypreferredthatthedecisionwasmadebytheinvolvedmedicalteam87.

Bothhealthcareprovidersandparentsplayanactiveroleinend-of-lifedecision-making42.Theviewpoint of parents who focus on their specific case, and healthcare providers who have a

24

multitudeof(end-of-life)experiences,isfundamentallydifferentandwhenrecommendationsforthe improvement of future practice are made, both should be considered. However, in thisdissertation,we chose to solely focusonproviding the viewpoint of healthcareproviders.Bydrawingupontheirmultitudeofexperiences,wehopetobeabletocomparewhatdifferentiatesbetweenagoodandabadexperience,insteadofjustbeingabletodiscussthedecision-makingprocessofasingleinfant,aswouldbethecaseinbereavedparents.

In a NICU setting, neonatologists and neonatal nurses work closely together. They fulfilfundamentallydifferentbutcrucialrolesinprenatalandneonatalend-of-lifedecision-making.Physiciansareexpertsinunderstanding theprognosisandpossibleoutcomesofthe infants58.Theyareoftentheoneswhoultimatelydecideontheend-of-lifedecisionandthereforetakefinalresponsibility88.DependingonthecountryoreventheNICUward,nursescansometimesalsobeinvolvedintheseend-of-lifediscussions89.Oneofthemostimportanttasksofinvolvednursesistakingcareofthechildduringtheend-of-lifephase89.Becauseoftheirpresenceatthebedsideoftheinfants,theyareoftenmoreemotionallyinvolvedwithparentsandinfantsthanphysicians61,and nurses are often the first to recognize and accept the possibility of impending death64.Physiciansandnursesthushaveuniqueandimportantrolesintheprenatalandneonatalend-of-life decision-making process, making a reflection on roles, attitudes and viewpoints of bothhealthcareprovidersessential.

1.6 Studyobjectivesandresearchquestions

Themain focus of this dissertation is end-of-life decision-making in stillbirths, neonates andinfantsonapopulationlevel,acrosscentres,patientsandphysicians.The followingtwoaims,eachwithspecificresearchquestions,guidedthisdissertation:

Thefirstaimistoexamineend-of-lifedecisionsinstillbirths,neonatesandinfantsinFlanders,Belgiumonapopulationlevel.Thefollowingresearchquestionswillbeanswered:

1. Whichmethodologycanbeusedtoreliablystudytheprevalenceofvariousend-of-lifedecisions,takenbeforeandafterbirth?Whichpopulation-leveldatabasescanbeusedtostudybothprenatalandneonatal end-of-lifedecisions, andhowcanweanonymouslycontactthephysicianinvolvedinthesestillbirthordeathcases?

2. Whatistheprevalenceofvariousend-of-lifedecisionsmadeintheneonatalperiod?Didtheprevalencechangeovertimecomparedtothepreviousdata-collectionin1999-2000?What are the clinical and demographic characteristics of infants whose death wasprecededbyvarioustypesofend-of-lifedecisions?Whichcircumstancesareassociatedwithvarioustypesofend-of-lifedecisionsinneonates?

Thesecondaimofthisdissertationistomaptheattitudes,viewsandexperiencesofinvolvedhealthcare providers, namely neonatologists and neonatal nurses, on neonatal end-of-lifedecision-making.Thefollowingresearchquestionswillbeansweredwithinthisaim:

3. Whatare theattitudesofneonatologistsandneonatalnursesconcerningprenatalandneonatalend-of-lifedecision-making?Whatarethedifferencesbetweenphysiciansand

25

nurses inattitudes towards thesedecisions?Which attitudes concerningprenatalandneonatalend-of-lifedecisionsandwhichdemographiccharacteristicsareassociatedwithpossibletreatmentoptionsthatareconsideredacceptableinahypotheticalcase?

4. Whichfactorsinvolvedinthedecision-makingprocesscan,accordingtoexperiencesfromneonatologistsandneonatalnurses,facilitateorimpedetheneonatalend-of-lifedecision-makingprocessinaFlemishneonatalintensivecareunit?

5. In what way are neonatologists and neonatal nurses supported by colleagues,psychologistsandthehospitalwardduringthedifficultprocessofend-of-lifedecisionsina Flemish neonatal intensive care unit? How sufficient is the current psychologicalsupportforcaregiversworkinginaFlemishneonatalintensivecareunit?

1.7 Methodsusedinthisdissertation

Toanswertheresearchquestionsandstudyobjectivesofthisdissertation,severaldata-collectionmethodsanddatasourceswereused,namelyamortalityfollow-backsurvey,anattitudesurveyandaqualitativestudywithface-to-facesemi-structuredinterviews.

1.7.1 Themortalityfollow-backsurvey

Thissectiononlytouchesbrieflyonthemortalityfollowbacksurvey-methodsincetheresearchprotocolforthisdata-collectionmethodisexplainedindetailinchapter2ofthisdissertation.Ashortsummaryofthemortalityfollowbacksurvey-methodwillbeprovided.Thisdata-collectionmethod was used to examine end-of-life practices and decisions in stillbirths, neonates andinfantsinFlanders,Belgiumonapopulationlevel.

Thissurveyfollowsthedesignofamortalityfollow-backsurveyonapopulation-levelbasedonall death certificates of stillborns from 22weeks of gestation or a birthweight of 500 gramonwards,andneonatesorinfantswhodiedbeforetheageofoneyear.AllincludedstillbirthsordeathsoccurredinFlandersorBrusselsandconcernedfoetusesorinfantswhosemotherwasaFlemishresidentatthetimeofdeath.ThedesignofthisstudywasidenticaltoasurveyconductedfromAugust1999toJuly200019,withtheexceptionofalongerinclusionperiodfromSeptember2016toDecember2017(12monthsin1999-2000versus16monthsin2016-2017).

Withinthreemonthsafterdeath,everycertifyingphysicianreceivedafour-pagequestionnairethrough theFlemishAgency forCareandHealthwho is responsible forprocessing thedeathcertificates with an introductory letter containing patient identification characteristics. Toguaranteeanonymity,alawyerservedasanintermediarybetweentherespondingphysicians,theFlemishAgencyforCareandHealth,andtheresearchers90.Theintermediaryensuredthatcompletedquestionnairescouldneverbelinkedtospecificpatients,physiciansorhospitals.

Two separate questionnaires were used during the survey namely one questionnaire toaccompanydeathcertificatesthatcertifiedastillbirthandonequestionnairetoaccompanydeathcertificatesthatcertifiedthedeathofaninfantbeforetheageofoneyear.Thequestionnairesusedinthesurveyaimedtoinquireaboutpossibleprenatalandneonatalend-of-lifedecisionsthatprecededthedeathorstillbirthreportedonthedeathcertificate.Avalidatedquestionnaire

26

usedtosurveyneonatalend-of-lifedecisionmakingdevelopedinthe1999-2000studywasusedasthebasis for thecurrent2016-2017questionnaires toensurecomparabilityofdata19.Bothquestionnairesfirstaskedwhetherthedeathoftheneonatehadbeensuddenandunexpected.Ifanswerednegatively,anend-of-lifedecisionwasconsideredpossibleandphysicianswereaskedwhether any end-of-life decisions preceded the death or stillbirth. The questionnaires neveraskedabout the end-of-lifedecision categories asdenoted in our conceptual framework, butrather classified themedical decisions based on a series of core questions following the twodimensionsoftheconceptualframework(seetheconceptualframeworkonpage15-16),namely:

1. Whichactoromissionwasused(themedico-technicaldimension)2. Which life-shortening intentionwas associatedwith the act or omission (themedico-

ethicaldimension)

Whenmore than one end-of-life decisionwith the same life-shortening intentionwas noted,administrationofdrugs(active)prevailedoverwithholdingorwithdrawingtreatment(passive).Incaseofanend-of-lifedecision,follow-upquestionswereaskedsuchas:byhowmuchtimewasthelifeoftheinfantshortened,whatwasthemostimportantreasonfordecidingontheend-of-lifedecision,andwhowasincludedinthedecision-makingprocess.Demographicinformationfromthedeathcertificateswasanonymouslylinkedwiththeirrespectivequestionnairedataafterdata-collectionwasfinished.Theusedquestionnaires(inDutch)canbefoundinAppendix1and2.

EthicalapprovalwasobtainedbytheethicscommitteeoftheUniversityHospitalofGhentandadditionally from theNational Privacy Commission (CBPL), the Sectoral Committee of SocialSecurityandHealth,andtheNationalDisciplinaryBoardofPhysicians.

1.7.2 Theattitudeandpsychologicalsupportsurvey

Inorder toexamine theattitudesandperceivedpsychologicalsupportof involvedhealthcareprovidersinneonatalend-of-lifedecision-making,afullpopulationmailsurveywassetupinallneonatologistsandneonatalnursesworkinginaFlemishneonatalintensivecareunit.AllFlemishneonatalintensivecareunitsparticipatedinthisstudy.Theseneonatalintensivecareunitsweresituated in the following hospitals: Ghent University Hospital, Brussels University Hospital,LeuvenUniversityHospital,AntwerpUniversityHospital,AZSint-JanBrugge-Oostende,HospitalOost-LimburgGenk,HospitalGZAStAugustinusandZNAMiddelheim.

DatawascollectedbetweenMay1standMay31stof2017.Thegatekeepermethodwasused,wherearepresentativephysicianworkingineachneonatalintensivecareunithandedoutthequestionnaire to every neonatologist and every neonatal nurse in their respective ward.Physiciansandnurseswereinvitedtofilloutthequestionnaireandsenditbackbymeansofaprepaidenvelopetotheresearcherswithintheperiodofonemonth.

The questionnaire used in this survey was developed based on an existing Flemish attitudequestionnairefromtheyear2000onneonatalend-of-lifedecisions19,andanAmericanstudyoncompassion fatigue, burnout and compassion satisfaction of neonatologists in a neonatal

27

intensive care setting91. A multidisciplinary team consisting of three sociologists, twopsychologists, three neonatologists and one gynaecologist developed the final questionnaire.Afterwards,thisquestionnairewascognitivelytestedonfiveneonatologistsfromfourseparatehospitals,threeneonatalnursesfromtwoseparatehospitalsandonegynaecologistinordertoensurevalidityoftheitems.Thequestionnaireconsistedofsixseparateparts:

1. Socio-demographicinformationoftheparticipantsThe questionnaire consisted of seven socio-demographic questions, including gender,age,yearsofexperienceworkinginaneonatalintensivecareunit,educationlevel,religionorbelief,whetherornottheirreligionorbeliefplaysaroleinend-of-lifedecision-making,andwhetherornottheylostsomeoneclosetothem.

2. Attitudesconcerningend-of-lifedecisionsinneonatesWithinthequestionnaire, sixattitudeitems focussedonneonatalend-of-lifedecisions,meaningend-of-lifedecisionsthatcanbemadeinalive-borninfantwithasevereorlethaldiagnosis.Attitudesweremeasuredbyindicatingwhetherornotparticipantsagreedwiththestatements, scoredona five-pointLikertscaleranging from ‘totallydisagree’until‘totallyagree’.

3. Attitudesconcerningprenatalend-of-lifedecisionsfromaviableageonwardsSix attitude items regarding prenatal end-of-life decisions were added to thequestionnaire.Withinthesesixitems,wegagedtheirattitudestowardsterminationofpregnancy fromaviableageof the foetusonwards, incaseofasevereor lethal foetaldiagnosis.Similartotheneonatalattitudeitems,attitudesweremeasuredonafive-pointLikertscale.

4. Hypothetical caseswhere prenatal and/or neonatal end-of-life decisions could beconsideredTwohypotheticalcaseswerepresented.Oneregardedaprenatalhypotheticalcasewithsevere spina bifida at 25weeks of gestation, leading to paralysis of the lower limbs,incontinence,andpossiblyacognitivedeficit.Thesecondcaseregardedaneonatalcaseofaninfantbornat27weeksofgestationwithsevereadditionalcomplications.Inbothcases, participants were given several treatment options, including continuing life-prolonging treatment and considering several end-of-life decisions. Participantswereaskedto indicatewhether theywould consider each treatmentoptionon a four-pointLikert scale ranging from ‘not a good treatment option’ until ‘a very good treatmentoption’.

5. End-of-lifepoliciesofthewardandthehospitalAfifthpartofthequestionnaireconsistedofsevenquestionsregardingtheexistenceofformalor informalpoliciesonprenatalandneonatalend-of-lifedecisionswithin theirward,andwhetherornottheysupportedthesepolicies.Similartotheotherpartsofthequestionnaire, participants indicated whether or not they agreed with the providedstatementsonafive-pointLikertscalerangingfrom‘totallydisagree’until‘totallyagree’.

6. Psychologicalsupportforhealthcareprovidersregardingfoetal-infantileend-of-lifedecisionsThe final part of the questionnaire consisted of seven questions regarding perceivedstress,professionalpsychologicalsupportprovidedbytheneonatalintensivecareunitandpsychologicalsupportprovidedbycolleagues.Thepsychologicalsupportstatements

28

werescoredona5-pointLikertscale,rangingfrom‘totallydisagree’until‘totallyagree’.Threeofthesevenquestionsdifferedbetweenneonatologistsandnursesbecausewhilephysiciansaretheonesdecidingtheend-of-lifedecisions,nursesareoftennotincludedinthedecision-makingprocessandareconsequentlyonlyinvolvedinimplementation

Ethical approval for this attitude survey was obtained from the Ethical Committee of GhentUniversityHospital.Respondentstookpartinthesurveyonavoluntarybasis.Sendingbackafilled-outquestionnairewasseenasgivinginformedconsent.Amoredetaileddescriptionofthemethods can be found in chapter 4 and 6, used questionnaires (in Dutch) can be found inAppendix3and4.

1.7.3 Face-to-facesemi-structured interviewswithneonatologistsandneonatalnurses

Inordertoexaminewhichfactors,involvedinthedecision-makingprocess,thatcouldfacilitateor impedeneonatal end-of-lifedecision-making according tohealthcareproviders,weused aqualitativeapproach.Thisinordertofullycoverthecomplexityandsubtletyoftheindividualexperiencesofhealthcareprovidersintheend-of-lifedecision-makingprocess.

In-depth,semi-structuredinterviewswereusedbecauseitallowedparticipantstotelltheirstoryfreelywithoutinterruptionorfearofnotbeingabletospeakopenlyingroupconversationssuchas focus groups. Furthermore, semi-structured interviews are flexible, allowing us to collectricherqualitativedata,whilststillbeingabletoprovidesomestructureinordertomakesurethatkeyresearchquestionswereformulatedsimilarlyineveryinterview.

Two groups of healthcare providers were included in the study, namely neonatologists andneonatalnursesworkinginaneonatalintensivecareunit.Neonatologistswererequiredtobearesidentphysicianataneonatalintensivecareunitinoneofthefourparticipatinghospitals,asopposed to an assistant physician in training. Furthermore, they should have been theattending/treatingphysician to at least one child thatdied at thewardwhere thedeathwasprecededbyanend-of-lifedecisioninthepastyear.Neonatalnurseswerealsorequiredtoworkin a neonatal intensive care unit in one of the four participating hospitals. Additionally, theyshouldhavebeenthemostinvolvednursetoatleastonechildthatdiedatthewardwherethedeathwasprecededbyanend-of-lifedecisioninthepastyear.

Recruitmenttookplaceatfourhospitals,namelyGhentUniversityHospital,BrusselsUniversityHospital,LeuvenUniversityHospitalandAZSint-JanBrugge-Oostende.Allneonatologistsandnursesinthefourneonatalintensivecareunitswerenotifiedofthestudybyarecruitmentletterfromtheresearchers,distributedbyarepresentativeneonatologistattheirownward.Purposefulsamplingwasusedtoselectparticipants.Bothneonatologistsandnurseswerefreetoparticipateonavoluntarybasis.ParticipantswererecruitedandinterviewedbetweenDecember2017andJuly 2018. Interviews took approximately one hour on average and took place either at theneonatalward in a secludedmeeting room, or in the comfort of their own home. Data wascollecteduntilnonewinformationemergedforbothneonatologistsandnursesseparately,anddatasaturationwasachieved.

29

Aconcisetopicguidewasusedduringallinterviews.Thistopicguide(inDutch)canbefoundinAppendix5and6.Aftereveryinterviewanevaluationtookplacesothatnecessaryalterationstothetopicguidecouldbemade,sincethisisinherenttoqualitativeresearch.Atthebeginningofeveryinterview,participantswereaskedtofilloutsomequestionsregardingtheirdemographiccharacteristicssuchasgender,age,yearsofexperienceworkinginaneonatalintensivecareunitandeducationlevel.

Interviewswereaudiotapedandtranscribedverbatim.Tworesearcherscodedthe interviewsindependently and openly by means of inductive coding during which they searched forfacilitatorsandbarriersthatinfluencedtheend-of-lifedecision-makingprocess.Thefirsteightinterviewswerecodedbybothresearchers.Afterfiveinterviewsafirstdiscussiononcodenodesandtreesoccurred.Theother22interviewswerecodedbyoneoftheresearchers.Codenodesandtreeswerediscussedamongstbothresearchersatregularmeetings,andduringtwoseparatemeetingsafterwardswithall co-authors.Whencodingdiscrepanciesoccurred,consensuswassought.

Ethicalapprovalwasobtainedfromtheethicalcommitteesoftheparticipatinghospitals,namelyGhentUniversityHospital,BrusselsUniversityHospital,LeuvenUniversityHospitalandAZSint-JanBrugge-Oostende.Amoredetaileddescriptionofthemethodscanbefoundinchapter5.

1.8 Dissertationoutline

Thisdoctoraldissertationisdividedintofourparts.PartIstartswithageneralintroductiontoend-of-life decision-making in the prenatal and neonatal context. Following this generalintroduction,theobjectivesandresearchquestionsofthisdissertationarestated,includingthedifferentmethodologiesusedtoexaminethem.Chapters2-6arebasedonarticleswhichhavebeenpublished, acceptedor submitted forpublication.All of those chapters canalsobe readindependently.

Thetwomainresearchaimsofthisdissertationareaddressedintwoseparateparts(partIIandpartIII).Eachpartconsistsofacoupleofchaptersthatanswerthespecificresearchquestionsofeachaim.

PartIIoffersanoverviewonthemethodologyusedtostudyfoetal-infantileend-of-lifedecisionsonapopulationlevelinFlanders,andadetaileddescriptionoftheprevalenceandtrendsoftheseend-of-lifedecisionsinneonates.Thispartaims todescribe themortality followbacksurvey-methodinthefoetal-infantileperiodasmentionedonpage23,andtoanswerresearchquestions1-2asmentionedonpage24.

PartIIIconsistsofanexplorationintoattitudes,viewsandexperiencesofhealthcareprovidersinvolved in neonatal end-of-life decisions. Within this part, we transcend frequencies andprevalence,andattempttodescribehowinvolvedhealthcareprovidersactuallyexperiencetheseend-of-lifedecisionsindailypractice.PartIIIcoversresearchquestions3-5,asdescribedonpage24-25.

30

PartIVofthisdissertationconsistsofathoroughreflectiononitsstrengthsandlimitations,anoverviewofthemainfindingsofthedoctoralstudy,anoveralldiscussionofthefindings,andadiscussionontheimplicationsofthesefindingsforprenatalandneonatalpractice,policy,andfutureresearch.

Lastly, an English and Dutch summary of the main findings, curriculum vitae and list ofpublications,andappendicesconcludethedissertation.

31

1.9 Referencelistchapter1

1. DevliegerR,MartensE,GoemaesR,CammuH.PerinataleActiviteiteninVlaanderen2017.StudvoorPerinatEpidemiol.2018.

2. Humblet P,Amerijckx G, Aujean S, et al. Les jeunes enfants à Bruxelles : d’une logiqueinstitutionnelleàunevisionsystémique.Brusselsstudies.2015;2015(0).

3. Paneth N, Thompson T. Neonatal and Perinatal Epidemiology. Avery’s Dis Newborn.2018;ContentRe:1-10.

4. Bijma HH, Schoonderwaldt EM, Van Der Heide A, Wildschut HIJ, Van Der Maas PJ,Wladimiroff JW. Ultrasound diagnosis of fetal anomalies: An analysis of perinatalmanagementof318consecutivepregnanciesinamultidisciplinarysetting.PrenatDiagn.2004;24(11):890-895.

5. BijmaHH,vanderHeideA,WildschutHI.Decision-MakingafterUltrasoundDiagnosisofFetalAbnormality.ReprodHealthMatters.2008;16(31SUPPL.):82-89.

6. Kind&Gezin.[TheChildinFlanders2017](inDutch).;2017.7. MalliaP.WASP(WriteaScientificPaper):Specialcasesofselectivenon-treatmentand/or

DNR.EarlyHumDev.2018;124:62-64.

8. WilkinsonD.Who should decide for critically ill neonates and how? The grey zone inneonataltreatmentdecisions.In:WhenDoctorsandParentsDisagree:Ethics,Paediatrics&theZoneofParentalDiscretion.;2016:57-79.

9. CuttiniM, NadaiM,KaminskiM, et al. End-of-life decisions in neonatal intensive care:physicians’self-reportedpracticesinsevenEuropeancountries.EURONICStudyGroup.Lancet.2000;355:2112-2118.

10. VerhagenE,SauerPJJ.TheGroningenprotocol--euthanasiainseverelyillnewborns.NEnglJMed.2005;352(10):959-962.

11. BergerTM,HoferA.Causesandcircumstancesofneonataldeathsin108consecutivecasesover a 10-yearperiod at the Children’sHospital of Lucerne, Switzerland.Neonatology.2009;95(2):157-163.

12. ChervenakFA,McCulloughLB.Nonagressiveobstetricmanagement:anoptionforsomefetalanomaliesduringthethirdtrimester.Jama.1989;261(23):3439-3440.

13. RennieJM,LeighB.Thelegalframeworkforend-of-lifedecisionsintheUK.SeminFetalNeonatalMed.2008;13(5):296-300.

14. FischerN,SteurerMA,AdamsM,BergerTM.Survivalratesofextremelypreterminfants(gestationalage<26weeks)inSwitzerland:impactoftheSwissguidelinesforthecareofinfantsbornatthelimitofviability.ArchDisChild-FetalNeonatalEd.2009;94(6):F407-F413.

15. BolandRA,DavisPG,Dawson JA,DoyleLW.Outcomesof infantsbornat22–27weeks’gestation in Victoria according to outborn/inborn birth status. Arch Dis Child - FetalNeonatalEd.2017;102(2):F153-F161.

16. ProvoostV,DeliensL,CoolsF,etal.Aclassificationofend-of-lifedecisionsinneonatesandinfants.ActaPaediatr.2007;93(3):301-305.

17. ProvoostV,DeliensL,CoolsF,etal.Aclassificationofend-of-lifedecisionsinneonatesandinfants.ActaPaediatr.2004;93(3):301-305.

18. ChambaereK,BilsenJ,CohenJ,Onwuteaka-PhilipsenBD,MortierF,DeliensL.Trendsinmedical end-of-life decision making in Flanders, Belgium 1998-2001-2007.Med Decis

32

Making.2011;31(3):500-510.19. ProvoostV,CoolsF,MortierF,etal.Medicalend-of-lifedecisionsinneonatesandinfants

inFlanders.Lancet.2005;365(9467):1315-1320.

20. ChervenakFA,McCulloughLB.Ethicaldimensionsofnon-aggressive fetalmanagement.SeminFetalNeonatalMed.2008;13(5):316-319.

21. Dombrecht L, Beernaert K, Roets E, et al. A post-mortem population survey on foetal-infantileend-of-lifedecisions :aresearchprotocol.BMCPediatr.2018:1-9.

22. PapiernikE,ZeitlinJ,DelmasD,etal.Terminationofpregnancyamongverypretermbirthsanditsimpactonverypretermmortality:ResultsfromtenEuropeanpopulation-basedcohortsintheMOSAICstudy.BJOGAnIntJObstetGynaecol.2008;115(3):361-368.

23. BosmaJM,vanderWalG,Hosman-BenjaminseSL.LateterminationofpregnancyinNorthHolland.BrJObstetGynaecol.1997;104(4):478-487.

24. Garne E, Khoshnood B, LoaneM, Boyd PA, DolkH. Termination of pregnancy for fetalanomalyafter23weeksofgestation:AEuropeanregister-basedstudy.BJOGAnIntJObstetGynaecol.2010;117(6):660-666.

25. Lawonterminationofpregnancy[inDutch].BelgianlawGaz.1990.26. ChambaereK,BernheimJ.End-of-LifeDecisions.In:TextbookofPalliativeCare.;2018:1-

21.

27. ChambaereK.Medicalend-of-lifepracticesinFlandersandBrussels,Belgium(Doctoralthesis).2010.

28. WatsonR.Belgiumextendseuthanasialawtochildren.BMJ.2014;348:g1633.29. BoydPA,DeViganC,KhoshnoodB,etal.SurveyofprenatalscreeningpoliciesinEurope

forstructuralmalformationsandchromosomeanomalies,andtheirimpactondetectionandterminationratesforneuraltubedefectsandDown’ssyndrome.BJOGAnIntJObstetGynaecol.2008;115(6):689-696.

30. BlondelB,CuttiniM,Hindori-MohangooAD,etal.HowdolateterminationsofpregnancyaffectcomparisonsofstillbirthratesinEurope?AnalysesofaggregatedroutinedatafromtheEuro-PeristatProject.BJOGAnIntJObstetGynaecol.2018;125(2):226-234.

31. Berer M. Abortion Law and Policy Around theWorld: In Search of Decriminalization.HealthHumRights.2017;19(1):13-27.

32. Sauer PJJ, and the members of the working group. Ethical Dilemmas in Neonatology:Recommendations of the Ethics Working Group of the Cesp. Eur J Pediatr.2001;160(6):364-368.

33. MinisterievanVeiligheidenJustitieenMinisterievanVolksgezondheidWelzijnenSport.Regelingbeoordelingscommissie late zwangerschapsafbreking en levensbeëindigingbijpasgeborenen.Staatscourant.2016:3145.

34. Francis N. Neonatal deaths under Dutch Groningen Protocol very rare despitemisinformationcontagion.JournalofAssistedDying.2016;11(11):7-19.

35. JacquemynY(ConsensusgroepVNenM. [RecommendationsPerinatalCareConcerningViabilityinFlanders](inDutch).TijdschrGeneeskd.2014;70:20:1159-1167.

36. FeltmanDM,DuH,LeuthnerSR.Surveyofneonatologistsattitudestowardlimitinglife-sustainingtreatmentsin theneonatalintensivecareunit. JPerinatol.2012;32(11):886-892.

37. RicklessSC.Thedoctrineofdoingandallowing.PhilosRev.1997;106(4):555-575.

33

38. PinterAB.End-of-lifedecisionbeforeandafterbirth:changingethicalconsiderations. JPediatrSurg.2008;43(3):430-436.

39. NorcrossA.IntroductiontotheSecondedition.In:SteinbockB,NorcrossA,eds.KillingandLettingDie:SecondEdition.FordhamUniversityPress;1994:1-23.

40. McMahanJ.Killing,lettingdie,andwithdrawingaid.Ethics.1993;103(2):250-279.41. LarcherV.Ethicalconsiderationsinneonatalend-of-lifecare.SeminFetalNeonatalMed.

2013;18(2):105-110.

42. PartridgeJC,MartinezAM,NishidaH,etal.InternationalComparisonofCareforVeryLowBirth Weight Infants : Parents ’ Perceptions of Counseling and Decision-Making.2005;116(2):28-38.

43. AujoulatI,HenrardS,CharonA,etal.End-of-lifedecisionsandpracticesforverypreterminfantsintheWallonia-BrusselsFederationofBelgium.BMCPediatr.2018;18:1-10.

44. EUROCATregister.[Availablefrom:www.eurocat-network.eu].45. RoyR,AladangadyN,CosteloeK, LarcherV.Decisionmakingandmodesofdeath ina

tertiaryneonatalunit.ArchDisChild-FetalNeonatalEd.2004;89(6):F527-F530.

46. BartonL.TheContributionofWithholdingorWithdrawingCare toNewbornMortality.Pediatrics.2005;116(6):1487-1491.

47. Verhagen AAE, Janvier A, Leuthner SR, et al. Categorizing Neonatal Deaths: A Cross-Cultural Study in the United States, Canada, and The Netherlands. J Pediatr.2010;156(1):33-37.

48. RebagliatoM,CuttiniM,KaminskiM,Persson J,ReidM,SaracciR.NeonatalEnd-of-LifeDecisionMaking:Physicians’attitudesandrelationshipwithself-reportedpracticesin10Europeancountries.Jama.2000;284(19):2451-2459.

49. Cuttini M. The European Union Collaborative Project on Ethical Decision Making inNeonatal Intensive Care (EURONIC): findings from 11 countries. J Clin Ethics.2001;12(3):290-296.

50. VanDerHeideA,VanDerMaasPJ,VanDerWalG,etal.Medicalend-of-lifedecisionsmadeforneonatesandinfantsintheNetherlands.Lancet.1997;350(9467):1329-1331.

51. ten Cate K, van de Vathorst S, Onwuteaka-Philipsen BD, van der Heide A. End-of-lifedecisions for children under1 year of age in theNetherlands: decreased frequency ofadministrationofdrugstodeliberatelyhastendeath.JMedEthics.2015;41(10):795-798.

52. RausK.TheExtensionofBelgium’sEuthanasiaLawtoIncludeCompetentMinors.JBioethInq.2016:305-315.

53. Rüegger C, Hegglin M, Adams M, Bucher HU. Population based trends in mortality,morbidity and treatment for very preterm- and very low birthweight infants over 12years.BMCPediatr.2012;12(1):17.

54. WilkinsonDJC,TruogRD.Theluckofthedraw:Physician-relatedvariabilityinend-of-lifedecision-makinginintensivecare.IntensiveCareMed.2013;39(6):1128-1132.

55. AzoulayÉ,MetnitzB,SprungCL,etal.End-of-lifepracticesin282intensivecareunits:DatafromtheSAPS3database.IntensiveCareMed.2009;35(4):623-630.

56. BurnsJP,MitchellC,GriffithJL,TruogRD.End-of-lifecareinthepediatricintensivecareunit:attitudesandpracticesofpediatriccriticalcarephysiciansandnurses.CritCareMed.2001;29(3):658-664.

57. RebagliatoM,CuttiniM,KaminskiM,Persson J,ReidM,SaracciR.NeonatalEnd-of-LifeDecisionMaking.2000;284(19):2451-2459.

34

58. BucherHU,KleinSD,HendriksMJ,etal.Decision-makingatthelimitofviability:Differingperceptions and opinions between neonatal physicians and nurses. BMC Pediatr.2018;18(1):81.

59. Garten L, Daehmlow S, Reindl T, Wendt A, Münch A, Bührer C. End-of-life opioidadministration on neonatal and pediatric intensive care units: Nurses’ attitudes andpractice.EurJPain.2011;15(9):958-965.

60. Miquel-Verges F, Woods SL, Aucott SW, Boss RD, Sulpar LJ, Donohue PK. Prenatalconsultation with a neonatologist for congenital anomalies: parental perceptions.Pediatrics.124(4):e573-e579.

61. GarelM,CaeymaexL,CuttiniM,KaminskiM.Ethicallycomplexdecisionsintheneonatalintensive careunit : impact of thenewFrench legislationonattitudes andpracticesofphysiciansandnurses.2011:240-244.

62. Brinchmann BS. Ethical Decisions about Neonates in Norway. J Nurs Scholarsh.1999;31(3):1998-1999.

63. Brinchmann BS, Nortvedt P. Ethical decisionmaking in neonatal units--the normativesignificanceofvitality.MedHealthCarePhilos.2001;4:193-200.

64. Wocial LD.Life SupportDecisions Involving Imperiled Infants. J PerinatNeonatalNurs.2000;14(2):73-86.

65. BrinchmannBS,FørdeR,NortvedtP.Whatmatterstotheparents?Aqualitativestudyofparents’ experienceswith life-and-death decisions concerning their premature infants.NursEthics.2002;9(4):17-19.

66. BrootenD,YoungblutJ,SeagraveL,etal.Parent’sperceptionsofhealthcareprovidersactionsaroundchildICUdeath :whathelped,whatdidnotNIHPublicAccess.AmJHospPalliatCare.2013;30(1):40-49.

67. CurrieER,ChristianBJ,HindsPS,etal.ParentPerspectivesofNeonatalIntensiveCareattheend-of-life.JPediatrNurs.2016;31(5):478-489.

68. EpsteinEG.End-of-lifeexperiencesofnursesandphysiciansinthenewbornintensivecareunit.JPerinatol.2008;28:771-778.

69. EpsteinEG.Moralobligationsofnursesandphysiciansinneonatalend-of-lifecare.NursEthics.2010;17(5):577-589.

70. CatlinA,CarterB.CreationofaNeonatalEnd-of-LifePalliativeCareProtocol.JPerinatol.2002;22:184-195.

71. Lee KJ, Dupree CY, Ph D, Fellow CBHD. Staff Experienceswith End-of-Life Care in thePediatricIntensiveCareUnit.JPalliatMed.2008;11(7):986-990.

72. Kaplan LJ. Toward aModel of Caregiver Grief: Nurses’ Experiences of Treating DyingChildren.Omega-JDeathDying.2000;41:187-206.

73. Mol MMC Van, Kompanje EJO, Benoit DD, Bakker J, Nijkamp MD. The Prevalence ofCompassionFatigueandBurnoutamongHealthcareProfessionalsinIntensiveCareUnits :ASystematicReview.PLoSOne.2015:1-22.

74. BellieniCV.,RighettiP,CiampaR,LacoponiF,CovielloC,BuonocoreG.Assessingburnoutamongneonatologists.JMaternNeonatalMed.2012;25(10):2130-2134.

75. ProfitJ,SharekPJ,AmspokerAB,etal.BurnoutintheNICUsettinganditsrelationtosafetyculture.BMJQualSaf.2014;23:806-813.

76. HallSL,CrossJ,SelixNW,etal.RecommendationsforenhancingpsychosocialsupportofNICUparentsthroughstaffeducationandsupport.JPerinatol.2015;35(S1):S29-S36.

35

77. NationalAssociationofPerinatalSocialWorkers.StandardsforSocialWorkServicesintheNewbornIntensiveCareUnit.

78. HynanMT,SteinbergZ,BakerL,etal.RecommendationsformentalhealthprofessionalsintheNICU.JPerinatol.2015;35(S1):S14-S18.

79. GoldKJ.Navigatingcareafterababydies :asystematicreviewofparentexperienceswithhealthproviders.JPerinatol.2007;27:230-237.

80. WilliamsC,MunsonD, Zupancic J,KirpalaniH. Supportingbereavedparents :practicalsteps in providing compassionate perinatal and neonatal end-of-life care - A NorthAmericanperspective.SeminFetalNeonatalMed.2008;13:335-340.

81. EnglerAJ,CussonRM,BrockettRT,etal.NeonatalStaffandAdvancedPracticeNurses’Perceptions of Bereavement/End-of-Life Care of Families of Critically Ill and/orDyingInfants.AmJCritCare.2004;13(6).

82. DareT.Parentalrightsandmedicaldecisions.PaediatrAnaesth.2009;19(10):947-952.83. CaeymaexL,SperanzaM,VasilescuC,etal.Livingwithacrucialdecision:Aqualitative

studyofparentalnarrativesthreeyearsafterthelossoftheirnewbornintheNICU.PLoSOne.2011;6(12).

84. McHaffieHE,LaingIA,ParkerM,McMillanJ.Decidingforimperillednewborns:Medicalauthorityorparentalautonomy?JMedEthics.2001;27(2):104-109.

85. Einarsdóttir J. Emotional experts: Parents’ views on end-of-life decisions for preterminfantsinIceland.MedAnthropolQ.2009;23(1):34-50.

86. McHaffieHE,LyonAJ,HumeR.Decidingontreatmentlimitationforneonates:Theparents’perspective.EurJPediatr.2001;160(6):339-344.

87. CarnevaleFA,CanouïP,HubertP,etal.Themoralexperienceofparentsregardinglife-supportdecisionsfortheircritically-illchildren:ApreliminarystudyinFrance.JChildHealCare.2006;10(1):69-82.

88. Dombrecht L, Deliens L, Chambaere K, et al. Neonatologists and neonatal nurses havepositive attitudes towards perinatal end-of-life decisions, a nationwide survey. ActaPaediatr.2019:apa.14797.

89. BelliniS,DamatoEG.Nurses’knowledge,attitudes/beliefs,andcarepracticesconcerningdonotresuscitatestatusforhospitalizedneonates.JOGNN-JObstetGynecolNeonatalNurs.2009;38(2):195-205.

90. ChambaereK,BilsenJ,CohenJ,etal.Apost-mortemsurveyonend-of-lifedecisionsusingarepresentativesampleofdeathcertificatesinFlanders,Belgium:researchprotocol.BMCPublicHealth.2008;8(1):299.

91. WeintraubAS,GeithnerEM,StroustrupA,WaldmanED.Compassionfatigue,burnoutandcompassionsatisfactioninneonatologistsintheUS.NatPublGr.2016;36(11):1021-1026.

36

37

Part2Medicalend-of-lifedecisionsinstillbirths,neonatesandinfantsinFlanders

Chapter2:Apost-mortempopulationsurveyonfoetal-infantileend-of-lifedecisions:aresearchprotocol

Chapter3:End-of-lifedecisionsinneonatesandinfants:arepeatedpopulation-levelmortalityfollow-backstudy

38

Apost-mortempopulation

surveyonfoetal-infantileend-of-lifedecisions:aresearch

protocol

LaureDombrecht,KimBeernaert,EllenRoets,KennethChambaere,FilipCools,LindeGoossens,GunnarNaulaers,LucDeCatte,JoachimCohen,LucDeliensonbehalfoftheNICUconsortium

PublishedinBMCPediatrics2018,18(1):260.

39

Abstract

Background:Thedeathofachildbeforeorshortlyafterbirthisfrequentlyprecededbyanend-of-life decision (ELD). Population-based studies of incidence and characteristics of ELDs inneonatesandinfantsarerare,andthoseinthefoetal-infantileperiod(>22weeksofgestation–one year) including both neonates and stillborns, are non-existent. However, importantinformation ismissedwhen decisionsmade before birth are overlooked. Our study protocoladdressesthisknowledgegap.

Methods:First,anewandencompassingframeworkwasconstructedtoconceptualiseELDsinthefoetal-infantileperiod.Next,apopulationmortalityfollow-backsurveyinFlanders(Belgium)was setupwithphysicianswho certifiedalldeath certificatesof stillbirths from22weeksofgestation onwards, and infants under the age of a year. Two largely similar questionnaires(stillbirthsandneonates)weredeveloped,pilottestedandvalidated,bothincludingquestionsonELDsandtheirprecedingdecision-makingprocesses.Eachdeathrequiresapostalquestionnairetobesenttothecertifyingphysician.Anonymityofthechild,parentsandphysicianisensuredbya rigorous mailing procedure involving a lawyer as intermediary between death certificateauthorities, physicians and researchers. Approval by medical societies, ethics and privacycommissionshasbeenobtained.

Discussion: This research protocol is the first to study ELDs over the entire foetal-infantileperiod on a population level. Based on representative samples of deaths and stillbirths andapplying a trustworthy anonymity procedure, the research protocol can be used in othercountries,irrespectiveoflegalframeworksaroundperinatalend-of-lifedecision-making.

40

2.1 Background

Recent decades have seen an increase in possible medical and technical interventions forcriticallyillneonatesandinfants1.However,inFlanders,Belgiumabout8.7perthousandchildrenstilldieduringthefoetal-infantileperiod,i.e.fromfoetusesofmorethan500gramsor22weeksof gestationupuntil oneyearafterbirth2. This is comparablewithdeath rates reported, forinstance, in the United States3. Many of these deaths occur at neonatal intensive care units(NICUs)andareprecededbyapossiblylife-shorteningend-of-lifedecision(ELD)4–6.Inneonates,these include non-treatment decisions such as withholding or withdrawing life-sustainingtreatment, intensification of alleviation of pain and/or other symptomswith a potential life-shorteningeffectandintentionallyendinglifewithlethaldrugs7.Additionally,prenataldiagnostictechniques(genetictechniques,prenatalimagingtechniques)haveevolvedconsiderably,leadingtoanincreasingnumberofcongenitalmalformationsbeingdiagnosedprenatallyinsteadofafterbirth8,9. Some decisions such as abstinence from treatment8–10 or termination of pregnancy(TOP)8,9canbemadeduringgestationincasesofthedetectionofseriousabnormalities11–13.Forstillbirthsfrom22weeksofgestationandonwards–whichisconsideredasthedefinitionofaviable foetus by theWHO – TOPs are considered late terminations. Stillborns and deceasedneonates cannotbe seen as separatepatientpopulations, since they are in essence the samepatientwhereanELDcanbemadeeitherbeforeorafterbirth.Theonlydifferenceisthereforetheoccurrenceofbirthandnotnecessarilyadifference indisordersorcongenitalanomalies.Research into end-of-life decision-making on a population level should therefore take intoaccount the foetal-infantileperiod in its entirety (insteadof bothperiods separately).This isneededtoprovidereliableincidenceratesandinformationonthedecision-makingprocessinthisvulnerablepopulation.EvaluationandmonitoringofELDpracticeintheentirefoetal-infantileperiodcouldleadtobetterunderstandingofcurrentprenatalandneonatalhealthcareanddetectpointsofimprovementsincetherehavebeennoall-inclusiveguidelinesuptothepresent.

Population-basedstudies(i.e.withalldeathcasesasthefocus)areidealtostudytheincidenceandcharacteristicsofELDs,butsuchstudiesarerare inneonatesand infants14–16and, toourknowledge,non-existentinstillborns.Inneonates,resultsaremostlybasedonreviewsofmedicalrecordsofaNICUataparticularhospital.Inthesestudies40%to93%ofdeathsinaNICUfollowwithdrawal of life-sustaining treatments6,17–19. The larger scaleEURONIC studywasbasedonphysicians’ self-reported practices within 143 European NICUs in the 1990s20. The onlypopulation-basedstudiesarefromtheNetherlands(in2014)15andBelgium(in2000)14.Thesestudies foundanELDbeingmade in60%of all deathsofneonates and infants. In stillborns,previous studies in 200311 and in 2000-2005 13 have only looked at the prevalence of lateTOP11,13,21.Notmuch is known about the entiretyof end-of-life practices (includingdecisionsother than TOP) and their decision-making process, or about patient characteristics besidesgestationalageandthepresenceoffoetalanomalies.

WedevelopedastudydesigntoevaluateandmonitorELDsandtheirdecision-makingprocessacross the entire foetal-infantile period in Flanders, Belgium. The study design involves thedevelopmentofavalidatedconceptualframeworkofELDsspanningtheentirefoetal-infantileperiod (based on existing frameworks) and the development of a survey methodology thataddressestheparticulardifficultiesincapturingandsurveyingstillbirthsandneonataldeaths,andprovidesopportunitiesforcomparisonofELDpracticesbetweenhospitals.

41

2.2 Methods

This population study has the design of a mortality follow-back survey based on all deathcertificatesofstillbirthsandneonates.Questionnairesareeithersenttothecertifyingphysiciansby post or are provided at maternity wards. In order to develop these questionnaires,adjustmentstoanexistingneonatalELDframeworkneededtobemade.

2.2.1 Conceptualframeworkoffoetal-infantileELDs

PrenatalELDsshouldbetakenintoaccountwhenpresentingareliableandcompletepictureonfoetal-infantileELDpractices.However,todatetheseprenatalELDshavenotbeenincludedinacomprehensiveframeworkwithneonatalELDs.Weadjustedapreviouslyexistingandvalidatedframework of ELDs in neonates7 in order to include both prenatal and neonatal ELDs. Thisframework7 includes three dimensions: ‘medico-technical‘, ‘medico-ethical’ and ‘consultationwithparents’.Thedimension‘consultationwithparents’wasexcludedfromourownframeworksince no decision can be made prenatally without at least the mother consenting to anintervention.Furthermore,thedimension‘consultationwithpatients’isalsoexcludedfromtheadultELDframeworkwherethemedicaldecisionanditsintentionaretheonlydeterminantsofanELD.However,thisdimensionisstillveryimportantwhichiswhyconsultationwithparentswill stillbeaddressed indetailbymeansofadditionalquestionsoutside theELD framework.Theseencompassthefollowing:

1. Themedico-technicalclassificationormedicalacts7,22:- non-treatment decisions such as withholding or withdrawal of life-sustainingtreatment

- administeringdrugsormedicalinterventions2. Themedico-ethicalclassificationorthelife-shorteningintentionofthephysiciancan

be7,14:- nointentionbuttakingintoaccountapotentiallylife-shorteningeffect- the potentially life-shortening effect is not the main goal but partly intended (co-intention)

- anexplicitlife-shorteningintention.

Tocoverallpossibledecisionsthatcouldpossiblyinfluencethedeathofafoetusorinfant,bothdimensionsshouldbetakenintoconsideration.Asasidenote,intentionallyendingthelifeofachildisillegal,meaningthatinthiscase,themedico-ethicaldimensionisconsideredtobeallthemoreimportantsincenoemphasisisputonthemedico-technicalclassificationspecifically.

Wepresentedthisframeworkforvalidationtogynaecologistsineightindividualinterviewsandtwoexpertpanelsrepresentingsevendifferenthospitals.ThegynaecologistswereaskedtogiveclinicalexamplesforallpossibleELDcategoriesappliedtotheprenatalcontext,andtoaddmorecategoriesincaseanyweremissing.Assoonasarealisticexamplewasgivenandagreedonbyothers, that ELD was considered possible and included in the framework (Table 2.1). Theresultingfoetal-infantileELDframeworkwasthenthoroughlyreviewedbythreeneonatologists.

42

Literature on end-of-life practices prior to stillbirth distinguishes between non-aggressiveobstetricmanagementandTOP9,10,23.Non-aggressiveobstetricmanagement(orabstinencefromtreatment)isthedenialofinterventionswhichareneededtosustainthelifeofthefoetusbecauseofapoorfoetalprognosis8–10.TOPhowever,activelyendsfoetal life8,9bypreterminductionoflaboureitherwithorwithoutfeticide(administeringmedicationtointentionallyendthelifeofthefoetusbeforebirth)priortothetermination24.

Medical-technicalclassification

Life-shorteningIntention

End-of-lifepracticespriortostillbirth

End-of-lifepracticesinneonatesandinfants

Non-treatmentdecisions

Nointention e.g.notocolysisinpretermlabourat24weeks’gestation

e.g.noantibiotics

Co-intention e.g.nocardiacsurgeryExplicitintention e.g.nointrauterine

transfusionforananaemicfoetus

e.g.withdrawingventilation

Drugadministrationormedicalintervention

Nointention e.g.start-upofanticonvulsivetherapy

Co-intention e.g.administeringmorphine

Explicitintention e.g.terminationofpregnancywithfeticide

e.g.administeringmusclerelaxant

Table2.1Acomprehensiveframeworkofend-of-lifepracticesinthefoetal-infantileperiod.

2.2.2 Questionnaires

Basedonthisadjustedframework,twoseparatebutsimilarquestionnairesweredevelopedforELDs in stillborns and ELDs inneonates respectively, since both populations have their ownspecificities.BothquestionnairesincludequestionsaboutELDs,thedecision-makingprocess,theinvolvementofparentsinthisprocess,theinvolvementofcolleaguesandexperts,andtheELDpolicyofthehospital.

Forneonatesandinfants,previouslyvalidatedquestionnairesthatfocusonend-of-lifedecisionsinminorsandneonates14,16,25wereusedasthebasisforourquestionnaire.Wemainlyfocusedonupdating the terms and grammar used, term ambiguity, length of the questionnaire andcomparability to the previous ELD study14. The resulting questionnairewas thoroughly pilottested and validated with eight neonatologists who represented all eight Flemish NICUs,researchersinthefieldofend-of-lifecareandanethicist.

For ELDs in stillborns a new questionnaire was developed based on previously validatedquestionnairesonTOPafter22weeks11,12,questionnairesonELDsinminorsandneonates14,16,26,andthenewlydevelopedframeworkforend-of-lifepracticesinthefoetal-infantileperiod.Thisquestionnaire was thoroughly pilot tested and validated with eight gynaecologists, three

43

neonatologists,researchersinthefieldofend-of-lifecare,anethicistandalawyerinthefieldofend-of-lifecare.

NeitherquestionnaireasksdirectlyaboutcategoriesofELDsbutclassifiesthesebasedonaseriesofcorequestionsfollowingthetwodimensionsoftheconceptualframeworkabout1)whichactoromissionwasused(medico-technical),and2)whichlife-shorteningintentionwasassociatedwiththeact(medico-ethical).Additionalquestionswereaskedaboutthewaysinwhichparentswereinvolvedinthedecision-makingprocess(parentconsultation).

2.2.3 Populationandsetting

Thepopulationincludes:allstillbirthsfrom22weeksofgestationormoreand/orabirthweightof500gorhigher(i.e.theinternationallyacknowledgedlimitofviabilityofthefoetus24,27,28)andalldeceasedneonatesandinfantsundertheageofoneyearoccurringinFlandersandBrusselswherethemotherisaFlemishresident.Nosampleisdrawn;thefullpopulationisincludedoveradatacollectionperiodof12monthsforstillbirthsand16monthsforneonatesandinfants.Thelongerobservationperiodforneonateandinfantdeathswaschosenbecausethesedeathsarelesscommonthanlateterminationstillbirths2andwewantedtoobtainapopulationlargeenoughtomakereliableprevalenceestimatesofend-of-lifepractices.

Deaths tobe included in the studyare identifiedusing thedeath certificate. Everydeathof aFlemishresidentinFlandersandBrusselsmustbedeclaredbymeansofadeathcertificatetotheFlemishAgencyforCareandHealthoftheMinistryoftheFlemishCommunityortheBrusselsHealth and Social Observatory respectively. The physician, in our study most probably aneonatologist,paediatricianorgynaecologist, completes themainpartof thedeathcertificatewhichindicatesthesexofthechild,thedateofbirthandthedateofdeath,medicalinformationsuchasthecauseofdeath,whetherornotthechildwasaliveatthetimeofbirth,andthetimeand place of death29. The physician then signs the certificate and adds his or her medicalregistrationnumber.Thedeathcertificateisthensenttothecivilregistrarofthemunicipalitywherethedeathtookplacewhereadditionalinformationiscompletedonthedeathcertificatesuch as socio-demographic information about the child and its parents. Certificates are thenprocessedbytheprovincesbeforebeingsenttothecentraladministrationauthorities.Itcantakeuptothreemonthsfordeathcertificatestoreachtheseadministrationauthorities.

2.2.4 Designandprocedures

Amortalityfollow-backprocedureisfollowed,slightlymodifyingwell-establishedproceduresinadults22 and minors26. Modifications concern a more stringent anonymity procedure and analternative identification procedure for stillbirths between 22 and 26 weeks. As for theanonymity procedure, ethical and legal considerations (criminal prosecution is possible forreported illegal ELDs) make it necessary to pay greater attention to the protection ofconfidentialityofthephysician,totheprivacyofthedeceased,theparentsandtherelatives,andtothesecurityofthedatathatwillbeobtainedinthesurvey.Byensuringtotalanonymity,boththeresponserateandthereliabilityoftheresponsescanbeimproved.Thedifferentstagesofthe

44

survey i.e. themailing, receiving and processing of the questionnaireswill be separated andperformedbyfourseparateentities(seefigure2.2).

1. Thedeathcertificateadministrationauthorities(namelyFlemishAgencyforCareandHealth of theMinistry of Welfare, Health and Family of the Flemish Government) isresponsibleforconstructionandmanagementofthemailingdatabaseandthemailingofthequestionnaires.Eachcaseisascribedauniquecodednumberderivedfromthedeathcertificatenumber.Theseuniquenumbersareusedattheendof thestudyto linkthequestionnairestothedemographicandmorbiditydata(suchasICD-10codesofthecauseofdeath)ofthedeceased,derivedfromthedeathcertificates,inadatabaseprovidedbytheadministrationauthorities.Anaccompanyingletterisincludedwiththequestionnaireprovidingthephysicianwithenoughpatientcharacteristicstoidentifythepatient.Theseincludesex,dateof(still)birth,dateofdeathandmunicipalityofdeath;forstillbornsthedateofdeathisreplacedbythedateofbirthofthemother.Whenthelawyer(seebelow)receives thequestionnaireheorshereportsback to theFlemishAgency forCareandHealth;allidentifiabledatarelatedtothepatientandthephysicianinquestionisthenremovedfromthestudydatabase.Afollow-upmailingofthreeremindersisperformed14,28and42daysaftertheinitialquestionnairewassent(followingtheTotalDesignMethod30).

2. The physician identifies the deceased or stillborn child based on the patientcharacteristicsprovided,fillsoutthequestionnairesandreturnsthesetoalawyerusingapostagepaidenvelope.Incasethecertifyingphysicianisnotthetreatingorattendingphysicianheorsheisgivenspecificinstructionstopassthequestionnairetothetreatingphysicianifpossible.

3. The lawyer, who is bound by confidentiality, safeguards the anonymity of thequestionnaires.Heorshecodestheparticipatinghospitalwardssothatcomparisonscanbe made, and removes any possible identifying information of hospital, physician orpatient,removestheuniquenumbersandreportsthesetotheadministrationauthorities.Additionally, place of death will not be sent to the researchers in order to ensureanonymityof theparticipatinghospitals.The lawyer links thequestionnaireswith theinformationonthedatabasefromthedeathcertificateadministrationauthorities,andattheendofthedatacollectionsendsthelinkeddatabasetotheresearchergroupinwhichall identifiers will be removed and information can no longer be traced back to thecorrespondingdeathcertificate.

4. The researchgroup receivesquestionnairesandensures thatboth inprocessingandanalysingthedatabaseitwillnotbepossibletodeterminetheidentityofthepatientorthephysician.

An alternative identification procedure for stillbirths between 22 and 26 weeks is includedbecausethedeathcertificatemethodprovestobechallengingforstillbirthsinthatagegroup.Fillingindeathcertificatesofstillbirthsbetween22and26weeksofgestationbyaphysicianisnotmandatory,whichmakesthedeathcertificatesapotentiallyincompletesamplingframework.We provided questionnaires to the ten biggestmaternitywards inFlanders and the FlemishhospitalsofBrusselssothatphysicianscanfilloutthisquestionnaireforeverystillbirthfrom22weeks of gestation onwards and/or child with a birthweight from 500 g onwards. ThesematernitywardswerechosenbasedonthepresenceofaNICUatthehospital,becauseofahigh

45

birthrateand/orbecausetheyaretertiarycentresforprenataldiagnostics.Foreachstillbirthforwhichaquestionnaireiscompleted,thephysicianisalsoaskedtofilloutadeathcertificate.Thismakes it possible for the lawyer to link the answers in the questionnaire to the clinical anddemographiccharacteristicsofthestillbornchild(foraschematicoverviewofthisprocedure,seefigure2.3).Thephysician sends thequestionnaire, togetherwith a separate letter containingpatientidentificationdetailstothelawyerandsendsthecertificatetotheofficialdeathcertificateagency.Becausethelattersendspatientidentificationdetailsofdeathcertificatesforstillbirthstothelawyer,thelawyercanthendeterminewhetheraquestionnairehasalreadybeenreceivedfor that death and notify the Flemish agency for Care andHealth via email. In this case, noquestionnaires are sent by the death certificate agency. The separate letter with patientidentificationdetails isdestroyedassoonas thequestionnaire is linked to thecorrespondingdeathcertificate.Ifaphysiciandidnotfilloutthequestionnaireavailableinthematernitywardbutdidfileadeathcertificate,theywillstillreceiveaquestionnairethroughtheregularpostalsurvey.

2.2.5 Improvingresponserates

To increase response,we follow theTotalDesignMethod (TDM)30.Therefore, physicianswillreceiveamaximumofthreefollow-uppostalmailings.Inordertofurtherimprovetheresponseratebothinstillbornsandinneonatesandinfantswewilladdanadditionalgeneralfollow-up.EverythreetofourmonthsoneoftheresearcherswillvisitalleightFlemishNICUsandalltenparticipatingmaternitywardstoinquireaboutthecourseofthestudy.Duringvisits,physicianswillbeabletoaskquestions,voiceconcernsorgivegeneralfeedback.Thesevisitsarealsomeantto counter responder fatigue by stimulatingmotivation for the study duration. Furthermore,three consortium meeting will be organised to discuss the progress of the study withrepresentativesofeveryNICU(onebeforethestart,onehalfwaythroughandoneattheendofthestudy).Lastly,thestudyisalsopresentedatrelevantconferencesandmeetings.

2.2.6 Ethicalaspectsanddataprotection

Thesensitivityoftheresearchpopulationandthedelicatenatureofourquestionnairemakesitnecessarytofollowarigorousethicalapprovalprocedure.Ethicsapprovalwasobtainedfromtheethics committee of the University Hospital of Ghent and additionally from the PrivacyCommission (CBPL), the Sectoral Committee of Social Security and Health, and the NationalCounciloftheOrderofPhysicians.Forourparallelprocedureinthetenbiggestmaternitywards,weobtainedethicsapprovalfromtheethicscommitteesofallparticipatinghospitals.

Toensureprivacyandanonymity,aswellastheprecautionsthathavealreadybetakenbyusingalawyer,westrivetoensurefulldataprotection.Thedataarealwayspasswordprotectedandstoredonaprotectedserver.Thedatabaseisnotreplicatedorsharedwiththirdparties;allcopiesneededforanalysisaredestroyedafterwards.

46

2.2.7 Data-analysis

AnSPSS24.0(SPSSInc.)fileissetupbytheresearchgroupwithacodingschemeforacertifieddatamanagementcompanythatwillenterthedata.Theresearcherswillperformalldatacleaningthrough SPSS syntax operations. Data will be analysed with descriptive statistics (validpercentages),bivariateandmultivariateassociationstatistics.

Figure2.2Schematicoverviewofthemailingandanonymityprocedure.

Physician - Signs death certificate

- Fills in questionnaire

Flemish Agency for Care and Health - Provides the death certificates - Conducts survey/ executes mailings

- Manages survey database - Prepares database of stillborn

characteristics - Provides ICD-10 codes

Lawyer - Links the questionnaires with the clinical

characteristics of the stillborns - Replaces sample number on questionnaire with new

number - Replaces hospital number with new number - Manages database of received sample numbers and

corresponding new numbers - Keeps received questionnaires until end of survey - Replaces sample numbers in the database of

stillborn characteristics with corresponding new numbers

- Replaces hospital numbers in the database of stillborn characteristics with corresponding new numbers

- Assists in small cells risk analysis before the database is sent to the research group

1 2 Questionnaire + letter with

the stillborn characteristics Reminders (if necessary)

3 Completed

questionnaire

4

5 Received sample numbers

Database of the stillborn

characteristics

Death certificate

Research group - Combines questionnaire data and corresponding

stillborn characteristics based on new numbers

6 Questionnaires + database of stillborn characteristics + anonymized NIC/MIC hospital codes

47

Figure2.3Schematicoverviewoftheparallelprocedureinthetenbiggestmaternitywards.

Physician - Signs death certificate

- Fills out questionnaire + letter with separate patient identification details

Flemish Agency for Care and Health - Samples every received

death certificate indicating a stillbirth

- Sends patient identification details of every certified stillbirth together with the unique and anonymous sample number to the lawyer

- Manages survey database - Prepares database of

stillborn characteristics

Lawyer - Save patient identification details in a

separate database until the data can be anonymized

- Receive identification details of every certified stillbirth from the Flemish Agency for Care and Health

- Send unique sample number of already received questionnaires back to the Flemish Agency for Care and Health

- Writes new number (corresponding with unique sample number) on questionnaire

- Manages database of received sample numbers and corresponding new numbers

- Keeps received questionnaires until

end of survey - Replaces unique sample numbers in

the database of stillborn characteristics with corresponding new numbers

- Assists in small cells risk analysis before the database is sent to the research group

1B. Completed questionnaire + letter with patient identification details

2. Patient identification details + unique sample number

1A. Death certificate

Research group - Combines questionnaire data and corresponding

stillborn characteristics based on new numbers

5. Questionnaires + database of stillborn characteristics

Stillbirth in one of the 10 selected hospitals

3. Unique sample numbers of already received questionnaires

4. Database of stillborn characteristics

48

2.3 Discussion

Theobjectivesof thispopulation studyare to evaluate andmonitorELDsand theirdecision-makingprocessinthefoetal-infantileperiodincludingELDsinthefoetalandtheneonatalperiod.Thisstudydesignhasseveralpotentialstrengthsaswellassomelimitationsassociatedwiththestudypopulationandthesurveymethod.

2.3.1 Strengths

Ourstudyisthefirsttoexaminefoetal-infantileELDsintheirentirety.Theresultswillbroadenknowledge onwhichmedical decisions aremade in cases of congenital anomalies or severedisorders fromthemomentofviability, regardlessofwhetherornot thechildhasbeenborn.Even though ELDs have been researched both prenatally11,13,21 and in neonates14,16,31, thecontinuityof care and theoverarchingdecision-makingprocesshasbeenmissed inpreviousstudiesandthereforekeyelements(suchaswhethertheELDwasmadeprenatallybutperformedafterthechildisborn)couldbeoverlooked.

Even though there are some studies comparing late TOP practices across Europeancountries11,13,21, not much is known about the full scope of end-of-life practices before birth(includingnon-treatmentdecisions)andtheirdecision-makingprocess.However,non-treatmentdecisions such as non-aggressive obstetricmanagementwith orwithout explicit intention toshortenthelifeofthefoetuscanalsooccur.OneofthestrengthsofourstudyisthereforetheinclusionofalltypesofpossibleELDsinneonatesandalsobeforebirth.Furthermore,evenwhenthechilddiedpostnatallyweinquireaboutdecisionsbeingmadeprenatallyandtherebyprovideafulloverviewofELDswithoutpriorfocusononespecificELD.

Most research on ELDs in prenatal11,13 and neonatal6,17,20 settings is limited to single centrestudiesandbasedonreviewsofmedicalrecords.Population-basedstudiesbasedonofficiallyregistereddeathcertificates,likeours,arehoweverfarmorecapableofobtainingrobustdataandreliableincidenceratessinceanationwidescopeensuresthattheentirepopulationisincluded.Thesecouldinturnleadtobetterunderstandingofcurrentend-of-lifecareanddetectpointsofimprovementtobenefitfutureparentsandchildrenwithseveredisorders.Theonlypopulation-basedstudyonBelgianneonatalELDsdatesbackto200014andsincethen,importantsocietalchanges such as questioning futile medical end-of-life care and refuting the idea of curativetreatmentasbeingnecessarilybeneficialcouldpossiblyhavehadaneffectonend-of-lifepracticeinunbornbabiesandneonates32.

Asidefrompopulationspecificstrengths,somestrengthscanbeattributedtothedeathcertificatemethod in particular. These include international comparability, lack of patient burden andconsequentattritionrates,reliabilityofthedata,anonymity,andexclusionofpossibleselectionbiasbyselectingcertainphysicians for thestudy.Anoverviewof thestrengthsrelatedto thedeath certificatemethod,which has successfully been implemented in adults33,minors26 andneonates14,canbefoundintheresearchprotocolofChambaereetal.(2008)29.

49

2.3.2 Limitations

Oneoftheweaknessesofthestudyisthatthedeathcertificatemethodprovidesachallengeinthecaseofstillbirthbetween22and26weeksofgestationbecausecompletingadeathcertificateisnotmandatoryat thisage.Despiteouraddeddatacollectionmethod,wecannotguarantee100% coverage of stillbirths. Nevertheless, the reports from the Flemish centre of PerinatalEpidemiology, which registers every birth, will be available after the study andwill make itpossible to estimate the number of missing cases. Furthermore, despite the additional datacollectionmethodthereisalsonowaytoensurethatphysicianswillalwayscompleteadeathcertificate (as it is not obligatory), even when they fill out the questionnaire. It is thereforepossiblethatwewillreceivequestionnaireswhichwearenotabletolinktoadeathcertificatewhichwillthereforebeunusableforthisstudy.

Delays in the processing of death certificates can reach up to four months before thequestionnaire issentto thephysician in the firstmethod29.Therefore,arecallbiascannotbeexcluded.However,nootherregistrationofdeathsuptotheageofoneyearexistsandtheonlyother registration of all births (live and stillbirths) occurs at the Flemish centre of PerinatalEpidemiology.Thisconsistsoffewermissingcases,however,andthedelayinprocessingthesedocumentscanbeuptooneyearwhichwoulddrasticallydecreasethereliabilityoftheresponses.Furthermore,thismethodofregistrationisduetobemergedwiththeexistingdeathcertificateregistration,makingourmethodthemostreliableforfuturetrendresearch.

We include all stillbirths from22weeks of gestation onwards because this is internationallyacknowledgedtobethelimitofviabilityofthefoetus24,27,28.However,somecongenitalanomaliescanbedetectedbeforethisviabilitythresholdsowecannotexcludeanELDhavingbeenmadebeforethe22weekscut-offusedinthisstudy.Furthermore,mostFlemishneonatologywardsonlyconsiderviabilityfrom24,25oreven26weeksofgestationwhichcouldalsohaveanimpactonwhetherornotadeathcertificateisfilledout.

2.3.3 Implicationsforfutureresearchandpractice

Regularrepetitionofthisstudyinthefutureisneededinordertomonitorandevaluatechangesinend-of-lifepracticesinthefoetal-infantilegroup.Becausethisstudydesignallowsapplicationinothercountries,werecommendinternationalcomparativestudiestoprovideuswithbetterinsightintofoetal-infantileend-of-lifepracticesandincidenceratessothatinternationalfoetalandneonatalcareattheendoflifecanbeoptimised.

This can eventually aid thedevelopment of obstetrical, neonatal and paediatric guidelines tosupportanethicalend-of-lifedecision-makingprocess.

50

2.4 Declarations

2.4.1 Ethicsapprovalandconsenttoparticipate

Fortheneonatalpartofthisstudy,approvalwasobtainedfromtheEthicsCommitteeofGhentUniversity (Belgian Registration Number B670201628795), the Privacy Commission (CBPL,registration number SA3/VT005071970), the National Council of the Order of Physicians(registrationnumberBD/wc/89997)andtheSectoralCommitteeofSocialSecurityandhealth(registration number SCSZG/16/234). For the prenatal section of this study, approval wasobtainedfromtheCentralEthicsCommitteeofGhentUniversity(BelgianRegistrationNumberB670201730997), the Local Ethics Committees of the participating hospitals in Flanders, thePrivacyCommission(CBPL,registrationnumberSA3/VT005071970),theNationalCounciloftheOrderofPhysicians(registrationnumberBD/wc/89997)andtheSectoralCommitteeofSocialSecurityandhealth(registrationnumberSCSZG/17/029).Sendingbackthequestionnairewasseenasconsenttoparticipate.

2.4.2 Funding

ThisstudyisfundedbytheResearchFoundationFlanders(FWO)andthespecialresearchfundof Ghent University (BOF). K. Beernaert is Postdoctoral Fellow of the Research FoundationFlanders(FWO).

2.4.3 Acknowledgements

Wewould like to thanktheFlemishAgency forCareandHealth,and theBrusselsHealthandSocialObservatory.Wearealsodeeplyindebtedtoallphysicianswhoparticipatedintestingandvalidating the questionnaire, the participating hospitals (UZ Gent, UZ Brussel, UZ Leuven, UZAntwerpen, Ziekenhuis Oost-Limburg Genk, GZA St Augustinus, AZ St Jan Brugge, ZNAMiddelheim,StLucasGent,andAZGroeningeKortrijk),andtothelawyerandtheirpersonnelwhoactasatrustedthirdpartybetweenrespondentsandresearchers.Lastly,wewouldliketothankJaneRuthvenforherlanguageediting.

51

2.5 Referencelistchapter2

1. Rüegger C, Hegglin M, Adams M, Bucher HU. Population based trends in mortality,morbidity and treatment for very preterm- and very low birthweight infants over 12years.BMCPediatr.2012;12(1):17.

2. Devlieger R, Martens E, Martens G, Van Mol C, Cammu H. Perinatale activiteiten inVlaanderen2014.StudvoorPerinatEpidemiol.2014;28:56.

3. BarfieldWD,COMMITTEEONFETUSANDNEWBORN.StandardTerminology forFetal,Infant,andPerinatalDeaths.Pediatrics.2016;137(5):e20160551-e20160551.

4. CuttiniM, NadaiM, KaminskiM, et al. End-of-lifedecisions in neonatal intensive care:physicians’self-reportedpracticesinsevenEuropeancountries.EURONICStudyGroup.Lancet.2000;355:2112-2118.

5. VerhagenE,SauerPJJ.TheGroningenprotocol--euthanasiainseverelyillnewborns.NEnglJMed.2005;352(10):959-962.

6. BergerTM,HoferA.Causesandcircumstancesofneonataldeathsin108consecutivecasesover a 10-year period at the Children’sHospital of Lucerne, Switzerland.Neonatology.2009;95(2):157-163.

7. ProvoostV,DeliensL,CoolsF,etal.Aclassificationofend-of-lifedecisionsinneonatesandinfants.ActaPaediatr.2007;93(3):301-305.

8. Bijma HH, Schoonderwaldt EM, Van Der Heide A, Wildschut HIJ, Van Der Maas PJ,Wladimiroff JW. Ultrasound diagnosis of fetal anomalies: An analysis of perinatalmanagementof318consecutivepregnanciesinamultidisciplinarysetting.PrenatDiagn.2004;24(11):890-895.

9. BijmaHH,vanderHeideA,WildschutHI.Decision-MakingafterUltrasoundDiagnosisofFetalAbnormality.ReprodHealthMatters.2008;16(31SUPPL.):82-89.

10. ChervenakFA,McCulloughLB.Nonagressiveobstetricmanagement:anoptionforsomefetalanomaliesduringthethirdtrimester.Jama.1989;261(23):3439-3440.

11. PapiernikE,ZeitlinJ,DelmasD,etal.Terminationofpregnancyamongverypretermbirthsanditsimpactonverypretermmortality:ResultsfromtenEuropeanpopulation-basedcohortsintheMOSAICstudy.BJOGAnIntJObstetGynaecol.2008;115(3):361-368.

12. BosmaJM,vanderWalG,Hosman-BenjaminseSL.LateterminationofpregnancyinNorthHolland.BrJObstetGynaecol.1997;104(4):478-487.

13. Garne E, Khoshnood B, LoaneM, Boyd PA, DolkH. Termination of pregnancy for fetalanomalyafter23weeksofgestation:AEuropeanregister-basedstudy.BJOGAnIntJObstetGynaecol.2010;117(6):660-666.

14. ProvoostV,CoolsF,MortierF,etal.Medicalend-of-lifedecisionsinneonatesandinfantsinFlanders.Lancet.2005;365(9467):1315-1320.

15. VanDerHeideA,VanDerMaasPJ,VanDerWalG,etal.Medicalend-of-lifedecisionsmadeforneonatesandinfantsintheNetherlands.Lancet.1997;350(9467):1329-1331.

16. ten Cate K, van de Vathorst S, Onwuteaka-Philipsen BD, van der Heide A. End-of-lifedecisions for children under1 year of age in theNetherlands: decreased frequency ofadministrationofdrugstodeliberatelyhastendeath.JMedEthics.2015;41(10):795-798.

17. RoyR,AladangadyN,CosteloeK, LarcherV.Decisionmakingandmodesofdeath inatertiaryneonatalunit.ArchDisChild-FetalNeonatalEd.2004;89(6):F527-F530.

52

18. BartonL.TheContributionofWithholdingorWithdrawingCare toNewbornMortality.Pediatrics.2005;116(6):1487-1491.

19. Verhagen AAE, Janvier A, Leuthner SR, et al. Categorizing Neonatal Deaths: A Cross-Cultural Study in the United States, Canada, and The Netherlands. J Pediatr.2010;156(1):33-37.

20. Cuttini M. The European Union Collaborative Project on Ethical Decision Making inNeonatal Intensive Care (EURONIC): findings from 11 countries. J Clin Ethics.2001;12(3):290-296.

21. BoydPA,DeViganC,KhoshnoodB,etal.SurveyofprenatalscreeningpoliciesinEuropeforstructuralmalformationsandchromosomeanomalies,andtheirimpactondetectionandterminationratesforneuraltubedefectsandDown’ssyndrome.BJOGAnIntJObstetGynaecol.2008;115(6):689-696.

22. ChambaereK,BilsenJ,CohenJ,Onwuteaka-PhilipsenBD,MortierF,DeliensL.Trendsinmedical end-of-life decision making in Flanders, Belgium 1998-2001-2007.Med DecisMaking.2011;31(3):500-510.

23. PinterAB.End-of-lifedecisionbeforeandafterbirth:changingethicalconsiderations. JPediatrSurg.2008;43(3):430-436.

24. RennieJM,LeighB.Thelegalframeworkforend-of-lifedecisionsintheUK.SeminFetalNeonatalMed.2008;13(5):296-300.

25. ProvoostV,DeliensL,CoolsF,etal.Aclassificationofend-of-lifedecisionsinneonatesandinfants.ActaPaediatr.2004;93(3):301-305.

26. Pousset G, Bilsen J, Cohen J, Chambaere K, Deliens L, Mortier F. Medical End-of-LifeDecisionsinChildreninFlanders,Belgium.2017;164(6):547-553.

27. BolandRA,DavisPG,Dawson JA,DoyleLW.Outcomesof infantsbornat22–27weeks’gestation in Victoria according to outborn/inborn birth status. Arch Dis Child - FetalNeonatalEd.2017;102(2):F153-F161.

28. FischerN,SteurerMA,AdamsM,BergerTM.Survivalratesofextremelypreterminfants(gestationalage&lt;26weeks)inSwitzerland:impactoftheSwissguidelinesforthecareofinfantsbornatthelimitofviability.ArchDisChild-FetalNeonatalEd.2009;94(6):F407-F413.

29. ChambaereK,BilsenJ,CohenJ,etal.Apost-mortemsurveyonend-of-lifedecisionsusingarepresentativesampleofdeathcertificatesinFlanders,Belgium:researchprotocol.BMCPublicHealth.2008;8(1):299.

30. Hoddinott SN, Bass MJ. The dillman total design survey method. Can Fam Physician.1986;32(November):2366-2368.

31. ProvoostV,CoolsF,BilsenJ,etal.Theuseofdrugswithalife-shorteningeffectinend-of-lifecareinneonatesandinfants.IntensiveCareMed.2006;32(1):133-139.

32. Carpentier N, Van Brussel-De Vriendt L. On the Contingency of Death. A Discourse-TheoreticalPerspectiveontheConstructionofDeath.CritDiscourseStud.2012;2(9):99-116.

33. BilsenJ,CohenJ,ChambaereK,etal.Medicalend-of-lifepracticesundertheeuthanasialawinBelgium.NEnglJMed.2009;361(11):1119-1121.

53

54

End-of-lifedecisionsinneonatesandinfants:a

repeatedpopulation-levelmortalityfollow-backstudy

LaureDombrecht, KimBeernaert, Kenneth Chambaere, Filip Cools, Linde Goossens, GunnarNaulaers,LucCornette,SabineLaroche,ClaireTheyskens,ChristineVandeputte,HildeVandeBroek,JoachimCohen,LucDeliens

Submittedforpublication.

55

Abstract

Background:Neonatologyhasundergoneimportantclinicalandlegalchangesinthepasttwodecades and the implications for end-of-life decision-making in seriously ill neonates areunknown.Weaimedtoexaminechangesinneonatalend-of-lifedecisions(ELDs).

Methods:Weperformedamortalityfollow-backsurveyforafull-populationcohortofdecedentsunder the age of one year between August 1999 and July 2000 and September 2016 andDecember 2017 in Flanders, Belgium. For each death, physicianswere asked to complete ananonymousquestionnaireaboutELD-makingprecedingdeath.Questionnairedatawerelinkedtoclinicalandsociodemographicinformationfromthedeathcertificates.

Findings: The response rate was 87% in 1999-2000 (253/292) and 83% in 2016-2017(229/276). The proportion of deaths of infants born before 26 weeks of gestation wassignificantlyhigher(14%vs34%,p=0.001).PrevalenceofELDsremainedstableat60%,withnon-treatmentdecisions occurring in about 35%of all deaths andpotentially life-shorteningintensifiedadministrationofmedicationinabout15%.Theuseofmedicationwithanexplicitlife-shortening intention was prevalent in 7% to 10% of all deaths (p=0.15). However, in earlyneonataldeath(<7daysold)theadministrationofmedicationwithanexplicitlife-shorteningintentionwas lower (12%versus6%);while in lateneonataldeath (7-27daysold) thiswashigher(0%versus26%).

Interpretation:Overa17-yearperiodtheprevalenceofneonatalELDshasremainedrelativelystableataboutthreeinfiveofalldeceasedneonates.Inbothstudyperiodsanon-negligiblegroupof neonatal and infant death was preceded by a decision to intentionally hasten death byadministrating medication. These findings call for an open debate and ethical and juridicalreflectionbetweenhealthcareprofessionals,ethicistsandpolicymakers.

Funding:ThisstudyreceivedgrantsfromtheResearchFoundationFlanders(FWO;G041716N)andthespecialresearchfundofGhentUniversity(BOF;01J06915).

56

3.1 Introduction

Despiteariseinprenataldiagnosticproceduresandneonatalinterventions1,2,aboutonein100live-bornchildrenindevelopedcountriesdiesbeforetheageofoneyear3,4.Manyofthesedeathsareprecededbyapossiblylife-shorteningend-of-lifedecision(ELD)5–7.NeonatalELDsincludenon-treatment decisions such as withholding or withdrawing potentially life-prolongingtreatment, or administering pain and/or symptom relief medication, all with a potential orexplicit life-shortening effect7,8. This raises theneed for ethical andmoral discussions amonghealthcareprofessionals andparents regarding thepotential benefits ofaggressive treatmentversusreducingsufferingbynotunnecessarilyprolonginglife9.

Despiteattitudesurveysindicatingthathealthcareprofessionalsareinsomecasespreparedtomake neonatal ELDs7,10–12, studies on actual prevalence are rare.Most research is limited tosingle-centrestudies,showingthatbetween40%and93%ofdeathsfollowwithdrawaloflife-sustaining treatments6,13–15. The larger scale international EURONIC studies reported non-treatmentdecisionsaspartofregularpracticeattheendoflifeofaneonate,andadministrationofdrugswithexplicitlife-shorteningintention,asrareinmostEuropeancountries5,16,17.However,inthesestudies,thephysicianistheunitofstudymeaningreliableestimatesofprevalenceareimpossible;toprovidethem,population-basedstudiesarerequired7,18,19withthetotalpopulationofneonataldeathsasdenominator.Previouslytheonlypopulation-basedstudieswerefromtheNetherlandsandBelgium7,19.TheyshowthatamajorityofneonataldeathswereprecededbyanELD,themostcommonbeinganon-treatmentdecision7,19,andthatintentionallyhasteningdeathbymeansofmedicationdoesoccurinneonatalclinicalpractice5,7.

Asmedicalpracticecontinuedtoevolveandnewandimprovedtreatmentoptionshavebecomeavailable2, end-of-life decision-making in neonates might have changed, suggesting newinformation is required.Theonly trend figures available are fromdeath certificate studies in1995,2001,2005and201019intheNetherlands,indicatingshiftsintheprevalenceoftypesofELDssuchasanincreaseinthepercentageofnon-treatmentdecisionsandadecreaseindrugadministration with explicit life-shortening intention19. However, these findings might havelimitedexternalvalidityduetotheexistenceoftheGroningenprotocol,underwhichintentionallyhasteningdeathinneonateswithasevereconditionunderstrictguidelines20isnotprosecuted,creatingauniquecontextthatmightnotberelevanttoothercountries20.Therefore,anevaluationofchangestoandprevalenceofneonatalELDsindifferentjurisdictionsiswarranted.

Inthispaper,weaddressthefollowingresearchquestions:1)towhatextenthastheprevalenceofdifferentELDsinneonatesinFlanders(Belgium)changedovertime2)havetheunderlyingreasonsfortheELDchangedovertimeand3)havethesocio-demographicandclinicalprofilesofinfantswhosedeathwasprecededbytheseELDschangedovertime.

57

3.2 Methods

3.2.1 Design

Weconductedapopulation-levelmortalityfollow-backsurveybasedonacohortofall infantsundertheageofoneyearresidinginFlanderswhodiedbetweenAugust1999andJuly20007(wave1)andSeptember2016andDecember2017(wave2).Thedesignofbothstudieswasidentical.STROBEguidelinesforreportingcross-sectionalresearchwereusedi.

3.2.2 Settingandparticipants

AllinfantsagedunderonewhodiedwithintheinclusionperiodsinFlandersorBrusselswhosemotherwasaFlemishresident.FlandersandBrusselsare twoof the threesemi-autonomousregionsofBelgiumwithautonomyoverthequalityofhealthcare.Alldeathsoccurringinbothregionsareprocessedbythesamecentraladministrationauthority(theFlemishAgencyforCareandHealth).WeincludedonlydeathsofFlemishresidentstoprovideprevalenceratesinasetpopulation within one semi-autonomous region. The longer inclusion period in wave 2 wasspecificallychosentoascertainapopulationlargeenoughtoensuresignificantpowertoprovidereliabletrendanalysis,basedoninformationfromthecentraladministrationauthority21.

Allcaseswereidentifiedthroughthedeathcertificatessignedwithintheinclusionperiod.Deathsmustbedeclaredbymeansofadeathcertificate.Thephysicianfillsoutthemainpart,indicatingdemographicinformation(sex,dateofbirth,dateofdeath)andrelevantclinicalinformatione.g.causeofdeath8.Foreachdeath,withinfourmonthsofitsoccurrencetheattendingphysicianwasasked to complete a questionnaire. The study design, mailing and anonymity procedure aredescribedelsewhere8.Toensurereliabilityandavoidsociallydesirableanswers,arobustmethodwasimplementedusingatrustedthirdpartyasintermediarytoensureanonymity8.TheTotalDesignMethodwasfollowed,includingamaximumofthreefollow-uppostalmailings22.

iTheStrengthening theReportingofObservational studies inEpidemiology (STROBE)guidelinesweredeveloped to improvequality in reportingobservational studies. It isa checklist forauthors toensureadequatereporting(whatwasplanned,done,found,andconcluded)aswellasassessmentofthestrengthsandweaknessesofthestudy.30

58

3.2.3 Questionnaireandvariables

Thevalidatedquestionnairedeveloped in the1999-2000studywasusedasabasis toensurecomparability of data7. Both questionnaires first asked whether death had been sudden andunexpected;ifnot,anELDwasconsideredpossibleandphysicianswereaskedwhethertheyhad:

- withheldorwithdrawnlife-prolongingmedicaltreatmenttakingintoaccountorexplicitlyintendinghasteningdeath

- intensifiedadministrationofmedication,takingintoaccountorco-intendinghasteningthedeathor

- prescribed,suppliedoradministeredmedicationwiththeexplicitintentionofhasteningdeath.

AnELDwasthusdefinedasamedicaldecisionwiththepotentialorcertaineffectofhasteningdeath.WhenmorethanoneELDwasnoted,thatwiththemostexplicitlife-shorteningintentionwasdeemedmostimportant;ifmorethanoneELDwiththesamelife-shorteningintentionwasnoted,administrationofdrugs(‘active’)prevailedoverwithholdingorwithdrawingtreatment(‘passive’).Follow-upquestionswereasked,includingwhatwasthemostimportantreasonfortheELD.

Thesamedemographicandclinicalpatientdata(placeofdeath,sex,ageatdeath,gestationalageatbirthandcauseofdeath)wereobtainedfromthedeathcertificatesin2016-2017asin1999-2000.Weusedadeterministiclinkageproceduretolinkdeathcertificatewithquestionnairedata,and small cells analysis to ensure that linked death certificate data would preventreidentification.

Aclinicallyrelevantcategorizationforthecauseofdeathwasdevelopedtoachievehomogenousgroupswithasimilarcauseofdeathwithoutrevealingdetailedcase-specificinformation.Thiscategorization (see Table 3.1 for a description) was evaluated, in terms of completeness toclassifyallpossible causesofdeathand clarityofdescriptions, by fourphysiciansworking inneonatalandprenatalcare.Casesweresortedintooneofsevencategoriesbyaneonatologist(FC)andaresearcherwithexperienceinneonatalend-of-lifecareresearch(LDm)basedontheunderlyingcauseofdeath,denotedbyICD-10codes,onthedeathcertificate.Whenmaincauseofdeathwasinconclusive,ICD-10codesofotherassociatedcausesofdeathweretakenintoaccount.Categoriesaremutuallyexclusive.

59

Thefollowingcauseofdeathcategorieswereidentified:- Prematurity and related disorders: Death due to a direct cause of prematurity,

immaturity or disorders related to prematurity. For example, necrotizingenterocolitis,intraventricularhaemorrhage,respiratorydistresssyndrome,ordeathdueto(extremely)lowbirthweightorlowgestationalage.

- Congenital anomalies - singular: Death due to a single congenital anomalywith adefectinoneorganororgansystem.Forexample,acongenitalmalformationoftheheartoraspinabifida.

- Congenitalanomalies-multipleorsystemicdisorders:Deathduetothepresenceofmultiplecongenitalanomaliesindifferentorgansystems,orduetoadisorderthataffects multiple organ systems. For example, chromosomal disorders, multiplecongenitalmalformationsdiagnosedinoneinfant,oraninbornerrorofmetabolism.

- Complications of pregnancywith repercussions on foetal growth or development:Infantdiedduetocomplicationsofpregnancythathadaninfluenceonthegrowthorthe health of the baby prenatally. For example, a cytomegalovirus infection withcongenitalinfectionofthefoetus,orpre-eclampsiawithsevereintrauterinegrowthrestriction.

- Acute complicationsofpregnancy and/orbirth inapreviouslyhealthy foetus. Forexample,aplacentalabruptionorbirthtraumacausingoxygendeprivation.

- Disordersacquiredafterbirth:Deathduetoanon-congenitaldisorder,acquiredafterbirth of a previously healthy baby. For example, infectious diseases resulting inmultipleorganfailure.

- Other:Causeofdeathwassudden,withoutpreviousdiagnoses.Examplesaresuddeninfantdeathsyndrome,accidentsortrauma.

Table3.1:Causeofdeathcategoriesinneonatology.

3.2.4 Statisticalanalysis

Demographicvariables(placeofdeath,ageatdeath,sex,gestationalageatbirth,andcauseofdeath)ofallcaseswitharesponsefromthephysicianwerecomparedseparatelyforbothstudyperiodsbymeansofchi-squaretests,Fisher’sexacttestsorKruskalWallisteststoexaminenon-responsebias.

Chi-squaretestsandtwo-tailedFisher’sexacttestswereusedtocomparechangesovertimeintheprevalenceofdifferenttypesofELDsandthesocio-demographicandclinicalcharacteristicsassociatedwithdifferenttypesofELDs(sex,ageatdeath,gestationalageatbirth,andcauseofdeath).MultivariablebinarylogisticregressionwasperformedwithELD(yes/no)asdependentvariable,andstudyperiod,ageatdeath,gestationalageatbirthandcauseofdeathasindependentvariable to account for possible confounding of the demographical variables. Additionally, amultivariablebinarylogisticregressionmodelwiththesemaineffectsandtheinteractioneffectsofthesewiththestudyperiodwasperformedtoexamineshiftsinprevalenceofanELDincertaindemographical groups over both periods, controlling for confounding of study period, age atdeath,gestationalageatbirthandcauseofdeath.MultivariableanalysisfortheseparatetypesofELDswerenotmadeduetosmallsamplesizes.

60

3.2.5 Ethicalapproval

Ethical approval was obtained from the Ethics Committee of Ghent University (BelgianRegistration Number B670201628795), the Privacy Commission (CBPL, registration numberSA3/VT005071970), the National Council of the Order of Physicians (registration numberBD/wc/89997)andtheSectoralCommitteeofSocialSecurityandhealth(registrationnumberSCSZG/16/234).ThisstudywassupportedbyalleightFlemishNeonatalIntensiveCareUnits.

3.2.6 Roleofthefundingsource

ThisstudywasfundedbyagrantfromtheResearchFoundationFlanders(FWO)andthespecialresearchfundofGhentUniversity(BOF).K.BeernaertisaPostdoctoralFellowoftheResearchFoundationFlanders(FWO).Thefundingsourceshadnoroleintheconceptionanddesignofthestudy,nor in thedata-collection,analysisand interpretationof thedata,or thewritingof themanuscript.

3.3 Results

We received 229 completed questionnaires for276 deaths between September 1st 2016andDecember31st2017(83%responserate),and253questionnairesfor292deathsbetweenAugust1st 1999 and July 31st 2000 (87% response rate). No significant differences in demographiccharacteristicsbetweendeathswithandwithoutaresponsewasfoundforbothsurveywaves,thereforeweighingofresultswasnotnecessary.

The2016-2017and1999-2000cohortsweresimilarintermsofplaceofdeath(respectively92%vs89%inhospital),ageatdeath(55%vs50%inthefirstsevendaysoflife)andsex(Table3.2).Statistically significant differences between both cohorts were found for gestational age(proportionofinfantdecedentsbornbefore26weeksofgestationwashigherinwave2[34%]thanwave1[14%];p-value=0.001)andcauseofdeath(higherproportionofcomplicationsofpregnancywith [12 to17%]andwithoutrepercussions for the foetus [8 to15%], less ‘other’causesofdeath[16to7%]inwave2thanwave1;p-value=0.01).

61

1999-200012months

2016-201716months

Allinfantandneonataldeathsinstudyperiod

292 276

Allinfantandneonataldeathsforwhicharesponsewasreceived(responsepercentage)

253(87%) 229(83%)

N % N % P-valuePlaceofdeatha 0.34a Hospital 225 89 210 92 NICU N/A 115 50 Otherhospitalward N/A 95 41 Home 18 7 15 7 Other 10 4 4 2 Ageatdeathb 0.11 Earlyneonataldeath(<7days) 127 50 125 55 Lateneonataldeath(7-27days) 34 13 43 19 Postneonataldeath(>27days) 92 36 61 27 Sexc 0.46 Male 147 58 135 59 Female 106 42 94 41 Gestationalageatbirthb 0.001 <26weeks 36 14 72 34 26-28weeks 38 15 28 13 29-31weeks 19 8 10 5 32-36weeks 57 23 25 12 ³37weeks 101 40 76 36 Causeofdeathd,e 0.01 Prematurityandrelateddisorders 47 19 47 21

Congenitalanomaliessingular 39 16 38 17 Congenitalanomaliesmultipleor

systemicdisorders48 19 34 15

Complicationsofthepregnancywithrepercussionsonfoetalgrowthordevelopment

30 12 40 17

Acutecomplicationsofpregnancyand/orbirthinapreviouslyhealthyfoetus

20 8 34 15

Disordersacquiredafterbirth 26 10 19 8 Other 41 16 17 7 Percentagesarecolumnpercentagescalculatedwithallcasesforwhicharesponsewasreceivedasthedenominator·Missingvalueswerelimited:datasetof1999-2000:gestationalage,n=2(0.8%),causeofdeath,n=2(0.8%).Dataset2016-2017:gestationalage,n=18(7.9%).Percentageswerecalculatedwithoutthesemissingcases.aDifferentiationbetweenNICUandotherhospitalwardswasonlypossibleinthe2016-2017dataset.Chi-squareanalysiswereperformedwiththreecategories(hospital,home,other).bKruskalWallistestswereusedtocomparedifferencesforageatdeathandgestationalageatbirthbetweenbothtimeperiodscTwo-tailedFisher’sexacttestswereusedtocomparedifferencesinsexbetweentimeperiods.dPearsonChi-squaretestswereusedtocomparedifferencesincauseofdeathbetweentimeperiods.eSeeTable3.1fordescriptionofthecauseofdeathcategoriesN/A:notasked.Table3.2:Demographicandclinicalcharacteristicsofdeceasedneonatesandinfantsin1999-2000and2016-2017

62

NostatisticallysignificantdifferencesinprevalenceoftypesofELDswerefoundbetweenwavesforthefullpopulationofdeceasedinfantsbothinunivariableandmultivariableanalysis(Table3.3).AnELDwasmadein61%ofalldeathsinwave2and57%inwave1.ThemostcommonELDinbothstudyperiodswasanon-treatmentdecision(37%ofdeathsinwave2and34%inwave1). Administration of medication taking into account a possible life-shortening effect wasintensifiedin14%ofdeathsinwave2and16%inwave1.Medicationwithanexplicitintentiontoshortenlifewasadministeredin10%(24cases)ofalldeathsinwave2and7%(17cases)inwave 1. In wave 2, opioids were used in 20 cases, in 11 of those an additional sedative(barbiturates or benzodiazepines) and in four an additional muscle relaxant. In two cases asedativeonlywasgivenandinoneamusclerelaxantonly(notintable,typeofdruginfowasmissingforonecasein2016-2017).Inwave1opioidswereusedin14cases;infiveamusclerelaxantwasadministeredinassociation.Inthreecases,potassiumchloridewasused.

Table3.3:Prevalenceofend-of-lifedecisions(ELDs)inneonatologyinFlanders,Belgiumin2016-2017vs1999-2000

Significantandsubstantialchangesoccurredwithinsubpopulationsdependingonageatdeath(Table3.4).Inwave2ELDsweremadesignificantlylessoftenthaninwave1ininfantsundertheageofsevendays(p-value=0.01;55%vs72%).Inthosewhodiedbetweensevenand27daysandthoseover27days,ELDsweremadesignificantlymoreoftenin2016-2017thanin1999-2000(74%vs50%and64%vs38%respectively,p-values=0.03and0.003).Theincreasecanmostlybeseeninthosewhodiedafterwithdrawaloftreatment(26%vs9%between7-27daysand 31% vs 16% over 27 days old) and those who received intensified administration ofmedicationwithanexplicitlife-shorteningintention(26%vs0%between7-27daysand10%vs

1999-2000(12months)n=253b

2016-2017(16months)n=229b

p-valuea

N % N %NoELDpossible(deathentirelysuddenandunexpected)

59 23 46 20 0.23

ELDpossible,butnotmade(deathnon-sudden)

51 20 43 19 0.73

ELDmade 143 57 140 61 0.31

Non-treatmentdecision 86 34 85 37 0.51 Withholdingtreatment 32 13 27 12 0.78 Withdrawingtreatment

54 21 58 25 0.33

Useofdrugs 57 23 55 24 0.75 Medicationwithhasteningdeath

takenintoaccountorco-intended40 16 31 14 0.52

Medicationwithanexplicitintentiontohastendeath

17 7 24 10 0.15

WhenmorethanoneELDwasnotedbyphysicians,onlythemostimportantdecisionwasused.Themostimportantdecisionisthedecisionwiththemostexplicitlife-shorteningintention.WhenmorethanoneELDwiththesamelife-shorteningintentionwasnoted,administrationofdrugs(active)prevailedoverwithholdingorwithdrawingtreatment(passive).aFisher’sexacttest:independentvariable=studyperiod,dependentvariable=ELDtypepresentyes/no.bColumnpercentages:percentageofcasesinthatstudyperiodwiththattypeofELDcategory.

63

2%over27days).Thestatisticallysignificantdifferencesbetweenthetwostudyperiodsintheprevalence of ELDs depending on the age of the infantwere confirmedwhen controlling forpossibleconfoundinginthemultivariablebinarylogisticregression(datanotshown).

Univariate analysis revealed that, in infants born at full term (>37weeks of gestation), thedecisiontowithdrawtreatmentwasmademoreofteninwave2(34%)thaninwave1(20%,p-value=0.04)(table3.4).NootherdifferencesinsociodemographicorclinicalpatternsforthespecificELDswereobservedbetweencohorts.

Comparisonbetweenboth studywaves in reasons formakinganELDwas invaliddue to thepossibilityofindicatingmultiplereasonsinwave2andonlyoneinwave1.In60%ofallELDcases inwave2, ‘norealchanceofsurvival’was indicatedandin50% ‘nohopeofabearablefuture’(notintable).Wheretreatmentwaswithheldorwithdrawn,ormedicationwithoutanexplicitlife-shorteningintentionwasgiven,themainreasongivenwas‘norealchanceofsurvival’(62%,76%and62%respectively).Wheremedicationwasadministeredwithan explicit life-shorteningintention,themainreasonwas‘nohopeofabearablefuture’(91%).

AnyELDa Non-treatmentdecisiona Useofdrugsa Withholding Withdrawing Medicationwithapotentially

life-shorteningeffectMedicationwithexplicitintentiontohastendeath

1999-2000

2016-2017

p-valueb 1999-2000

2016-2017

p-valueb 1999-2000

2016-2017

p-valueb 1999-2000

2016-2017

p-valueb 1999-2000

2016-2017

P-valueb

SexMale 58% 64% 0.33 14% 13% 0.86 20% 26% 0.32 16% 13% 0.40 7% 12% 0.15Female 55% 57% 0.78 10% 10% 0.99 23% 24% 0.87 15% 15% 0.99 7% 9% 0.79AgeatdeathEarlyneonataldeath(<7days)

72% 55% 0.01 18% 18% 0.99 28% 22% 0.31 13% 10% 0.43 12% 6% 0.12

Lateneonataldeath(7-27days)

50% 74% 0.03 15% 2% 0.08 9% 26% 0.08 26% 21% 0.60 0% 26% N/A

Postneonataldeath(>27days)

38% 64% 0.003 4% 7% 0.71 16% 31% 0.05 15% 16% 0.99 2% 10% 0.06

Gestationalageatbirth<26weeks 61% 57% 0.84 25% 19% 0.62 11% 18% 0.41 17% 10% 0.35 8% 10% 0.9926-28weeks 74% 71% 0.99 16% 11% 0.72 29% 21% 0.58 18% 21% 0.77 11% 18% 0.4829-31weeks 68% 80% 0.67 16% 0% N/A 11% 30% 0.31 21% 30% 0.66 21% 20% 0.9932-36weeks 49% 56% 0.64 9% 8% 0.99 28% 32% 0.79 11% 4% 0.43 2% 12% 0.08³37weeks 51% 64% 0.07 9% 7% 0.78 20% 34% 0.04 17% 14% 0.84 5% 9% 0.37CauseofdeathPrematurityandrelateddisorders

64% 60% 0.83 15% 11% 0.76 23% 23% 0.99 15% 17% 0.99 11% 9% 0.99

Congenitalanomaliessingular

72% 74% 0.99 18% 16% 0.99 23% 34% 0.32 23% 16% 0.57 8% 8% 0.99

Congenitalanomaliesmultiple

75% 71% 0.80 25% 12% 0.17 27% 29% 0.99 15% 21% 0.56 8% 9% 0.99

Complicationsofthepregnancywithrepercussionsforthefoetus

67% 68% 0.99 13% 23% 0.37 17% 20% 0.77 30% 10% 0.06 7% 15% 0.45

Table3.4:ELDprevalenceindifferentpatientgroupsbysociodemographicandclinicalcharacteristicsovertime;1999-2000versus2016-2017

Acutecomplicationsofthepregnancyand/orbirthinahealthyfoetus

75% 56% 0.24 10% 6% 0.62 30% 32% 0.99 30% 9% 0.06 5% 9% 0.99

Disordersacquiredafterbirth

38% 63% 0.14 0% 5% N/A 27% 16% 0.48 8% 16% 0.64 4% 26% 0.07

Other 10% 12% 0.99 0% 0% N/A 7% 12% 0.62 0% 0% N/A 2% 0% N/ADatawasanalysedbymeansofindividualchi-squaretestsforeachdemographiccharacteristic(exampleallfemales)withstudyperiodasindependentvariableandtheprevalenceofthetypeofELD(anyELD,withholdingtreatment,withdrawingtreatment,medicationwithapotentiallylife-shorteningeffectandmedicationwithanexplicitlife-shorteningeffect)asdependentvariable.aRowpercentages.PercentageofinfantswiththatsociodemographicorclinicalcharacteristicthatreceivedthattypeofELDwithineachstudyperiod.Example:percentageofmaleinfantsin1999-2000thatdiedwithoutanELD.bP-valuesrepresentthesignificanceofdifferenceoftheChi-squaretest.Whensignificant,thepercentageofcaseswiththatclinicalorsociodemographiccharacteristic(ex.Male)issignificantlydifferentinthatcategoryofELD(includingnoELD)in1999-2000comparedto2016-2017.Missingvaluesingestationalage:2casesin1999-2000and18casesin2016-2017.Missingvaluesincauseofdeath:2casesin1999-2000.N/A:notapplicable,oneofthecellsinthecomparisonwasequaltozero.

66

3.4 Discussion

Thispopulation-levelmortalityfollow-backsurveyindicatesthat,whencomparingacohortofallinfantdecedentsundertheageofoneyearbetweenAugust1999andJuly2000andSeptember2016 and December 2017, relatively modest changes have occurred in end-of-life decision-makingpractices.Despitechangesintheclinicalprofileofthedecedents(e.g.theproportionofneonataldeathsofextremelyprematureinfantsincreased),deathsprecededbyanELDremainedat about 60%, with non-treatment decisions being about 35%. In both study waves a non-negligible group of deaths was preceded by a decision to intentionally hasten death byadministrating medication (7% in 1999-2000, 10% in 2016-2017). Prevalence of ELDs hassubstantiallydecreasedinearlyneonataldeath(<7days),andsubstantiallyincreasedafterthefirstsevendaysoflifenowcomparedto17yearsago.

3.4.1 Strengthsandlimitations

Despitethesensitivityofthetopic,weachievedhighresponserates(83%and87%)byusingarobust design with a rigorous follow-up procedure, making conclusions valid for the entirepopulationofdeceasedinfantsundertheageofoneyearirrespectiveofcaresettingordiagnosis.Thequestionnairewasdevelopedbasedonexistingandpreviouslyvalidatedquestionnairesonend-of-lifedecisionsinneonates7,19,minors23andadults24,25,ensuringcomparabilityovertime,settings, countries andage groups. Sociallydesirable answersorunwillingness toparticipatewere reduced by ensuring anonymity. Comparison of the response and non-response groupsrevealed no significant differences, indicating that results are generalizable to the entirepopulationofFlemishdeceasedneonates.

Inamortalityfollow-backstudyrecallandmemorybiascannotbeexcludedsincequestionnaireswere filled out up to fourmonths after death. However, a death certificate is the populationregister with the shortest processing delay, making it the best method to study ELDs on apopulationlevel.Althoughotherimportantactorsinthedecision-makingprocesssuchasparentsor nurses can provide useful information, we deemed the physician perspective as mostimportanttoreportonthemedicaldecisionsmade.Aquestionnairewithclosedmultiple-choiceanswersislesssuitableforin-depthstudyofthedecision-makingprocess,asitfailstoreflectthedepthandreasoningbehindadecision;however,itisthemostreliablemethodofstudyingtheprevalenceofELDswithoutdemandingtoomuchofrespondingphysicians.

3.4.2 Generaldiscussion

An interesting finding inour study is thenon-negligibleproportionof infantswhodiedafteradministrationofmedicationwithanexplicitlife-shorteningintent,namely7%ofalldeaths17yearsagoand10%now.Thisresultcontrastssharplywiththedecreaseintheuseofmedicationwithexplicitlife-shorteningintentionwhichwasseenintheNetherlandsfrom8%ofallneonataldeathsin2005to1%in201019.IntheNetherlands,intentionallyhasteningdeathinextremelyillneonates is not prosecuted under strict guidelines in the Groningenprotocol20. Evaluation ofwhetherallduecarecriteriawereappliedinaspecificcasehappensretrospectively,afterwhichadecision ismadewhetherornotaprosecution iswarranted20. InBelgiumsuchaprotective

67

frameworkislackingandintentionallyhasteningdeathbymeansofmedicationisthusnotlegallypermissible. Possibly, the Groningen protocol, by setting up specific and detailed rules andprocedures, has discouraged physicians in the Netherlands from engaging in practices tointentionallyhastendeath.Alternatively,itmighthaveledtoadifferentunderstandingofwhatconstitutesdeath-hasteningmedical interventionswithanexplicit life-shortening intention inextremecasesandadministeringmedicationtakingintoaccountapossiblelife-shorteningeffectin other cases. It is also possible that Flemish and Dutch physicians make differentrationalizationswhenreportingonidenticalquestions.

Increaseddosesofsedativesandopioidswerereportedinthemajorityofcaseswheremedicationwithanexplicitlife-shorteningintentionwasadministered.Wecanpresumethatthismedicationwasadministeredtorelievethesufferingoftheneonatesandinfantsforwhomtherewasnohopeofabearablefutureorthosewhowouldnotsurvivewithoutlifesustaininginterventions,evenwhendeathwasherebyhastened.Thisclinicalpracticefitswithinapalliativecarecontextanddecisions are probably made in the best interest of the child. It should be noted that theprevalenceofadministeringmedicationwithanexplicitlife-shorteningintentioninFlandersisconsiderablyhigherinneonatesandinfants(10%)ascomparedtominors(8%)23andadults(6%,includingeuthanasia)26,raisingthequestionofwhetherthispracticeneedstobemonitoredandevaluatedmorecloselyinsuchavulnerablepatientgroup.

In 2016-2017 the prevalence of administering medication with an explicit life-shorteningintentionandthewithdrawaloflife-sustainingtreatmentwashigherafterthefirstsevendaysoflifethanin1999-2000.Availabilityofimprovedmedicaltreatments2mighthaveledtoamoreactive initial therapeutic approach in severely ill neonates, who previouslywould have diedshortlyafterbirth.Itseemsthatthesetreatmentsprovetobesuccessful,asbirthratesinFlandersincreasedby4%inwave23,27,yetneonatalandinfantmortalityinbothperiodsremainedstabledespitealongerinclusionperiodin2016-2017.However,asnotallinfantscontinuetobenefitfromthisactiveapproachandriskhavingalifewithpoorqualityifintensivecareiscontinued,itmighthaveresultedinan increase indecisions towithdraw life-sustainingtreatmentorevenadministermedicationwithanexplicitlife-shorteningintentafterthefirstweekoflife.Thisiscorroborated by our data indicating an increase in the use of medication with explicit life-shorteningintentionduetopoorexpectedqualityof lifein2016-2017,ratherthanduetothechildnothavinganyrealchanceofsurvival.

ThedecreaseoftheuseofELDsinthefirstweekoflifecanpossiblyberelatedtothenoticeablechange inthepopulationofneonataldeathswitha largerproportionofdecedentsnowbeingextremelypremature(<26weeksofgestation).Incontrastwith17yearsago,intensivecarecannowbeofferedtoinfantsbornat24weeks’gestationinFlanders28.Mortalityintheseextremelyprematureinfantsisstillhigh,however29,andtheyoftendieduringthefirstweekoflifedespiteactivetreatment,withoutanELD.

Despite changes in the prevalence of ELDs in specific age groups, the overall prevalence of ELDs stays relatively stable at about 60% of all neonatal deaths in Flanders, which is similar to that in the Netherlands (63%)19, the only other country with reliable population-based prevalence rates. Similarly, the proportion of ELDs is higher in neonates than in children aged one to 17 years (36%)23 and adults (48%)26. This is not surprising since deaths occurring in adults and minors are more often sudden and

68

unexpected, such as accidents or trauma, making ELDs impossible. The most prevalent neonatal ELDs are non-treatment decisions (35%), specifically withdrawing of life-prolonging treatment, which occurs in about one in four cases, a prevalence estimate comparable with non-population-based studies across Europe6,13–15. In a similar percentage of deaths in both study periods, intensive life-saving treatment was thus started only to be foregone at a later stage.

3.4.3 Conclusion

Over a 17-year period, the proportion of infant deaths in Flanders preceded by an ELD hasremainedrelativelystableataboutthreeinfive,confirmingthatnon-treatmentdecisionsaswellasintentionallyhasteningdeathbymeansofmedicationcontinuetobeanintegralpartofmedicalpractice in severely ill neonates. The difference between a non-negligible prevalence ofadministeringmedicationwithanexplicitlife-shorteningintentioninFlanders-aregionwherethispracticeiscurrentlynotregulatedbymeansofaprotocoloralaw-andalowprevalenceofthesamepracticeintheNetherlands-whichprovidesguidelinesandregulationsforbestpracticewiththeGroningenprotocol–isremarkable.Itmayprovideinputforasocietaldebateabouttheneedforrevisedguidelines,protocolsorlawstoshapeend-of-lifepracticeinneonatalandinfantcareandabouttheneedforfurtherresearchandevaluationtomonitorandunderstandthesedecisions.

3.5 Acknowledgements

WewouldliketothankallphysiciansandNICUwardsthatparticipatedinthisstudy,aswellasthephysiciansandexpertswhoaidedintestingandvalidatingthequestionnaire.WearegratefulforthesupportandcooperationoftheFlemishAgencyforCareandHealthwithoutwhomdata-collectionwouldnothavebeenpossible,andfortheaidoflawyerWimDeBrockandProf.Dr.RobertVanderSticheleinensuringanonymityofallparticipants.WewouldliketothankRoosColmanandEllenDeschepperfortheirstatisticalexpertise,andJaneRuthvenforherlanguageediting.

69

3.6 Referencelistchapter3

1. BijmaHH,vanderHeideA,WildschutHI.Decision-MakingafterUltrasoundDiagnosisofFetalAbnormality.ReprodHealthMatters.2008;16(31SUPPL.):82-89.

2. Rüegger C, Hegglin M, Adams M, Bucher HU. Population based trends in mortality,morbidity and treatment for very preterm- and very low birthweight infants over 12years.BMCPediatr.2012;12(1):17.

3. DevliegerR,GoemaesR,LaubachM.PerinataleActiviteiteninVlaanderen2018.[inDutch].StudvoorPerinatEpidemiol.2019.

4. DraperES,GallimoreID,KurinczukJJ,etal.MBRRACE-UKPerinatalMortalitySurveillanceReport,UKPerinatalDeathsforBirthsfromJanuarytoDecember2016.LeicesterInfantMortalMorbStudDepHealSciUnivLeicester.2018.

5. CuttiniM, NadaiM, KaminskiM, et al. End-of-lifedecisions in neonatal intensive care:physicians’self-reportedpracticesinsevenEuropeancountries.EURONICStudyGroup.Lancet.2000;355:2112-2118.

6. BergerTM,HoferA.Causesandcircumstancesofneonataldeathsin108consecutivecasesover a10-yearperiod at theChildren’sHospital of Lucerne, Switzerland.Neonatology.2009;95(2):157-163.

7. ProvoostV,CoolsF,MortierF,etal.Medicalend-of-lifedecisionsinneonatesandinfantsinFlanders.Lancet.2005;365(9467):1315-1320.

8. Dombrecht L, Beernaert K, Roets E, et al. Apost-mortem population survey on foetal-infantileend-of-lifedecisions :aresearchprotocol.BMCPediatr.2018:1-9.

9. FeltmanDM,DuH,LeuthnerSR.Surveyofneonatologistsattitudestowardlimitinglife-sustaining treatments intheneonatal intensivecareunit. JPerinatol.2012;32(11):886-892.

10. Dombrecht L, Deliens L, Chambaere K, et al. Neonatologists and neonatal nurses havepositive attitudes towards perinatal end-of-life decisions, a nationwide survey. ActaPaediatr.2019:apa.14797.

11. BucherHU,KleinSD,HendriksMJ,etal.Decision-makingatthelimitofviability:Differingperceptions and opinions between neonatal physicians and nurses. BMC Pediatr.2018;18(1):81.

12. RebagliatoM,CuttiniM,KaminskiM,Persson J,ReidM,SaracciR.NeonatalEnd-of-LifeDecisionMaking:Physicians’attitudesandrelationshipwithself-reportedpracticesin10Europeancountries.Jama.2000;284(19):2451-2459.

13. RoyR,AladangadyN,CosteloeK, LarcherV.Decisionmakingandmodesofdeath inatertiaryneonatalunit.ArchDisChild-FetalNeonatalEd.2004;89(6):F527-F530.

14. BartonL.TheContributionofWithholdingorWithdrawingCare toNewbornMortality.Pediatrics.2005;116(6):1487-1491.

15. Verhagen AAE, Janvier A, Leuthner SR, et al. Categorizing Neonatal Deaths: A Cross-Cultural Study in the United States, Canada, and The Netherlands. J Pediatr.2010;156(1):33-37.

16. BoizeP,BorrhomeeS,MichelP,BetremieuxP,HubertP,MorietteG.Neonatalend-of-lifedecision-making almost 20 years after the EURONIC study : A French survey. ArchPédiatrie.2019;26(6):330-336.

70

17. SchneiderK,MetzeB,BührerC,CuttiniM,GartenL.End-of-LifeDecisions20YearsafterEURONIC:Neonatologists’ Self-ReportedPractices,Attitudes,andTreatmentChoices inGermany,Switzerland,andAustria.2019.

18. VanDerHeideA,VanDerMaasPJ,VanDerWalG,etal.Medicalend-of-lifedecisionsmadeforneonatesandinfantsintheNetherlands.Lancet.1997;350(9467):1329-1331.

19. ten Cate K, van de Vathorst S, Onwuteaka-Philipsen BD, van der Heide A. End-of-lifedecisions for children under1 year of age in theNetherlands: decreased frequency ofadministrationofdrugstodeliberatelyhastendeath.JMedEthics.2015;41(10):795-798.

20. VerhagenE,SauerPJJ.TheGroningenprotocol--euthanasiainseverelyillnewborns.NEnglJMed.2005;352(10):959-962.

21. Teamgegevensverwerkingenresultaatsopvolging.DeathRatesonBabies(Foeto-InfantileMortality) [in Dutch]. Brussels: Flemish Agency for Care and Health, Department ofInformationandSupport.;2014.

22. Hoddinott SN, Bass MJ. The dillman total design survey method. Can Fam Physician.1986;32(November):2366-2368.

23. Pousset G, Bilsen J, Cohen J, Chambaere K, Deliens L, Mortier F. Medical End-of-LifeDecisionsinChildreninFlanders,Belgium:APopulation-BasedPostmortemSurvey.ArchPediatrAdolescMed.2010;164(6):547-553.

24. ChambaereK,BilsenJ,CohenJ,etal.Apost-mortemsurveyonend-of-lifedecisionsusingarepresentativesampleofdeathcertificatesinFlanders,Belgium:researchprotocol.BMCPublicHealth.2008;8(1):299.

25. RobijnL,CohenJ,RietjensJ,DeliensL.TrendsinContinuousDeepSedationuntilDeathbetween2007and2013 :ARepeatedNationwideSurvey.PLoSOne.2016;June:1-11.

26. ChambaereK,VanderSticheleR,MortierF,CohenJ,DeliensL.RecentTrendsinEuthanasiaand Other End-of-Life Practices in Belgium. New England Journal of Medicine.2015;372(12):1179-1181.

27. Cammu H, Martens G, Coen K De, Mol C Van, Defoort P. Perinatale Activiteiten inVlaanderen2005.[inDutch].StudvoorPerinatEpidemiol.2006.

28. JacquemynY(ConsensusgroepVNenM. [RecommendationsPerinatalCareConcerningViabilityinFlanders](inDutch).TijdschrGeneeskd.2014;70:20:1159-1167.

29. GlassHC, CostarinoAT, Stayer SA, BrettC, Cladis F, Davis PJ. Outcomes for ExtremelyPrematureInfants.AnesthAnalg.2015;120(6):1337-1351.

30. CuschieriS.TheSTROBEguidelines.Saudijournalofanaesthesia.2019;13(Suppl1):S31.

71

Part3Attitudes,viewsandexperiencesofhealthcareprovidersonneonatalend-of-lifedecision-making

Chapter4:Neonatologistsandneonatalnurseshavepositiveattitudestowardsperinatalend-of-lifedecisions,anationwidesurvey

Chapter5:Barrierstoandfacilitatorsofend-of-lifedecision-makingbyneonatologistsandneonatalnursesinneonates:aqualitativestudy

Chapter6:Psychologicalsupportinend-of-lifedecision-makinginneonatalintensivecareunits:fullpopulationsurveyamongneonatologistsandneonatalnurses

72

Neonatologistsandneonatal

nurseshavepositiveattitudestowardsperinatalend-of-life

decisions,anationwidesurvey

LaureDombrecht,LucDeliens,KennethChambaere,SaskiaBaes,FilipCools,LindeGoossens,GunnarNaulaers,EllenRoets,VeerlePiette,JoachimCohen,KimBeernaertonbehalfoftheNICUconsortium

PublishedinActaPaediatrica2019,March28th.

73

Abstract

Aim: Perinatal death is often preceded by an end-of-life decision (ELD). Disparate hospitalpolicies,complexlegalframeworksandethicallydifficultcasesmakeattitudesimportant.ThisstudyinvestigatedattitudesofneonatologistsandnursestowardsperinatalELDs.

Methods:AsurveywashandedouttoallneonatologistsandneonatalnursesinalleightneonatalintensivecareunitsinFlanders,BelgiuminMay2017.RespondentsindicatedagreementwithstatementsregardingperinatalELDsonaLikert-scaleandsentbackquestionnairesviamail.

Results:Theresponseratewas49.5%(302/610).Mostneonatologistsandnursesfoundnon-treatment decisions such as withholding or withdrawing treatment acceptable (90-100%).Terminationofpregnancywhenthefoetusisviableincasesofsevereorlethalfoetalproblemswasconsideredhighlyacceptableinbothgroups(80-98%).Physiciansandnursesdonotfinddifferent ELDs equally acceptable, e.g. nursesmore often than physicians (74% versus 60%,p=0.017)agreethatit isacceptableincertaincasestoadministermedicationwiththeexplicitintentionofhasteningdeath.

Conclusion: There was considerable support for both prenatal and neonatal ELDs, even fordecisions that currently fall outside the Belgian legal framework. Differences betweenneonatologists’andnurses’attitudesindicatethatbothopinionsshouldbeheardduringELD-making.

74

4.1 Background

Despiteincreasedpossibilitiestodetectandtreatcongenitalanomalies1,perinataldeathsstillrangefrom1.1to4.8per1,000birthsacrossEuropeancountries2.Manyofthesedeathsoccureitheratmaternitywardsorneonatalintensivecareunits(NICUs)andareoftenprecededbyanend-of-life decision (ELD) 3–5, such as withholding or withdrawing life-sustaining treatment,possibly life-shorteningalleviationofpainand,or,othersymptomsordeliberatelyendinglifewithalethaldoseofdrugs6orthirdtrimesterorlateterminationofpregnancy(TOP)7,8.TheseELDscanhastendeath,inthisstudythisincludesboththepassivedecisiontonotprolonglifeandthe activedecision to (possibly) shorten life.The ethicaldilemma in someof these situationsbetweensaving the infant’s lifeandnotknowingwhat theburdenofsufferingwillbe9needsthoughtfulandprofessionaldeliberationofallpartiesinvolvedinthedecision-makingprocess.ThesedecisionsarefurthercomplicatedbydisparateNICUpolicies,evenwithincountries10,andcomplex legal frameworks, making the attitudes towards prenatal and neonatal ELDs of theprofessionalsinvolvedintegraltotheprocess.

AlthoughtheneonatologistactuallycarriesresponsibilityfortheELD,nursesarealsoinvolvedinend-of-lifediscussionsandtheprovisionofcareforthechildandthefamily11.Physiciansandnursesarekeyfigureswhohaveaninfluencebothoneachotherandontheparentsduringanend-of-lifedecision-makingprocess.Previousresearchhasshownthat,eveninnew-bornswiththe same pathology, there is variance between types of ELDs taken 12,13. As well as thecharacteristicsoftheNICUstaff12,theirattitudesmayplayacrucialroleinend-of-lifedecision-making3,14–16.Evenwithinacareteamworkingcloselytogether,importantdifferencesbetweenphysiciansandnursesinattitudestowardsELDshavebeenfound17.

PreviousstudiesonattitudesonELDsforseverelyillneonatesinNICUs14,15,18–20arelimitedinseveralways.SomepopulationstudiesaboutattitudestowardsneonatalELDsdatebackalmosttwo decades making it impossible to assess attitudes under current medical practice andlegislation14,15,otherswerelimitedtosinglecentrestudies15,20.Inotherstudiesonattitudesinperinatalcare,thescopeofthestudywaslacking.Firstly,somestudiesonlyincludedattitudesonappropriatetreatmentornon-treatmentforinfantsbornatthelimitofviability(16,21,22)whichfails to cover ELDs when a life-limiting foetal condition is diagnosed or when extremely illneonatesarebornattermlimitingthescopetoaveryspecificgroupofinfants.Asecondgroupofstudies focusses on only prenatal or neonatal ELDs separately. Since attitudes and decisionsbeforeorafterbirthcouldpossiblyinfluenceeachother,andneonatologistsareoftenconsultedinprenatalELDs21,wefeellikeattitudesonbothprenatalandneonatalELDsshouldbeincludedinto one study. This iswhy our study focusses on attitudes on all perinatal ELDs instead offocussingoneitherprenatalorneonatalELDsseparately.

Therefore, this study addressed following researchquestions:whatare theattitudes towardsprenatalandneonatalend-of-lifedecision-makingofneonatologistsandneonatalnurses?Whatarethedifferencesbetweenneonatologistsandneonatalnursesintheirattitudes?Andwhatistheinfluenceofsex,age,professionandattitudesofneonatologistsandneonatalnursesonthedecisionstheywouldconsideraspossibleoptionsinahypotheticalneonatalcase?

75

4.2 Methods

4.2.1 Designandparticipants

Weperformedafull-populationmailsurveyofallneonatologistsandneonatalnursesinalleightNICUsinFlanders,Belgium;83physiciansand527nurseswereidentifiedintotalbymeansofpersonnelfilesateachNICU.

4.2.2 Datacollection

ArepresentativeworkingateachNICUhandedoutthequestionnairetoeveryneonatologistandneonatalnurseintheirrespectiveNICUon1May2017andinvitedthemtofillitoutandsenditbackbymeansofaprepaidenvelopetotheresearchersbefore31May2017.

4.2.3 Questionnaire

ThequestionnairewasbasedonanexistingFlemishattitudequestionnairefromtheyear2000onneonatalELDs14,addingquestionsaboutprenatalELDsanddescribingahypotheticalandmedically complex case. A multidisciplinary team consisting of three sociologists, twopsychologists,threeneonatologists,onegynaecologistdevelopedthefinalquestionnairewhichwascognitivelytestedonfiveneonatologistsfromfourseparatehospitals,threeneonatalnursesfromtwoseparatehospitalsandonegynaecologisttoensurecontentvalidityoftheitems.

4.2.4 Measures

Thequestionnaireconsistedofsevensocio-demographicquestionsand12 itemsonperinatalELDs.SixoftheseattitudeitemsfocussedonneonatalELDsandsixitemsfocussedonprenatalELDs(lateTOP).Attitudesweremeasuredby indicatingwhetherornot theyagreedwith thestatements,scoredonafive-pointLikertscale.TheseELDstatementscanbeclassifiedbasedontwodimensions.The firstdimension isamedico-technical classificationof themedicalactaseitheranon-treatmentdecision,theadministrationofdrugsortheimplementationofmedicalinterventions 22.Theseconddimensionisamedico-ethicalclassificationof the life-shorteningintention.Wealsopresentedahypothetical caseof a foetusbornat27weeks gestationwithadditional complications;participantswere givensevenpossible treatmentoptionsandwereasked to indicate whether they would consider each option on a four-point Likert scale.Furthermore, it is important tonote the legal context ofELDs inFlanders,Belgium,which isrepresentedinTable4.1.

76

InBelgium,terminationofpregnancyafter12weeksofgestationispossiblewhen:

- completingthepregnancypresentsaseriousthreattothewoman’shealth

- thechildwillsufferfromaparticularlysevereailment,acknowledgedtobeincurable

atthetimeofdiagnosis

DeliberatelyendingthelifeofaneonateisnotlegallypossibleinBelgium.Theonlycountry

thatcurrentlylegallycondonesactivelyendingthelifeofaneonateunderstrictconditionsis

theNetherlands,inthefollowingthreedistinctcases23:

- physiologicfutilityoftreatmentinnewbornswithnochanceofsurvival

- infants who may survive after a period of intensive treatment, but their actual or

foreseensufferinginthenearfutureissevereandunbearable

- infants with an extremely poor prognosis who do not depend on technology for

physiologicalstabilitybutwhosesufferingissevereandcannotbealleviated

Table4.1:Legalframework

4.2.5 Statisticalanalyses

The data on attitude items were analysed with separate Kruskal-Wallis tests with group(neonatologistsandneonatalnurses)as independentvariableandthesixELDorsix lateTOPattitudeitemsasdependentvariablesusingSPSS24.0(SPSSInc.,Chicago,Illinois).Apost-hocFriedmantestwasperformedtoexaminethedifferencesinacceptanceofthedifferenttypesofneonatalELDs.ThisFriedmantestwasperformedinneonatologistsandnursesseparatelyandwasadjustedformultipletestingbymeansofaBonferronicorrection.TheLikertscaleitemswererescaled from a five-point to a three-point scale, indicating disagreement, neutrality andagreement.Next,aPrincipalComponentAnalysis(PCA)wasperformedonallattitudeitemstorevealtheunderlyingstructureofattitudes.

Forthehypotheticalcase,weranKruskal-Wallistestswithgroupasindependentvariableandthe six treatment options as dependent variables. All answers were dichotomised into notconsideringtheoptionwhichincludesnotagoodoptionandlessgoodoptionandconsideringtheoptionwhichincludesgoodoptionandverygoodoption.Aseparatemultivariableordinallogisticregression(PoLytomousUniversalModels)wasfittedforeachtreatmentoptionwitha4-pointLikertscaleinordertoestimatetheirassociationwithsociodemographiccharacteristicsand standardised scores on the PCA attitude components. Nonsignificant variables wereeliminatedfromthefinalmodelbymeansofabackwardsstepwiseapproach,significancelevelswere set at 0.05. Since age and years of experience working in an NICU setting are highlycorrelated,weoptednot to includeboth into the samemodel and tested twoalternative fullmodelswiththesevariables.Whenbothstepwiseeliminationsforeachstatementdidnotresultinthesameresults,weoptedforthemodelwitheitherageoryearsofexperience,dependingonwhichprovidedthebestfit.Whenthesameresultwasobtained,thatwasconsideredasthemodel

77

with the best fit. Odds ratios (OR) and 95% confidence intervals (CI) were provided. Aprofessionalstatisticianwasconsulted.

4.2.6 Ethicalconsiderations

Ethical approval was obtained from the Ethics Commission of Ghent University Hospital(Registrationnumber:B670201731709). If a filled-outquestionnairewas sentback, thiswasseenasgivingconsenttoparticipatinginthisstudy.

4.3 Results

Across all eight NICUs, the response rate was 63% (52/83) for neonatologists and 46%(250/527)fornurses.AnoverviewofalldemographiccharacteristicscanbefoundinTable4.2.

NeonatologistsN=52(%)

NeonatalnursesN=250(%)

Pvaluea

Sex Female Male

37(71.2)15(28.8)

237(95.2)12(4.8)

<0.001

Age <30 30-39 40-49 ≥50

12(23.1)15(28.8)11(21.2)14(26.9)

75(30.2)65(26.2)53(21.4)55(22.2)

0.73

YearsofexperienceworkinginaNICU <5years 5-10years 11-20years >20years

22(42.3)8(15.4)9(17.3)13(25)

58(23.3)34(13.7)77(30.9)80(32.1)

0.02

Functionofphysicians Neonatologist Specialistintraining

39(75)13(25)

N/A N/A

Degreenursesb Graduate Bachelor Master Extraspecialisationc

N/AN/AN/A

3(1.2)229(92.3)16(6.5)94(37.9)

N/A

Missingvalues:variedfrom0.4%forsexandyearsofexperienceto0.8%forageanddegreeinnurses.Therewerenomissingvaluesinneonatologists.aPearsonchi-squarebCategoriesarenotmutuallyexclusivecOverviewofthespecificspecialisations:87.2%advancedbachelorneonatologyandpaediatrics,2.1%advancedbacheloremergencyandintensivecare,4.3%professionaltitleinneonatologyandpaediatrics,3.2%midwifery,3.2%postgraduate

Table4.2:Demographiccharacteristicsofthestudyparticipants.

78

4.3.1 AttitudestowardsneonatalELDs

Overall, acceptability of all types of neonatal ELDs in certain cases of neonates with severeconditionsishighinbothneonatologistsandnurses(60-100%)(Table4.3).Allneonatologistsand90.4%ofnursesagreedthatnotinitiatingtreatmentforaneonate,takingintoaccountthepossibility that this could hasten the end of life, is acceptable (p=0.023). Acceptance of theadministeringofmedicationtakingintoaccountthatitcouldhastentheendoflifewashigherinneonatologists (96.2%) than neonatal nurses (83.6%; p=0.024). Acceptance of administeringmedicationwith theexplicit intention tohasten theendof lifewashigher inneonatalnurses(73.6%)thaninneonatologists(59.6%;p=0.017).

Fewerneonatologistsagreetoactivelyadministeringmedicationwiththeexplicit intentionofhastening theendof life than theyagree towithholding(p=0.013)orwithdrawing treatment(p=0.013)takingintoaccountthepossibilitythatitcouldhastentheendoflife(Table4.4).Thiswasalsofoundinneonatalnurses.Furthermore,neonatologistsagreemorewithadministeringmedication taking into account the possibility that it could hasten the end of life than withadministeringmedicationwiththeexplicitintentionofhasteningtheendoflife(p=0.042),whileno significant difference between the two options was found for nurses. Other differencesbetweentheattitudestowardsneonatalELDsofneonatologistsandnurseswerenotsignificant.

Item Group Disagree(%)

Neutral(%)

Agree(%)

Pvalue(KruskalWallis)

Attitudestowardsneonatalend-of-lifedecisions

Incertaincasesofnewbornswithsevereconditionsitisacceptable:nottoinitiatetreatment,takingintoaccountthepossibilitythatthiscouldhastentheendoflife

Neonatologist

0.0

0.0

100

0.023

Neonatalnurse 4.0 5.2 90.4

nottoinitiatetreatmentwiththeexplicitintentionofhasteningtheendoflife Neonatologist 0.0 0.0 100 0.081Neonatalnurse 1.6 4.0 94.4

towithdrawtreatment,takingintoaccountthepossibilitythatthiscouldhastentheendoflife Neonatologist 5.8 5.8 88.5 0.154Neonatalnurse 7.6 12.8 79.6

towithdrawtreatmentwiththeexplicitintentionofhasteningtheendoflife Neonatologist 7.7 5.8 86.5 0.992Neonatalnurse 3.6 10.4 86.0

toadministermedication,takingintoaccountthepossibilitythatthiscouldhastentheendoflife Neonatologist 3.8 0.0 96.2 0.024Neonatalnurse 4.0 12.4 83.6

toadministermedicationwiththeexplicitintentionofhasteningtheendoflife Neonatologist 21.2 19.2 59.6 0.017Neonatalnurse 7.2 19.2 73.6

Attitudestowardsprenatalend-of-lifedecisionsandlateterminationofpregnancy

Terminationofpregnancyinthecaseofaviablefoetusshouldbecompletelyprohibited Neonatologist 100.0 0.0 0.0 0.002 Neonatalnurse 84.4 10.8 4.8

Terminationofpregnancyinthecaseofaviablefoetusattherequestofthemotherisacceptable Neonatologist 44.2 26.9 28.8 0.476 Neonatalnurse 33.6 39.6 26.0

Ifthemotherishealthy,terminationofpregnancyataviablestageisacceptableinthecaseofalethalfoetalmedicalproblem

Neonatologist 0.0 1.9 98.1 0.201Neonatalnurse 1.2 5.2 93.6

Ifthemotherishealthy,terminationofpregnancyataviablestageisacceptableinthecaseofaseverefoetalproblem

Neonatologist 1.9 5.8 92.3 0.041Neonatalnurse 4.8 14.8 80.4

Ifthefoetusishealthy,terminationofpregnancyataviablestageisacceptablewhenthelifeofthemotherisindanger

Neonatologist 19.2 17.3 63.5 0.688Neonatalnurse 13.6 21.6 64.4

Ifthefoetusishealthy,terminationofpregnancyataviablestageisacceptablewhenthemotherhasaseverepsychologicalproblem

Neonatologist 61.5 23.1 15.4 0.474Neonatalnurse 54.8 30.0 15.2

AllattitudeitemsweretranslatedbyalanguageeditorTable4.3:Attitudesofneonatologistsandneonatalnursestowardsprenatalandneonatalend-of-lifedecision-making.

Notinitiatingtreatmenttakinglife-shorteningintoaccount

Notinitiatingtreatmentwithexplicitlife-shorteningintention

Withholdingtreatmenttakinglife-shorteningintoaccount

Withholdingtreatmentwithexplicitlife-shorteningintention

Administeringmedicationtakinglife-shorteningintoaccount

Administeringmedicationwithexplicitlife-shorteningintention

Neonatologistschi-square(pvalue)

Nurseschi-square(pvalue)

Neonatologistschi-square(pvalue)

Nurseschi-square(pvalue)

Neonatologistschi-square(pvalue)

Nurseschi-square(pvalue)

Neonatologistschi-square(pvalue)

Nurseschi-square(pvalue)

Neonatologistschi-square(pvalue)

Nurseschi-square(pvalue)

Neonatologists(chi-square)

Nurses(chi-square)

Notinitiatingtreatmenttakinglife-shorteningintoaccount

-0.346(1) -0.353(0.526) 0(1) -0.122(1) -0.385(1) -0.131(1) -0.125(1) -0.217(1) -1.221(0.013) -0.542

(0.018)

Notinitiatingtreatmentwithexplicitlife-shorteningintention

0.346(1) 0.353(0.526) -0.346(1) -0.476

(0.068) -0.038(1) -0.223(1) -0.221(1) -0.137(1) -0.875(0.256) -0.189(1)

Withholdingtreatmenttakinglife-shorteningintoaccount

0(1) 0.122(1) 0.346(1) 0.476(0.068) -0.385(1) -0.253(1) -0.125(1) -0.339

(0.645) -1.221(0.013) -0.665(0.001)

Withholdingtreatmentwithexplicitlife-shorteningintention

0.385(1) 0.131(1) 0.038(1) 0.223(1) 0.385(1) 0.253(1) -0.260(1) -0.086(1) -0.837(0.339) -0.412(0.211)

Administeringmedicationtakinglife-shorteningintoaccount

0.125(1) 0.217(1) 0.221(1) 0.137(1) 0.125(1) 0.339(0.645) 0.260(1) 0.086(1) -1.096(0.042) -0.325

(0.785)

Administeringmedicationwithexplicitlife-shorteningintention

1.221(0.013) 0.542(0.018) 0.875(0.256) 0.189(1) 1.221(0.013) 0.665

(0.001) 0.837(0.339) 0.412(0.211) 1.096(0.042) 0.325

(0.785)

Allsignificantresults(pvalue<0.05)areindicatedinbold.Table4.4:FriedmantestneonatalELDs.

81

AttitudestowardslateTerminationofPregnancy

AllneonatologistsdisagreedwiththestatementthatTOPwhenthefoetuswasviableshouldbeprohibited,thiswasmorethan84.4%ofneonatalnurses(p=0.002).AlmostallphysiciansandnursesagreedontheacceptanceoflateTOPincasesofalethalfoetalmedicalproblem(98.1vs93.6%respectively).Neonatologists foundTOPwhentheviable foetushasasevereproblemmoreoftenacceptable(92.3%)thannursesdid(80.4%;p=0.041).Whentheviablefoetuswashealthybutthemothersufferedfromseverepsychologicalproblems62%ofneonatologistsand55%ofnursesdisagreedwithterminationofpregnancy.

4.3.2 ELDattitudecomponents

ThePCAresultedinfourcomponents.Itemswithaloadingonacomponenthigherthan0.5wereretainedinthatcomponent.AfirstcomponentincludedfavourabilitytowardneonatalELDswithexplicit intention of hastening the end of life. The second component indicated favourabilitytowardsneonatalELDswherethepossibilitythattheend-of-lifecouldbehastenedistakenintoaccount.ThethirdincludedfavourabilitytowardsTOPataviablestageandthelastcomponentincludedfavourabilitytowardslateTOPforreasonsconcerningthemother(Table4.5).

Item FavourabletowardsneonatalELDswithexplicitintentiontohastentheendoflife

FavourabletowardsneonatalELDstakingintoaccountthepossibilitythatitcouldhastentheendoflife

Favourabletowardslateterminationofpregnancywhenthefoetusisviable

Favourabletowardslateterminationofpregnancyformaternalreasons

Incertaincasesofneonateswithsevereconditionsitisacceptablenottoinitiatetreatment,takingintoaccountthepossibilitythatthiscouldhastentheendoflife

- 0.741 - -

Incertaincasesofneonateswithsevereconditionsitisacceptabletowithdrawtreatmentwiththeexplicitintentionofhasteningtheendoflife 0.767 - - -

Incertaincasesofneonateswithsevereconditionsitisacceptablenottoinitiatetreatmentwiththeexplicitintentionofhasteningtheendoflife - 0.699 - -

Incertaincasesofneonateswithsevereconditionsitisacceptabletowithdrawtreatmentwiththeexplicitintentionofhasteningtheendoflife 0.814 - - -

Incertaincasesofneonateswithsevereconditionsitisacceptabletoadministermedication,takingintoaccountthepossibilitythatthiscouldhastentheendoflife

- 0.604 - -

Incertaincasesofneonateswithsevereconditionsitisacceptabletoadministermedicationwiththeexplicitintentionofhasteningtheendoflife 0.790 - - -

Terminationofpregnancyinthecaseofaviablestageattherequestofthemotherisacceptablea - - 0.451 -

Terminationofpregnancyinthecaseofaviablestageshouldbecompletelyprohibited

- - 0.732 -

Ifthefoetusishealthy,terminationofpregnancyataviablestageisacceptablewhenthelifeofthemotherisindanger - - - 0.806

Ifthefoetusishealthy,terminationofpregnancyataviablestageisacceptableincaseofawherethemotherhasaseverepsychologicalproblem - - - 0.783

Ifthemotherishealthy,terminationofpregnancyataviablestageisacceptableinthecaseofalethalfoetalmedicalproblem - - 0.724 -

Ifthemotherishealthy,terminationofpregnancyataviablestageisacceptableinthecaseofaseverefoetalproblem - - 0.642 -

Cronbach’salpha 0.818 0.734 0.596 0.544Standardisedscoreswerecalculatedbyattributingaweightequaltothecomponentloadingtoeachsalientvariable.Ahigherstandardisedscoreindicatesmoreagreementwiththeitemsincludedinthecomponent.Theonlyexceptiontothisruleistheitem‘Terminationofpregnancyinthecaseofaviablefoetusshouldbecompletelyprohibited’,whichwasrescaledindicatingthatahigherscoreforthisitemcorrespondswithlessagreement.*thisitemloadedequallyhighontwocomponentswhichiswhywemadetheexecutivedecisiontoplaceitwithinthecomponentthatbestmatchedthecontentofthatitem.Also,thisitemwasrescaledtomatchtheotheritemsinthePCA,indicatingthatahigherscoreforthisitemcorrespondswithlessagreementtothisitem.

Table4.5:Principalcomponentanalyses:componentloadings.

83

4.3.3 ELDtreatmentoptionsinahypotheticalneonatalcase

In a neonatal case of a premature neonate with complications leading to severe long-termmorbidity(Table4.6),81%oftheneonatologistsand87%ofthenursesdidnotfindstartingorcontinuinglife-prolongingtreatmentagoodtoverygoodtreatmentoptionforthempersonally(p=0.34).Wefoundsignificantdifferencesbetweenwhatneonatologistsandnursesconsideredasgoodoptionsforthefollowingpossibletreatmentoptions:notinitiatingtreatmentbothwithexplicit intention (75% and 50.4% respectively) and taking into account the possibility ofhastening the end of life of the neonate (88.5% and 63.3% respectively) and withdrawingtreatmentbothwithexplicitintention(67.3%and52.2%respectively)andtakingintoaccountthe possibility of hastening the end of life (82.7% and 66.8% respectively). No significantdifferenceswerefoundforadministeringmedication.Administeringmedicationwiththeexplicitintentionof hastening the endof life of the foetuswas indicated asa goodoptionby29%ofneonatologists and 39% of nurses (p=0.16). While no large differences can be seen in thepercentage of physicians and nurses who found non-treatment decisions (withholding andwithdrawingtreatment)acceptablewith(86%and100%)andwithout(80%and100%),explicitlife-shortening intention,wedoseea lowerpercentageofbothphysiciansandnurses findingnon-treatmentdecisionswith explicit life-shortening intention tobe a goodtreatmentoption(50%and75%)thannon-treatmentdecisionswithoutanexplicitlife-shorteningintention(63%and89%).

Hypotheticalneonatalcaseofprematurelyborninfantwithadditionalcomplications

Asidefromwhatparents(andphysiciansa)want,whichofthefollowingoptions

arepossibleoptionsforyoupersonallyregardingthiscase?Group

Notagood

orlessgood

option(in

%)

Goodor

verygood

option(in

%)

Pvalue

(Kruskal

Wallis)b

Startingorcontinuinglife-prolongingtreatmentforthechild Neonatologist 86.5 13.5 0.338

Neonatalnurse 80.9 19.1

Notinitiatingtreatmenttakingintoaccountthepossibilitythatthiscouldhasten

theendoflifeofthepatient

Neonatologist 11.5 88.5 <0.001

Neonatalnurse 36.7 63.3

Notinitiatingtreatmentwiththeexplicitintentionofhasteningtheendoflifeof

thepatient

Neonatologist 25.0 75.0 0.001

Neonatalnurse 49.6 50.4

Withdrawingtreatmenttakingintoaccountthepossibilitythatthiscouldhasten

theendoflifeofthepatient

Neonatologist 17.3 82.7 0.024

Neonatalnurse 33.2 66.8

Withdrawingtreatmentwiththeexplicitintentionofhasteningtheendoflifeof

thepatient

Neonatologist 32.7 67.3 0.047

Neonatalnurse 47.8 52.2

Administeringmedicationtakingintoaccountthepossibilitythatthiscould

hastentheendoflife

Neonatologist 31.4 68.6 0.064

Neonatalnurse 45.5 54.5

Administeringmedicationwiththeexplicitintentionofhasteningtheendoflife

ofthepatient

Neonatologist 71.2 28.8 0.159

Neonatalnurse 60.7 39.3

Casedescription:Lizaisatwin,bornat27weekswithextremeintra-uterinegrowthretardation.Herbirthweightwasonly500g.Thefirstfewdaysofherlifeareremarkablyuneventful:she

breathesindependentlywiththehelpofnon-invasiverespiratorysupportandenteralnutritionisintroducedcarefully.Theultrasoundscanofherbrainiscompletelynormal.Whensheiseightdays

old,however,shehasagastricperforationleadingtoseveresepticshockwithmultipleorganfailure.Thesituationstabilisesafterafewdaysandherorgansstartfunctioningagain.Sheappearsto

haveenteredarecoveryphase,whichiscomplicated,however,byseveredehiscenceoftheabdominalwound,exposingtheintestines.Thiswillcertainlyneedasurgicalintervention(ifnotseveral),

butthisisnotyetpossibleatthisstage.Inaddition,thebrainultrasoundperformedafewdayslater,showsrapidlyprogressingmulticysticleukomalaciasuggestiveofwidespreadwhitematter

damage.Duringamultidisciplinarydiscussion,thedoctorsagreethatthiswillcertainlyleadtoseverespasticquadriparesis.Atthispoint,sheisthreeweeksold,breathingautonomouslywithnon-

invasiverespiratorysupportandhaemodynamicallystable,butobviouslyfedparenterally.

aThiswasaddedforquestionnairesofneonatalnursesonly.

bDifferencebetweenneonatologistsandneonatalnurses.

Table4.6:Treatmentoptionsneonatologistsandneonatalnursesconsidertobegoodoptionsinahypotheticalneonatalcase.

85

4.3.4 Relationshipofattitudes,demographicsandhypotheticaltreatmentoptions

Thoseunder30yearsoldandthosebetween30and39,morethanthoseover50,indicatethatcontinuinglife-prolongingtreatmentisanacceptabletreatmentoptioninthehypotheticalcase(OR3.45,95%CI1.82-6.54andOR1.91,95%CI1.01-3.61,respectively).Inthiscase,menmorethanwomenagreethatwithdrawingtreatmenttakingintoaccountthatthiscouldhastendeathisanacceptabletreatmentoption(OR5.72,95%CI1.32-24.83)(Table4.7).

Byusing thePCAattitude components,we found that generalattitudeswere associatedwithwhichtreatmentoptionswereconsideredasgoodtoverygoodinaconcretehypotheticalcase(Table4.7).AhigherscoreonthefirstPCAcomponent,‘favourabilitytoneonatalELDswiththeexplicit intention of hastening the end of life’, indicates that respondents find ELDswith anexplicitlife-shorteningintentionmoreacceptablethanothers.Thosewithahighscoreonthisfirstcomponentarelesslikelytoconsiderthetreatmentoptionoflife-prolongingtreatmentthanothers(OR0.75,95%CI0.66-0.85).Thisgroupisalsomoreinclinedtoconsidernotinitiatingtreatment(OR1.50,95%CI1.32-1.70),withdrawingtreatment(OR1.60,95%CI1.41-1.82)andadministeringmedicationwiththeexplicitintentionofhasteningtheendoflife(OR1.59,95%CI1.39-1.82) as possible treatment options. A higher score on the third PCA component,‘favourabilityofterminationofpregnancywhenthefoetusisviable’,indicatesmoreacceptanceoflateTOPthanthosewithalowscore.Thegroupwhoscoreshighonthisthirdcomponentismorelikelytoconsideradministeringmedicationtakingintoaccountthepossibilitythatitcouldhasten the end of life as a possible treatment option (OR 1.22, 95% CI 1.01-1.47). All otherdemographiccharacteristicsdidnothaveasignificantrelationtowhichtreatmentoptionswereconsideredacceptableinthehypotheticalcase.

Continuinglife-

prolonging

treatmentc

Notinitiating

treatment,no

explicitintentiond

Notinitiating

treatment,explicit

intention

Withdrawing

treatment,no

explicitintentiond

Withdrawing

treatment,explicit

intention

Administering

medication,no

explicitintentiond

Administering

medication,explicit

intention

Predictor OR 95%CI OR 95%CI OR 95%CI OR 95%CI OR 95%CI OR 95%CI OR 95%CI

Function

Neonatologist(vsnurse,ref

category)

a a 3.65 (1.48-9.00) 2.86 (1.59-5.16) a a a a a a 0.56 (0.32-0.98)

Sex

Male(vsfemale,refcat) a a a a a a 5.72 (1.32-

24.83)

a a a a a a

Age(refcat50+yrs)

<30 3.45 (1.82-6.54) a a a a a a a a a a a a

30-39 1.91 (1.01-3.61) a a a a a a a a a a a a

40-49 1.65 (0.84-3.25) a a a a a a a a a a a a

AcceptanceofneonatalELDs

withexplicitintentione0.75 (0.66-0.85) a a 1.50 (1.32-1.70) a a 1.60 (1.41-

1.82)

a a 1.59 (1.39-1.82)

AcceptanceofneonatalELDs,

noexplicitintentione

a a 1.37 (1.11-1.70) a a 1.44 (1.17-1.78) a a 1.47 (1.18-

1.84)

a a

AcceptanceofTOPe a a a a a a a a a a 1.22 (1.01-

1.47)

a a

Modelfittinginformation,

pseudoR-squaref0.13

0.11

0.18

0.10

0.19

0.11

0.20

OR=oddsratio.

PresentedfiguresareORsand95%Cis.Independentvariablesthathavenosignificantrelationshipsarenotpresentedinthetable.ExperienceinaNICUandattitudestowardsacceptabilityoflateterminationof

pregnancyformaternalreasonswereenteredintheregressionbutwerenotsignificantforanyofthestatementsandwere,therefore,eliminatedfromthetable.Separateordinalregressionmodelswereperformed

foreachdependentvariable.ThefulldescriptionofthestatementsispresentedinTable5,afulldescriptionofthePCAattitudefactorscanbefoundinTable4.aEnteredintheregressionbutnotsignificantandconsequentlyeliminatedfromthemodel.bReferencecategory.cThresholdfrom‘lessgoodoption’to‘goodoption’isnotsignificantinthismodel.

dDuetoaviolationoftheparallellinesassumptioninmultivariateordinallogisticregression(theregressionlineswerenotparallelforeachlevelofthedependent)wecombined‘notagoodoption’and‘lessgood

option’,and‘goodoption’and‘verygoodoption’intoabinarylogisticregression.eComponentsasaresultofthePCA,seetable6.fNagelkerke.

Table4.7:FactorspredictingacceptanceofpossibletreatmentoptionsinthehypotheticalneonatalELDcase.

87

4.4 Discussion

Inthisfull-populationsurveystudywedistributedattitudequestionnairesconcerningperinatalend-of-lifedecisionsamongstallneonatologistsandneonatalnursesworkinginFlemishNICUs.Themajorityofbothgroupsaccept(acceptancerateofover60%)bothprenatalandneonatalend-of-lifedecisions.However,somedifferencescanbenoted,suchasahigheracceptanceforactively ending the life of a neonate bymeans of medicationwith an explicit life-shorteningintention (activeELDs)bynurses compared tophysicians.Moreover,we found thatattitudestowardslateTOPandneonatalend-of-lifedecisionshaveasignificantimpactonthetreatmentoptionstheywouldconsiderinahypotheticalneonatalcase.

Actively administering medication with explicit life-shortening intention was consideredacceptable by more than half of neonatologists and three quarters of nurses and was evenconsideredasagoodtreatmentoptioninthehypotheticalcaseinathirdofneonatologistsandtwofifthsofnurses.ThisindicatesahighacceptanceofanELDthatcurrentlyfallsoutsidethelegalframeworkinBelgiumandmostothercountries.ApossiblehypothesisisthatNICUstaffmightprefernot toprolongunnecessaryneonatalsufferingbyadministering a lethal doseofmedicationevenwhenthismighthavelegalcomplications14.Thiscorroboratespreviousstudiesreporting the occurrence of hasteningdeath in neonates takingplace both in Belgium 14 andacross Europe 3 even though the only countrywhere actively ending the life of a neonate iscurrently legislatively condoned under strict conditions is the Netherlands 24. However,physiciansandnursesinourstudyweresignificantlymoreacceptabletowardsnon-treatmentdecisionswithapotentiallylife-shorteningeffectthantheyaretowardsactivelyendingthelifeofaneonatewithmedication.Inouropinion,neonatologistsandnursesinourstudywouldpreferanon-treatmentdecision,withorwithoutextracomfortcare,whenpossible.However,insomecases,theintentiontoreducesufferingbyshorteningthelifeofachildwithaseverelylife-limitingdiagnosis cannot always be achieved solely by a non-treatment decision. In these cases, asindicatedinourresults,threeoutoffiveneonatologistsandthreequartersofnursesagreethat,insomecases,shorteningthelifewithalethaldrugisacceptable.ThepositiveattitudetowardstheseactiveELDsofasubstantialproportionofpeoplecaringforextremelyillneonatesandtheiroccurrenceacrossEuropecanbeseenassupport,orcanbe thebasis foranethicaland legaldiscussionofinstallingalegislationsimilartothatoftheNetherlands.

Wefoundthatahigherproportionofnursesthanphysiciansaccepttheuseofmedicationwithanexplicitintentiontohastendeath.ApossibleexplanationcanbethatphysiciansadoptamorecautiousapproachtowardsELDsfallingoutsidethelegislationbecausetheyareultimatelystilltheoneswhoarelegallyresponsibleformedicaldecisionsattheendofaneonate’slife.Similarresultswerefoundinresearchinadultend-of-lifedecision-makingthatshowsthatphysiciansare less in favourof euthanasia(i.e. intentionallyending a life byaphysicianat thepatient’sexplicit request) than nurses 25. Another possible explanation could be thatnurses aremoreexposed to the suffering of the infant and the resultingdiscomfort in parents, since they arepresent at the bedside of neonate for more extended periods of time compared to theneonatologist.TheycouldthereforebepressedtolimitthissufferingasmuchaspossiblewhilephysiciansmightprefertoattemptadditionaltreatmentorlessinvasiveELDsfirst.

88

Non-treatmentdecisionsregardlessoflife-shorteningintentionwereconsideredacceptableby80%ormoreofallNICUphysiciansandnurses.Also,lateTOPincaseofasevereorlethalfoetalanomaly when the mother is healthy was considered acceptable by over 80% of bothneonatologistsandneonatalnurses.ThishighacceptanceoflateTOPinthesecasescouldpartlybeascribedtothepossibilityoflateabortionincasesofseverefoetalmalformationsinBelgianlegislation,seeTable4.126–29.Thispositiveattitudecouldthusbelessfrequentincountrieswithamore limitedlateTOP lawsuchasMalta,where lateTOP isprohibited,or ItalyandFinlandwhereitisonlylegalbefore28weeksofgestation30.Additionally,incountrieswherelateTOPispossiblewithoutgestationalagelimitwhentheseverecongenitaldisordersarelethalorwhenthedisorderswould lead to severe and incurable impairment such as theNetherlands 30,wemightexpecttoseesimilarattitudes.

Whilenolargedifferencescouldbefoundbetweentheproportionofphysiciansandnurseswhofound non-treatment decisions acceptable either with or without explicit life-shorteningintention,wecouldhoweverseedifferencesinthehypotheticalcase;theproportionofphysiciansandnurseswhofoundnon-treatmentdecisionstobeagoodoptioninthehypotheticalcasedoesseemtobelowerwhenthelife-shorteningintentionisexplicitratherthanimplicit.Thiscouldbeduetothespecificnatureofthehypotheticalcase,butitisimportanttonotethatthesegeneralattitudesmaynotalwaysbereflectedintheactualmedicaldecision-makingprocess.Eveninahypotheticalcase,eventhoughgeneralattitudeshadaninfluenceonwhichtreatmentoptionswereconsideredasgoodoptions,bothphysiciansandnursesaremorecautiouswhentheexplicitshorteningofthelifeoftheneonateisintended.

OurresultsshowadifferenceininfluenceofthetwoneonatalELDattitudecomponents,namelyfavourability towards ELDs with and without explicit life-shortening intention, on whichtreatmentoptionsneonatologistsandnurseswouldconsiderinahypotheticalneonatalcase.Wecan thusdistinguish adifference in influenceof attitudes towardsELDs thathave an explicitversusapossiblelife-shorteningintention.Thisreflectsboththeimportanceofthedimensionofintention in the classification of neonatal ELDs 6 and the importance of attitudes of bothphysicians and nurses in clinical practice because of their impact on the possible treatmentoptions theywouldconsiderinactualend-of-lifepractice.Similarlytoresults fromstudiesonend-of-lifedecision-making in adults 31, it is possible that thewillingness to consideror evenperformneonatalELDsmightnotonlybeamatterofwhetherornotthereisalegalpossibilitybutcouldalsopartlydependontheattitudesoftheattendingNICUstaff.Additionally,weseeanassociation between having a high acceptance for late TOP because of a foetal condition onconsidering administration ofmedicationwithout explicit life-shortening intention as a goodpossible treatment option in the neonatal case. This supports our suggestion of consideringprenatalandneonatalELDsasonegroupunderperinatalELDs,sinceattitudesofphysiciansandnursesonprenatalELDsarerelatedtowhichtreatmentoptionsareconsideredtobeacceptableinneonates.

We also found that NICU staff under 39 years of age were more inclined to consider life-prolongingtreatmentasapossibletreatmentoptioninthecasethanthoseovertheageof50.Furthermore,menaremorelikelythanwomentoconsiderwithdrawingtreatmenttakingintoaccount the possibility that this could hasten the end of life. We did not find any otherdemographicvariablesassociatedwithdecision-makinginthehypotheticalcase.Inadditionto

89

previous research indicating the importance of social, cultural or religious attitudes of NICUphysicians12,wefoundthatgenderandagecouldalsobeimportantcharacteristicsrelatedtoend-of-lifedecision-making.

4.4.1 Strengthsandlimitations

Whereasmoststudiesinneonatalend-of-lifecareandlateTOParelimitedwithregardtosamplesize 15,20, by targeting the entire Flemish population of neonatologists and neonatal nursesworking in an NICU we received a response from half of the entire population (63% forneonatologists and 46% for nurses). Our study adds to the existing attitude literature byprovidingattitudescoveringallfoetal-infantileELDs.Thisstudyalsohaslimitations.Wedonothavedemographic information aboutphysicians andnurseswhodidnotparticipate, or theirreasonsfornotdoingso.Demographicvariablessuchasreligiousbeliefs,whichmighthaveaninfluenceonattitudes,werenotincludedinourquestionnaire.Nodefinitionoftheconceptofhastening death was provided in the questionnaire, and despite thorough cognitive testingshowinginterpretationof‘hasteningdeath’wascleartorespondents,responsescouldbesubjecttointerpretationfromtherespondentsasbeingeitherpassivelydecidingtonotprolonglifeoractivelydecidingtoshortenlife.Duetoethicalconsiderations,wewerebothunabletocomparethe attitudes of all eight Flemish NICUs with each other and unable to link the attitudes ofneonatologistsandneonatalnurseswithactualmedicaldecisionsmadeinclinicalpractice.Wecould only use the hypothetical neonatal case to examinewhich treatment optionswould beconsideredinarealisticsituationwithoutactuallymeasuringbehaviouralintentions,whichcouldleadtogeneralisationofresults.

4.4.2 Conclusion

Our study found a large acceptance of both prenatal and neonatal end-of-life decisions inneonatologistsandneonatalnurses,evenfordecisionsthatcurrentlyfalloutsidetheBelgianlegalframework. However, physicians and nurses differed slightly in their acceptance of differenttypesofend-of-lifedecisions,bothatanabstractlevelandinahypotheticalneonatalcase.Ourfindings regarding the impact of attitudes in considering actual medical decision-makingindicatedtheimportanceofinvolvingbothphysiciansandnursesinclinicalpractice.

4.5 Acknowledgements

We would like to thank all neonatologists and nurses who participated in this study, theirparticipatinghospitals(UZGent,UZBrussel,UZLeuven,UZAntwerpen,ZiekenhuisOost-LimburgGenk,GZAStAugustinus,AZStJanBruggeandZNAMiddelheim)andallphysiciansandnurseswho aided in testing and validating the questionnaire. We would also like to thank ourcollaboratorsintheNICUconsortium:SabineLaroche,ClaireTheyskens,ChristineVandeputte,Luc Cornette and Hilde Van de Broek. Lastly, we would like to thank Roos Colman for herstatisticalexpertiseandJaneRuthvenforherlanguageediting.

90

4.6 Referencelistchapter4

1. Rüegger C, Hegglin M, Adams M, Bucher HU. Population based trends in mortality,morbidity and treatment for very preterm- and very low birthweight infants over 12years.BMCPediatr.2012;12(1):17.

2. de JongeA,BaronR,WesternengM,Twisk J,HuttonEK.Perinatalmortalityrate in theNetherlands compared to other European countries: A secondary analysis of Euro-PERISTATdata.Midwifery.2013;29(8):1011-1018.

3. CuttiniM, NadaiM, KaminskiM, et al. End-of-lifedecisions in neonatal intensive care:physicians’self-reportedpracticesinsevenEuropeancountries.EURONICStudyGroup.Lancet.2000;355:2112-2118.

4. VerhagenE,SauerPJJ.TheGroningenprotocol--euthanasiainseverelyillnewborns.NEnglJMed.2005;352(10):959-962.

5. BergerTM,HoferA.Causesandcircumstancesofneonataldeathsin108consecutivecasesover a 10-year period at the Children’sHospital of Lucerne, Switzerland.Neonatology.2009;95(2):157-163.

6. ProvoostV,DeliensL,CoolsF,etal.Aclassificationofend-of-lifedecisionsinneonatesandinfants.ActaPaediatr.2007;93(3):301-305.

7. Bijma HH, Schoonderwaldt EM, Van Der Heide A, Wildschut HIJ, Van Der Maas PJ,Wladimiroff JW. Ultrasound diagnosis of fetal anomalies: An analysis of perinatalmanagementof318consecutivepregnanciesinamultidisciplinarysetting.PrenatDiagn.2004;24(11):890-895.

8. BijmaHH,vanderHeideA,WildschutHI.Decision-MakingafterUltrasoundDiagnosisofFetalAbnormality.ReprodHealthMatters.2008;16(31SUPPL.):82-89.

9. GastmansC,NaulaersG,VanholeC,DenierY.FromBirthtoDeath?APersonalistApproachtoEnd-of-LifeCareofSeverelyIllNewborns.JChristBioeth.2013;19(1):7-24.

10. Sauer PJJ, Dorscheidt JHHM, Verhagen AAE, Hubben JH. Medical practice and legalbackgroundofdecisionsforseverelyillnewborninfants:viewpointsfromsevenEuropeancountries.ActaPaediatr.2013;102(2):e57-63.

11. BelliniS,DamatoEG.Nurses’knowledge,attitudes/beliefs,andcarepracticesconcerningdonotresuscitatestatusforhospitalizedneonates.JOGNN-JObstetGynecolNeonatalNurs.2009;38(2):195-205.

12. WilkinsonDJC,TruogRD.Theluckofthedraw:Physician-relatedvariabilityinend-of-lifedecision-makinginintensivecare.IntensiveCareMed.2013;39(6):1128-1132.

13. AzoulayÉ,MetnitzB,SprungCL,etal.End-of-lifepracticesin282intensivecareunits:DatafromtheSAPS3database.IntensiveCareMed.2009;35(4):623-630.doi:10.1007/s00134-008-1310-6

14. ProvoostV,CoolsF,MortierF,etal.Medicalend-of-lifedecisionsinneonatesandinfantsinFlanders.Lancet.2005;365(9467):1315-1320.

15. RebagliatoM,CuttiniM,KaminskiM,Persson J,ReidM,SaracciR.NeonatalEnd-of-LifeDecisionMaking:Physicians’attitudesandrelationshipwithself-reportedpracticesin10Europeancountries.Jama.2000;284(19):2451-2459.

16. BucherHU,KleinSD,HendriksMJ,etal.Decision-makingatthelimitofviability:Differingperceptions and opinions between neonatal physicians and nurses. BMC Pediatr.2018;18(1):81.

91

17. BurnsJP,MitchellC,GriffithJL,TruogRD.End-of-lifecareinthepediatricintensivecareunit:attitudesandpracticesofpediatriccriticalcarephysiciansandnurses.CritCareMed.2001;29(3):658-664.

18. Inghelbrecht E, Bilsen J, Pereth H, Ramet J, Deliens L. Medical End-of-Life Decisions:Experiences and Attitudes of Belgian Pediatric Intensive Care Nurses. Am J Crit Care.2009;18(2):160-168.

19. Miljeteig I, Markestad T, Norheim OF. Physicians’ use of guidelines and attitudes towithholding andwithdrawing treatment for extremelyprematureneonates inNorway.ActaPaediatrIntJPaediatr.2007;96(6):825-829.

20. Garten L, Daehmlow S, Reindl T, Wendt A, Münch A, Bührer C. End-of-life opioidadministration on neonatal and pediatric intensive care units: Nurses’ attitudes andpractice.EurJPain.2011;15(9):958-965.

21. Miquel-Verges F, Woods SL, Aucott SW, Boss RD, Sulpar LJ, Donohue PK. Prenatalconsultation with a neonatologist for congenital anomalies: parental perceptions.Pediatrics.124(4):e573-e579.

22. Dombrecht L, Beernaert K, Roets E, et al. Apost-mortem population survey on foetal-infantileend-of-lifedecisions :aresearchprotocol.BMCPediatr.2018:1-9.

23. MinisterievanVeiligheidenJustitieenMinisterievanVolksgezondheidWelzijnenSport.Regelingbeoordelingscommissie late zwangerschapsafbreking en levensbeëindigingbijpasgeborenen.Staatscourant.2016:3145.

24. Verhagen AAE. The Groningen Protocol for newborn euthanasia; which way did theslipperyslopetilt?JMedEthics.2013;39(5):293-295.

25. Gielen J, Van Den Branden S, Broeckaert B. Attitudes of European Physicians TowardEuthanasia and Physician-AssistedSuicide: aReview of the Recent Literature. J PalliatCare.2008;24(3):173-192.

26. PapiernikE,ZeitlinJ,DelmasD,etal.Terminationofpregnancyamongverypretermbirthsanditsimpactonverypretermmortality:ResultsfromtenEuropeanpopulation-basedcohortsintheMOSAICstudy.BJOGAnIntJObstetGynaecol.2008;115(3):361-368.

27. BosmaJM,vanderWalG,Hosman-BenjaminseSL.LateterminationofpregnancyinNorthHolland.BrJObstetGynaecol.1997;104(4):478-487.

28. Garne E, Khoshnood B, LoaneM, Boyd PA, DolkH. Termination of pregnancy for fetalanomalyafter23weeksofgestation:AEuropeanregister-basedstudy.BJOGAnIntJObstetGynaecol.2010;117(6):660-666.

29. Lawonterminationofpregnancy[inDutch].BelgianlawGaz.1990.

30. BoydPA,DeViganC,KhoshnoodB,etal.SurveyofprenatalscreeningpoliciesinEuropeforstructuralmalformationsandchromosomeanomalies,andtheirimpactondetectionandterminationratesforneuraltubedefectsandDown’ssyndrome.BJOGAnIntJObstetGynaecol.2008;115(6):689-696.

31. LöfmarkR,NilstunT,CartwrightC,etal.Physicians’experienceswithend-of-lifedecision-making:Surveyin6EuropeancountriesandAustralia.BMCMed.2008;6:1-8.

92

Barrierstoandfacilitatorsofend-of-lifedecision-makingbyneonatologistsandneonatal

nursesinneonates:aqualitativestudy

LaureDombrecht,VeerlePiette,LucDeliens,FilipCools,KennethChambaere,LindeGoossens,GunnarNaulaers,LucCornette,KimBeernaert,JoachimCohenonbehalfoftheNICUconsortium

PublishedinJournalofPainandSymptomManagement2019,October19th.

93

Abstract

Context: Making end-of-life decisions in neonates involves ethically difficult and distressingdilemmas for healthcare providers. Insight into which factors complicate or facilitate thisdecision-makingprocesscouldbeanecessaryfirststepinformulatingrecommendationstoaidfuturepractice.

Objectives:Thisstudyaimedtoidentifybarrierstoandfacilitatorsoftheend-of-lifedecision-makingprocessasperceivedbyneonatologistsandnurses.

Methods:Weconductedsemi-structuredface-to-faceinterviewswith15neonatologistsand15neonatalnurses,recruitedthroughfourneonatalintensivecareunitsinFlanders,Belgium.Theywereaskedwhat factorshad facilitatedandcomplicatedpreviousend-of-lifedecision-makingprocesses.Tworesearchersindependentlyanalysedthedata,usingthematiccontentanalysistoextractandsummarizebarriersandfacilitators.

Results:Barriersandfacilitatorswerefoundatthreedistinctlevels:thecase-specificcontext(e.g.uncertainty of the diagnosis and specific characteristics of the child, the parents and thehealthcareproviderswhichmakedecision-makingmoredifficult),thedecision-makingprocess(e.g. multidisciplinary consultations and advance care planning (ACP) which make decision-makingeasier),andtheoverarchingstructure(e.g.lackofprivacyandcomplexlegislationmakingdecision-makingmorechallenging).

Conclusions:Barriersandfacilitatorsfoundinthisstudycanleadtorecommendations,somesimpler to implement than others, to aid the complex end-of-life decision making process.Recommendations include establishing regular multidisciplinary meetings to include allhealthcare providers and reduce unnecessary uncertainty, routinely implementing ACP inseverely ill neonates tomake importantdecisionsbeforehand, creatingprivacy forbad-newsconversationswithparentsandreviewingthecomplexlegalframeworkofperinatalend-of-lifedecision-making.

94

5.1 Background

Despitemedicaladvancesoverthelastdecades,asubstantialnumberofchildrendiebeforetheyreachtheageofoneyear1–3.Manyofthesedeathsareprecededbyanend-of-lifedecision(ELD)with a potentially life-shortening effect, such as withholding or withdrawing medication oractivelyendinglifewithlethalmedication4–7.Themedicalandethicaldilemmasduringend-of-life(EoL)decision-makingcausesignificantdistressinneonatologists,nursesandparents8.Inmostcountries,includinginBelgium,activelyendinglifewithlethalmedicationisillegal,thoughprevious research shows that some healthcare providerswould consider actively ending lifeacceptableinsevereorlethalcases9,andthatitdoeshappeninclinicalpractice3,10.Thismightmakethedecision-makingprocessevenmoredifficult.Therefore,researchintowhatcouldmakethisprocesslessdistressingisimperative.

BothhealthcareprovidersandparentsplayanactiveroleinEoLdecision-making11.However,healthcareprovidershavearangeofEoLexperienceswhichmakesthemideallyplacedtoreflectonwhatmakessuchdecision-makingeasierormoredifficult,whereasparentsusuallyhaveonlytheoneuniquelypersonalandtragicexperience.Sincetheviewpointofparentsisfundamentallydifferentfromthatofhealthcareproviders,butstillcrucialtoneonatalEoLdecision-making,aforthcomingpaperwillfocussolelyontheirexperiences.Fromahealthcareproviderperspective,physiciansareexpertsinunderstandingtheprognosisandpossibleoutcomes12,whilenursesarecontinuallypresentatthebedsideandoftenhaveacloserpersonalbondwiththeparents,makingthemkeyfiguresinbuildingatrustingrelationshipwiththeparents13,14whichiscrucialinEoLdecision-making.Theythushaveauniqueandimportantroleinthedecision-makingprocess,makinginvestigationofbothviewpointsessential.

Toourknowledge,nostudiesexistthatdescribebarrierstoandfacilitatorsofELDsinneonatesfrom a healthcare provider perspective. However, previous studies with a broader focus onaspectsofELDsinneonatesmentionfactorsinfluencingdecision-making:1)aFrenchinterviewstudyonattitudesandELDpracticesrevealedthatnursesoftenexperience the timebetweengraspingtheseverityofthesituationandactuallytakingadecisionasextremelydifficultastheyareconstantlyconfrontedwithsufferingofthechild13;2)arecentonlinesurveyinneonatologistsandnursesinSwitzerlandondecision-makingatthelimitofviabilityidentifiedseveralcrucialdifficultiessuchasprognosticuncertainty,difficultiesininterpretingtheattitudeoftheparents,insufficienttimefordecision-making,legalconstraintsandconflictsbetweentheirownprinciplesandunitpolicy12.

Furthermore, factors influencingdecision-makingarementioned in studies examiningoverallEoLcareinneonates.InonestudyonEoLexperiences,physiciansindicatedthatabondoftrustwithparentsmakescommunicatingbadnewseasier8;another,onmoralobligationsexperiencedbyhealthcareproviders,revealsthatanuncertainprognosisandambivalenceaboutincludingparentswhilewantingtoshieldthemfromtheburdenofdecision-makingarekeydifficulties15.

ThesestudiesrevealedsomeinfluencingfactorsonEoLdecision-makinginneonates,butdidnotexplicitlyfocusonbarriersandfacilitators,makingitpossiblethatkeyfactorsmayhavebeenoverlooked.WethereforeexaminebarriersandfacilitatorsintheEoLdecision-makingprocess

95

inneonates,asexperiencedbyneonatologistsandnurses.Hereby,wefocusedonwhatmakesiteasierormoredifficultintheprocesstocometoortomaketheend-of-lifedecision.Weaimedtostudythesebarriersandfacilitators,intheexpectationthatinsightintothemcanusefullyshapefutureEoLdecision-making.

5.2 Methods

5.2.1 Studydesign

A qualitative study was conducted using semi-structured face-to-face interviews withneonatologistsandneonatalnursesworkinginaFlemishneonatalintensivecareunit(NICU).Wechose a qualitative research methodology to cover the complexity, subtlety and individualspecificity of experiences in the end-of-life decision-making process regarding neonates thatwouldbemissedbyaquantitativeapproach.Becauseofthesensitivityofthesubjectweoptedforindividualinterviews.CriteriaforreportingqualitativeresearchfromtheCOREQguidelineswereused16.

5.2.2 Settingandparticipants

We recruited neonatologists working as resident physicians at one of four Flemish NICUs(University hospitals of Ghent, Brussels and Leuven, and general hospital Sint-Jan Bruges)betweenDecember2017andJuly2018whohadbeentheattending/treatingphysiciantoatleastonechildwhohaddiedattheNICUwhereanELDwasmadeinthepastyear,andnurseswhohadbeenthemostinvolved.Noexclusioncriteriawereused.

5.2.3 Recruitment

Aneonatologistofeachparticipatinghospital(FC,LG,GNandLC)informedallneonatologistsandnurseswithintheirrespectiveNICUofthepurposeofthestudy,andprovidedcontactdetailsofthosewillingtoparticipate.ResearcherscontactedthemandsetupadatefortheintervieweitherattheirNICUorattheirhomeresidency.Purposefulsamplingwasusedtoselectparticipants.

5.2.4 Datacollection

A topic guide (Table 5.1) was developed by a multidisciplinary team of nine experiencedresearchersinthefieldsofend-of-lifecareandneonatology.ParticipantswereaskedwhatmadeiteasierormoredifficulttomakeELDsintheNICU.Beforetheinterview,ashortquestionnairewas administered to collect socio-demographic data. LDm (female, MSc in experimentalpsychology;DoctoralResearcher)andVP(female,MAinneurolinguisticsandBScinpsychology;DoctoralResearcher)performedallinterviewswiththeparticipants.Datawerecollecteduntilnonewbarriersandfacilitatorsemergedforbothneonatologistsandnursesseparately,anddatasaturationwasachieved.

96

Questiontype Question PromptsIntroduction Iwanttodiscussthedifficulttopicof

end-of-lifedecisionsandwouldliketostartwithwhichdecisionsaresometimesbeingmadeinthisNICU?

Transition(onlyfornurses)

Inwhatwayareyou,asanurse,involvedintakingtheseend-of-lifedecisions?

Key - Whatmakesiteasierforyoutodecideonend-of-lifedecisions?

- Whatmakesitmoredifficulttodecideonend-of-lifedecisions?

- Doyoufeelsupportedbycolleaguesorparentsduringthisdecision-makingprocess?

Wewouldliketofocusonyourownroleasphysician/nurse(andnotonwhatmakesiteasier/moredifficultfortheparents).Otherpromptsinclude:- Whydoesthatmakeiteasier/moredifficultforyou?

- Howdidthatmakeyoufeel?- Canyougiveaspecificexampleofacasewherethishappened?Andwhatdecisionwaseventuallymadeinthisinstance?

Table5.1:semi-structuredinterviewguide.

5.2.5 Dataanalysis

Interviewswereaudiotapedandtranscribedverbatim.NVivo12wasusedforstructuringthedata and thematic content analysis17 was used to analyse it. Two researchers coded theinterviewsindependentlyandopenlybymeansofinductivecodingduringwhichtheysearchedfor facilitators and barriers that influenced the end-of-life decision-making process. The firsteightinterviewswerecodedbybothresearchers.Afterfiveinterviewsafirstdiscussiononcodenodesandtreesoccurred.Theother22interviewswerecodedbyoneoftheresearchers.Codenodesandtreeswerediscussedamongstbothresearchersatregularmeetings,andduringtwoseparate meetings afterwards with all co-authors. When coding discrepancies occurred,consensus was sought. Data saturation was reachedwhen no new codes emerged for threeconsecutiveinterviewsinneonatologistsandnursesseparately,andwhenasimilarnumberofparticipants from each participating hospital were recruited. The final model of factorsinfluencingEoLdecision-makinginneonateswasagreeduponbyallauthors.

5.3 Results

Weconducted15interviewswithneonatologistsand15withneonatalnursesfromfourNICUs(Table5.2),lastingaboutanhoureach.Identifiedthemesregardingbarriersandfacilitatorsonthe EoL decision-making process were classified into three discrete levels: 1) context level,themesrelatedtothespecificEoLcase;2)processlevel,themesrelatedtocharacteristicsofthedecision-makingprocess itself;and3)structure level, themesrelated tocharacteristicsof the

97

overarchingdeterminantsofoverallpolicyandpracticeintheNICUwardorinthewidersociety(Table5.3).

Neonatologists NeonatalnursesNumberofinterviewedcaregivers

15 15

StaffinNICU… A B C D

4344

4443

Sex Male 7 0 Female

8 15

Age <30years 0 3 30-39years 7 5 40-49years 6 4 >50years 2 3

YearsofexperienceinaNICU <5years 2 5 5-10years 5 1 11-20years 4 3 >20years 4 6

Table5.2:demographiccharacteristicsparticipants

5.3.1 Contextlevel

Accordingtotheinterviewees,thecharacteristicsofkeyplayerssuchasthechild,parentsandhealthcareproviderscanhaveaninfluenceonthedecision-makingprocess.

5.3.1.1 Childcharacteristics

Physiciansandnursesmentionedtheinfluenceofseveralchildcharacteristicsonthedecision-makingprocessincludinggestationalage,prognosisandpossibleELDoptions.

When the child is born at full term, healthcare providers indicated that the decision totransitionfromcurativecaretoanELDismoredifficultbecauseahealthy,fulltermbabyhadahighchanceofsurvivalearlyon,whilethesurvivalchancesofanextremelyprematurebabywerealreadylowermakingeveryonepreparedforbadnews.

98

“ItturnsoutthatIfinditmoredifficultwithchildrenbornattermthanwitha24-25weekbaby.WiththelatterIfeellike,let’sgiveitachancebutthennaturedecidesthatitwon’twork.That’sdifferenttochildrenwhoaredoingreallywellupto38weeksinthewomb,andthentheyarebornandgetseriousinfections.Iftheyhadbeen delivered by caesarean a week earlier, you’d have a perfect child. With aprematurebabythere’ssolittleyoucandowhenlabourstarts.”-Nurse12

Both neonatologists and nurses indicate that decisions are easier to make when a badprognosisbecomesevidentquicklyandiscertain,whilefluctuationsinhealthleadtodoubtsaboutlife-expectancyand/orfuturequalityoflife.

“Itoftenhastodowithpathology,andyouknowthetypeofdiscussionyoucanhaveabout ‘how certain is yourprognosis?’ That’s especially the casewithprematurebabieswithextensivebrainhaemorrhages.Ifinditeasyiftheyhavealreadybeenfairlyintensivelytreatedandyounoticethat,well,it’snotreallyworking.Andthenthere’sabrainhaemorrhageontopofalltherest.Thenyouthink,right,well,thisreallydoesn’tlookgood.But,well,ifyouhearthefigures,andtheymainlyhavetodowithextremelyprematurebabies,at25or26weeks,thereisquitealotofdebateaboutthat.[…]thatdoesleadtoquiteadifferenceinopinions.”-Doctor2

Lastly,oftheinterviewees,onlyneonatologistsdiscusstheimportanceofbeingsurethatalloptionshavebeenexploredfirst,beforeanELDisconsidered.Whenallcurativeoptionsfailed,andan ELD is the onlyway to ensure an end to the suffering of the child, the decision isdescribedasbeingeasierthanwhenothertreatmentoptionsarestillpossible.Furthermore,when an EoL decision is made, physicians indicate that it is easier if withholding orwithdrawing treatment is sufficient rather than when the only possible option involvesactivelyendinglifewithlethalmedication.

5.3.1.2 Parentcharacteristics

Neonatologists and nurses indicate the same barriers and facilitators in terms of parentcharacteristics, including cultural and language differences, socio-economic status andtherapeuticrelationshipswithparents.

In general, healthcareproviders indicate thatEoL decision-making is easierwhenparentshavethesamecultureandlanguageasthephysiciansandnursesinvolved.Translationsmakehealthcareprovidersfeellessabletoconveythedepthandnuancesneededtodescribethediagnoses and (EoL) treatment options. A difference in cultural background betweenhealthcareprovidersandparentsmakesneonatologistsandnurses feel theyarelimited toonlydiscussingcertainELDs.

“…averydifficultcontextisforexampleparentswithaMuslimbackground,whowanteverythingtobedonefortheirchildnomatterthecost,eventhoughthereisnopossibilityofdoinganythinguseful.Andyoustillhavetocontinueon,thatyouhavetodoafutilemedicalact.Thatmakesitmoredifficult.”–Doctor11

99

Alowersocio-economicstatuswasalsoindicatedasanimportantinfluencingfactor.Whenachildwillsufferaseverehandicapinfuture,anditisjudgedthatparentswillnotbeabletoprovideasafeenvironmentforthechildfinanciallyoremotionally,healthcarepractitionersfinddecidingonanELDeasierthanwhenthechildwillbecaredforandbothparentsarewell-resourcedfinanciallyandemotionally.Theformerincludeunstablehouseholdsituationswithe.g. drug abuse, criminal history, teenage pregnancies and extreme debt. The healthcareproviders indicate they find theseunstable situations facilitate end-of-lifedecision-makingbecausetheytakeintoconsiderationtheextremesufferingofthechildinfuture,duetotheirmedical condition, combinedwith adifficult family life.While someparticipants struggledwiththefactthatsocio-economicstatuswasindicatedasaninfluencingfactor,reflectionontheethicalramifications,othersstateditasamatteroffactandrationalizedthisasoneofmanyinfluencingfactorsindecision-making.

Lastly,bothneonatologistsandnursesindicatethattheEoLdecision-makingprocessiseasierwhenatherapeuticrelationshipisestablishedwiththeparents.

5.3.1.3 Healthcareprovidercharacteristics

Previous experience with EoL decisions was mentioned as a factor in making the EoLdecision-makingprocesseasier,becausehealthcareprovidersarebetterabletoanticipatethechild’sfuturecondition.Furthermore,somenursesindicatedthatexperiencewiththedisabilityandsufferingoftreatedchildrenlater intheir livesmakesEoLdecision-makingeasier,becausetheywerebetterabletoenvisagethechild’sfuturequalityoflife.

“Ithinkexperiencedoeshelp...certainlyinthelearningprocesssurroundingend-of-lifedecisions.IfIthinkbacknowtoabout14yearsago,thefirsttimeIcaredforafamilywithadyingchild,well,youstillreallydon’tknowwhatyouaresupposedtoaskparents,orsuggesttothem.AndnowIreallyhavedonequitealotandthenyoudoenduplearning.”-Nurse5

Lastly,physiciansandnursesmentionedtheeffectoftheirownabilitytorelatetothespecificcase;whentheyhavechildrenoftheirownortheirfamilysituationissimilar,decidingonanELDismoredifficult.

5.3.2 Processlevel

Accordingtoneonatologistsandnursesthecommunicationbetweenallinvolvedactors(parents,neonatologists,nurses,psychologists,etc.),divergenceofopinionandadvancecareplanningarekeyelements.

5.3.2.1 Communicationandmultidisciplinaryconsultations

Healthcare providersmentioned that communication amongst all actors, debriefings afterdeathandformalsecondopinionsarecrucialfactorsduringEoLdecision-makinginneonates.

100

Intensecommunicationbetweenhealthcareprovidersandparents is imperative inmakingELDs.Allactorsshouldbeawareofthemostrecentupdatesonthechild,andofeachother’sviewsandopinions.

“Whencommunicationgoesbadly,Ithinkthatthosecasesarethemostdifficult.Iamthinkingaboutachildthatwasillforalongtime[…]whattheparentswishedandhowthephysicianinterpretedthisdidnotmatch.”–Nurse5

Healthcarepractitionersalsomentionedcommunicationbetweenpractitionersboth insideandbeyondtheteamashelpfulduringthedecision-makingprocess,eitherformallyduringmultidisciplinaryteammeetingsordebriefings,orinformally.Multidisciplinarymeetingswiththeentireteam,includingphysiciansandnurses,ensuresthatdecisionsaresupportedbyallandthateveryoneisincludedinthedecision-makingprocess.Whenneonatologists,ormorefrequentlynurses,areexcludedfromthisdecision-makingprocess,butarelaterrequiredtoimplementthedecision,theEoLdecision-makingprocesswasexperiencedasbeingharder.

“I wasn’t involved then, actually, and then it was difficult at that point, if thedecisionshavealreadybeenmade,well, to gobackon them.As anoutsider, youmightsay,althoughofcoursewehaddiscusseditwitheachotherbeforehand.Buthowitactuallyhappened.Andifthechildhasdied,thenyouthinkohdear,wedoneedtositdownwitheveryoneassoonaspossibleanddiscussitandtoseewhatweneedtododifferentlynexttime.”-Doctor1

Onlyneonatologistsexpressedthe importanceofasking fora formalsecondopinionbyanindependent physician either within their own hospital (e.g. other disciplines such ascardiology)orfromanotherhospital.

“...thenIthinkthesecondopinionsystemisagoodsystem.Ifwehaveasituationlikethat,Iphone(name)andIsay:(name)wearegoingtoreferthatchildthrough,givemeafreshopinion.”-Doctor3

Formalandpre-setdebriefingsamongsthealthcareprovidersafterachilddiedwereindicatedashelpful in futureEoLdecision-makingprocesses.Debriefingsprovidereflectiononwhatwent well and what could be improved while an absence of debriefings can leave othermembersofthemedicalteamwithunresolvedquestions.

5.3.2.2 Divergenceofopinion

Whenoneoftheinvolvedactors(parents,nurses,neonatologists)wantstocontinuecurativetreatmentandothersoptforanELD,compromisesneedtobemade.Differingopinionscanput pressure on any one of them to change their minds, making EoL decision-makingextremelydifficult.

“IfI’mnotonthesamewavelengthastheparents,thatmakesitdifficultforme.Soitcangotwoways.Iftheparentsasktostop(thetreatment),butI’mnotyetready

101

forthatmyselforIthinkitisn’tclearenoughyet.Thosearethethingsthatmakeitdifficult.IfIbelievethatthereisnopoint,andtheparentsdon’tagree,I findthatdifficulttoo.”-Doctor2

5.3.2.3 Advancecareplanning(ACP)/mappingofpossibleactions

According to neonatologists and nurses, ACP is a crucial factor in EoL decision-making.Consideringinadvancetogetherwithhealthcareprovidersandparentsallthedirectionsthechild’sconditioncantakeanddecidingonwhichmedicalresponseswillbemadeineachleadsto easier decision-making than when rushed decisions have to be made due to acutedeteriorationwhereACPdidnotorcouldnottakeplace.

“...theparentscanalreadyindicatedirectlyatthatpointthat,yesbutdoctor,ifmychildisbornat24weeksandyouaretalkingabouthaemorrhagesthatcanhappen,ifthatisthecase,Iwanttobecertainyouwon’tintervene.OrotherwiseIwant,ifitturnsoutthatyouexpectmychildwillhavecertaindisabilitiesinthefuture,Idon’twant that. In the theoretical situation, then, that makes it easy to go backafterwards,whenwhatyoudiscussedactuallyhappensandthatyouhavealreadydiscusseditwiththeparentsyourself.”-Doctor7

WhenanELDisdiscussedduringtheACPprocess,thedyingprocesscanbeplannedaccordingto the wishes of parents and the advice of the healthcare providers. Planning includesreservingaprivateroom,makingsuretheparentsarepresent,thatdeathisnotrushed,andcreatingmemorieswithparents.

“Irememberacasewherethedeathwasfairlysudden,inareanimationsetting,andthedoor<oftheconsultationroom>wasopen.Andthenursesfortheotherchildrendidn’treallyrealisethatthechildwasdyingandthefathersaid:oneimagestillsticksinmymind:thatisthoselaughingnurseswalkingpastthedesk.AndthatwasverydifficultforhimandIalsoreportedthatbacktothenurseshere,andtheydecidedtoput a lamp on the desk and to use that, actually, as a signal that serenitywasneeded.”-Doctor11

5.3.3 Structurelevel

Athird important level includes factorsrelating to theoverarchingstructureof theward, thehospital and the broader society that could make decision-making and the decision-makingprocesseasierormoredifficult.

5.3.3.1 Emotionalandpracticalsupportattheward

Accordingtohealthcareproviders,emotionalsupport(orlackthereof)fromcolleaguesisacrucial facilitator (or barrier) in EoL decision-making in neonates. This includes being ‘ashouldertocryon’andbeingapersontoconfirmdiagnosesortreatmentoptionswith.Most

102

neonatologistsandnursesmentionedthelackofpsychologicalsupportforteammembersattheNICU.

“Ithinkthatweneedapsychologist,wellwehaveapsychologistattheward.SheistherefortheparentsandIthinkthatshecouldmeanmuchmoretoourward.[…]She<thepsychologist>isnotthereforus,andweseethat,thatshe’snotthereforus.”–Nurse1

Participants indicate thepositiveeffectofaward thatpromotescollegialityandteamworkcultureduringEoLcare.Whenothernursescantakeoversomeofyourdailytasksoraidincaringforlesscriticalpatients,orphysicianscancoverforeachothersothattheyhavethetimetoallocatesolelytotheparents,EoLcareforadyingneonateisindicatedtobeeasier.

“Ifmyotherchildrenaretakenoverbycolleagues,soIonlyhavetoconcernmyselfwiththatbaby.Intermsoftheteam,ifitreallystartstobeacriticaltime,notyetleadinguptobutiftheyarestillstablebutiftheparentsaretherethen,forexample,then I could just concentrate on those people quietly onmy own.My colleagueswouldtakeovermywork,infact.Thatisverypracticalbutveryimportant.”–Nurse1

5.3.3.2 NICUpolicy,practiceandexpertise

Healthcareprovidersmentionedthenegativeeffectof lackofaseparateroomforprivacy,shortageofavailabletrainedpersonnelanddifferencesinexpertiseacrossNICUs.

“Thisistheonlyinterviewroomwehaveforeveryone,foreverything,forwhoeveritis.To talkaboutgoinghome, release fromhospital, follow-upconversationswithnurses,trainees.Itallhappenshere.Andpeoplejustwanderinandout.Thatisn’tvery pleasant, you just want to be alone with the parents at that point andconcentrateonthem.Leaveyourphonewithsomeoneelsesothatyoucandevoteallyourattentiontothosepeopleandthatstory.”-Doctor14

AnotherimportantaspectofNICUpracticementionedbybothhealthcareprovidersisthatashortageofneonatologistsandnursesexperiencedinEoLcareleadstoahigherburdenonqualifiedstaff.

OnlyneonatologistsmentionedthatdifferencesinknowledgeofcertaindiagnosesbetweendifferentNICUsandtheiraccompanyingstandardtreatmentplansare,withoutadequatewaystodisseminatethisknowledge,animportantbarriertoprovidingthebestpossiblecareattheendofaneonate’slife.

“Ithinkthatgettinganideaofhowitisdoneinotherhospitalsisalreadyabigthing.Becauseyoudon’tfindoutfromeachotherhowotherhospitalsdothings.Whattheircriteriaare, forexample, forstoppingtreatmentinachildwithsevereperipartalasphyxia.With seriousneurological abnormalities. I’d like just to beable to talk

103

aboutthatopenly.Becauseeveryonecangetholdoftheliterature.Butthereisstilladifferencebetweenreadingastudyanddoingitforrealinyourdepartment.”-Doctor13

5.3.3.3 Legalframework

ThecurrentBelgianlegislationwasalsomentionedbysomeneonatologistsandnurses.Whenmentioned,theystatedthatthelackofalegalframework-activelyendingthelifeofaneonateiscurrentlynotallowed-isseenasanimportantbarrierincontrasttopregnancy,wherethereistheoptiontoterminateassoonasalife-limitingfoetalabnormalityhasbeendiagnosed.

“But,well,ifthechildhasn’thadanyacutesituationsorcomplicationsyet,there’snothingyoucando.Andthosecasesarerare,buttheydoexist.Andthenifyoualsohaveparentswhoarereallyaskingurgentquestionsaboutendingthings,well,thereisactuallynothingyoucandoasadoctorandIfindthattough.”&“Butsomethingthatconcernshealthcareprovidersisthediscrepancyinthelegalsituationbetweentheprenatalandpostnatalperiod.[...]Thisimpliesthatprenatal,withlotsofthingsthat you can see and know, that you can also go quite a long way towardsterminating the pregnancy and that there is probably an even bigger differencetherethaninwhatgoesonneonatally?”-Doctor10

104

Theme DescriptionWhenthethemeisonlyabarrierorafacilitatoritwillbeindicatedbya(b)oran(f).Whenthethemecanbeseenasafacilitatorandtheoppositecanbeseenasabarrier,onlythefacilitatororbarrierismentionedwhichwillbeindicatedby(f;opposite=b)or(b;opposite=f)

Mentionedby…Neonatologists Nurses

Context Characteristicsofthespecificcasethatinfluencetheend-of-lifedecision-makingprocess

Childcharacteristics - Medicaldiagnosisofthechild:ð Certaintyofthediagnosis(f;opposite=b)ð Abadprognosisisquicklyevident(f;opposite=b)

xx

xx

- Babyisbornatfullterm(b;opposite=f)- Theinfantlookshealthy(b)- Medicaloptions:

ð Everycurativeoptionwasexploredbeforeconsideringanend-of-lifedecision(f)

ð Onlypossibleend-of-lifedecisionisactivelyendingthelifeofaneonate(b;opposite=f)

xxxx

xx

Parentcharacteristics - Culturaldifferencesbetweenparentsandhealthcareproviders(b;opposite=f)

- Differentlanguage(b;opposite=f)

xx

xx

- Lowersocio-economicstatus(f;opposite=b)- Havingatherapeuticrelationshipwiththeparents(f;opposite=b)

xx

xx

Healthcareprovidercharacteristics

- Experienceð Experiencewithend-of-lifedecisions(f;opposite=b)ð Experiencewithdisabilityandsufferingofchildrenlaterinlife(f)

x

xx

- Personalcharacteristicsð Havingchildrenofyourown(b)ð Being/havingbeeninasimilarpersonalsituation(b)

xx

xx

Process Characteristicsofthedecision-makingprocessitself Communicationand(multidisciplinary)consultations

- Formal(organised)andinformal(e.g.hallwayencounter)communication:ð Clear,efficient,andregularcommunicationbetweenparentsand

healthcareprofessionals(f;opposite=b)ð Healthcareprofessionalsamongstthemselvescommunicateclearly,

efficientlyandregularly(f;opposite=b)ð Formaldebriefingsafterdeathtoimprovetheend-of-lifedecision-

makingprocessinthefuture(f;opposite=b)

xxx

xxx

- Formalandorganisedcommunicationwith(external)healthcareproviders:ð Asecondopinionaboutthediagnosisand/ortheend-of-lifedecision(f)ð Multidisciplinarymeetings(f)ð Beingincluded/consultedduringtheend-of-lifedecision-making

process(f;opposite=b)

xxx

xx

Divergenceofopinion - Betweenparentsandhealthcareproviders(b;opposite=f)- Betweenhealthcareprofessionalsamongstthemselves(b;opposite=f)

xx

xx

Advancecareplanning/mappingofpossibleactions

- Planningthedifferentpossibleoutcomesandtreatmentoptionswithhealthcareprovidersandparents(f) x x

- Healthcareprovidersknowthenorms,valuesandwishesoftheparents(f;opposite=b) x x

- Planningthedyingprocess:Finalmomentsareplanned(who,howandwhen)andserene(f;opposite=barrier)

x

x

Structure Characteristicsoftheoverarchingstructure(society,NICUwardpolicyandpractice)

Emotionalandpracticalsupportattheward

- Emotionalsupportfromcolleagues(f;opposite=b) x x- SupportfromapsychologistattheNICUisnotavailable(b) x x- CultureintheNICUofcolleaguesworkingtogether,takingovertasksor

assistingeachotherduringthedyingprocessofaninfant(f) x x

NICUpolicy,practiceandexpertise

- VaryingknowledgebetweenthedifferentNICUs(b) x - Notenoughhealthcareprofessionalstrainedinend-of-lifecare(b) x x- Absenceofseparateroomtoaccommodateparentsandinfantsduringthe

decision-makingprocessandbefore,duringandafterdeath(b)x x

Legalframework - Activelyendingthelifeofaneonatewithlethaldrugsisnotincludedinlegalframework(b) x x

- Discrepanciesbetweenthelegislationprenatallyandpostnatally(b) x Table5.3:barriersandfacilitatorsoftheneonatalend-of-lifedecision-makingprocess

105

5.4 Discussion

InthisqualitativeinterviewstudywithneonatologistsandnursesworkinginaNICUwefoundfactors thatmayhinderor facilitate end-of-lifedecision-making inneonateson threedistinctlevels, namely the case-specific context level, the decision-making process level and theoverarching structure level. Key barriers and facilitators identified relate to specificcharacteristicsoftheinvolvedactors(suchasculturalandlanguagedifferences,atherapeuticbond with parents and the experience of the healthcare practitioners), uncertainty of theprognosis,ACPandtheinfluenceofpolicy,legislationandmedicalpractice.

5.4.1 Strengthsandlimitations

Byusingthequalitativeapproachofface-to-faceinterviewswithbothneonatologistsandnurseswewere able to give a viewofwhatmakesEoLdecision-making inneonates easier ormoredifficult for them. We believe parents could have crucial additional insights which will bereflectedoninaforthcomingseparatepublication;howevertheexperienceofbereavedparentsfitslesswellintothisstudy,whosefocusisthetheoreticalgeneralizabilityofELDexperiencesinneonatesandhowthiscancontributetorecommendationsonthestandardEoLdecision-makingprocessinneonates.

5.4.2 Generaldiscussion

Our results show there are some modifiable factors which may aid the complex end-of-lifedecision-makingprocess,thoughsomecouldbeconsideredmorepossibletoachievethanothers.

Thelackofprivacyandseparateroomsforbad-newsconversationswasmentionedbyhealthcareproviders as a barrier to the EoL decision-making process. Creating privacy for bad-newsconversationssothatdifficultELDscanbemadewithoutunnecessaryinterferencecouldaidbothhealthcareprovidersandparents, indicating thatsmallchangescouldpotentiallyhavealargeimpact.Therearesimilarfindingsinpreviousresearchintothepaediatricintensivecareunit,indicatingthattheintensivecareunitisnotseenasanidealenvironmentforEoLdecision-makingandbroaderEoLcaresinceprivacycannotbeassured18.

Bothneonatologistsandnursesmentionedtheimportanceofbuildingintodailypracticebothmultidisciplinaryteammeetingsanddebriefingsafterthedeathofaneonate.Previousresearchhasalreadysuggestedmakinguseofthecollectivewisdomofexperiencedhealthcareprovidersto reduceuncertainty ina general intensive care setting19. Especially inneonates, prognosticuncertainty is a key theme20. Regular multidisciplinary meetings could provide healthcareproviderswithahigherdegreeofinvolvementwithintheseELDsandwithafeelingofcertaintythatdecisionsarecarriedbytheentireteam,reducingunnecessaryuncertainty.

RespondentsemphasizedtheimportanceofACPinneonateswithasevereprognosis.PreviousresearchalreadyindicatesthebenefitsofroutineuseofanindividualizedsymptommanagementplanforneonatesduringEoLcare21.Inadults,ACPisknowntodecreasedecisionalconflictforsurrogatedecision-makers, since they aremore likely to knowthepatient’swishes22.Also, in

106

adolescents, ACP leads to better communication between adolescent, parents and healthcareproviders23. Aside from these possible effects of routinely implementing ACP in severely illneonates,ourresultsalsoindicatethefacilitatingeffectofhavingpreviouslyplannedcoursesofaction for all possible outcomes on the EoL decision-making process for the healthcarepractitionersinvolved.

Another significant factor includes promoting emotional support and a team-work cultureamongststaffinaNICU.MakingitpossibletoswitchtasksduringEoLcaretorelieveotherssotheycanfocusonthedyinginfant,orprovidingtheopportunityforstaffmemberstoindicatewhetherornottheyarewillingtobepartofanEoLdecision-makingprocessatthattime,canhaveaninfluenceontheoverallwellbeingofhealthcarepractitionersthemselves24.Debriefingsandevaluationstodiscussemotionalwellbeingofstaffbefore,duringorafteranEoLdecision-makingprocesscouldfurtherpromoteopportunitiesforthemtosupporteachother.

Theneedformoreexperience,andtheneedformorehealthcareproviderstrainedinneonatalEoLcarementionedbyphysiciansandnursescanbelinkedtogetherunderamoregeneralneedforeducationandtraininginEoLcareandELDs.PreviousresearchindicatedthatahighnumberofstudieshavereportedasimilarneedforformaltraininginbothbereavementcareandoverallEoL care communication skills, allowing time to learn from others25. Including amodule onneonatal death and EoL decision-making in standard curricula for healthcare practitionersincreasesclinicalexperienceandEoLcommunicationskillsearlyonintraining,whichleadstoenhancedconfidenceandfewernegativeexperienceswithEoLcareintheNICU25.

Although parental involvement in EoL decision-making is currently common practiceinternationally11,neonatologistsandnursesindicatedthatwhenparentshaveadifferentculturalbackgroundtoorspeakadifferentlanguage fromthehealthcareproviders,difficulties inEoLdecision-making may arise. Cultural differences can result in misunderstandings and/orfundamentallydifferentviewsontheacceptabilityofcertainELDs.As inadults,we thinkthatperinatalpalliative care teams shouldbe consulted tomediateas they are trained indifficultconversations26.However,nocurrentBelgianperinatalpalliativecareteamsexist.

Some respondentsmentioned the difficulty of theEoL decision-making processwhen severefuturesufferingisforeseenwherewithholdingorwithdrawingtreatmentwouldnotresultinthedeathof theneonate.This isbecauseactivelyending the lifeofaneonate is illegalwithintheBelgian legislation which therefore limits the possible options in such cases. Furthermore,previous studies indicate the occurrence of these active ELDs in Flanders27 and the positiveattitudeofahighnumberofneonatalhealthcarepractitionerstowardsthesetypesofELDs9.Ourresultscanbethebasisforanethicalandlegaldiscussionaboutinitiatinglegislationsimilartothat in the Netherlands where actively ending the life of a neonate is currently legislativelycondonedunderstrictconditions28.NothavingthisoptioniscurrentlyseenasabarrierindifficultEoLdecision-makingprocesses.BecauseBelgiumhasbothaeuthanasialawincompetentminorsandadults,anda lawallowinglate terminationofpregnancy incaseofsevereor lethal foetalanomalies, we can state that the ethical climate in Belgium concerning ending life could beconsideredfairlypermissivecomparedtoothercountries.Possibly,theseexperiencedbarrierscouldbedifferentincountrieswithalesspermissiveclimate.

107

Finally, someof the influencing factors found inourstudyarenot inthepowerofhealthcarepractitioners tomodify, including thegestationalageof theneonate, lowersocio-economicorunfortunate household situations, and the effect on relating to a specific case because ofsimilaritieswith theirownsituations.Beingawareof these influencing factorsduringanEoLdecision-makingprocessinneonatescanbeseenasacrucialfirststeptowardsaneasierdecision-making process. In-depth research is needed on the effect of these risk-factors, which couldpossibly increase theoddsof thedeathof the infant following an end-of-lifedecision, on theprognosis of the child. Furthermore, providing healthcare providers with concrete tools inimprovingcommunicationduringdifficultend-of-lifediscussions,especiallywithregardtotheserisk-factors, should be included in nurse and physician education. Additionally, thoughparticipants did not indicate it themselves, training healthcare providers in ethical decision-makingmightaidinprovidingclaritywhendealingwiththesecomplicatedsituations29.

5.4.3 Conclusion

Ourqualitativeinterviewstudyrevealedbarriersandfacilitatorsduringtheend-of-lifedecision-makingprocess inneonates as reportedbyhealthcarepractitioners. Somemodifiable factorswere identified to improve theprocess, such as creatingprivacy forbad-news conversations,regularmultidisciplinarymeetingsanddebriefingstoreduceuncertainty,routinelysettingupanadvance care plan, promoting emotional support and team-work culture amongst healthcareproviders,aneedformoreexperienceinend-of-lifecare,awaytodealwithculturalorlanguagedifferences,andnavigatingadifficultlegalframework;thesepossiblyrequiremorefundamentalchangesinNICUpolicyoroverallsocietyinordertofacilitatetheend-of-lifedecisionprocessinclinicalpractice.

5.5 Acknowledgements

Wewouldliketothankallneonatologistsandnurseswhoparticipatedinthisstudy,andtheirparticipatinghospitals(UZGent,UZBrussel,UZLeuven,andAZStJanBrugge).Furthermore,wewould also like to thank our collaborators in the NICU consortium: Sabine Laroche, ClaireTheyskens,ChristineVandeputteandHildeVandeBroek.Lastly,wewouldliketothankHelenWhitefortranslatingthequotesusedinthisarticle,andJaneRuthvenforherlanguageediting.

108

5.6 Referencelistchapter5

1. Rüegger C, Hegglin M, Adams M, Bucher HU. Population based trends in mortality,morbidity and treatment for very preterm- and very low birthweight infants over 12years.BMCPediatr.2012;12(1):17.

2. DevliegerR,MartensE,GoemaesR,CammuH.PerinataleActiviteiteninVlaanderen2017.StudvoorPerinatEpidemiol.2018.

3. ProvoostV,CoolsF,MortierF,etal.Medicalend-of-lifedecisionsinneonatesandinfantsinFlanders.Lancet.2005;365(9467):1315-1320.

4. CuttiniM, NadaiM, KaminskiM, et al. End-of-lifedecisions in neonatal intensive care:physicians’self-reportedpracticesinsevenEuropeancountries.EURONICStudyGroup.Lancet.2000;355:2112-2118.

5. VerhagenE,SauerPJJ.TheGroningenprotocol--euthanasiainseverelyillnewborns.NEnglJMed.2005;352(10):959-962.

6. BergerTM,HoferA.Causesandcircumstancesofneonataldeathsin108consecutivecasesover a 10-year period at the Children’sHospital of Lucerne, Switzerland.Neonatology.2009;95(2):157-163.

7. KeenanHT, Sheetz J, Bratton SL. Differences in Characteristics of Dying ChildrenWhoReceiveandDoNotReceivePalliativeCare.2013:72-78.

8. EpsteinEG.End-of-lifeexperiencesofnursesandphysiciansinthenewbornintensivecareunit.JPerinatol.2008;28:771-778.

9. Dombrecht L, Deliens L, Chambaere K, et al. Neonatologists and neonatal nurses havepositive attitudes towards perinatal end-of-life decisions, a nationwide survey. ActaPaediatr.2019:apa.14797.

10. ten Cate K, van de Vathorst S, Onwuteaka-Philipsen BD, van der Heide A. End-of-lifedecisions for children under1 year of age in theNetherlands: decreased frequency ofadministrationofdrugstodeliberatelyhastendeath.JMedEthics.2015;41(10):795-798.

11. PartridgeJC,MartinezAM,NishidaH,etal.InternationalComparisonofCareforVeryLowBirth Weight Infants : Parents ’ Perceptions of Counseling and Decision-Making.2005;116(2):28-38.

12. BucherHU,KleinSD,HendriksMJ,etal.Decision-makingatthelimitofviability:Differingperceptions and opinions between neonatal physicians and nurses. BMC Pediatr.2018;18(1):81.

13. GarelM,CaeymaexL,CuttiniM,KaminskiM.Ethicallycomplexdecisionsintheneonatalintensive careunit : impact of thenewFrench legislationonattitudes andpracticesofphysiciansandnurses.2011:240-244.

14. Wocial LD.Life SupportDecisions Involving Imperiled Infants. J PerinatNeonatalNurs.2000;14(2):73-86.

15. EpsteinEG.Moralobligationsofnursesandphysiciansinneonatalend-of-lifecare.NursEthics.2010;17(5):577-589.

16. Tong A, Sainsbury P, Craig J. Consolidated criterio for reporting qualitative research(COREQ): a 32- item checklist for interviews and focus group. Int J Qual Heal Care.2007;19(6):349-357.

17. Anderson R. Thematic Content Analysis (TCA) - Descriptive Presentation of QualitativeData.;2007.

109

18. Howes C. Caring until the end : a systematic literature review exploring PaediatricIntensiveCareUnitend-of-lifecare.2014;20(1):41-51.

19. FisherM,RidleyS.Uncertaintyinend-of-lifecareandshareddecisionmaking.CritCareResusc.2011;13(3):81-87.

20. Brinchmann BS. Ethical Decisions about Neonates in Norway. J Nurs Scholarsh.1999;31(3):1998-1999.

21. GilmourD,DaviesMW,HerbertAR.Adequacyofpalliativecareinasingletertiaryneonatalunit.JPaediatrChildHealth.2017;53:136-144.

22. Chiarchiaro J,BuddadhumarukP,ArnoldRM,WhiteDB.Prioradvancecareplanning isassociatedwithlessdecisionalconflictamongsurrogatesforcriticallyillpatients.AnnAmThoracSoc.2015;12(10):1528-1533.

23. LyonME,JacobsS,BriggsL,ChengYI,WangJ.Family-centeredadvancecareplanningforteenswithcancer.JAMAPediatr.2013;167(5):460-467.

24. CatlinA,CarterB.CreationofaNeonatalEnd-of-LifePalliativeCareProtocol.JPerinatol.2002;22:184-195.

25. ShoreyS,AndréB,LopezV.InternationalJournalofNursingStudiesTheexperiencesandneedsofhealthcareprofessionalsfacingperinataldeath :Ascopingreview.IntJNursStud.2017;68:25-39.

26. MartyCM,CarterBS.SeminarsinFetal&NeonatalMedicineEthicsandpalliativecareintheperinatalworld.SeminFetalNeonatalMed.2018;23(1):35-38.

27. Dombrecht L, Beernaert K, Roets E, et al. Apost-mortem population survey on foetal-infantileend-of-lifedecisions :aresearchprotocol.BMCPediatr.2018:1-9.

28. Verhagen AAE. The Groningen Protocol for newborn euthanasia; which way did theslipperyslopetilt?JMedEthics.2013;39(5):293-295.

29. Schneiderman LJ, Gilmer T, Teetzel HD, et al. Effect of Ethics Consultations onNonbeneficial Life-Sustaining Treatments in the Intensive Care Setting A RandomizedControlledTrial.Jama.2003;290(9):1166-1172.

110

Psychologicalsupportinend-

of-lifedecision-makinginneonatalintensivecareunits:fullpopulationsurveyamongneonatologistsandneonatal

nurses

LaureDombrecht, JoachimCohen,FilipCools,LucDeliens,LindeGoossens,GunnarNaulaers,KimBeernaert,KennethChambaereonbehalfoftheNICUconsortium

PublishedinPalliativeMedicine2019,November19th.

111

Abstract

Background:Moraldistressandburn-outrelatedtoend-of-lifedecisions(ELDs)inneonatesiscommon in neonatologists and nurses working in Neonatal Intensive Care Units (NICUs).Attentiontotheiremotionalburdenandpsychologicalsupportinresearchislacking.

Aim: To evaluate perceived psychological support in relation to ELDs of neonatologists andnursesworkinginFlemishNICUs,andwhetherornotthissupportissufficient.

Design/participants: A self-administered questionnaire was sent to all neonatologists andneonatalnursesofalleightFlemishNICUs(Belgium)inMay2017.Theresponseratewas63%(52/83)forneonatologistsand46%(250/527)fornurses.Respondentsindicatedtheirlevelofagreement(5-pointLikertscale)withsevenstatementsregardingpsychologicalsupport.

Results:70%ofneonatologistsandnursesreportedexperiencingmorestressthannormalwhenconfrontedwithanELD;86%ofneonatologists feelsupportedby theircolleagueswhentheymakeELDs,45%ofnursesfeelthatthetreatingphysicianlistenstotheiropinionwhenELDsaremade. About 60% of both neonatologists and nurses would likemore psychological supportofferedbytheirdepartmentwhenconfrontedwithELDsand41%ofneonatologistsand50%ofnursesstated theydidnothaveenoughpsychologicalsupport fromtheirdepartmentwhenapatientdied.Demographicgroupsdidnotdifferintermsofperceivedlackofsufficientsupport.

Conclusions:EventhoughNICUcolleaguesgenerallysupporteachotherindifficultELDs,thepsychologicalsupportprovidedbytheirdepartmentiscurrentlynotsufficient.Professionaladhoc counselling or standard debriefings could substantially improve this perceived lack ofsupport.

112

6.1 Background

Neonatologists andnursesworking inneonatal intensive careunits (NICUs)often experiencemoraldistress1,2especiallywhenaninfantintheircarecannolongerbenefitfromtreatmentandalife-shorteningend-of-lifedecision(ELD)ismade1,3.TheemotionalimpactonparentsoflosingachildandthesupportneededfrombothNICUandpsychologicalsupportstaffhavepreviouslybeen studied4,5 and guidelines on supporting them have been developed by severalorganisations3,6,7.However,researchonprofessionalsupportforNICUstaffandtheircopingandemotionalburdenhasbeenlacking.

Healthcare professionals often experience suffering and grief as well as moral distress andemotional exhaustion8,9. Becauseof this, ICUhealthcareprofessionals in generalareprone todevelopingcompassionfatigueandburnout10,11.InNICUs,surveystudiesestimatetheprevalenceof burnout tobe30% inneonatologists12and7.5-54.4% innurses13.Developingburnout andcompassionfatiguedoesnotonlyhaveanimpactontheirpersonallifebutalsoaffectstheirabilitytocareforpatientsandtohaveempathyforgrievingparents6,11,12whichcouldreducethequalityofcareoverall.Despitetheseknownrisks,onlyonestudy,afterreviewingneonatalend-of-lifeprotocols, recommended colleague andprofessionalpsychological support around end-of-lifecare forNICUstaffmembers3.Actual researchonperceivedpsychologicalsupportbyand forNICUprofessionalsislacking.

OurstudyevaluatesstressinrelationtoELDs,perceivedcolleagueandprofessionalpsychologicalsupportandwhetherornotthissupportissufficientinneonatologistsandnursesworkinginNICUs and examines whether psychological support differs between socio-demographic orprofessionalgroups.

6.2 Methods

6.2.1 Designandparticipants

Weperformedafull-populationmailsurveyofallneonatologistsandneonatalnursesinalleightFlemishNICUs,withfullcooperationfromallunits.Atotalof83neonatologistsand527nurseswereidentifiedbymeansofpersonnelfiles.

6.2.2 Datacollection

ArepresentativeworkingateachNICUhandedoutthequestionnairetoeveryneonatologistandnurseintheirunitinMay2017(gatekeepermethod)invitingthemtofillitoutanonymouslyandsenditbackinaprepaidenvelopewithinonemonth.Thismethodwaspreferredtosendingaquestionnairedirectlytoeveryneonatologistandnurseinordertomaximisetheirmotivationtoparticipate.Sendingbackafilled-outquestionnairewasseenasinformedconsent.Weobtainedethical approval from the ethical review board of Ghent University Hospital (Registrationnumber:B670201731709).

113

6.2.3 Questionnaire

Thequestionnaireitemsusedinthisreportconsistedofsevensocio-demographicquestions(seeTable1)andsevenquestionsconcerningcolleagueandprofessionalpsychologicalsupport.Thepsychological supportquestionswerebasedonastudy fromWeintraubetal. on compassionfatigue,burnoutcompassionsatisfactioninaneonatalintensivecareunitintheUnitedStates11,andweretranslatedandamendedtotheFlemishcontextbyamultidisciplinaryteamconsistingof sociologists, psychologists, neonatologists and a gynaecologist. The questionnaire wascognitivelytestedwithfiveneonatologists(fromfourseparatehospitals),threeneonatalnurses(from two separate hospitals) and one gynaecologist, leading to only minor adjustments inwording.

6.2.4 Measures

The questionnaire included statements about perceived stress, professional psychologicalsupportprovidedbytheNICUandpsychologicalsupportprovidedbycolleagues.WeincludedastatementontheoptionofexpressingprotestconcerninganELD,whichcouldbeanadditionalsourceofdistresswhenthisisdiscouraged.Thestatementswerescoredona5-pointLikertscale.Threeofthesevenquestionsdifferedbetweenneonatologistsandnursesbecause,intheFlemishhealthcaresetting,physiciansarethemaindecision-makerswhenitcomestomakingend-of-lifedecisionsfortheirpatients,mostlyduringphysicianteammeetings.Thiswhilenursesareoftennot involved in this decision-making process, but they are however involved in theimplementationofthemedicaldecisions.

6.2.5 Statisticalanalysis(SPSS24.0)

Percentagesofdisagreement(‘totallydisagree’and‘disagree’),neutralityandagreement(‘agree’and‘totallyagree’)werecalculatedforneonatologistsandnursesseparately.

114

6.3 Results

Across all eight NICUs, the response rate was 63% (52/83) for neonatologists and 46%(250/527) fornurses. In our sample, 71%ofneonatologists and95%ofnurseswere female(Table6.1).

NeonatologistsN=52(%)

NeonatalnursesN=250(%)

Sex Female Male

37(71.2)15(28.8)

237(95.2)12(4.8)

Age <30 30-39 40-49 ≥50

12(23.1)15(28.8)11(21.2)14(26.9)

75(30.2)65(26.2)53(21.4)55(22.2)

YearsofexperienceworkinginaNICU <5years 5-10years 11-20years >20years

22(42.3)8(15.4)9(17.3)13(25)

58(23.3)34(13.7)77(30.9)80(32.1)

Functionofphysicians Neonatologist Specialistintraining

39(75)13(25)

N/A

Degreenurses Graduate Bachelor Master

N/A

3(1.2)229(92.3)16(6.5)

Religionorbeliefs Religious Notreligious

28(53.8)24(46.2)

164(66.1)84(33.9)

BeliefthattheirreligionorbeliefhasimpactontheirattitudestowardsELDs

Yes No

13(25.5)38(74.5)

45(18.4)200(81.6)

Missingvalues:variedfrom0%forsex,age,yearsofexperience,functionandto1.9%intheimpactofreligioninneonatologists(n=52)andfrom0.4%insexandyearsofexperienceto2%intheimpactofreligioninneonatalnurses(n=250)

Table6.1:demographicsofneonatologistsandneonatalnurses

115

MostneonatologistsandnursesagreedthatmakinganELD(neonatologists)orbeingconfrontedbyone(nurses)inneonatescausesmorestressthanusual(72.5%and70.2%respectively,Table6.2). During the decision-making process,most neonatologists (86.3%) agreed that they feelsupported by their colleagues. Fewer than half the neonatal nurses (44.6%) agreed thatphysicianslistentotheiropinionsinmakinganELD.Whilemostneonatologists(88.2%)agreedthattheirNICUprovidessufficientopportunitytoexpressprotestaboutcertainELDs,only31.6%ofnursesagreedwiththisstatement.AlmostallneonatologistsandnursesagreedthattheycantalktotheircolleagueswhensomethingisbotheringthemaboutanELD(neonatologists,94.1%,nurses,92.4%).WhentheydonotagreewithanELDthathasbeenmade,halfofneonatologists(52.9%)and65%ofnursesagreedthattheycanopttonolongerbeinvolvedinthatcase;57%ofneonatologistsand60%ofneonatalnursesagreedthattheywouldprefertheirNICUtoprovidemorepsychologicalsupportforstaffmemberswhentheyarebeingconfrontedwithELDs.About40% of neonatologists and half of neonatal nurses agreed that they receive sufficientpsychologicalsupportfromtheirNICUafterapatientdies.

Forbothgroupssex,age(<40yearsand≥40years),yearsofexperience(£10years,>10years),whetherornottheyarereligiousandwhethertheybelievetheirreligionhasanimpactontheirattitudes towards ELDs were added. Additionally, we included function for neonatologists(residentorintraining)anddiplomafornurses(bachelor,mastersorgraduatedegree).Noneofthedemographicvariableshadasignificantinfluence(notintable).

116

Item Group Disagree(%)

Neutral(%)

Agree(%)

StressTakingdecisionsabouttheendoflifecausesmemorestressthanusual

Neonatologist 6(11.8) 8(15.7) 37(72.5)Neonatalnurse N/A N/A N/A

Beingconfrontedwithanend-of-lifedecisionforanewbornbabyinmydepartmentcausesmemorestressthanusualc

Neonatologist N/A N/A N/ANeonatalnurse 44(17.7) 30(12.1) 174(70.2)

PsychologicalsupportbycolleaguesIfeelthatIambeingsupportedbymycolleaguesinthedecisionsImakeaboutmypatients’endoflife

Neonatologist 0(0) 7(13.7) 44(86.3)Neonatalnurse N/A N/A N/A

Ihavethefeelingthatthetreatingphysician(s)listentomyopinionwhenanend-of-lifedecisionistakenaboutanewbornbabywithaseriousconditionb

Neonatologist N/A N/A N/ANeonatalnurse 68(27.3) 70(28.1) 111(44.6)

ThereareadequatepossibilitiesofferedbythedepartmenttoexpressanyprotestsImighthaveaboutend-of-lifedecisionsd

Neonatologist 2(3.9) 4(7.8) 45(88.2)Neonatalnurse 95(38.5) 74(30) 78(31.6)

Ifsomethingisbotheringmeabouttakinganend-of-lifedecision,Icantalktomycolleaguesaboutit

Neonatologist 0(0) 3(5.9) 48(94.1)Neonatalnurse N/A N/A N/A

Ifsomethingisbotheringmeaboutadecisionmadeaboutapatient’sendoflife,Icantalktomycolleaguesaboutita

Neonatologist N/A N/A N/ANeonatalnurse 8(3.2) 11(4.4) 231(92.4)

IfIdon’tagreewiththeoutcomeofacertaindecisionaboutapatient’sendoflife,Icanopttonolongerbeinvolvedinthatcasea

Neonatologist 10(19.6) 14(27.5) 27(52.9)Neonatalnurse 23(9.2) 65(26) 162(64.8)

ProfessionalpsychologicalsupportIwouldlikemydepartmenttooffermorepsychologicalhelptostaffwhentheyareconfrontedwithend-of-lifedecisionsc

Neonatologist 6(11.8) 16(31.4) 29(56.9)Neonatalnurse 38(15.3) 61(24.6) 149(60.1)

Ireceivesufficientpsychologicalsupportfrommydepartmentafterapatienthasdiedinourdepartmenta

Neonatologist 13(25.5) 17(33.3) 21(41.2)Neonatalnurse 85(34) 40(16) 125(50)

AllitemsweretranslatedbyalanguageeditorOneneonatologisthadmissingsonallpsychologicalsupportitemsandwasthusexcludedfromanalysis.aNomissingvaluesinnursesb0.4%missingvaluesinnursesc0.8%missingvaluesinnursesd1.2%missingvaluesinnursesTable6.2:proportionofneonatologistsandneonatalnursesagreeingwithpsychologicalsupportitems

6.4 Discussion

In this survey study concerning stress and perceived psychological support by colleagues orprofessionals during the neonatal end-of-life decision-making process, we found that bothneonatologistsandneonatalnursesworkinginaFlemishNICUexperiencemorestressthanusualwhendealingwithELDs.Eventhoughalmostallfeelsupportedbycolleagues,onlyabouthalffeelthatthepsychologicalsupporttheyreceiveissufficient.Lastly,wecouldnotidentifyasubgroupbasedondemographiccharacteristicsthathadahigherneedforpsychologicalsupportwithinourpopulation.

MostneonatologistsandnursesreportedhavingmorestressthanusualwhentheymakeorareconfrontedwithanELD.Theygenerallyfeltthattheycantalktotheirpeerswhensomethingisbothering them regarding an ELD. However, this support from colleagues does not seemsufficient.Ourfindingsshowthatother,professional,supportisoftenlackingsinceabout60%ofneonatologistsandnurseswouldliketheirdepartmenttoprovidemorepsychologicalsupport

117

whentheyareconfrontedwithanELD,andonlytwooutoffiveneonatologistsandhalfofnursesfeelthattheyreceivesufficientpsychologicalsupportfromtheirdepartmentwhenoneoftheirpatientsdies.Aswedidnotspecifywhichpsychologicalsupporttheparticipantswouldliketoreceive or which support they are currently lacking, we consulted available studies andrecommendationsonvaryingtypesofpsychologicalsupportinaNICUsuchasdebriefingsandcounselling sessions. However, future studies should inquire about the specific nature andcontent of the psychological support that is currently lacking for Flemish neonatologists andneonatalnurses.Existingguidelinesonneonatalend-of-lifeandpalliativecarealreadyprovidesuggestions forstaffsupport,namelyregulardebriefingsandcounsellingsessions inorder toprevent and counteract the negative consequences of stress3. This could not only benefit thepersonalandprofessionallivesofstaffbypreventingburnoutandcompassionfatigue6,butmightalsoimprovetheirabilitytocarefor,andshowempathytowards,bothneonatesandparents12,thusimprovingthecareandsupporttheyprovide13.

Sinceonly45%ofnursesfeltthatthetreatingphysicianslistentotheiropinionregardingELDsand only 32% felt they can express any objections theymight have, our study indicates thatnursesoftenfeelexcludedfromthedecision-makingprocess.Webelievethatincludingnursescouldincreasethequalityofthesedecisions,becausetheyoftenhavemoreinteractionwiththeinfantandfamilythanphysiciansdo,andarethereforemorefamiliarwiththeirwishesregardingthecareanddeathofthechild1,14.AnotherstudyindicatedthathigherlevelsofstressinnursescomparedwithphysicianscouldpossiblybeduetothemhavinglessimpactonELDs15.WethushypothesisethatincludingnursesininterdisciplinaryELDteammeetingscouldpossiblybenefitthenursesthemselvesbyreducingmoraldistresscausedbybeingexcludedfromthedecision-making.

6.4.1 Limitationsofthestudy

OurstudycontactedallneonatologistsandneonatalnursesworkinginallFlemishNICUs,whichis a strength. However, only about 50% completed our questionnaire and we do not havedemographicinformationaboutthosewhodidnotparticipate,ortheirreasonsfornotdoingso.Usinga5-pointLikert-scaleleavesthemotivationoftheparticipantstoanswerinthatmanneropen to interpretation, causing evaluations on the reason behind the lack of support to behypothetical.Duetoethicalconsiderations,wewereunabletoidentifytheNICUsinwhichtherespondentsworked and are thus not able to identifywhich do or do not provide adequatesupporttotheirstaff.Lastly,wedidnotexaminewhetherdifferenttypesofend-of-lifedecisionssuch as non-treatment decisions or drug administration with or without an explicit life-shortening intention are associated with different perceived stress levels or needs ofpsychologicalsupport.Wethereforerecommendfutureresearchtoexaminewhetherdifferenttypesof end-of-lifedecisionsbring forthdifferences in stress levels andwhetherornot theywarrantdifferentmeansofpsychologicalsupport.

6.5 Acknowledgements

Wewouldliketothankallneonatologists,nursesandNICUwardsthatparticipatedinthisstudy,aswellastheneonatologistsandnurseswhoaidedintestingandvalidatingthequestionnaire.

118

Furthermore,wewouldliketothankSaskiaBaesforherhelpindevelopingthequestionnaireandcollectingthedata.Lastly,wewouldliketothankRoosColmanforherstatisticalexpertise,Helen White for translating the items used in our questionnaire and Jane Ruthven for herlanguageediting.

119

6.6 Referencelistchapter6

1. EpsteinEG.End-of-lifeexperiencesofnursesandphysiciansinthenewbornintensivecareunit.JPerinatol.2008;28:771-778.

2. EpsteinEG.Moralobligationsofnursesandphysiciansinneonatalend-of-lifecare.NursEthics.2010;17(5):577-589.

3. CatlinA,CarterB.CreationofaNeonatalEnd-of-LifePalliativeCareProtocol.JPerinatol.2002;22:184-195.

4. WilliamsC,MunsonD, Zupancic J,KirpalaniH. Supportingbereavedparents :practicalsteps in providing compassionate perinatal and neonatal end-of-life care e A NorthAmericanperspective*.SeminFetalNeonatalMed.2008;13:335-340.

5. GoldKJ.Navigatingcareafterababydies :asystematicreviewofparentexperienceswithhealthproviders.JPerinatol.2007;27:230-237.

6. HallSL,CrossJ,SelixNW,etal.RecommendationsforenhancingpsychosocialsupportofNICUparentsthroughstaffeducationandsupport.JPerinatol.2015;35(S1):S29-S36.

7. NationalAssociationofPerinatalSocialWorkers.StandardsforSocialWorkServicesintheNewbornIntensiveCareUnit.

8. Lee KJ, Dupree CY, PhD, Fellow CBHD. Staff Experienceswith End-of-Life Care in thePediatricIntensiveCareUnit.JPalliatMed.2008;11(7):986-990.

9. MeltzerLS,HuckabayLM.CriticalCareNurses’PerceptionsofFutileCareandItsEffectonBurnout.AmJCritCare.2004;13(3):202-207.

10. Mol MMC Van, Kompanje EJO, Benoit DD, Bakker J, Nijkamp MD. The Prevalence ofCompassionFatigueandBurnoutamongHealthcareProfessionalsinIntensiveCareUnits :ASystematicReview.PLoSOne.2015:1-22.

11. WeintraubAS,GeithnerEM,StroustrupA,WaldmanED.Compassionfatigue,burnoutandcompassionsatisfactioninneonatologistsintheUS.NatPublGr.2016;36(11):1021-1026.

12. BellieniCV.,RighettiP,CiampaR,LacoponiF,CovielloC,BuonocoreG.Assessingburnoutamongneonatologists.JMaternNeonatalMed.2012;25(10):2130-2134.

13. ProfitJ,SharekPJ,AmspokerAB,etal.BurnoutintheNICUsettinganditsrelationtosafetyculture.BMJQualSaf.2014;23:806-813.

14. Kavanaugh K, Moro TT, Savage TA. How nurses assist parents regarding life supportdecisions for extremely premature infants. JOGNN - J Obstet Gynecol Neonatal Nurs.2010;39(2):147-158.

15. HamricAB,Blackhall LJ.Nurse-Physicianperspectiveson the careof dyingpatients inintensive care units: collaboration, moral distress, and ethical climate. Crit Care Med.2007;35(2):422-429.

120

121

Part4Generaldiscussionandconclusion

Chapter7:Generaldiscussionandconclusion

122

Generaldiscussionand

conclusion

123

7.1 Introduction

Theoverarchingaimofthisdissertationwastoexamineend-of-lifedecision-makinginstillbirths,neonatesandinfantsonapopulationlevel,acrosscentresandclinicians.Firstly,wedevelopedamethodology,namelythemortalityfollow-backsurvey,thatcouldbeusedtoreliablystudytheprevalenceofvariousprenatalandneonatalend-of-lifedecisionsbasedontheregisterofdeathcertificates.Using thismethodology,wealso compared theprevalenceofneonatal end-of-lifedecisions in Flanders from 1999-2000 with 2016-2017. In part two of this dissertation, weexploredindepthhowhealthcareprovidersinneonatalintensivecareunitsfeeltowardstheseend-of-lifedecisions,andhowneonatalend-of-lifedecisionsareexperiencedindailypracticebymeansofanattitudesurveyandface-to-faceinterviews.Hereby,wefocussedontheattitudesofneonatologistsandneonatalnursesconcerningprenatalandneonatalend-of-lifedecisions,thefactorsinvolvedindecision-makingthatcanmakeend-of-lifedecisionseasierormoredifficultforthem,andtheamountofperceivedpsychologicalsupporttheyreceiveduringthisdifficultend-of-lifedecision-makingprocess.

Inthisdiscussionsection,themainfindingsoftheincludedstudiesarediscussed.First,abriefsummary of the main findings of the dissertation is given, followed by a discussion of themethodologicalstrengthsandlimitationsoftheincludedstudy-designs.Next,ourfindingswillbediscussed and reflected upon in-depth in relation to current research and clinical practice.Additionally,we formulate anumber of implications and recommendations for future policy,practice,andresearch.Finally,anoverallconclusionofthisdissertationwillbeformulated.

7.2 Summaryofthemainfindings

The main findings in relation to each of the research questions of this dissertation aresummarizedbelow.

7.2.1 Examining end-of-life decisions in stillbirths, neonates and infants inFlanders,Belgiumonapopulationlevel

7.2.1.1 Developingamethodologytostudytheprevalenceofend-of-lifedecisionsbeforeandafterbirth

In chapter 2we present a study design aimed to evaluate andmonitor the prevalence ofprenatalandneonatalend-of-lifedecisionsonapopulationlevelinFlanders,Belgium.Thisstudydesign involved thedevelopmentof a validated conceptual frameworkof end-of-lifedecisions across the entire foetal-infantile period and the development of a surveymethodology tostudy these foetal-infantileend-of-lifedecisions independentof thesettingwithinwhichthedeathorstillbirthtookplace.

Wecreatedanew,all-encompassingframeworktoconceptualizeend-of-lifedecisionsintheentirefoetal-infantileperiod,includingbothdeathsbeforebirthfromaviableageofthefoetusonwards(from22weeksofgestationorabirthweightof500gramormore1)andliveborn

124

neonates who died before the age of one year. Two dimensions were deemed important,namelythemedico-technicaldimensionthatclassifiedthemedicalactthatwasperformed,and the medico-ethical classification that classified the life-shortening intention of thephysicianassociatedwiththatmedicalact.Intermsofmedicalacts,adistinctionwasmadebetween non-treatment decisions such as withholding or withdrawing life-supportingtreatment, and administering drugs or performing active medical interventions with apossiblelife-shorteningeffect.Thelife-shorteningintentionofthephysicianontheotherhandcouldeitherbe:1)nointentiontoshortenlife,yetapotentiallylife-shorteningeffectwastakenintoaccount,2)apotentiallylife-shorteningeffectwaspartlyintended,yetnotthemainaimofthemedicalact,and3)thelife-shorteningintentionwasexplicit.Basedonthisframework,we developed two separate but similar questionnaires in order to examine end-of-lifedecisionsinstillbornsandneonatesrespectively.

Ouraimwastostudyend-of-lifedecisionsinlive-borninfantsandstillbornsfrom22weeksofgestationonwardsonapopulationlevel.Basedonpreviousexperienceinneonates2,children3andadults4,5,usingdeathcertificatesasthebasisforsendingoutquestionnaireswasdeemedideal. For stillbirths between 22 and 26 weeks of gestation, this procedure proved to bechallengingasadeathcertificateisnotmandatoryunder26weeksofgestationandthusoursamplingframeworkbymeansofdeathcertificatescouldpotentiallybeincomplete.However,usingtheonlyotherregistryofallstillbirths,namelythebirthregistry(livebornandstillborn)oftheStudycentreforPerinatalEpidemiology(SPE),woulddrasticallydecreasethereliabilityof our responses since delays in processing these documents could take up to one year.Therefore, we chose to rely on the robust mortality follow back survey-method for bothdeceased neonates and stillbirths, with some minor adjustments to improve coverage ofstillbirthsinthecrucialperiodbetween22and26weeksofgestation.TheFlemishAgencyofCareandHealth,whichprocessesalldeathcertificates, startedencouragingregistrationofstillbirthsfrom22weeksonwardsforepidemiologicalpurposesduringthedata-collectionofour study. In addition to this method, we provided our questionnaires to the ten largestmaternitywards inFlanders so thatphysicianswere encouraged to fill themout for eachstillbirth from22weeksof gestationonwards, inaddition to fillingout the accompanyingdeathcertificate.

Physiciansfilledoutthemainpartofadeathcertificateforeveryneonataldeathorstillbirth,whichincludeddemographicandmedicalinformation.Afterwards,thecentraladministrationauthorities, in our case the Flemish Agency of Care and Health, received the filled-outcertificates.TheAgencywasresponsibleforsendingoutquestionnairesandaccompanyingletterswithpatientinformationtophysiciansforeachdeathcertificatedenotingthedeathofaneonateorastillbirth from22weeksofgestationonwards.Thephysician identified theinfant, according to the information on the accompanying letter, and filled out thequestionnaire. All filled-out questionnaires were sent to a lawyer, who was bound byconfidentiality and thus safeguarded the anonymity of the physician, patient, parents andhospitals.Afterdatacollectionwasfinished,thelawyerlinkeddatafromthequestionnaireswithinformationonthedeathcertificates.

Thedevelopedresearchprotocol is the firsttostudyend-of-lifedecisions instillbornsanddeceasedneonatesand infantsunder theageofoneyearonapopulation levelwithinone

125

studydesign.Weareconvincedthatregularrepetitionofthisstudyinthefutureisneededinordertomonitorandevaluatechangesinend-of-lifepracticesundereverchangingsocietal,legalandclinicalinfluencesinavulnerablegroupoffoetusesandinfantswhoareunabletospeak for themselves. By basing inclusion of all deaths and all stillbirths on the deathcertificates,thisresearchmethodcanbeusedinothercountries,irrespectiveofdifferentlegalframeworks regarding perinatal end-of-life decision-making, making internationalcomparativestudiespossible.Providingtheseprevalenceestimates,notonlyinFlanders,caneventuallyaidthedevelopmentofobstetric,neonatalandpaediatricguidelinestosupportaverydifficultethicalend-of-lifedecision-makingprocessindailypractice.

7.2.1.2 Theprevalenceofend-of-lifedecisionsintheneonatalperiod

Chapter3ofthisdissertationfocusedonprovidingpopulationestimatesoftheprevalenceofend-of-lifedecisionsinneonatesandinfantsinFlandersovertwostudyperiods(1999-2000and2016-2017).Theseestimateswereexaminedbymeansofthedevelopedpopulation-levelmortalityfollow-backsurveywedescribedinthepreviousparagraphs.

Atotalnumberof276neonatesandinfantsdiedbetweenSeptember1st2016andDecember31st2017(229filled-outquestionnairesreceived,83%responserate);and292neonatesandinfants died between August 1st 1999 and July 31st 2000 (253 filled-out questionnairesreceived,87%responserate).Studyresultsshowedthattheprevalenceofneonatalend-of-lifedecisionshasstayedrelativelystableacrossbothtime-pointsatabout60%ofneonatalandinfantdeathsbeingprecededbyanend-of-lifedecision.Non-treatmentdecisionsarestillthemostprevalentat34%ofallneonatalandinfantdeathsin1999-2000,comparedto37%in2016-2017.Withholdingtreatmentoccurred in13%ofallneonataland infantdeathsin1999-2000and12%in2016-2017,whilewithdrawing treatmentwasprevalent in21%ofcasesin1999-2000and25%in2016-2017.Administeringmedicationwithapotentiallylife-shorteningeffectstayedrelativelystableat16%in1999-2000comparedto14%in2016-2017, while the prevalence of administering medication with an explicit life-shorteningintentionoccurred inasimilargroupof7%in1999-2000and10%in2016-2017.Despitestableprevalenceratesoverall,importantshiftsinthetypeofend-of-lifedecisionsbeingmadeindifferentagegroupswerenoted.End-of-lifedecisionswerenowsignificantlymoreoftentakenafterthefirstweekoflife(74%ofdeathsbetween7and27daysoldwasprecededbyanend-of-lifedecisionin2016-2017comparedto50%in1999-2000,p=0.03;64%ofdeathsafter 27 days of life in 2016-2017 compared to 38% in 1999-2000, p=0.003). In deathsoccurring in the firstweekof life, prevalenceof end-of-lifedecisions significantlydropped(55%ofdeathsin2016-2017comparedto72%in1999-2000,p=0.01).Afterthefirstweekoflife, end-of-life practice in Flanders considerably changed compared to 17 years ago, asdecisionstowithdrawlife-sustainingtreatmentoradministermedicationwithanexplicitlife-shorteningintentionbecomenoticeablymoreprevalent.In1999-20009%ofdeathsbetween7and27dayswasprecededbyadecisiontowithdrawtreatmentandtherewerenocaseswheremedicationwithanexplicitlife-shorteningintentionwasadministered,whilein2016-2017 in the same age group, withdrawing treatment and administering medication withexplicitlife-shorteningintentionwereeachprevalentin26%ofcases.Afterthefirst27daysoflife,theprevalenceofwithdrawingtreatmentrosefrom16%in1999-2000to31%in2016-

126

2017,andtheprevalenceofadministeringmedicationwithexplicitlife-shorteningintentionrosefrom2%to10%.

This chapter shows that end-of-life decisions continue to be an integral part of medicalpracticeinextremelyillneonatesandinfants,withthreeinfivedeathsbeingprecededbysuchdecisions,whichindicatestheneedtodiscusstheirpermissibilityandrequirementsforgoodclinicalpracticeamongsthealthcareproviders.

7.2.2 Attitudes, views and experiences of healthcare providers involved inneonatalend-of-lifedecision-making

7.2.2.1 Attitudesofneonatologistsandneonatalnursesconcerningperinatalend-of-lifedecisions

In chapter4we present the attitudes of neonatologists and neonatal nursesworking in aneonatalintensivecareunittowardsperinatalend-of-lifedecisions,examinedbymeansofafull-populationmailsurvey.

We found that overall, considerable support for both prenatal and neonatal end-of-lifedecisionscouldbenotedamongstFlemishneonatalhealthcareproviders.Intermsofprenatalend-of-life decisions, between 80 and 98% of neonatologists and nurses consideredtermination of pregnancy at a viable term acceptable in case of severe or lethal foetalanomalies.Whenthefoetusishealthy,yetthelifeofthemotherisindanger,morethan60%of neonatologists and nurses found termination of pregnancy at a viable term acceptable.However,when the foetus is healthy but themother has a severe psychological problem,acceptance rates drop to 15% in both physicians and nurses. In extremely ill livebornneonates,between80and100%ofallparticipatinghealthcareprovidersfoundnon-treatmentdecisionssuchaswithholdingorwithdrawingtreatmentacceptable,regardlessofwhetherthelife-shorteningeffectwassolelytakenintoaccountorexplicitlyintended.Asidefromgeneralconsensus between neonatologists and neonatal nurses on the abovementioned types ofprenatal and neonatal end-of-life decisions, some differences in attitudes between bothhealthcare providers could be noted. Administering medication with a potentially life-shorteningeffectwasconsideredacceptablebythemajorityofbothhealthcareproviders,yetneonatologistsweresignificantlymorelikelytoagreetothispractice(96%)thannursesdid(84%,p=0.02).Conversely,thoughmorethanhalfofbothhealthcareprovidersfoundactivelyadministeringmedicationwithanexplicit life-shortening intentionacceptable, thepracticewasmoreoftenconsideredacceptablebynurses(74%)thanbyneonatologists(60%,p=0.02).

Ourstudythusfoundalargeacceptanceofbothprenatalandneonatalend-of-lifedecisionsinneonatologistsandneonatalnurses,evenfordecisionsthatcurrentlyfalloutsidetheBelgianlegal framework. However, physicians and nurses differed slightly in their acceptance ofdifferenttypesofend-of-lifedecisions,whichcouldpossiblyberelatedtonursesnotcarryingthefinallegalresponsibilityofthesemedicaldecisions.Thesefindingsindicatetheimportanceofincludingbothperspectivesinthesedifficultdecisionsattheendofaninfant’slife.

127

7.2.2.2 Barriersandfacilitators forneonatologistsandneonatalnursesregardingneonatalend-of-lifedecision-making

Inchapter5,weexploredbarriersandfacilitatingfactorsexperiencedbyneonatologistsandneonatalnursesduringtheend-of-lifedecision-makingprocessinaneonatalintensivecareunit.Hereby,weaimedtoprovideinsightonthecomplexityofneonatalend-of-lifedecisionsindailypractice,andtheindividualnatureofpersonalexperiencesonthistopic.

Somebarriersandfacilitatorsarelinkedwiththecharacteristicsofthespecificcase.Thesefactors relate to either the ill neonate, the parents or the involved healthcare providers.Decisions seemedeasierwhen thebadprognosiswas evident fairlyquicklyas opposed towhenthereisalotofprognosticinsecurity,andexploringallcurativetreatmentoptionsfirsttoensurethattheend-of-lifedecisionistheonlyavailableoptionlefttoreducesufferingofthechildhelpedmakedecisionseasier.Healthcareprovidersindicateaneasierdecision-makingprocess when parents have the same culture and language as the physicians and nursesinvolved.Previousexperienceofhealthcareproviderswithend-of-lifedecisionsisconsideredacrucialinfluencingfactor,astheyarebetterabletoanticipatethechild’sfuturecondition.

Onaprocesslevel,weconsiderfactorsthatarerelatedtocharacteristicsofthedecision-makingprocessitself.Intensecommunicationbetweenhealthcareprovidersandparentsisimperative for an easier end-of-lifedecision-makingprocess. Furthermore, communicationamongsthealthcareprovidersisessential,forexamplebyplanningregularmultidisciplinaryconsultationsordebriefings.Additionally, decidingon an end-of-lifedecision canbemadeeasierbyconsideringalldirectionsthechild’sconditioncantakeinadvanceduringoneormore advance care planning conversations between parents and healthcare providers.Hereby, decisions regarding themedical responses in each situation canbemadewithoutbeingrushedbyanacutedeteriorationofthechild.

A final level includes factors relating to the overarching structure of the ward, thehospital and the broader society that could influence decision-making. Emotional andpracticalsupportfromcolleaguesattheward,orlackthereof,iscrucialinend-of-lifedecision-makinginneonates.Additionally,thelackofseparateroomstoensureprivacyduringbad-newsconversations,andtheshortageofavailabletrainedpersonnelinend-of-lifecarewereclearly identified as barriers for end-of-life decision-making. Lastly, the current Belgianlegislation was mentioned as an influencing factor. When mentioned, neonatologists andnursesstatedthattheyexperiencethelackofalegalframeworktoallowforactivelyendingthe life of aneonate in extreme cases tobe an importantbarrier, especially in contrast toduring the pregnancy, where the option to terminate as soon as a life-limiting foetalabnormalityisdiagnosedisavailable.

Ourqualitativeinterviewstudyrevealedbarriersandfacilitatorsduringneonatalend-of-lifedecision-makingwhich could lead to recommendations for improving thisprocess indailypractice.Theserecommendationsincludeestablishingregularmultidisciplinarymeetingstoincludeallhealthcareprovidersandreduceunnecessaryuncertaintyregardingtheprognosisand the best possible course of action, routinely implementing advance care planning inseverelyillneonatestomakeimportantdecisionsbeforehand,creatingprivacyforbad-news

128

conversationswithparentsandreviewingthecomplexlegalframeworkofperinatalend-of-lifedecision-making.

7.2.2.3 Psychologicalsupportinend-of-lifedecision-makingforneonatologistsandneonatalnurses

Chapter 6 of this dissertation focussed on the perceived stress that neonatologists andneonatalnursesexperienceduringanend-of-lifedecision-makingprocessintheirneonatalintensivecareunit,andtheirperceivedpsychologicalsupportbothfromcolleaguesandfromprofessionals.Thiswasexaminedbymeansofafull-populationmailsurvey.

The majority of neonatologists and nurses agreed that making an end-of-life decision(neonatologists)orbeingconfrontedbyone(nurses)causedmorestressthanusual(73%and70%respectively).Duringthedecision-makingprocessfortheseend-of-lifedecisions,mostphysicians(86%)indicatedthattheyfeltsupportedbytheircolleagues.However,fewerthanhalfoftheneonatalnurses(45%)agreedthatthephysicianslistenedtotheiropinionswhenthesedecisionswerebeingmade.Whilemostneonatologists(88%)agreedthattheirneonatalintensive careunit provides sufficient opportunity to express any reservations theymighthave about certain end-of-life decisions, only 32% of nurses agreed with this statement.Almostalloftheparticipatingneonatologistsandneonatalnursesagreedthattheycantalktotheircolleagueswhensomething isbotheringthemregardinganend-of-lifedecision(94%and92%respectively).Furthermore,whentheydidnotagreewithanend-of-lifedecisionthathadbeenmade,halfofneonatologists(53%)and65%ofnursesagreedthattheycouldopttonolongerbeinvolvedinthatparticularcase.Despitethefactthatbothgroupsofhealthcareprovidersindicatedthattheycouldtalktotheircolleagueswhensomethingregardingend-of-life decision-making bothered them, 57% of neonatologists and 60% of neonatal nursesindicated that they would prefer their neonatal intensive care unit to provide morepsychologicalsupport forstaffmemberswhentheywerebeingconfrontedwithend-of-lifedecisions. Furthermore, only 41% of neonatologists and 50% of nurses agreed that theyreceivedsufficientpsychologicalsupportfromtheirneonatalintensivecareunitafteroneoftheirpatientsdied.

Our findings seem to suggest that neonatal intensive care units need professional ad hoccounselling or standard debriefings, as we believe they could substantially improve theperceivedlackofsupportindicatedbycliniciansworkingattheNICU.Furthermore,webelievethatincludingnursesininterdisciplinaryend-of-lifediscussionscouldnotonlyincreasethequalityofthesedecisions,butcouldpossiblyalsobenefitthenursesthemselvesbyreducingmoraldistresscausedbybeingexcludedfromthedecision-making.

7.3 Methodologicalconsiderations,strengthsandlimitations

Weusedthreedifferenttypesofstudymethodologiesinordertoanswerourresearchquestions.Firstly, a population-basedmortality follow-back studywas performed, aiming to include allstillbirthsfrom22weeksofgestationandonwards,andalldeathsofinfantsundertheageofoneyearduringasetperiod(chapter2and3ofthisdissertation).Secondly,afullpopulationmail

129

survey was performed to study attitudes on prenatal and neonatal end-of-life decisions ofhealthcareproviders(chapter4),andtheamountofpsychologicalsupporttheyreceived(chapter6).Thirdly,weperformedaqualitativestudytodeterminethebarriersandfacilitatingfactorsthat neonatologists and neonatal nurses experience during the neonatal end-of-life decision-makingprocess (chapter5).We summarized themethodsusedbelowandelaboratedon themethodologicalstrengthsandchallengesofthestudydesigns.

7.3.1 General methodological considerations, strengths and limitations of thedissertation

Ingeneral,themainstrengthofthisdissertationistheuseofamixedmethodology,combiningstrongquantitativedataonneonatalend-of-lifedecisionsandattitudesofhealthcareproviderson a population level, with elaborate and in-depth qualitative data on the viewpoints andexperiencesofthosesamehealthcareproviders.Thismulti-methodapproachprovidesuswithhighlydetailedpersonalexperiencescombinedwithrobustprevalenceestimates,givingamorecompleteoverviewofdailypracticethanwouldbeachievedbyasinglemethodology.Anotherimportantmethodological strength is the supportof allFlemishneonatal intensive careunitsduring the courseof thisdissertation.Becauseof this support, experts inneonatal carewereinvolvedineverystepofthepresentedstudies:fromthedevelopmentphasewheretheirinputwas crucial to ensure content validity of our questionnaires and topic guides; to the data-collectionphasewherethesupportoftheneonatalintensivecareunitsaidedinensuringhighresponseratessothatconclusionscouldbegeneralizedacrosstheentirepopulationofdeceasedneonatesand involvedhealthcareproviders;and finally to thediscussionphasewhereyearlyconsortium meetings with representatives from all Flemish neonatal intensive care unitsprovided much needed clinical feedback on the implications of the collected data. Theirinvolvementthroughoutthisentiredissertation,acrossallstudies,ensuredthathypothesesandimplicationsweregroundedindailypractice,andthatconclusionsweresupportedbyexperts,makingthemrelevantforclinicalpracticebothnationallyandinternationally.

A general limitationof thisdissertation is that all included studies focuson the viewpoint ofhealthcare providers. Input from (bereaved) parents is missing from the narrative of thisdissertation. The choice to focus solely on healthcare providers in this dissertation wasdeliberate, as their viewpointwasboth crucial and sufficient to answerour aims.Within theattitudesurveyandthequalitativestudy,physiciansandnurseswereincludedbecausetheyhaveamultitudeofexperiencesandexpertise,makingthemideallyqualifiedtoevaluateend-of-lifedecisionsindependentofaspecificcase.Inthemortalityfollowbacksurvey,theonlypossiblepoint of contact was the certifying physician, which was ideal as physicians are ultimatelyresponsibleforthemedicalcareprovidedattheendofthatpatients’life,andtheyarethusideallypositionedtoreportontheintentofthedecisionsmadeandtheimpactofthesemedicaldecisionsonlifeexpectancyofthechild.Parentalviewsareacrucialperspectivethatdeservesitsownin-depthstudy.Weincludeditinaseparatequalitativestudythatfallsoutsideofthescopeofthisdissertation.

130

7.3.2 Themortalityfollow-backsurvey

Themethodofthemortalityfollow-backsurveybasedondeathcertificateshasproventoresultin reliable estimations of incidence rates of end-of-life decisions in neonates2, minors6 andadults4,7.Byusingthisthoroughlytestedandrobustmethod,wewerenotonlyabletoprovidereliableestimatesofend-of-lifedecisionsinneonatesandinfants,butwewerefurthermoreabletocomparetheseincidenceratesacrosstimeandagegroups.Hereby,evolutionsandtrendsinend-of-lifepracticesinBelgiumcouldbeinvestigatedfrombeforeaninfantisbornupuntiladultsdiefromoldage.Additionally,themortalityfollow-backsurveybasedondeathcertificateshasbeen used internationally8,9, allowing for reliable international comparisons. Due to rigorousfollow-upproceduressuchasuseoftheTotalDesignMethod10andregularvisitstoallFlemishneonatal intensive care units and the ten biggest Flemish maternity wards in order tocontinuouslymotivateallphysicianstoparticipateinthestudy,ahighresponserateof83%wasachieved.

7.3.2.1 Usingdeathcertificates

Useofdeathcertificatesoffersamyriadofmethodologicaladvantages.Firstly,wewereabletoincludetheentirepopulationofneonateswhodiedbeforetheageofoneyear,independentof the care settingwithinwhich the death occurred, and the cause of death of the infant.Includingtheentirepopulationofdecedentsprovidesuswithunbiasedandreliableincidencerates,asopposedtomorewidelyusedsingle-centrestudies11–13whichfocusononeormore,oftenhighlyspecialized,hospitalunitswhere incidenceratesof end-of-lifedecisionscouldpossiblybeskewed.Secondly,deathcertificatesincludecontactinformationforthecertifyingphysicianofthatspecificdeathcase,makingiteasytosendquestionnairestothepersonwhoisbestpositionedtoprovideinformationregardingtheend-of-lifedecision-makingprocessofthedeceasedinfant.Thirdly,usingdeathcertificatesasthebasisforsendingquestionnairesallowsus to link informationon theoccurrenceof end-of-lifedecisions from the filled-outquestionnaires to socio-demographical data available on the death certificates. This way,information such as sex, age, gestational age, cause of death and place of death is readilyavailablewithoutincludingadditionalquestionstothequestionnaire,herebyaimingtoavoidincompletequestionnairesornon-responseduetolengthyquestionnaires.

However,usingdeathcertificatesalsopresentssomelimitations.Firstly,delaysinprocessingthedeathcertificatescanreachuptofourmonthsbeforequestionnairescouldbesenttothephysicians4. Therefore, a recall andmemory bias cannot be excluded. Recall bias includesphysiciansnotbeingabletorecalleverydetailwithregardtotheend-of-lifedecision-makingprocessofaninfantwhodiedfourmonthspriortoreceivingthequestionnaire.Memorybiasontheotherhand,includesashiftinthecontentoftherecalledmemorycausingphysicianstorememberfactsdifferently,especiallyinregardtovalue-ladenmemoriessuchasthedeathofaninfant.Boththeissueofrecallbiasandmemorybiaswerepresentinpreviousstudiesusingthemortalityfollowbacksurvey-method2–4.However,similartostudyingend-of-lifedecisionsinminors,wedoexpecttherecallbiastohaveplayedasmallerrolecomparedtostudyingend-of-lifedecisionsinadults14.Thisisbecausethedeathofaninfantorminorisafarmorerareand intense event for involved healthcare providers, leading us to expect that physicianswouldrecallcircumstancesoftheirdeathsmoreclearly.Memorybiasontheotherhandmight

131

beexpectedtobeequallypresent,ifnotmore,thaninadultssincethedeathofaneonateatthebeginningoftheirlifemightbeevenmorevalue-laden.Tomitigaterecallbiasandmemorybias,physicianswereencouragedtoconsultthemedicalfileoftheinfantinquestionwhenfillingoutthequestionnaires.Secondly,thephysicianmentionedonthedeathcertificatewasnotalwaystheattendingphysician.Toovercomethisproblem,physicianswereaskedtopassonthequestionnairetothephysicianwhowasbestsuitedtoanswerquestionsregardingthatparticularpatientintheletteraccompanyingthequestionnaire.Thirdly,contrarytomortalityfollowbacksurveysinadults,inthemajorityofneonataldeathsbeforetheageofoneyear,theinfant was treated by a very select number of physicians specialized in neonatology andpaediatrics.Somephysicianshavethereforealmostcertainlycertifiedmultipledeathswithinourstudypopulation.Since thenumberofquestionnairesperphysicianwasnotlimited inordertoinvestigatetheentirepopulationofinfantdeathsundertheageofoneyear,responderfatiguemayhaveoccurred,whichcouldpossiblyleadtonon-response.However,asallFlemishneonatalintensivecareunitsparticipatedinthestudyandwerethushighlymotivated,andresponserateswerehigh,wedonotexpectthistohaveplayedamajorrole.

7.3.2.2 Thequestionnaires

Some strengths of the mortality follow back survey-method can be related to thequestionnaires used to study prenatal and neonatal end-of-life decisions. Firstly, thequestionnairesusedinthemortalityfollowbacksurvey-methoddescribedinthisdissertationwere developed based on existing and previously validated questionnaires on end-of-lifedecisions in stillborns15,16, neonates2,8, minors6 and adults4,17 ensuring comparability overtime, settingsandage groups. Furthermore, thequestionnairesusedwere rigorouslypilottested and validated in order to ensure content validity. As was the case in previousquestionnaires on end-of-life decisions, our questionnaires used descriptive questions toidentifywhichend-of-lifedecisionsweretakeninthatspecificpatientinsteadofusingloadedconceptssuchaseuthanasia innew-borns,orabortion.As thesetermsareoftensubjecttodifferentinterpretationsandtheycaninciteastrongemotionaland/ormoralreaction,theymayleadtosociallydesirableanswersorevenunwillingnesstoparticipateinthestudy.

7.3.2.3 Ethicalconsiderationsandanonymity

Prenatalandneonatalend-of-lifedecisionsareunderstandablyaverysensitivetopicamonginvolvedhealthcareproviderswithimmenseethical,moralandlegalweightattachedtothepractice.Wecanthereforenotignorethepossibilityofunderreportingcertainneonatalend-of-lifepractices,especially those thatarecurrentlynotconsideredlegalwithintheBelgianlegal framework. Furthermore, due to this sensitivity, social desirability bias cannot beexcluded.To account for this, a thoroughand rigorous anonymityprocedurewasused fordata-collection.A complexmailingprocedurewassetup involvinga sworn-in lawyer as atrusted third partywho acted as intermediary between the Flemish Agency for Care andHealth, the participating physicians and the researchers. The study methodology wasapprovedbytheethicscommitteeoftheUniversityHospitalofGhent,thePrivacyCommission(CBPL),theSectoralCommitteeofSocialSecurityandHealth,andtheNationalCounciloftheOrderofPhysicians.

132

7.3.3 Theattitudeandpsychologicalsupportsurvey

Contrarytomostsurveysinprenatalandneonatalend-of-lifedecisions18,19,ourstudytargetedtheentirepopulationofneonatologistsandneonatalnursesworkinginaneonatalintensivecareunit. Therefore, our survey did not include any selection bias by focusing for example onhealthcare providers within a single centre18,19. Additionally, instead of focusing solely onprenatal or neonatal end-of-life decisions separately within one study, attitudes of neonatalexpertswere examinedonboth end-of-lifedecisionsprenatally andneonatally, as this is thedomain where their expertise might be relevant. Furthermore, we included not only theperspectiveofphysiciansbutalsotheperspectiveofneonatalnurses,whoareanessentialpartof end-of-life decision-making20–22 but are often forgotten in research. Despite the fact thattargetingthefullpopulationofneonatologistsandneonatalnursesworkinginaFlemishneonatalintensive care unit is a strength, only about 50% filled out a questionnaire. We have noinformationaboutthosewhodidnotwish toparticipateandarethereforeunabletoexaminepossibledifferencesbetweentheresponseandnon-responsegroups.Furthermore,theirreasonsfor not participating in the survey are unclear and can therefore range from them beinguninterestedinthetopic,tothemnotfeelingliketheyhadenoughexpertisetoshare,orevensimplybeingabsentduringthetimeofdata-collection,sinceitonlyspannedonemonthintime.

By basing our questionnaire on an existing Flemish attitude questionnaire, and furtherdevelopingourquestionnaireinamultidisciplinaryteamconsistingofsociologists,psychologists,neonatologistsandagynaecologist,weensuredthatthequestionswerebothrelevanttoclinicalpractice and usable within a research context. Furthermore, the final questionnaire wascognitively testedwithinour targetpopulation, by interviewing fiveneonatologists from fourseparatehospitals,threeneonatalnursesfromtwoseparatehospitalsandonegynaecologistinordertoensurecontentvalidityoftheitems.Thisquestionnairecanthusalsobeusedtoexamineattitudesofhealthcareproviders,andtheirperceivedpsychosocialsupportinothercountriesorsettings,makinginternationalcomparisonsandcomparisonsacrosssettingspossible.However,questionnairesonlyallowalimitedpotentialtofullycapturethecomplexityofexperiencesandattitudes regarding prenatal and neonatal end-of-life decisions. By structuring answers ofhealthcareprovidersonafourorfive-pointLikertscale,wewereabletoensurestrongoptionsforcomparabilityoverhealthcareproviders,settingsandevencountries.Yetitfailstoencompassthefullscopeofwhatitmeanstodecideonanend-of-lifedecisionprenatallyorneonatally.Open-ended questions or even interviewswith participants would provide us withmore in-depthinformation,thoughitwouldincreaseworkloadonthehealthcareproviders,whichwouldthusnegativelyimpacttheparticipation-rate.

7.3.4 Theface-to-facesemi-structuredinterviewstudy

Ourinterviewstudywasthefirsttoexaminebarriersandfacilitatingfactorsregardingneonatalend-of-lifedecision-makinginhealthcareprovidersworkinginaneonatalintensivecareunit.Weincludednotonlyphysicianswhoaremostoftenresponsibleformakingtheend-of-lifedecision,butalsotheneonatalnurseswhoareofteninvolvedintheprovisionofthedecidedcaretothechildandthefamily20.Hereby,weincludedviewpointsofbothhealthcareproviderssothateveryfacetof thedecision-makingprocesscouldbeconsidered.By includingparticipants fromfourdifferenthospitalwardsacrossFlanders,weincludedvariabilityatthelevelofthehospitalward.

133

Hospitalsandwardscouldvaryonanumberoffactorsrangingfromtheirpolicyregardingend-of-lifedecision-making,tothediversityofthepatientpopulationwhofrequentsthosehospitalwards,oreventheavailabilityofcertaininfrastructureattheward.Bybroadeninginclusiontofour centres fromdifferent regions inFlandersand includingbothuniversityhospitalsandageneralhospital,insteadofperformingasingle-centrestudy,conclusionsweremorelikelytobeapplicablebroadlyacrosssettings.Webelievethatparentscouldhavecrucialadditionalinsights,whichiswhytheywereincludedinaseparatequalitativestudyonneonatalend-of-lifedecision-makingthatfallsoutsideofthescopeofthisdissertation.

Somemethodologicalconsiderationsarelinkedtothechosenstudydesignofusingface-to-facesemi-structured interviews.Firstly, theuseof individual interviewsallowsparticipants to telltheir story freely without interruption or fear of not being able to speak openly due to thepresenceoftheotherimportantactorsinthecareprocessoftheinfant,whichwouldbethecaseduringforexamplefocusgroups.Secondly,byusingaqualitativeapproach,wewereabletofullycapturethecomplexity,subtletyandindividualityoftheexperiencesofhealthcareprovidersintheneonatal end-of-lifedecision-makingprocess.Thequantitative approachof theother twodata-collection methods used in this dissertation are hereby supplemented by covering theindividualexperiencesandtheethical,moraloremotionalloadthatisoftenassociatedwiththeseexperiences by means of face-to-face interviews. However, this form of open recall, whereparticipantswereinvitedtotalkaboutmemoriesfreely,couldresultinonlyextremelypositiveornegativememoriesbeingrecalled.Whilethisishighlybeneficialtowardsformulatingbarriersandfacilitatingfactorsinrelationtoend-of-lifedecision-making,itmightresultinabiastowardsextremeswhilethemajorityofexperiencesarefarlessextreme,thusreducinggeneralizabilityindaily, regularpractice.Additionally, amemorybias couldoccur, especiallywhen the recalledmemory was highly value-laden, as is the case when remembering the death of an infant.Healthcare providers could in this case remember something different than what actuallyhappened.

Additionally,someconsiderationsarerelatedtothewaythedatawasanalysed.Thematiccontentanalysiswasusedtoextractcodesbymeansofabottom-upapproach,meaningthatnoaprioriframework was used as a basis for analysis and themes were identified as they emergedthroughout the interviews. Hereby, the researchers remained close to the experiences of theparticipants and results are thus an accurate reflection of daily clinical practice.Recommendationsthatfollowed,basedonconclusionsdrawninthestudy,arethereforereadilyapplicable in the daily care of extremely ill neonates. Furthermore, codingwas done by tworesearchersindependently,whichimprovedreliabilityofthecodes.Additionally,allcodesandinterpretationswerediscussedwith experts inneonatal andend-of-life careduringandafterdata-analysis occurred, ensuring that interpretations and recommendations are carried byclinicianswhoareconfrontedwiththesecasesonadailybasis.

Lastly,withinourinterviewstudy,aselectionbiasattheleveloftheparticipatingneonatologistsand nurses cannot be excluded. Healthcare providers who are less open to speaking aboutneonatal end-of-life decisions, or who have a conservative stance regarding these types ofmedicaldecisionsat the endof a child’s life, are less likely tobeheard.Thisbias is not onlyunavoidable,itisalsopreferableasintrinsicallymotivatedparticipantsprovidetherichestsourceofinformationinqualitativestudies.

134

7.4 Discussionofthefindings

Inthissection,themainfindingsofthisdissertationarediscussedin-depthinrelationtoeachother,andinrelationtothecurrentstateoftheart.First,wewillgiveasummarizedoverviewofwhatbeingpartofaneonatalend-of-lifedecision-makingprocesswithintheFlemishhealthcarecontext is like for healthcare providers, including evidence from all studies used in thisdissertation. Secondly,wewill examine these findings in relation to internationally availableevidence. Third, the impact of the specific Belgian legal framework on neonatal end-of-lifedecision-makingwillbeconsidered.Fourth,wewilldiscusstheimportanceofprovidingsupportforthehealthcareprovidersduringthestressfulneonatalend-of-lifedecision-makingprocess.Lastly,wewillreflectontheroleofpalliativecareinneonatalend-of-lifedecisionmaking.

7.4.1 Flemishneonatalend-of-lifedecision-making:prevalence,attitudes,viewsandexperiencesofhealthcareproviders

Infantmortality during the first year of life in Belgium is rather low. To put our prevalenceestimatesonneonatalend-of-lifedecisionsintoperspective,in2016and2017,theaveragebirthratepermonthinFlanderswas538623whileweregisteredatotalnumberof287deathsacross16months(anaverageof 18permonth).Within the small populationofneonatal and infantdecedents, prevalence of end-of-life decisions has remained fairly constant over the last fewdecadesatabout60%.Inthefollowingparagraphswewillextensivelydiscusstheevidenceonnon-treatment decisions and administration of medication with a potential or explicit life-shorteningeffectinFlanders,whichwascollectedacrossallstudieswithinthisdissertation.

7.4.1.1 Non-treatmentdecisionsarethemostcommonandmostacceptedneonatalend-of-lifedecision

MorethanfourinfivehealthcareprovidersworkinginaFlemishneonatalintensivecareunitfindnon-treatmentdecisionssuchaswithholdingorwithdrawinglife-sustaining treatmentacceptableinseverelyillneonatesandinfants,regardlessofwhetherornotthelife-shorteningintentionofwithholdingorwithdrawingwasexplicitratherthanimplicit24.Theseattitudeswere reflected within daily practice, as non-treatment decisions are the most commonneonatalend-of-lifedecisioninFlanderswithaprevalenceof37%ofalldeaths.Withinourqualitative interview study, physicians and nurses also indicated that withholding orwithdrawingtreatmentwhendeemednolongerbeneficialtothechildmakesforalessdifficultdecision-making process than administering medication to end suffering25. The ethicaldifferencebetweenallowingachildtodiebystoppingornotstartingfutiletreatmentandinessencelettingnaturerunitscourse,andactivelyintendingforthechildtodiebymeansofmedication26likelyplayacrucialroleinwhynon-treatmentdecisionsaremoreprevalent.Thisisespeciallythecasewhenprovidingtreatmentcouldpossiblyevencausesufferingforachildthatisalreadydying27,andchoosingtoforgothetreatmenttosparethechildbecomeseasier.

In research on end-of-life decisions, withholding and withdrawing treatment are oftengroupedtogetherundertheumbrellaofnon-treatmentdecisions.Withdrawaloftreatmentisessentially theremovalof intensive therapystarted inanattempt tosustain the lifeof the

135

infant.Withholdingtreatmentontheotherhandindicatesadecisionnottoinitiateanynewtherapeuticinterventions.Ethicistshavedefinedbothpracticesasbeingethicallyequivalent28.TheBritishMedicalAssociationalsostatedthat,althoughitmightpsychologicallybeeasiertowithholdtreatmentthantowithdrawtreatmentthathadalreadybeeninitiated,therearenolegalormoraldifferencesbetweenthetwopractices29.Despitethedifferenceinacceptancebetween the two being small and not statistically significant in our attitude survey,withholding treatment was in fact found acceptable by 90-100% of neonatal healthcareproviderswhilewithdrawingtreatmentshowedslightlyloweracceptancerates(between80and 89%). If acceptance of withholding treatment was found higher than withdrawingtreatment, this might raise the question why the prevalence of withholding treatmentprecedingneonataldeathinourstudyisn’thigherthanthatofwithdrawingtreatment.Thoughresultsof theattitudesurveydid indicatethatdecisionsindailypracticeare influencedbyattitudesofhealthcareproviders,eachclinicalcasehasitsownclinicalcharacteristicssuchaswhether or not treatment was initially started and whether withholding or withdrawingtreatmentisconsideredapossibility.Additionally,itisimportanttokeepinmindthat,withinourstudy,onlythedecisionthatwasdeemedmostimportantwasincludedinanalysisandprevalence estimates. When both withholding and withdrawing treatment were present,withdrawing treatmentprevailed andas such, total prevalence estimatesmightbehigher.Furthermore, in neonates, the prognosis of the child is often uncertain30 and thereforetreatmentisofteninitiallystartedinordertogivethechildthebenefitofthedoubt.Ourdatacorroboratestheseclaims,asintensivetreatmentwasstartedinabout60%ofallcases,andinonlyaboutoneinfiveneonatesthedecisiontoforgoalltypesofintensivetreatmentwasmade(datanot given in chapter3).Despite thepossibility thatwithholding treatmentmightbepsychologically easier forhealthcareproviders, inmost cases treatment is initially startedwith the intention to sustain or save the life of an extremely ill neonate. Data in thisdissertationshowedthat,whentreatmentwasdeemedfutile,itcanbe,andisoften,withdrawnto reduce suffering or because it has become futile, making withdrawal of life-sustainingtreatmentthemostoccurringneonatalend-of-lifedecisioninFlanders.

7.4.1.2 Medication with and without explicit life-shortening intention is also animportantpartofneonatalend-of-lifedecision-makinginFlanders

Aside fromnon-treatmentdecisions, thisdissertation showed that in aboutone in four ofdeceasedneonatesandinfants,themostimportantend-of-lifedecisionprecedingdeathwasadministeringmedicationwith an implicit or explicit life-shortening intention (chapter3).Prevalenceofthesetypesofend-of-lifedecisionsinFlandersisthussimilartothoseinminorsbetweentheageofoneand17yearsold6,yetslightlylowerthantheprevalenceintheadultpopulation(31%)31.

Administration of medication, taking into account the possible life-shorteningintention

Withinthecategoryofusingmedication,wedistinguishmedicationwherethelife-shorteningeffectwastakenintoaccountorco-intended,andadministeringmedicationwithanexplicitlife-shorteningintention.Overthepast17yearstheprevalenceofadministeringmedicationwithoutexplicitlife-shorteningintentionstayedrelativelystableatabout15%ofallneonataldeaths. Interestingly, in adults the prevalence of intensified alleviation of pain and other

136

symptomswithoutexplicitlife-shorteningintentionisfarhigherat24%31,withtheprevalenceinminors being somewhat in between (18%)6. In neonates, assessing pain and symptomburdenremainschallengingastheyareunabletoexpressiteitherverballyornon-verbally32,whichisverydifferentfromtreatingpaininminorsandadults.Measuresforassessingpaininneonates exist, but they arestillmoredifficult to clearly interpret32 thanadults orminorsindicatingtheirsufferingthroughverbalornon-verbalcues.Becauseofthis,whenhasteningdeath is not the main goal of the administered medication, and providing pain and/orsymptommanagementmedicationisthususedtorelievesufferingduringacurativeratherthanapalliativesituation,itisinfinitelymoredifficulttoassesstherightdosesthanisthecaseinminorsoradults.Furthermore,aslife-shorteningisnotthemaingoal,carefulconsiderationtowardsthedosesofforexampleopioids,benzodiazepinesorothersedativesormedicationsto relievepainand sufferingneeds tobe given ashighdoses could lead to anunintendedrespiratoryarrestorcardiovasculardistressinneonates33,34.Physiciansmightinthesecasesbemorehesitanttoincreasethesetypesofmedicationwhenallhopeofsavingtheinfantisnot yet lost, and they are adamant to avoid an unintended life-shortening effect. Thesedifficulties in assessing pain and determining the correct dosages of pain and symptommedicationarespecifictotheneonatalsetting,andthusitispossiblethattheyplayacrucialroleinwhyadministrationofmedicationwithoutanexplicitintentiontohastendeathislessprevalentthaninolderpatientgroups.

Administrationofmedicationwithanexplicitlife-shorteningintention

In2016-2017,administeringmedicationwithanexplicitlife-shorteningeffectoccurredinoneintenneonataldeaths,whichwassimilartotheprevalenceestimateof7%ofallneonatalandinfant deaths preceded by this type of end-of-life decision in 1999-2000 (chapter 3). Ourattitudesurvey(chapter4)revealedthat60%ofFlemishneonatologistsand74%ofneonatalnursesconsideredadministeringmedicationwithanexplicitlife-shorteningintentiontobeacceptable in certain cases of extremely ill infants24. Similarly, in the Walloon region ofBelgium,asurveyshowedthat77%ofneonatologistsworkinginaneonatalintensivecaresettingwouldconsiderperforming‘active’end-of-lifepracticesinthecontextofapalliativecare pathway35. Even though allowing for an infant todie bywithholding orwithdrawingtreatment, and administering medication to explicitly hasten death might morally andethicallybeconsideredequivalentastheendresultisthesame,ourqualitativestudyrevealedthat the decision-making process in these cases is far more difficult than when a non-treatmentdecisionwouldsufficetorelievesufferingbyallowingtheneonateorinfanttodie25.Thedecision-makingprocessisevenmorecomplexaswhetherornotactivelyendinglifeinneonates is legally possible within the Belgian legal framework can be debated (seeintroduction),yetourstudiesshowthatitisanimportantpartofneonatalend-of-lifepractice.

When comparing the practice of administering medication with explicit life-shorteningintentioninneonatesandinfantswitholderpatientgroups,somedifferencescanbenoted.Inadults,5%ofalldeathsareprecededbyeuthanasia(wherethephysicianactivelycarriesoutthe patients request to die), and less than one percent is preceded by physician-assistedsuicide(wherethephysicianmakesthelethalmeansavailabletothepatienttobeusedatatime of the patient's own choosing)31. Euthanasia and physician-assisted suicide are alsopossible in minors with decisional capacity in Flanders, yet the only study reporting onprevalenceratespredatestheadditionofcapableminorstotheeuthanasialawin2014,which

137

reportedaprevalenceestimateof0%6. Inneonatesand infants,euthanasiaandphysician-assistedsuicidearebothimpossibleinourpopulationofneonates,asnewbornsandinfantsdonothaveanydecisionalcapacityandthusadministeringmedicationwithanexplicitlife-shorteningintentionisbydefinitionalwayswithoutexplicitrequestfromthepatient(butnotnecessarily without explicit request from parents). The prevalence of explicitly hasteningdeathbyadministeringmedicationwithoutexplicitpatientrequestinadultsisestimatedat2%31,andinminorsthiswasestimatedat8%ofalldeaths6.Theprevalenceofthistypeofend-of-lifedecisionisthushigherinneonatesandinfants,whichcouldpossiblybeduetothefactthatinneonatesandinfantsdecisionalcapacitycanneverbeachieved,makingadministrationofmedicationwithout explicit patient request the only option to actively hasten death bymedication.Inminorswithdecisionalcapacityandadults,anexplicitrequestbythepatienttohasten death by means of a euthanasia or physician-assisted suicide is possible. Medicalsituationswhereactivelyhasteningdeathbymedicationisneededtorelievesufferingwithoutexplicitpatientrequestisthusfarmorerareinminorsandadultsthaninneonatesandinfants.Additionally,astheadministrationofadequatepainandsymptomcontrolinneonatesissodifficulttoassess32,thelinebetweenprovidingcomfortbycontinuouslyanddeeplysedatingasufferingchilduntildeathandactivelyhasteningtheendoftheirlifemightbecomeblurred.Itcouldthusbepossiblethatphysiciansincludecasesofcontinuousdeepsedationuntildeathwitha“welcomed”shorteningoflifeinneonatesandinfantsunderthiscategory.Inadults,theexistenceof a grey zonebetweenhasteningdeathandadequatepalliative sedation iswellknown17,36.Duetotheinherentdifficultiesinassessingthecorrectdosagetoeasesufferinginneonates,andthefinelinebetweenadequatepainreliefandhasteningdeath,wecanexpectthisgreyzoneandtheaccompanyingdifficultiesinlabellingmedicaldecisionsattheendoflifeisequallyorevenmoreinherentlydifficultinneonates.

7.4.2 Comparing Flemish neonatal end-of-life decision-making withinternationallyavailableevidence

Neonatalmortalityvarieswidelyacrosscountries.While3%ofallneonatesdiewithinthefirst28 days of life in low income countries, this number drops to 1% in upper-middle incomecountriesandeventolessthan0.5%inthehighestincomecountriesworldwide37.In2016inFlanders,neonatalmortalitywithin the first28daysof lifeoccurred in0.23%ofall live-borninfants23.Asidefromthesedifferencesinneonatalmortality,differencesinethicalperspectiveexistbetweencountriesintheacceptabilityanduseofmedicaldecisionsattheendofaninfant’slife37.Itisthereforenecessarytocomparetheinformationgatheredwithinthisdissertationonneonatalend-of-lifedecisionsinFlanders,Belgiumwithinternationallyavailabledatainordertounveilcountry-specificfactorsinfluencingdecision-making.

7.4.2.1 Internationalcomparisonofthepracticeofnon-treatmentdecisions

Inmosthigh-incomecountries,thedifferencesintreatmentpoliciesofextremelyillneonates,especially those born at the limit of viability, are not huge37 yet in the grey zone somedifferencesbetweenforexampleEuropeancountriescanbenoted.WhileinFlandersandtheNetherlandsinfantsbornbefore24weeksofgestationarenottreated37,38,inGermanytheyare,indicatingsmallinternationaldifferencesinwhenaninitialdecisiontoforgotreatmentdue toextremeprematurityismade.Generally though, inEurope,non-treatmentdecisions

138

such as withholding or withdrawing treatment are well accepted39,40, and themajority ofphysiciansworkinginneonatalintensivecarereporthavingbeeninvolvedinatleastonecaseinwhichlimitstointensivecarewereset41.Internationally,thelikelihoodoflimitingintensivetreatment in neonates is known to be dependent on the positive or negative attitude ofphysicians towards these types of end-of-life decisions18. The positive attitude of Flemishneonatalhealthcareproviderstowardsnon-treatmentdecisionsandthecorrespondinghighprevalenceofthesetypesofdecisionsintheentirepopulationofneonataldeathsbeforetheageofoneyearreportedinthisdissertationcorroboratethesefindings.Wecouldthereforehypothesizethatourprevalenceestimatesconcerningnon-treatmentdecisionscouldpossiblybecomparabletothoseofEuropeancountrieswithasimilarlypositivestanceontheseend-of-life decisions such as the UK and the Netherlands18. Physicians working in neonatalhealthcareinEuropeancountriessuchastheBalticstates,Italy,SpainandGermanyhaveastrongerpro-lifeattitude18.WhileFlandersandcountriessuchastheUKandtheNetherlandscouldthusbeconsideredtohaveanacceptingattitudetowardsnon-treatmentdecisionswithapotentiallylife-shorteningeffect,otherEuropeancountriesmightbemorerestrictive.

Theprevalenceofnon-treatmentdecisionsinFlandersin2016-2017was37%ofallneonataldeaths(chapter3).ThisisslightlyhigherthanthatoftheNetherlands,whichwasestimatedat31%in20108.ReportsfromneonatalintensivecarecentresintheUnitedStates,theUnitedKingdom,AustraliaandEuropeshowthatbetween40and93%ofneonataldeathsoccurafterwithholding orwithdrawing artificial ventilation or other life-sustaining treatments12,42–46.The difference between these population-based estimates (37% in our study) and theprevalence estimates of the number of deaths in specializedneonatal intensive care unitsinternationallybeingprecededbyanon-treatmentdecision(between40and93%)canberelatedtoseveralfactors.Firstly,somemethodologicaldifferencesinassessingtheprevalenceofend-of-lifedecisionscouldbenoted.Inthepopulation-basedstudies,thetypeofend-of-lifedecisionwasconsideredmutuallyexclusive,indicatingthatwhenmorethanoneend-of-lifedecision was noted for a death case, the decision with the most explicit life-shorteningintentionwasused.Furthermore,whenmorethanoneend-of-lifedecisionwiththesamelife-shorteningintentionwasnoted,administrationofmedication,asanactiveformoftreatment,prevailedoverwithholdingorwithdrawingtreatmentwhichwasconsideredamorepassiveactofhasteningdeath.Therefore,theprevalenceofallnon-treatmentdecisions,regardlessofwhetherornottheywerenotedincombinationwithdrugsthatcouldpossiblyshortenlifewashigherat56%ofallneonatalandinfantdeaths(datanotgiveninchapter3).Secondly,thedifferencebetweenthesinglecentrestudiesestimatesandthepopulation-basedestimatescanbeduetothespecializedsettingofneonatalintensivecareunits.Decision-makingattheendofaneonate’slifecanbeconsiderablyvariableinthesespecializedsettings,wherehealthcareprovidersareconfrontedwithextremelyillneonatesonadailylevelandhighlyspecializedtechnicalequipmentisreadilyavailable47.Asoneofthemainstrengthsofourmortalityfollowbacksurvey-methodologyistheinclusionofalldeathsregardlessofthetypeofhospitalunitthe infantswere admitted toprior todeath, thedifferences inprevalencemightreflectanactualdifferencebetweenend-of-lifedecision-makinginspecializedlevelthreeintensivecareunitscomparedtoinfantswhoweretreatedinhospitalslocatedintheperiphery,orevenathome.Toexaminethesedifferences,futurestudiesshouldaddressthedifferencesinneonatalend-of-lifedecision-makingwithregardtothecaresettinginwhichtheneonatesareadmitted.Additionally,aspopulation-levelstudiesarescarce,actualinternationalcomparisonsbetween

139

our population estimates and estimates from countries outside of the Netherlands wereimpossible,whichshoulddefinitelyberemediedinfuturestudiesinordertoexaminecountry-specific influences on clinical practice. The study design described in chapter 2 of thisdissertationwouldbeidealforthispurpose.

Estimatesontheprevalenceofnon-treatmentdecisionsusingamyriadofstudydesigns,asdescribedabove,are limited topractices inEurope,Australiaand theUnitedStates,whereclinicalresourcesforneonatesandinfantswithsevereconditionsarepracticallyunlimited48.According to our knowledge there isno evidenceon end-of-lifepractices inneonates andinfantswithinlesswell-developedhealthcaresettings.Oneretrospectivechartreviewstudysuggestedthatend-of-lifedecision-makinginunitswithinlessdevelopedhealthcaresettingswassimilarto thatindevelopedcountries48,yetthisstudyonlyconsideredonehospital inCuraçao, where all paediatricians and residents received training in the Netherlands.Therefore, no reliable conclusions can be drawn without comparing reliable prevalenceestimatesonapopulation-level.TheconsiderabledifferencebetweenthepositiveattitudeofFlemishhealthcareproviderstowardsnon-treatmentdecisionsinextremelyillneonatesandthe conservative attitude towards limiting life-sustaining treatment of for exampleArgentinianneonatologists49suggeststhatinthiscountry,asignificantlydifferentneonatalend-of-life culture exist than that of Flanders. Therefore, future international comparisonstudiesshouldincluderegionswiththesepotentiallydifferingend-of-lifeculturestofurtherexamine which country-specific factors, such as for example an accepting versus non-acceptingattitudeofhealthcareproviderstowardsdifferenttypesofend-of-lifedecisions,ortheexistenceofnationallawsorguidelines,haveimpactonactualprevalencerates.

7.4.2.2 International comparisons of the administration of medication with animplicitorexplicitlife-shorteningintention

The findingswithin thisdissertation indicate thathealthcareproviders inFlandershaveafairly accepting attitude towards more active types of end-of-life decisions such asadministeringmedicationwith a potential or explicit life-shortening intention, evenwhenthese decisions currently fall outside of the Belgian legal framework. In the followingparagraphs,wewillgo intodetail onwhetheror not this accepting climate is comparableinternationally, both in terms of attitudes of healthcare providers, and when looking atinternationalprevalenceestimates.

Weseethatthelife-shorteningintentionofadministeringmedicationbeingeitherimplicitorexplicit makes a crucial difference in whether the Flemish accepting attitude could becorroborated internationally. In Switzerland, 95% of physicians and nurses working in aneonatalintensivecareunitfoundadministeringsedativesoranalgesicsacceptable,evenifthis might cause respiratory depression and death50. However, when the life-shorteningintention of administering medication becomes explicit, acceptance rates of Swissneonatologistsandnursesdropto24%50.InCanada,asurveyontheattitudesofCanadianpaediatricians revealed a collective unease towards non-voluntary euthanasia in never-competentchildren51,52,suggestingthatCanadianpaediatriciansandneonatologistsmightbealotlessacceptingthanthoseinFlanders.InFrance,amultidisciplinaryworkinggrouponethical issues inperinatalmedicine even stated thatacts todeliberatelyhasten apatient’s

140

deatharebothlegallyandmorallyforbidden53,indicatinganevenmorerestrictiveattitude.Flemish neonatal healthcare providers might thus be much more accepting towardsadministeringmedicationwithapotentialorexplicitlife-shorteningintentioninextremelyillneonatesandinfantsthanhealthcareprovidersinothercountries.Inthisregard,theinfluenceoftheBelgianlegalcontextandtheaccompanyingmedicalcultureondeathanddyingshouldbetakenintoaccount.Ashortreflectionontheseinfluencescanbefoundinchapter7.4.3.

Asidefromareflectiononinternationalattitudestowardstheadministrationofmedicationwith a potential or explicit life-shortening intention, international prevalence estimatesshould be considered. A multi-national study (EURONIC) in eight European countries(Belgium not included) revealed that between 32% and 89% of physicians working in aneonatalintensivecareunithadpreviouslyadministeredpainandsymptomrelief,despitetherisk of respiratory depression and even death54. These numbers varied greatly betweencountries, with France, the Netherlands and Sweden reporting the highest number ofphysicianswithpreviousexperienceinadministeringsedativesandanalgesicsevenattheriskofhasteningdeath(86-89%),andItalybeingtheonlyreportedcountrywithratesunder50%(namely32%)54. Furthermore, that study revealed thatadministeringmedicationwith thepurposeofendinglifeinneonatesoccursveryrarelyinthemajorityofreportedEuropeancountries.Only2-4%ofphysiciansworkinginaneonatalintensivecareunitinItaly,Spain,Sweden, Germany and the UK reported ever having taken these types of decisions54.Interestingly,theonlyoutliersinthiscasewereFrance,where73%ofphysiciansindicatedthattheypreviouslyadministeredmedicationwithanexplicitlife-shorteningintention,andtheNetherlands,with reported ratesof 47%54.This indicates thatBelgium’sneighbouringcountries (FranceandtheNetherlands)mighthaveasimilarunderlyingcultureofmedicalpractice regarding these more active forms of hastening death in neonates than that ofFlanders.DatafromtheEURONICstudyhowever,datesbacktotheearly2000sindicatingthatmedicalpracticecouldpossiblyhavechangedsincethen.InthecaseofFranceforexample,morerecentdatashowedthatacceptanceofhealthcareproviderstowardsactiveterminationoflifeinneonatesdroppedfrom73%to39%55.Inarecentfollow-upstudyoftheEURONICstudyin2016inGermany,SwitzerlandandAustria56,97%ofphysiciansreportedinanonlinesurvey thattheyhaveadministeredsedativesandanalgesicsevenattheriskofpotentiallyhasteningdeathatleastonce.Whenconsideringtheadministrationofdrugswithanexplicitlife-shortening intention however, the proportion of physicians ever having made thisdecision in Germany, Switzerland and Austria drops to 4%56. We can conclude thatinternationally,somevariabilityinprevalenceestimatesofadministeringmedicationwithapotentialorexplicitlife-shorteningintentioncanbenoted.

IntheEURONICstudies,physicianswereaskedwhetherornottheyhadeverpreviouslytakenspecific typesofneonatalend-of-lifedecisionsand thereforewehaveno indicationofhowfrequent thesedecisionsactuallyare.Providingaclearcomparisonbetweendata fromtheEURONICstudiesandourprevalenceestimatesofadministeringmedicationwithapotentialor explicit life-shortening intentionwithin an entire population of deceased neonates andinfants during a set period of time therefore proves to be difficult. The only availableprevalenceestimatesregardingtheuseofmedicationtoimplicitlyorexplicitlyhastendeathin neonates is from the Netherlands8, indicating that medical practice might not be ascomparable toFlanders aspreviously suspected.Where ourprevalence estimates indicate

141

that 14% of all neonatal deaths in 2016-2017 were preceded by the administration ofmedicationwithapotentiallylife-shorteningeffect,theNetherlandsreportedaprevalenceofjust4%in20108.Whenlookingattheprevalenceofadministeringmedicationwithanexplicitlife-shorteningintention,thedifferencebetweenthe10%prevalenceestimatesofFlandersin2016-2017 with the Dutch estimate of 1% in 20108 is even more striking, particularlyconsideringthattheNetherlandshavealegalframeworkwhichpermitssuchdecisionsinrarecasesofextremelyillinfantswhereFlandersdoesnot.Availabilityofalegalframeworkthusnotnecessarilyleadstoanincreaseintheprevalenceofthispractice.However,thelackofinternationalpopulationestimatesmakesitimpossibletodrawrobustandvalidconclusionsontheimpactofthesedifferentlegislativechoicesregardingthepermissibilityandrulesforgoodclinicalpracticeofvariousneonatalend-of-lifedecisionsontheiractualprevalence.

7.4.3 The possible impact of the Belgian legal context on neonatal end-of-lifedecisions

The “permissive” climate towards neonatal end-of-life decisions amongst Belgian neonatalhealthcareprovidersmustbeviewedwithinthebroadercontextoftheBelgianlegalandmedicalculture.AsBelgiumhasbothafairlyliberallawonterminationofpregnancycomparedtosomeothercountriesandalawoneuthanasiainadultsandcompetentminors(seechapter1),itcouldbedebatedthattheBelgiummedicalandlegalcultureasawholecouldbeconsideredasmoreaccepting of certain decisions at the end of a person’s life regardless of their age than isinternationallythecase.Eventhoughneonatesthemselvesdonotfallunderthejurisdictionofthementionedlaws,weshouldconsider ifandhowtheirpresenceand implementationcouldinfluenceneonatalend-of-lifedecision-making.

Oneofthemainargumentsagainstthelegalisationofeuthanasiaworldwideistheslipperyslopeargument,whichstatesthatlegalisingeuthanasiawill leadtoerror,abuseandtheviolationofrightsofvulnerablepeopleforwhichthelawwasnotintended.Asneonatesarenotcompetenttorequest,andreceive,euthanasia,theycouldthusbeclassifiedassuchavulnerablegroupthatfallsoutsidethejurisdictionoftheeuthanasialaw,butyetcouldexperienceaninfluenceofitsimplementation. If this is the case, the implementation of the euthanasia law for adults andcompetentminorsshouldleadto1)anincreaseofdeliberatelyendingthelifeofneonateswithsevereconditions,and2)thisincreaseshouldbeattributedtophysiciansfeelingmoreateasewiththepracticeexplicitlyduetotheexistenceorextensionoftheeuthanasialaw.Prevalenceestimatesprovidedinthisdissertationcanonlyprovideinsightintothefirstclaim,namelythatthe prevalence of administering medication with an explicit life-shortening intention stayedrelativelyconstantat7%twoyearsbefore;and10%ofthetotalpopulationofdeceasedneonatesandinfants15yearsafterimplementationoftheeuthanasialawinadults(threeyearsaftertheadditionofcompetentminors).Furthermore,acausalrelationbetweentheimplementationofthe euthanasia law for adults and competent minors, and the considerable prevalence ofadministrationofmedicationtointentionallyhastendeathinneonates,asmentionedinpoint2,cannot be proven by the data presented in this dissertation. The highly positive attitude ofneonatologists andneonatal nurses towards these types of end-of-life decisions,which couldpossiblyidentifyFlandersasauniqueregionwithregardtoneonatalend-of-lifedecision-making,wasneverlinkedtotheexistenceoftheeuthanasialawbyparticipantsinthestudiesdiscussedinthisdissertation.ContrarytothepossibilitythattheBelgianeuthanasialawhasanunintended

142

facilitatingeffectonend-of-lifedecision-makinginneonates,theexistenceofthelawcouldhaveanimpedinginfluence.Herewecouldarguethattheexistenceofaclearlawonwhenactivelyending life inminors andadults is legally allowed, this alsoprovidesphysicianswith a cleardemarcationonwhenusinglethaldrugsisconsideredillegal.Asneonatesareundoubtedlynevercompetent,theprevalenceofdrugadministrationwithanexplicitlyintendedeffecttoshortenlifein neonates should in this case decrease because they fall outside of the due care criteriamentionedinthelaw.Ourdatashowsthatthisisnotthecase,andthuswecanhypothesizethatthe existence of the euthanasia law has no impeding effect on neonatal end-of-life decision-making.

Whileapossibleinfluenceoftheeuthanasialawonneonatalend-of-lifedecision-makingcanbecontested,theinfluenceofanotherBelgianlawonneonatalend-of-lifedecisionswasexplicitlymentionedinourqualitativestudy,namelythelawonterminationofpregnancy(seechapter5).Late terminationofpregnancy is legallypossible inBelgiumwhen completing thepregnancypresentsaseriousthreattothewomen’shealthorwhenit isestablishedthat,whenborn,thechildwillsufferfromaparticularlysevereailment,acknowledgedtobeincurableatthetimeofdiagnosis15,16,57,58. During interviews with neonatologists and nurses, several participantsmentionedthatthelackofalegalframeworktointentionallyhastendeathinneonateswithasevere condition is identified as a barrier in neonatal end-of-life decision-making, especiallycomparedtotheexistenceofthelawonterminationofpregnancywhichallowsforintentionallyhasteningdeathininfantswithasevereconditionaslongasthechildisnotyetborn.SincetheBelgianlawonterminationofpregnancyduetoasevereorlethalfoetalconditiondoesnothaveagestationalagerestriction58,itcanbeconsideredmoreliberalthancountriesthatdoinvokeagestationalagerestriction59,orevencountrieswhere late terminationofpregnancy for foetalanomalies is illegal regardless of the gestational age60. As healthcare providers specificallymentiontheinfluenceofthislawonneonatalend-of-lifedecision-making,andmorespecificallytherestrictionsanduneasinesstheysometimesfeelwhentheyareunabletointentionallyhastendeathinsufferingneonatesandinfantswhenapregnancycouldbeterminatedforexactlythesamediagnosisinanunbornfoetus,itmightbepossiblethattheexistenceofaliberalterminationofpregnancylaw(comparedtoothercountriesinternationally,seeintroductionforanoverview)hasaninfluenceondecision-makingafterbirth.However,astheBelgianlawonterminationofpregnancywas instated in1990, prevalence estimatesonneonatal end-of-lifedecisions from1999-2000and2016-2017discussedinthisdissertationwillthusnotbeabletopointoutanychangesfollowingtheimplementationofthislaw.Furthermore,whetherthepermissiveattitudeofhealthcareprovidersinperinatalcarefollowsratherthanprecedestheimplementationoftheterminationofpregnancylawcanbedebated.

Wefoundthatahighnumberofneonatologistsandnursesfoundadministeringmedicationwithanexplicitlife-shorteningintentionacceptableinextremelyillneonatesandinfants,andthatthispracticeoccursinaboutoneinteninfantswhodiedbeforetheageofoneyear,despitethefactthatthelegalityofthesedecisionscanpossiblybecontested.Furthermore,internationalstudiesindicatedthatthispracticeoccursacrossEurope41,61,whiletheNetherlandsiscurrentlytheonlycountrywherethisislegallycondonedunderstrictsubstantiveandproceduralduecarecriteria(seechapter1).BeforedebatingonwhetherornotthisdataindicatesthatlegislativechangesinBelgiumarewarranted,weshouldthusconsiderwhyphysiciansinFlanders(andinternationally)performsuch acts indaily clinical practice.During interviewsand consortiummeetingswith

143

Flemishneonatologists,somephysiciansemphasizethattheyinterpretdeliberateadministrationoflethaldrugsasillegal,sinceneonatesarenotcapableofdeterminingtheirwishesregardingactiveterminationoflife.However,duetothelegalhazinessexplainedintheintroductionofthisdissertation, other physicians contest the fact that allowing aneonate todie in suchmannerwouldcountasillegal.Itwouldthusbeshort-sightedtostartfromthepresumptionthatin10%of deceased neonates, illegal medical acts (as interpreted by the physician) preceded death.Additionally,fromamoralpointofview,wecanevencontestthatthereisaninherentdifferencebetweenanon-treatmentdecisionwithanexplicitlife-shorteningintentionandadministeringlethaldosesofmedication,astheintendedendresultisthesamenamelytoendthesufferingofthechildbyhasteningdeath62.

When comparing prevalence estimates of administering medication with an explicit life-shortening intention in Flanders - a regionwhere this practice is currently not regulated bymeans of a protocol or a law - with the Netherlands - who chose to provide guidelines andregulationsforbestpractice–,weseethatthispracticeoccursmoreoftendespitethelackofaclear regulation. Our prevalence estimates on a practice that is currently not clearly legallycondoned, combined with the knowledge that attitudes of Flemish neonatal healthcareprofessionals towards administering medication with explicit life-shortening intention arepermissible24,raisesthequestionofwhetherguidelines,protocolsorlawsareneededtomonitorthese decisions in such a vulnerable patient group. However, the existence of a permissiveattitudeofinvolvedhealthcareproviderstowardsthesedecisions,andtheexistenceofempiricalevidence indicating that these decisions are actually made in daily clinical practice do notautomatically warrant support towards these legislative changes.While our interview studyshowedthatneonatologistsandnursesfindthelackofalawallowingforactivelyhasteningdeathbymeansofmedicationinseverecasestobeabarrierindecision-making,theyalsoindicatedtobewaryofpossiblestandardisationbymeansofalaw.Additionally,aWalloonattitudesurveyrevealedthatonlyhalfofrespondentswereinfavourofallowingandstandardizingactiveend-of-lifepractices,withapreferenceforinstatingaprotocolratherthanalaw35.Furthermore,alargeproportionoftheirparticipantsindicatedfeelinguncertainaboutwhetherornottheywishfor a protocol or a law to be instated35, which is consistent with the hesitation felt duringconversationswith Flemish neonatologists in context of this dissertation. TheWalloon studyindicatesthatneonatologistsdonotwantdecisionsconcerninglifeanddeathsituations inanindividualcaseofneonatalsufferingtoberegulatedbyarestrictivelaw35.Ontheotherhand,thejustificationofplacing theultimatedecisionregarding lifeordeathsolely in thehandsof theinvolvedhealthcareproviderswithouttheavailabilityofclinicalandethicalguidelinescouldalsobequestioned.Whiletheexistenceofaprotocoloralawtolegallyallowthesedecisionsmightaiddecision-making,andcouldpossiblyprovideguidelinestowardswhatwouldbeconsideredbestpracticeinthesecases,cautioniswarranted.Thisextremelysensitiveissueneedsfurtherinterdisciplinarydebate,includingphysiciansfromboththeFlandersandtheWalloonregionandpossibly even ethicists and policy makers. The empirical evidence provided within thisdissertationrevealsthatend-of-lifedecisions,eventhosethatpossiblyfalloutsideoftheBelgianlegalframework,areprevalentindailypracticeandtherefore,theseinterdisciplinarydebatesontheirpermissibilityandrequirementsforgoodclinicalpracticearewarranted.

144

7.4.4 Supportforhealthcareprovidersduringtheneonatalend-of-lifedecision-makingprocess

Akeyfindingofthisdissertationisthatpsychologicalandpsychosocialsupportforhealthcareprovidersworking inneonatalend-of-lifecare iscurrentlylacking(chapter5and6)25,63.Bothbeingpartofanend-of-lifedecision-makingprocessandexperiencingthedeathofaneonateintheircarecausesaconsiderableamountofstress for involvedphysiciansandnurses.Belgianneonatologists and nurses are no exception in this case. International evidence shows thathealthcareprovidersworkinginaNICUsettingarerecognizedaspronetothenegativeeffectsofexperiencingstresswithintheiroccupation,astheprevalenceofburnoutisestimatedtobeabout30%inneonatologists64andbetween7.5and54.4%inneonatalnurses65.Thisisunsurprisingashealthcareprovidersworkinginanintensivecaresettingareconfrontedwithacontinuoushigh-stressworkenvironmentonadailybasis66,whereend-of-lifeissues67,difficultethicaldecisions68,and observing continuous suffering of patients69 can have a debilitating effect on their ownqualityoflife.ThenegativeeffectoftheexperienceofworkinginanintensivecaresettingonadailybasisonstressandwellbeingofthehealthcareproviderscouldbeevenlargerinaNICUsetting,astheycareforextremelyillanddyinginfantswhodidn’tevengetthechancetostarttheirlivesproperly.Duringinterviews,neonatologistsandneonatalnursescontinuouslystressedhowdealingwithseverelyillnewbornscanweighontheiremotionalwellbeing,especiallywhentheinfantlookslikeahealthy,fulltermbaby,orwhenhealthcareprovidershaveyounginfantsoftheirownwhichcausesthemtoprojectthehardshipstheyviewandexperienceonthejobontheirownhouseholdsituation.Additionally,theyindicatedthatbeingpartofaneonatalend-of-life decision-makingprocess is never easy, and thatdiagnostic andprognostic insecurity canheavilyweighontheirstateofmind.

Tocopewiththeelevatedamountsofstressduetobeingconfrontedwithend-of-lifedecisionsandinfantdeathonaregularbasis,ourstudiesshowedthathealthcareprovidersturntotheirpeercolleaguesforsupport.Thelargemajorityofbothneonatologistsandnursesindicatedthattheywereabletotalktotheircolleagueswhensomethingisbotheringthemregardinganend-of-life decision made for their patients63. Furthermore, neonatologists and nurses within ourinterview study indicated a considerably easier end-of-life decision-making process forextremely ill neonateswhen they felt like they could counton their colleagues for emotionalsupport25.Additionally, not everyhealthcareprovider isprepared tobepart of an end-of-lifedecision-makingprocess.Whilesomenursesduringinterviewsindicatedthattheywouldoptoutofprovidingpalliativeorend-of-lifecareinfavorofcaringforinfantswithhighersurvivaloddsevery single time to spare themselves the emotional burden, others indicated that theirwillingness to provide end-of-life care was related to the available emotional reserves theythemselveshadatthatpointintime.Whenbothcolleaguesandtheneonatalwardareopentoallowingtheirstafftoswitchtasksandhealthcareproviderscanindicatewhetherornottheyarewillingtobepartofanend-of-lifedecision-makingprocessatthattime,thiscanhaveaninfluenceontheiroverallwellbeing70.However,wefoundonlyoneneonatalpalliativecareprotocolwhichmentionedthis in theirrecommendationssection70.Promoting thistypeofemotionalsupportamongstcolleaguesmorebroadlyinneonatalpalliativecareguidelines,andinstallingateam-workcultureintheNICUwards,couldprovetobebeneficialforthewellbeingoftheirstaff.

145

Though the positive impact of collegial support from peers on wellbeing of the healthcareprovidersshouldnotbeoverlooked,itisnotsufficienttocopewiththestressorsassociatedwithend-of-life decision-making and infant-death in the Flemish neonatal intensive care units.Counselling forbereavedparentsafterperinatal loss tohelp themcope ismuchmorereadilyavailable71 than it is forhealthcareproviders,as they areoftennot recognizedas abereavedpersonbysocietyortheirworkenvironment72.Asaresultofthis,mostrecommendationsandguidelinesonpsychosocial supportduringdeath andend-of-life care inneonates focussesonprovidingphysicians,nursesandotherhealthcareprofessionalswithconcretetoolstooptimallyattend to parents in their decision-making process and grief73–75. Caring for the healthcareprovidersinthiscasebecomessecondaryorevennon-existent,eventhoughtheaddedemotionaldistressofdealingwiththeseextremelydifficultdecisionsregularlycanprovetobemorethantheycancopewith.Healthcareproviderswhosufferfromemotionaldistressandevenburn-outarefurthermoreknowntohaveadiminishedcapacitytocarefor,andshowempathytowardstheillneonatesintheircareandtheirparents64,65.Caringforthehealthcareprovidersmightthusnotonlybenefit theirwellbeingon apersonal level, but itmightalso considerably improve theirability to care for the infants and support the families. The lack of professional support forhealthcareprovidersworkinginaneonatalintensivecareunit,asshownbyseveralstudiesinthis dissertation, should thus obviously be addressed and resolved. Suggested measures toincreasethisstaffsupportareregularformaldebriefingswiththeentireteamtodiscussdifficultend-of-lifecases,andcounsellingsessionsforhealthcareprovidersduringregularworkhoursinsteadofonavoluntarybasisorduringunpaidtime70.

7.4.5 Theroleofpalliativecareinneonatalend-of-lifedecisionmaking

In caring for extremely ill neonates and infants, deciding to either reorient care from cure-orientedandlife-extendingtocomfortandpalliativecare,ortoprovidebothcure-orientedandcomfortcareconcurrently, ispartofdailyclinicalpractice34.Additionally,whenalife-limitingcondition isdiagnosedbeforebirth, decisions tostartuppalliative careat birth canbemadeprenatally70.Inadults,whenpalliativecareisprovided–alongsidestandardcare-duringthefinaldaysoflife,ithasbeenshowntoreducesymptomburdenandincreasequalityoflifeforbothpeoplewhoaredyingandthoseclosetothem76.Anincreasingamountof(cluster)RCTsshowimprovements in outcomemeasures such as dying at home, reducing symptom burden, andimprovingpatientandcaregiversatisfactioninpeoplewithadvancedillnessessuchascancer,chronicobstructivepulmonarydisease,congestiveheartfailure,etc77,78.Inprenatalandneonatalpracticeontheotherhand,palliativecareisarelativelynewfieldwherethepossiblebeneficialimpactishypothesizedbuthasnotyetbeenevaluated79.Furthermore,thoughseveralreportsonperinatalandneonatalpalliativecareprotocols,teamsoreducationalinterventionsexist,theyarenotevidencebased,raisingthequestionofwhatexactlyaperinatalorneonatalpalliativecareprogramshouldentail80.Theroleofend-of-lifedecisions,andthepossibleimplicationsofdataprovidedwithinthisdissertation,withinsuchaperinatalorneonatalpalliativecareapproachisunclear.

When parents receive a life-limiting diagnosis for their child, it is extremely important thathealthcareprovidersprovidethemwithanempathetic,understandableandbalancedoverviewofalltreatmentoptions,includingactiveandcure-orientedinterventions,end-of-lifedecisions,andpalliativecare81,82.Asthepracticeofwithholdingorwithdrawingunnecessarilyinvasivelife-

146

supportingtreatmentinmodernneonatalintensivecareunitsgloballyiswellsupported34,56,83,and results of this dissertation corroborated their central role in neonatal end-of-life care(chapter3),we can expect these conversationsbetweenhealthcareproviders andparents toincludediscussingnon-treatmentdecisions.Furthermore,wecanexpectdecisionstowithholdorwithdrawunnecessarilyinvasivetreatmenttobeanintegralpartofashiftfromcurativecaretopalliativecare.Aprimeexampleof this is thepracticeofcompassionateextubation,whereassisted ventilation which is often vital for survival of the child is withdrawn to increasecomfort34. Additionally, an integral part of perinatal and neonatal palliative care comprisesallowingparentstobeinvolvedinnormalbabycare22andcreatingmemorieswheretherearenone74,whichisuniquetothepalliativesettingoflosingachildattheverybeginningoftheirlife.Allowing parents to bathe, care for and dress their child if they so wish, and providingopportunitiestotakephotographsasafamilyasawaytosaygoodbyetotheirchildbeforeheorshepassesawayisthuspartofgoodpracticeinperinatalandneonatalpalliativecare74.Withinourinterviewstudy(chapter5),welearnedthatanecessaryfirststepinordertoprovidetheseopportunities is to disconnect the dying child from life-support and mechanical ventilation.Duringadvancecareplanningconversations,whichispartoftheprovidedpalliativecareservice,thehealthcareproviderswalkthroughthedyingprocessofthechildstartingwithwithdrawalfromlife-supportingtreatment,tomakingthemostoftheirfinalmoments25.Furthermore,resultsfrom our mortality follow back survey-method showed that, now more than 17 years ago,prevalenceofend-of-lifedecisionsroseafterthefirstweekoflife(chapter3),herebyindicatinglessneedtorushthedyingprocessoftheinfantthanbefore,andallowingfortimetoadequatelysaygoodbye.

Asidefromotherimportantcomponentssuchasadvancecareplanningandsupportforparents,providingadequatepainreliefandcomfortareacrucialcomponentofperinatalandneonatal(palliative) care84. Withholding or withdrawing treatment is therefore usually followed byincreasing analgesics and sedatives to treat the dying neonates’ pain and suffering85. Painmanagementinnewbornsremainsverychallengingforhealthcareproviders,sinceinfantsareunable to express what they feel, and pain cues are difficult to interpret86. Furthermore,physiciansworry that high doses of opioids, benzodiazepines and other sedativeswhich areneededtoprovideadequatepainreliefduringbothcurativecareandpalliativecare,willleadtorespiratory or cardiovascular distress33,34. In neonates more than in minors or adults,differentiatingbetweenactivelyandintentionallyendingthelifeoftheneonatewithlethaldosesofmedication andproviding adequatepain reliefwithin apalliative care setting is thus verydifficult87.Aspainandsymptomcontrolaresuchcrucialcomponentsofpalliativecare,theyareinadvertently linkedand a clear demarcation between an end-of-life decision to increase theadministrationofpainmedicationwithoutexplicitintenttohastendeathandprovidingadequatepalliativecareisunnecessary.Yetthequestionbecomeswhetheranend-of-lifedecisionwheremedication is administered with an explicit life-shortening intention is compatible within apalliativecaresetting,orwhetherbothpracticesshouldbeseenasmutuallyexclusive.Thehighdosesofpainmedicationneededtoprovidereliefforsufferingneonateswithinthecontextofprovidinggoodpalliativecarecould,andareoften(chapter3)givenevenwhenalife-shorteningeffectwasforeseenorevenintended.Inadults,weseethatBelgiumprovidesauniquecontextwhere the practice of euthanasia and palliative care are integrated88–90, rather than beingconsideredasincompatiblepracticesasisthecaseabroad91.Thefactthatprovidingadequatepainreliefduringapalliativecareapproach,andadministeringmedicationwithanexplicitlife-

147

shorteningintentwerenotseenasseparatepracticesbythehealthcareprovidersinvolvedinthisdissertationseemtocorroboratethesefindings.

Embeddingtheend-of-lifedecision-makingprocesswithinaneonatalorperinatalpalliativecareapproachcouldbeusefultoaddressthecomplexfamilyneedsinanemotionallyturbulenttimebyprovidingafamily-centeredapproachwithafocusonparental(spiritualandcultural)values,memory making, and compassionate communication between parents and providers92–94. Asfamily-centered psychosocial support and bereavement care is central within a neonatal orperinatal palliative care approach, ample attention is given to the values, goals andneeds ofparents,siblingsandothersignificantrelatives82,84,94.Additionally,suchapalliativecareapproachcouldbenefithealthcareproviders,asexistingperinatalpalliativecareprotocolsoftenincludesectionsonpsychosocial staff support, used to improvequality of care and counteractmoraldistress,burnoutandcompassion fatigue80,95.Aneonatalorperinatalpalliativecareapproachthus includes not only adequate pain relief and comfort for the child, but also has a strongemphasis on compassionate communication and psychosocial support for parents, familymembersandinvolvedhealthcareproviders.Withinafollow-upprojectofthisdissertation,wewillthereforeaimtodevelopthefirstBelgianperinatalpalliativecareprogramtoaddressthismuchneededsupportforfetusesandinfantsattheendoftheirlife,aswellastheirfamiliesandinvolvedhealthcareproviders.Within thisproject,ampleattentionwillbegiventoend-of-lifedecision-making,andhowtosupportparentsandhealthcareprovidersduringtheoftendifficultdecision-makingprocess.

7.5 Implicationsandrecommendations

7.5.1 Implicationsandrecommendationsforpractice

The data providedwithin the qualitative interview study (chapter 5) lead to some concreterecommendations with the aim of making the end-of-life decision-making process itself lessdifficultforhealthcareproviders.Firstly,attentionshouldbegiventocreatingaprivateroomforbad-newsconversationsintheneonatalintensivecareunitandinotherhospitalwardsweresuchconversationsareprevalentandnecessary.Byensuringprivacy,difficultend-of-lifedecisionscanbe discussed between healthcare providers and parents without outside interference ordisturbance.During theseconversations,compassionatecommunicationbetweenparentsandhealthcareprovidersiswarranted.Secondly, installingaroutineuseofadvancecareplanningwith parents in neonates with a severe prognosis could aid difficult decisions. By planningpossible courses of action depending on the possible clinical situations of the ill neonatebeforehandwithparents,healthcareproviderswillbemoreabletobasetheirdecisionsontheparents’wishesandpreferences.Theseadvancecareplanningconversationscanthenbeusefulin times of acute deterioration of the child’s condition, where decisionsmight otherwise berushed or parents would previously be excluded from decision-making. Thirdly, prognosticuncertaintycanbereducedbyinstallingregularmultidisciplinaryteammeetingsanddebriefings,and routinelyasking fora secondopinion fromotherphysicians.By relyingon the collectivewisdomofmultiple,experiencedhealthcareproviders,asopposedtomakingmedicaldecisionsindividually,uncertaintyregardingtheprognosisandthebestpossiblecourseofactioncanbereduced.Lastly,difficultiesinworkingwithparentswhohaveadifferentculturalbackgroundorspeak a different language than that of the involved healthcare providers could possibly be

148

reduced by consulting a neonatal or perinatal palliative care team. As cultural and languagedifferencescanresultinmisunderstandingsorevenentailfundamentallydifferentviewsontheacceptability of certain types of end-of-life decisions, conversations between parents andhealthcareprovidersonthistopiccanbestrenuousandstressful.Insimilarsituationsinadults,apalliativecareteamcanbeconsultedtomediatetheseconversations,astheyaresupposedtohave ample experience in dealing with these difficult issues96. Additionally, neonatal andperinatal palliative care teams put ample emphasis on conversational training andcompassionate communication between parents and healthcare providers92–94, making themideallyplacedtomediateduringdifficultend-of-lifedecision-makingprocesses.Internationally,asmallbutgrowingamountofperinatalandneonatalpalliativecareteamshavebeeninstalled97,yetthedevelopmentoftheseteamsisstillinitsinfancy.InBelgiumtherearecurrentlynoofficialperinatalpalliativecareteamsavailable,despitethefactthattheycouldprovideneonatologistsandneonatalnurseswithcrucialandmuchneededsupportinend-of-lifesituations,includingbutnot limited to providing assistance when families and healthcare providers disagree on thecourseofactionfortheirdyingchild.

End-of-lifedecisionsandapossibleredirectionofcarefromcurativetopalliativeandcomfortcare is part of daily practice when working in a neonatal intensive care unit. Therefore,neonatologists,neonatalnursesandotherhealthcareprofessionalsworkinginthissettingshoulddevelop generalist palliative care skills. In Belgium, there is currently no formal training onneonatalpalliativecareavailabletoaidhealthcareprovidersinattainingtheseneonatalpalliativecareskills98.Includingamoduleonneonataldeathandend-of-lifedecision-makinginstandardcurricula for neonatologists andneonatal nurses increases clinical experience and end-of-lifecommunicationskillsearlyonintraining,whichleadstoenhancedconfidenceandfewernegativeexperienceswithend-of-lifecareintheneonatalintensivecareunits99.Furthermore,neonatalintensive care units could provide on-the-job training to newer/younger staff members bypairing them upwith more experienced colleagues during their first encounters with dyingneonates.Byensuringbasictrainingonneonatalend-of-lifecareandpalliativecareforhealthcareproviders,individualexperienceisraised,whichwasindicatedasafacilitatingfactorinneonatalend-of-life decision-making during our interview study. Additionally, when all availablepersonnelworkinginneonatalintensivecareunitsattainedbasicpalliativecareandend-of-lifecommunication skills, dividing the workload of caring for neonates with a poor prognosisamongstcompetentandtrainedhealthcareprovidersbecomeseasier.

As collegial support from peers in the neonatal intensive care unit was deemed crucial yetinsufficienttosupportneonatologistsandneonatalnursesduringstressfulend-of-lifecareandpalliativecaresituations,thelackofprofessionalpsychosocialsupportatthewardiscauseforconcern. As discussed extensively inprevious paragraphs,we suggest the implementation ofregularformaldebriefingswiththeentireteamresponsibleforcaringforaneonatewhodiedwithintheunit.Hereby,opportunitiesarecreatedtoreviewanddiscusswhatcouldhavebeenimproved,whichcouldaidinfutureend-of-lifecases.Furthermore,werecommendcounsellingsessions forhealthcareproviderswhowere involved inend-of-lifecasesduringregularworkhours,asopposedtothemattendingcounsellingsessionsonavoluntarybasisorduringunpaidtime70.Byprovidingadequateprofessionalpsychosocialsupportattheneonatalintensivecareunit, elevated levels ofwork-related stress and increased risk of burn-out can be avoided orremedied,whichcouldimprovejobsatisfactionandpersonalwellbeingofhealthcareproviders.

149

7.5.2 Implicationsandrecommendationsforpolicy

Someoftheresultsfromthe2016-2017studydescribedinchapter3wereagainstexpectationsof Flemish neonatologists and available trend figures from the Netherlands. Relying on thereports and experiences of healthcare providers to provide estimates ofwhich decisions aremadecouldthuscauseskewedresults.Monitoringdailypracticebymeansofpopulation-levelstudiesisthuscrucialtohaveareliable,continuousandup-to-dateoverviewofwhichdecisionsareactuallymadeinclinicalpracticeacrosssettings,andwhatthemainreasonsarefordoingso.Actualprevalence estimateson apopulation levelare invaluable as they enable an empiricalanalysisoftheplausibilityofcertaindecisionsbeingmade,andextenttowhichthesedecisionsarepracticed,especiallyinsensitivetopicssuchasneonatalandprenatalend-of-lifedecision-makingwhicharethetopicofmuchethicalandlegaldebate.Withouttheseprevalenceestimates,ethicalandlegaldiscussions,andevenlegislativedecision-making,arebasedonexperiencesandviewpointsofaselectnumberofconsultedexperts,whilepopulation-datacanprovideanactualempiricalbasisonifandhowoftenvariousend-of-lifepracticesoccurinthepopulation.Withinanever-changingsocietywherethereisacontinuousriseinmedicalpossibilitiestosavethelifeof neonateswith severe health concerns, systematicmonitoring of end-of-life decisions on apopulation-levelonaregularbasis isparamount.Policymakersshouldsupport thisrecurrentperiodicmonitoringinordertobeawareofanysignificantchangesindailypracticewhichmightwarrantlegislativeorpolicy-changes.

Thedatapresentedwithinthisdissertationshowedthatend-of-lifedecisionmakingiscommonandwe can thus assume that a largemajority of neonatologists andneonatal nurseswill beconfrontedwithanend-of-lifedecision-makingprocessnumeroustimesoverthecourseoftheircareer. As current national and international guidelines on the acceptability and adequateperformance of these neonatal end-of-life decisions are lacking, there might still be a lot ofuncertaintyamongphysiciansandnursesregarding theirpermissibilityandrequirements forgood clinical practice. Guidelines and protocols such as the Groningen protocol in theNetherlands,andeventheinstitutionoflawssuchastheeuthanasialawforadultsandcompetentminorscanservetoeliminatesomeofthecontroversiesthatareinherentinneonatalend-of-lifecare.However,asdiscussedbefore,weshouldatthesametimebewaryofoverregulatingend-of-life practices as each individual neonatal end-of-life case is unique. Furthermore,overregulationcanevenhaveanadverseeffect,causingdoubtandleadingtoinactionamongstphysiciansevenwhenitconcernsacceptableend-of-lifepractices.BearinginmindthestudyonattitudesofneonatologistsintheWalloonregionofBelgiumonend-of-lifedecisionsandpracticesfor verypreterm infants35, issuingpractice recommendationsandguidelines in the formof aprotocolmight be amore adequate solution than striving towards an actual law to regulateneonatal end-of-life decision-making. Our prevalence estimates can provide experts with astartingpointtodiscussthepossibleformulationoftheseguidelinesorlegislativealternativesfurther.Additionally,theprevalenceestimatesandpossiblebarriersandfacilitatorshealthcareprovidersexperienceduringaneonatalend-of-lifedecision-makingprocessdiscussedwithinthisdissertationmightbeanidealstartingpointtowardsformulatingaidsandguidelinestowardswhatisconsideredbestpracticeinthesecases.

150

7.5.3 Implicationsandrecommendationsforfutureresearch

Inputfrom(bereaved)parentswasmissingfromthenarrativeofthisdissertation,yetparentalviewsarecrucialtoprovideacomprehensivepictureofaneonatalend-of-lifedecision-makingprocess because they serve as the surrogate decision-maker for their child. A forthcomingpublicationoftheresearchgroupwillfocusonthebarriersandfacilitatorsforparentsduringtheneonatalend-of-lifedecision-makingprocess,yetfutureresearchshouldcontinuetofocusontheviews,attitudesandexperiencesofparentswithinneonatalend-of-lifedecision-making,andhowtheyrelatetothatofhealthcareproviders.

End-of-lifedecision-makinginneonatesisirrevocablyconnectedtoprenatalend-of-lifedecision-making,asalotofcongenitaldisordersoranomaliescanbediagnosedprenatally100.Ifaprenataldiagnosisismade,muchoftheadvancecareplanningcanbedonebeforethechildisborn101.Duringthistimeadecisiontoterminatethepregnancyoradecisiontoforgointensivetreatmentatbirthcanbemade.Whenlookingatshiftsintheprevalenceofend-of-lifedecisionsovertime,thedecisionsmadebeforebirthshouldthusnotbeoverlooked.Thestudymethodologydescribedinchapter2ofthisdissertationprovestobeidealtoexaminesuchshifts.Withinthisdissertation,solelyresultsinneonatesandinfantswerediscussedasthedatacollectedduringthecourseofthestudieswastooextensivetosummarizewithinonecomprehensivedoctorate.Weaimedtoprovidedetailedandin-depthinformationregardingmultipleaspectsoftheneonatalend-of-lifedecision-makingprocessas opposedto abrief descriptionof decision-makingacrossa largerscopeofthefoetal-infantileperiod,yetresultsonthisareforthcoming.Futureresearchshouldcontinue to include both prenatal and neonatal decisions, using the framework provided inchapter 2. Additionally, as shown in chapter 4, when examining attitudes, opinions andexperiencesofinvolvedhealthcareproviders,bothprenatalandneonatalshouldbeconsidered.

Themortalityfollow-backmethoddescribedinchapter2canbeusedinmanycountries,aslongasadeclarationsystemofneonataldeathandprenatalstillbirthisavailable.Itisthereforeidealtocompare theprevalenceon foetal-infantileend-of-lifedecisionsinternationally.As theonlyother currently available population-based trend figures are from the Netherlands8, futurestudies should focus on collecting population data internationally. International comparativeresearch can identify country-specific or even region-specific factors thatmight influence theoccurrence of end-of-life decisions and end-of-life practice. Furthermore, it could provideevidence of differing medical cultures concerning neonatal end-of-life care. Additionally,comparing internationalprevalence estimates in countrieswithdiffering guidelines and legalframeworksregardingneonatalend-of-lifedecision-makingcangivecrucialinsightinwhattheimpactofdecidingwhetherornottoregulatethesepracticescanbeonactualdailypractice.Thisinsightcaninturnbeusedinfuturenationalandinternationaldebatesonwhetherguidelines,protocolsorlawsareneededtomonitorthesedecisionsinsuchavulnerablepatientgroup.

Asstatedintheoverviewofthelimitationsofthemortalityfollow-backsurvey,questionnairesonlyprovidealimitedpotentialtofullycapturethecomplexityofaprenatalandneonatalend-of-lifedecision-makingprocess.Withinthisdissertationweaddedcrucialdataonexperiencesandattitudes of healthcare providers by means of attitude surveys and qualitative, in-depthinterviewsinordertoframeprevalenceestimateswithinsightintohowthesedecisionsaremadeindailypractice.However,futurestudiesshouldcontinuetofocusongatheringmoredetailed

151

informationonforexamplehowneonatalandprenatalend-of-lifedecisionsaremadeindailypractice, and how parents are involved in decision-making. Additionally, the academicclassificationoffoetal-infantileend-of-lifedecisionsprovidedwithinthisdissertationcandifferfromtheinterpretationofneonatologistsandneonatalnursesindailypractice.Thesedifferenceswillnotbenoticeablewithinour large-scalestudies,yetqualitativestudiesonasmallerscalewithin individual healthcare settings would be able to reveal this crucial information.Furthermore,theparentalviewsandexperiencesonneonatalend-of-lifedecision-makingwereexcludedfromthisdissertation,howevertheywouldprovideessentialadditionalinsightsinacomplex decision-making process wheremultiple actors are involved. Future studies shouldtherefore focus on embedding the parental views within the knowledge from healthcareprovidersavailablewithinthisdissertation.

Aspreviouslystated,neonatalend-of-lifedecisionsareembeddedinneonatalandevenperinatalpalliative care (PPC). Although crucial elements of a palliative care approach are alreadyimplemented in regular perinatal practice, the existence of actual perinatal palliative teamsinternationally is rare97.ExistingPPCprogramsmostlyoriginatedbottom-up fromneeds thatarose within daily practice without evidence-based support80. Reports on these perinatalpalliativecareteamsexistandindicationsoftheirpositiveeffectonthevulnerablepopulationofextremely ill infantsbeforeandafterbirthwere found internationally80,102,103,yetresearchontheirimplementationandeffectsontheprovidedcareislacking80.Asthisisarelativelynewandemergingresearchfieldthataddressesmuchneededsupportforextremelyillinfantsbeforeandafterbirthaswellasfortheirfamiliesandinvolvedhealthcareproviders,futureresearchshouldfocusonevaluatingthebestmodelofcarewithinthissetting.

7.6 Conclusions

This dissertation has revealed that end-of-life decision-making is an important part of dailyclinicalpracticewhencaring forneonatesand infantswithsevereconditions.ThemajorityofFlemishneonatologistsandneonatalnursesworkinginaneonatalintensivecareunitconsidernon-treatmentdecisionsand theadministrationofmedicationacceptable to relieve suffering,evenwhenthishasapotentialorexplicitlife-shorteningeffect.Consequently,theprevalenceofbothnon-treatmentdecisionsandtheadministrationofmedicationwithapotentialorexplicitlife-shorteningintentwithinthetotalpopulationofdeceasedneonatesandinfantsinFlandersisrelativelyhighataboutthreeinfive.Inoneintendeceasedneonatesandinfants,medicationwasadministeredwith an explicit intention tohastendeath,which is currentlynot clearly legallycondonedwithintheBelgianlegalframework.Despitetheircommonality,healthcareprovidersareconfrontedwithasignificantnumberofbarriersduring theneonatalend-of-lifedecision-makingprocesssuchasalackofprivacyforbad-newsconversations,prognosticanddiagnosticuncertainty, lack of training in palliative and end-of-life care, and difficult legal frameworks.Additionally,beingconfrontedwithbothanend-of-lifedecision-makingprocessandtheresultingdeathoftheirpatientscausesneonatologistsandneonatalnurseswithaconsiderableamountofstress,whichisnotalwaysadequatelyaddressedbytheneonatalward.Thesefindingscanleadto a number of readily implementable recommendations for daily practice, such as creatingprivacy for bad-news conversations, installing regular multidisciplinary meetings anddebriefings to reduce uncertainty, routinely setting up an advance care plan, and providingphysicians and nurses with appropriate psychosocial support during regular work hours.

152

Additionally,werecommendaddingamandatorymoduleonend-of-lifecareandpalliativecaretraining within standard curricula for neonatologists and neonatal nurses. Lastly, somerecommendationsforpolicycanbemadesuchasattentionforregularanddetailedmonitoringofthepracticeofend-of-lifedecisionsindailypractice,whichcouldbeusedtoevaluatepossiblelegislativechanges.

Severalotherimportantquestionsregardingneonatalend-of-lifedecision-makingremain,suchas the need for internationally comparable prevalence estimates to reveal country-specificfactors that influencedecision-making,orhowtheparentalnarrative fitswithin theprovideddataofthisdissertation.Furthermore,attentionshouldbepaidtowardsdevelopingasupportiveandencompassingperinatalpalliativecareapproachthatfitswithintheBelgianclinicalandlegalframework,whereinend-of-lifedecisionsmightplayanimportantrole.Therecommendationsbasedontheempiricalevidenceprovidedinthisdissertationwillhopefullyaidfutureinfants,parentsandhealthcareproviders,sothatadifficultend-of-lifedecision-makingprocesscanbemademorebearable.

153

7.7 Referencelistchapter7

1. DevliegerR,MartensE,GoemaesR,CammuH.PerinataleActiviteiteninVlaanderen2017.StudvoorPerinatEpidemiol.2018.

2. ProvoostV,CoolsF,MortierF,etal.Medicalend-of-lifedecisionsinneonatesandinfantsinFlanders.Lancet.2005;365(9467):1315-1320.

3. Pousset G, Bilsen J, Cohen J, Chambaere K, Deliens L, Mortier F. Medical End-of-LifeDecisionsinChildreninFlanders,Belgium.2017;164(6):547-553.

4. ChambaereK,BilsenJ,CohenJ,etal.Apost-mortemsurveyonend-of-lifedecisionsusingarepresentativesampleofdeathcertificatesinFlanders,Belgium:researchprotocol.BMCPublicHealth.2008;8(1):299.

5. BilsenJ,CohenJ,ChambaereK,etal.Medicalend-of-lifepracticesundertheeuthanasialawinBelgium.NEnglJMed.2009;361(11):1119-1121.

6. Pousset G, Bilsen J, Cohen J, Chambaere K, Deliens L, Mortier F. Medical End-of-LifeDecisionsinChildreninFlanders,Belgium:APopulation-BasedPostmortemSurvey.ArchPediatrAdolescMed.2010;164(6):547-553.

7. ChambaereK,BilsenJ,CohenJ,Onwuteaka-PhilipsenBD,MortierF,DeliensL.Trendsinmedical end-of-life decision making in Flanders, Belgium 1998-2001-2007.Med DecisMaking.2011;31(3):500-510.

8. ten Cate K, van de Vathorst S, Onwuteaka-Philipsen BD, van der Heide A. End-of-lifedecisions for children under1 year of age in theNetherlands: decreased frequency ofadministrationofdrugstodeliberatelyhastendeath.JMedEthics.2015;41(10):795-798.

9. DierickxS,Onwuteaka-philipsenB,PendersY,etal.Commonalitiesanddifferencesinlegaleuthanasia and physician-assisted suicide in three countries : a population-levelcomparison.IntJPublicHealth.2019;6.

10. Hoddinott SN, Bass MJ. The dillman total design survey method. Can Fam Physician.1986;32(November):2366-2368.

11. BergerTM,HoferA. Causes andCircumstancesofNeonatalDeaths in108ConsecutiveCases over a 10-Year Period at the Children ’ s Hospital of Lucerne , Switzerland.Neonatology.2009;95(2):157-163.

12. RoyR,AladangadyN,CosteloeK, LarcherV.Decisionmakingandmodesof death inatertiaryneonatalunit.ArchDisChild-FetalNeonatalEd.2004;89(6):F527-F530.

13. Cuttini M. The European Union Collaborative Project on Ethical Decision Making inNeonatal Intensive Care (EURONIC): findings from 11 countries. J Clin Ethics.2001;12(3):290-296.

14. PoussetG.End-of-lifecareanddecisions inchildren.Studiesofattitudesandpractices.(Doctoralthesis).2010.

15. PapiernikE,ZeitlinJ,DelmasD,etal.Terminationofpregnancyamongverypretermbirthsanditsimpactonverypretermmortality:ResultsfromtenEuropeanpopulation-basedcohortsintheMOSAICstudy.BJOGAnIntJObstetGynaecol.2008;115(3):361-368.

16. BosmaJM,vanderWalG,Hosman-BenjaminseSL.LateterminationofpregnancyinNorthHolland.BrJObstetGynaecol.1997;104(4):478-487.

17. RobijnL,CohenJ,RietjensJ,DeliensL.TrendsinContinuousDeepSedationuntilDeathbetween2007and2013 :ARepeatedNationwideSurvey.PLoSOne.2016;June:1-11.

154

18. RebagliatoM,CuttiniM,KaminskiM,Persson J,ReidM,SaracciR.NeonatalEnd-of-LifeDecisionMaking:Physicians’attitudesandrelationshipwithself-reportedpracticesin10Europeancountries.Jama.2000;284(19):2451-2459.

19. Garten L, Daehmlow S, Reindl T, Wendt A, Münch A, Bührer C. End-of-life opioidadministration on neonatal and pediatric intensive care units: Nurses’ attitudes andpractice.EurJPain.2011;15(9):958-965.

20. BelliniS,DamatoEG.Nurses’knowledge,attitudes/beliefs,andcarepracticesconcerningdonotresuscitatestatusforhospitalizedneonates.JOGNN-JObstetGynecolNeonatalNurs.2009;38(2):195-205.

21. GarelM,CaeymaexL,CuttiniM,KaminskiM.Ethicallycomplexdecisionsintheneonatalintensive careunit : impact of thenewFrench legislationonattitudes andpracticesofphysiciansandnurses.JournalofMedicalEthics.2011:240-244.

22. Wocial LD.Life SupportDecisions Involving Imperiled Infants. J PerinatNeonatalNurs.2000;14(2):73-86.

23. DevliegerR,GoemaesR,LaubachM.PerinataleActiviteiteninVlaanderen2018.[inDutch].StudvoorPerinatEpidemiol.2019.

24. Dombrecht L, Deliens L, Chambaere K, et al. Neonatologists and neonatal nurses havepositive attitudes towards perinatal end-of-life decisions, a nationwide survey. ActaPaediatr.2019:apa.14797.

25. DombrechtL,PietteV,DeliensL,etal.Barrierstoandfacilitatorsofend-of-lifedecision-making by neonatologists and neonatal nurses inneonates : a qualitative study. J PainSymptomManage.2019.

26. MalliaP.WASP(WriteaScientificPaper):Specialcasesofselectivenon-treatmentand/orDNR.EarlyHumDev.2018;124:62-64.

27. WilkinsonD.Who should decide for critically ill neonates and how? The grey zone inneonataltreatmentdecisions.In:WhenDoctorsandParentsDisagree:Ethics,Paediatrics&theZoneofParentalDiscretion.;2016:57-79.

28. VincentJ.Withdrawingmaybepreferabletowithholding.CritCare.2005;9(3):226-229.29. British Medical Association. Withholding and Withdrawing Life-Prolonging Medical

Treatment.GuidanceforDecisionMaking.ThirdEdition.JohnWiley&Sons;2008.

30. WeissEM,MunsonDA.ActionandUncertainty inNeonatal IntensiveCare.AmJBioeth.2016;16(5):31-33.

31. ChambaereK,VanderSticheleR,MortierF,CohenJ,DeliensL.RecentTrendsinEuthanasiaand Other End-of-Life Practices in Belgium. New England Journal of Medicine.2015;372(12):1179-1181.

32. StevensBJ,PillaiRiddellRR,OberlanderTE,GibbinsS.Assessmentofpaininneonatesandinfants.In:PaininNeonatesandInfants.;2007:67.

33. MorrisonAW,KangT.JudgingtheQualityofMercy :DrawingaLineBetweenPalliationandEuthanasia.Pediatrics.2014;133(133).

34. CarterBS, JonesPM.Evidence-basedcomfortcareforneonatestowards theendof life.SeminFetalNeonatalMed.2013;18(2):88-92.

35. AujoulatI,HenrardS,CharonA,etal.End-of-lifedecisionsandpracticesforverypreterminfantsintheWallonia-BrusselsFederationofBelgium.BMCPediatr.2018;18:1-10.

155

36. RobijnL,DeliensL,RietjensJ,PypeP,ChambaereK.BarriersintheDecisionMakingAboutand Performance ofContinuous SedationUntil Death inNursingHomes.Gerontologist.2019:1-10.

37. Lantos JD. Seminars in Fetal & Neonatal Medicine International and cross-culturaldimensions of treatment decisions for neonates. Semin Fetal Neonatal Med.2015;20(5):368-372.

38. JacquemynY(ConsensusgroepVNenM. [RecommendationsPerinatalCareConcerningViabilityinFlanders](inDutch).TijdschrGeneeskd.2014;70:20:1159-1167.

39. VerhagenE,SauerPJJ.TheGroningenprotocol--euthanasiainseverelyillnewborns.NEnglJMed.2005;352(10):959-962.

40. Sauer PJJ, and the members of the working group. Ethical Dilemmas in Neonatology:Recommendations of the Ethics Working Group of the Cesp. Eur J Pediatr.2001;160(6):364-368.

41. CuttiniM,CasottoV,OrzalesiM.Ethicalissuesinneonatalintensivecareandphysicians’practices :AEuropeanperspective.ActaPaediatr.2006;95:42-46.

42. BergerTM,HoferA.Causesandcircumstancesofneonataldeathsin108consecutivecasesover a 10-year period at the Children’sHospital of Lucerne, Switzerland.Neonatology.2009;95(2):157-163.

43. BartonL.TheContributionofWithholdingorWithdrawingCare toNewbornMortality.Pediatrics.2005;116(6):1487-1491.

44. Verhagen AAE, Janvier A, Leuthner SR, et al. Categorizing Neonatal Deaths: A Cross-Cultural Study in the United States, Canada, and The Netherlands. J Pediatr.2010;156(1):33-37.

45. HagenCM,HansenTWR.Deathsinaneonatalintensivecareunit:A10-yearperspective.ClinInvestig.2004;5(5):463-468.

46. WilkinsonDJ,FitzsimonsJJ,DargavillePA,etal.Deathintheneonatalintensivecareunit:changingpatternsofendoflifecareovertwodecades.ArchDisChildFetalNeonatalEd.2006;91(4):F268-71.

47. AmericanAcademyofPediatricsCommitteeonFetusAndNewborn.Levelsofneonatalcare.Pediatrics.2012;130(3):587-597.

48. VerhagenAAE,SpijkermanJ,MuskietFD,SauerPJ.Physicianend-of-lifedecision-makingin newborns in a less developed health care setting : insight in considerations andimplementation.ActaPaediatr.2007;96:1437-1440.

49. Limitationof life-sustaining treatment inNICU:physicians´ beliefs andattitudes in theBuenosAiresregion(underreview).EarlyHumDev.

50. BucherHU,KleinSD,HendriksMJ,etal.Decision-makingatthelimitofviability:Differingperceptions and opinions between neonatal physicians and nurses. BMC Pediatr.2018;18(1):81.

51. Davies D.Medical assistance in dying : Apaediatric perspective.Paediatr Child Health.2018;23(2):125-130.

52. OwenK.Takingcourage :Neonataleuthanasiaandethicalleadership.NursEthics.2018:1-5.

53. DagevilleC,BétrémieuxP,GoldF,SimeoniU,fortheWGonEIinPerinatology.TheFrenchSociety of Neonatology ’ s Proposals for Neonatal End-of-Life Decision-Making.Neonatology.2011:206-214.

156

54. CuttiniM, NadaiM, KaminskiM, et al. End-of-lifedecisions in neonatal intensive care:physicians’selfreportedpracticesinsevenEuropeancountries.Lancet.2000;6736(April2016):2112-2118.

55. BoizeP,BorrhomeeS,MichelP,BetremieuxP,HubertP,MorietteG.Neonatalend-of-lifedecision-making almost 20 years after the EURONIC study : A French survey. ArchPédiatrie.2019;26(6):330-336.

56. SchneiderK,MetzeB,CuttiniM,GartenL.End-of-LifeDecisions20YearsafterEURONIC:Neonatologists’Self-Reported Practices, Attitudes, and Treatment Choices in Germany,Switzerland,andAustria.JPediatr.2019;207:154-160.

57. Garne E, Khoshnood B, LoaneM, Boyd PA, DolkH. Termination of pregnancy for fetalanomalyafter23weeksofgestation:AEuropeanregister-basedstudy.BJOGAnIntJObstetGynaecol.2010;117(6):660-666.

58. Lawonterminationofpregnancy[inDutch].BelgianlawGaz.1990.59. BlondelB,CuttiniM,Hindori-MohangooAD,etal.Howdolateterminationsofpregnancy

affectcomparisonsofstillbirthratesinEurope?AnalysesofaggregatedroutinedatafromtheEuro-PeristatProject.BJOGAnIntJObstetGynaecol.2018;125(2):226-234.

60. BoydPA,DeViganC,KhoshnoodB,etal.SurveyofprenatalscreeningpoliciesinEuropeforstructuralmalformationsandchromosomeanomalies,andtheirimpactondetectionandterminationratesforneuraltubedefectsandDown’ssyndrome.BJOGAnIntJObstetGynaecol.2008;115(6):689-696.

61. SchneiderK,MetzeB,BührerC,CuttiniM,GartenL.End-of-LifeDecisions20YearsafterEURONIC:Neonatologists’ Self-ReportedPractices,Attitudes,andTreatmentChoices inGermany,Switzerland,andAustria.TheJournalofPediatrics.2019;207:154-160.

62. McMahanJ.Killing,lettingdie,andwithdrawingaid.Ethics.1993;103(2):250-279.63. DombrechtL,CohenJ,CoolsF,etal.Psychologicalsupportinend-of-lifedecision-making

in neonatal intensive care units : Full population survey among neonatologists andneonatalnurses.PalliatMed.2019.

64. BellieniCV.,RighettiP,CiampaR,LacoponiF,CovielloC,BuonocoreG.Assessingburnoutamongneonatologists.JMaternNeonatalMed.2012;25(10):2130-2134.

65. ProfitJ,SharekPJ,AmspokerAB,etal.BurnoutintheNICUsettinganditsrelationtosafetyculture.BMJQualSaf.2014;23:806-813.

66. Mol MMC Van, Kompanje EJO, Benoit DD, Bakker J, Nijkamp MD. The Prevalence ofCompassionFatigueandBurnoutamongHealthcareProfessionalsinIntensiveCareUnits :ASystematicReview.PLoSOne.2015:1-22.

67. Todaro-FranceschiV.CriticalCareNurses’PerceptionsofPreparednessandAbilitytoCarefor the Dying and Their Professional Quality of Life. Dimens Crit Care Nurs.2013;32(4):184-190.

68. TeixeiraC,RibeiroO,FonsecaAM,CarvalhoAS.Ethicaldecisionmakinginintensivecareunits :aburnoutriskfactor ?Resultsfromamulticentrestudyconductedwithphysiciansandnurses.JMedEthics.2014;40:97-103.

69. Todaro-FranceschiV.CompassionFatigueandBurnoutinNursing.EnhancingProfessionalQualityofLife.NewYork:Springer;2013.

70. CatlinA,CarterB.CreationofaNeonatalEnd-of-LifePalliativeCareProtocol.JPerinatol.2002;22:184-195.

71. KoopmansL,WilsonT,CacciatoreJ,FlenadyV.Supportformothers,fathersandfamiliesafterperinataldeath(Review).CochraneDatabaseSystRev.2013;(6).

157

72. Kaplan LJ. Toward aModel of Caregiver Grief: Nurses’ Experiences of Treating DyingChildren.Omega-JDeathDying.2000;41:187-206.

73. GoldKJ.Navigatingcareafterababydies :asystematicreviewofparentexperienceswithhealthproviders.JPerinatol.2007;27:230-237.

74. WilliamsC,MunsonD, Zupancic J,KirpalaniH. Supportingbereavedparents :practicalsteps in providing compassionate perinatal and neonatal end-of-life care - A NorthAmericanperspective.SeminFetalNeonatalMed.2008;13:335-340.

75. EnglerAJ,CussonRM,BrockettRT,etal.NeonatalStaffandAdvancedPracticeNurses’Perceptions of Bereavement/End-of-Life Care of Families of Critically Ill and/orDyingInfants.AmJCritCare.2004;13(6):489-498.

76. KavalieratosD,CorbelliJ,ZhangD,etal.Associationbetweenpalliativecareandpatientandcaregiveroutcomes:Asystematicreviewandmeta-analysis.JAMA-JAmMedAssoc.2016;316(20):2104-2114.

77. GomesB,CalanzaniN,CurialeV,MccroneP,IjH.Effectivenessandcost-effectivenessofhomepalliativecareservicesforadultswithadvancedillnessandtheircaregivers(Review) SUMMARY OF FINDINGS FOR THE MAIN COMPARISON. Cochrane Collab Publ byJohnWileySons,Ltd.2014;(6).

78. RosenwaxL,SpilsburyK,McNamaraBA,SemmensJB.Aretrospectivepopulationbasedcohortstudyofaccesstospecialistpalliativecareinthelastyearoflife:Whoisstillmissingoutadecadeon?BMCPalliatCare.2016;15(1):1-9.

79. CataniaTR,BernardesLS,BenuteGRG,etal.WhenOneKnowsaFetusIsExpectedtoDie :PalliativeCareintheContextofPrenatalDiagnosisofFetalMalformations.JPalliatMed.2017;20(9):1020-1031.

80. ParraviciniE.Neonatalpalliativecare.CurrOpinPediatr.2017;29(2):135-140.81. Sidgwick P, Harrop E, Kelly B, Todorovic A,WilkinsonD. Fifteen-minute consultation :

perinatalpalliativecare.ArchDisChildEducPract.2017;102:114-116.

82. WoolC.StateoftheScienceonPerinatalPalliativeCare.JObstetGynecolNeonatalNurs.2013;42(3):372-382.

83. Janvier A, MeadowW, Leuthner SR, et al. Whom are We Comforting? An Analysis ofComfortMedicationsDeliveredtoDyingNeonates.JPediatr.2011;159(2):206-210.

84. Balaguer A, Martín-ancel A, Ortigoza-escobar D, Escribano J, Argemi J. The model ofpalliative care in the perinatal setting : a review of the literature. BMC pediatrics.2012;25(March).

85. VerhagenAAE,DorscheidtJHHM,EngelsB,HubbenJH,SauerPJ.Analgesics,sedativesandneuromuscularblockersaspartofend-of-lifedecisionsinDutchNICUs.ArchDisChild-FetalNeonatalEd.2009;(October2005):434-438.

86. CarterBS,BrunkhorstJ.Neonatalpainmanagement.SeminPerinatol.2017;41(2):111-116.

87. VerhagenAAE.Neonataleuthanasia :LessonsfromtheGroningenProtocol.SeminFetalNeonatalMed.2014;19(5):296-299.

88. VandenBlockL,DeschepperR,BilsenJ,BossuytN,CasterenVVan,DeliensL.Euthanasiaandotherendoflifedecisionsandcareprovidedinfinalthreemonthsoflife:nationwideretrospectivestudyinBelgium.BMJ.2009;339:19-21.

89. SmetsT,CohenJ,BilsenJ,VanWesemaelY,RurupM,DeliensL.AttitudesandExperiencesof Belgian Physicians Regarding Euthanasia Practice and the Euthanasia Law. J PainSymptomManage.2011;41(3):580-593.

158

90. DierickxS,DeliensL,CohenJ,ChambaereK.Involvementofpalliativecareineuthanasiapractice inacontextof legalizedeuthanasia :Apopulation-basedmortality follow-backstudy.PalliatMed.2018;32(1):114-122.

91. RadbruchL, LegetC,BahrP, et al. Euthanasia andphysician-assisted suicide :AwhitepaperfromtheEuropeanAssociationforPalliativeCare.PalliatMed.2016;30(2):104-116.

92. McmahonDL,TwomeyM,0’ReillyM,DevinsM.Referralstoaperinatalspecialistpalliativecare consult service in Ireland , 2012 – 2015. Arch Dis Child - Fetal Neonatal Ed.2018;103:F573-F576.

93. CarterBS.PediatricPalliativeCareinInfantsandNeonates.Children.2018;5(21):1-9.94. Denney-koelschE,BlackBP,Côté-ArsenaultD,WoolC,SujeongK,KavanaughK.ASurvey

of Perinatal Palliative Care Programs in the United States : Structure, Processes, andOutcomes.JPalliatMed.2016;19(10):1080-1086.

95. Wallbank S, Robertson N. Predictors of staff distress in response to professionallyexperiencedmiscarriage,stillbirthandneonatalloss :Aquestionnairesurvey.IntJNursStud.2013;50(8):1090-1097.

96. MartyCM,CarterBS.SeminarsinFetal&NeonatalMedicineEthicsandpalliativecareintheperinatalworld.SeminFetalNeonatalMed.2018;23(1):35-38.

97. AmyKuebelbeck.PerinatalHospice&PalliativeCare.ContinuingYourPregnancyWhenYour Baby’s Life Is Expected to Be Brief. List of Perinatal Hospice & Palliative CareProgramsandSupport.www.perinatalhospice.org.Published2018.

98. Arias-CasaisN,GarraldaE,RheeJY.EAPCAtlasofPalliativeCareinEurope.Vilvoorde:EAPCPress;2019.

99. ShoreyS,AndréB,LopezV.InternationalJournalofNursingStudiesTheexperiencesandneedsofhealthcareprofessionalsfacingperinataldeath :Ascopingreview.IntJNursStud.2017;68:25-39.

100. Dombrecht L, Beernaert K, Roets E, et al. Apost-mortem population survey on foetal-infantileend-of-lifedecisions :aresearchprotocol.BMCPediatr.2018:1-9.

101. Leuthner SR.Palliative careof the infantwith lethal anomalies.Pediatr ClinNorthAm.2004;51:747-759.

102. SamselC,LechnerBE.End-of-lifecareinaregionallevelIVneonatalintensivecareunitafterimplementationofapalliativecareinitiative.JPerinatol.2015;35(3):223-228.

103. YoungeN,SmithPB,GoldbergRN,etal.Impactofapalliativecareprogramonend-of-lifecareinaneonatalintensivecareunit.JPerinatol.2015;35(3):218-222.

159

Summaries

160

Englishsummary

161

Introduction

Recent decades have seen an increase in possible medical and technical interventions forcriticallyillneonatesandinfants.However,inFlanders,Belgiumabout8.7perthousandchildrenstilldieduringthefoetal-infantileperiod,i.e.fromfoetusesofmorethan500gramsor22weeksofgestationupuntiloneyearafterbirth.Manyofthesedeathsoccuratneonatalintensivecareunits and are preceded by a possibly life-shortening end-of-life decision. In neonates, theseincludenon-treatmentdecisionssuchaswithholdingorwithdrawinglife-sustainingtreatment,intensificationofalleviationofpainand/orothersymptomswithapotentiallife-shorteningeffectand intentionally ending life with lethal drugs. Additionally, prenatal diagnostic techniques(genetic techniques, prenatal imaging techniques) have evolved considerably, leading to anincreasingnumberofcongenitalmalformationsbeingdiagnosedprenatallyinsteadofafterbirth.Somedecisions such asabstinence from treatmentor terminationof pregnancy canbemadeduringgestationincasesofthedetectionofseriousabnormalities.

Theethicaldilemmainsomeofthesecasesbetweensavingthelifeofthefoetusorneonate,andnot knowingwhat theburdenof sufferingwill be later onneeds thoughtful andprofessionaldeliberation of the parents and involved healthcare professionals. Even though these ethicaldilemmasneedtobeevaluatedonacase-by-caselevel,consideringthespecificcharacteristicsandmedicalsituationofthechild,populationdataonwhatoccursinsimilarsituationscouldbevaluablefortheinvolvedhealthcareprovidersincasesofuncertaintyordisagreementbetweeninvolvedactors.Currently,availableresearchbothwithintheBelgiancontextandabroadiseitherincompleteoroutdated,andthusnothelpfulasaguidetoaidhealthcareprovidersincurrentdailypractice.Withinthestudiesincludedinthisdoctoralthesis,wethereforefocusedonkeycharacteristicsofend-of-lifedecisionsinavulnerablepopulationofchildrenfromaviabletermofpregnancyupuntiltheyreachtheageofoneyear.Theaimofthisdissertationwastwofold:1)toprovideanaccountofwhathappensonapopulationlevelbymeansofprovidingprevalencerateson end-of-lifedecisionsand their clinicalanddemographic characteristics; and2) to godeeperintowhatitmeanstomakethesedecisionsindailypracticebymappingoutattitudes,viewsandexperiencesonneonatalend-of-lifedecision-makingofinvolvedhealthcareproviders,namelyneonatologistsandneonatalnurses,inordertoadequatelyframethesenumbersindailypractice.

Currentevidenceonaspectsof foetal-infantileend-of-lifedecisionsdiscussed inthisdissertation

Theprevalenceoffoetal-infantileend-of-lifedecisions

Population-basedstudies(i.e.withalldeathcasesasthefocus)areidealtostudytheincidenceandcharacteristicsofend-of-lifedecisions,butsuchstudiesarerareinneonatesandinfantsand, toourknowledge,non-existent instillborns. Inneonates, resultsaremostlybasedonreviewsofmedicalrecordsofaneonatalintensivecareunitataparticularhospital.Inthesestudies40%to93%ofdeathsinaNICUfollowwithdrawaloflife-sustainingtreatments.Thelarger scale EURONIC study was based on physicians’ self-reported practices within 143EuropeanNICUsinthe1990s.Theonlypopulation-basedstudiesarefromtheNetherlands(in

162

2014)andBelgium(in2000).Thesestudiesfoundanend-of-lifedecisionbeingmadein60%ofalldeathsofneonatesandinfants.Instillborns,previousstudiesin2003andin2000-2005haveonlylookedattheprevalenceoflateterminationofpregnancy.Notmuchisknownabouttheentiretyofend-of-lifepractices(includingdecisionsotherthanterminationofpregnancy)andtheirdecision-makingprocess,oraboutpatientcharacteristicsbesidesgestationalageandthepresenceoffoetalanomalies.

Priortothedevelopmentofthestudiesinthisdissertation,bothprenatalandneonatalFlemishhealthcare professionals stated the need for more recent, population-based data on theprevalence of end-of-life decisions. In light of ever changing societal, legal and clinicalinfluences,wethusbaseimportantclinicaldecisionsandrecommendationsindailypracticeonoutdatedpopulation-data.Importantsocietalchangestookplacethatcouldpossiblyimpactend-of-life practice, including in the unborn and newborn population. There was theimplementationoflawsonpatientrights,palliativecareandeuthanasiainadultsin2002,andthelawoneuthanasiaforchildrenwithdecisionalcapacityin2014.Neonatesdonotfallunderthiseuthanasialaw,whichislimitedtoadultsandcapableminors,yetapossibleimpactonprenatalandneonatalpracticecannotbeexcluded.Internationally,theGroningenprotocolintheneighboringNetherlandscouldpossiblyhaveanimpactonBelgianprevalencerates.Asidefrom legal changes, the rise inmedical treatmentoptions for extremely ill neonates couldpossiblyhavechangedmedicalpractice.Therefore,aneedforcurrentandreliableincidenceratesofFlemishend-of-lifedecisionsisindicated,notonlybyresearchersbutalsobyFlemishrepresentativesfromalleightneonatalintensivecareunits.Withinthisdissertation,wewillthereforeaimtoexaminetheseincidencesonapopulationlevel,ininfantswhodiedbeforetheageofoneyear.

Attitudesofhealthcareprovidersconcerningfoetal-infantileend-of-lifedecisions

Previous research showed that, even in newborns with the same pathology, variabilitybetweentypesofend-of-lifedecisionscanbenoted.Thisisbecauseend-of-lifedecisionscanbeinfluencedbyanumberofcontextualvariablessuchasavailablehospitalresourcesandtheparents’ and clinicians’ social, cultural and religious beliefs. Aside from these contextualvariables,attitudesofcaregiversplayacrucialroleinend-of-lifedecision-making.Andevenwithinacareteam,importantdifferencesbetweenphysicians’andnurses’attitudestowardsend-of-lifedecisionshavebeenfound.Personalcharacteristicsofhealthcareprovidersmaythusplayacrucialroleinend-of-lifedecision-makinginneonates.

Anattitudesurveystudyin10Europeancountriesin2000foundthatthelikelihoodoflimitinglife-supporting treatments inneonates isdependenton the countryof residence, reportedreligionof thephysician, their gender,whetherornot thephysicianhas children, and theamountofvery low-birth-weight infants thatareadmitted to theirneonatal intensivecareunit. Furthermore, a self-report questionnaire combinedwith retrospectivemedical chartreview revealed that an unintentional life-shortening effect of administering opioids isconsideredacceptableformostneonatalintensivecareandpaediatricintensivecarenurses.Thesestudiesarehoweverlimited,sincetheyfailtoincludeattitudestowardsdecisionsthatcouldhavebeenmadebeforethebabywasborn.Becauseattitudesanddecisionsbeforeorafter birth could possibly influence each other, and neonatologists are often consulted in

163

prenatalend-of-lifedecisions,attitudestowardsprenatalandneonatalend-of-lifedecisionsshould thus be included into one overarching study to make valid comparisons possible.Because of their relevance for clinical practice, a separate part of thisdissertationwill bedevotedtotheexaminationofattitudesregardingfoetal-infantileend-of-lifedecisionsofthemost involved healthcare providers in neonatal end-of-life decision-making, namelyneonatologistsandneonatalnurses.

Barrierstoandfacilitatorsoftheneonatalend-of-lifedecision-makingprocessforhealthcareproviders

Despitethesevereimpactofend-of-lifedecision-makingonNICUstaffmembers,fewstudieshavefocusedonwhattheinvolvedneonatologistsandneonatalnursesfindeitherhelpfulordifficultinmakingtheseend-of-lifedecisions.QualitativestudieswithNICUstaffmembersinNorwayondecidingwhetherornottocontinuelife-sustainingtreatmentshowthatthelackofcertaintyintheprognosisofthechildandwhattheirsufferingwillbelateroncanbeseenasanimportantbarrierindecision-making.Furthermore,theseNorwegianstudiesshowthattheambivalencebetweenwantingtoincludeparentsandwantingtosparethemsomeofthepain,cancauseindecisionregardingwhether,whenandhowcertaininformationabouttheprognosisneedstobegivenbythehealthcareproviderstotheparents.

Previousstudiesonthesebarriersandfacilitatorsinneonatalend-of-lifedecision-makingarelimited in that theymainly focus on specific end-of-life practices such aswithholding andwithdrawingoftreatmentratherthanfocusingontheentirespectrumofend-of-lifedecisions,orthattheymainlyfocusontheexperiencesofparents,herebyexcludinghealthcareprovidersas an important co-actor in the decision-making process. A separate chapter in thisdissertation will therefore focus on examining which factors neonatologists and neonatalnursesexperienceaseitherhelpfulordifficult inneonatalend-of-lifedecision-making inaNICU. Knowledge on which factors could either benefit or hinder the neonatal end-of-lifedecision-makingprocessfromtheviewpointofthemostinvolvedhealthcareproviderscouldbeacrucialstartingpointinformulatingrecommendationstoaidfuturepractice.

Psychologicalsupportinend-of-lifedecision-makingforhealthcareproviders

Neonatologists and neonatal nurses who work in a neonatal intensive care unit oftenexperience stressors and moral distress due to the high demands of their occupation.Especiallyintimeswhenaninfantintheircarecannolongerbenefitfromaggressiveorevenfutiletreatmentandanend-of-lifedecisionneedstobemade.Similarlytopaediatricintensivecareunitstaff,theyexperiencesadness,helplessnessandfrustrationwhentheyareunabletodomorewhenachilddies.Becauseofthisdistress,neonatologistsandneonatalnursesareprone to developing compassion fatigue or burnout when the emotional price of caringbecomestoohighforthemtocope.PsychosocialsupportforNICUstaffmembersiscurrentlyincluded in recommendations for NICU practices, however most recommendations andguidelines concerning this psychosocial support focus on providing neonatologist andneonatalnurseswithconcretetoolstooptimallyattendtoparentsintheirdecision-makingprocessandgrief.Furthermore,researchonhowsupportedtheyactuallyfeelislacking.

164

Toourknowledge,onlyonestudyincludedspecificrecommendationssolelyfocusingonthebenefittoNICUstaffmembersinaneonatalend-of-lifepalliativecareprotocol.Catlin&Carterrecommendedformalmeetingsorcounsellingsessionsaspartofregularworkhours,insteadof on a voluntarybasis orduringunpaid time. Furthermore, they recommended thatbothneonatologists andnurses shouldbe able toopt out of end-of-life careby takingonotherassignments. A last part of this dissertation therefore focusses on the experiencedpsychologicalsupportofhealthcareprovidersworkinginaneonatalintensivecareunitasanimportant aspect of the foetal-infantile end-of-life decision-making context. Caring for theonesresponsibleforthecareofcriticallyillinfantscouldbeacrucialsteptowardsprovidingbettersupportforbothpatientsandgrievingparentsinaneonatalintensivecareunit.

Studyobjectives

Themain focus of this dissertation is end-of-life decision-making in stillbirths, neonates andinfantsonapopulationlevel,acrosscentres,patientsandphysicians.The followingtwoaims,eachwithspecificresearchquestions,guidedthisdissertation:

Thefirstaimistoexamineend-of-lifepracticesanddecisionsinstillbirths,neonatesandinfantsinFlanders,Belgiumonapopulationlevel.Thefollowingresearchquestionswillbeanswered:

1. Whichmethodologycanbeusedtoreliablystudytheprevalenceofvariousend-of-lifedecisions,takenbeforeandafterbirth?Whichpopulation-leveldatabasescanbeusedtostudybothprenatal andneonatal end-of-lifedecisions, andhowcanweanonymouslycontactthephysicianinvolvedinthesestillbirthordeathcases?

2. Whatistheprevalenceofvariousend-of-lifedecisionsmadeintheneonatalperiod?Didtheprevalencechangeovertimecomparedtothepreviousdata-collectionin1999-2000?What are the clinical and demographic characteristics of infants whose death wasprecededbyvarioustypesofend-of-lifedecisions?Whichcircumstancesareassociatedwithvarioustypesofend-of-lifedecisionsinneonates?

Thesecondaimofthisdissertationistomaptheattitudes,viewsandexperiencesofinvolvedhealthcare providers, namely neonatologists and neonatal nurses, on neonatal end-of-lifedecision-making.Thefollowingresearchquestionswillbeansweredwithinthisaim:

3. Whatare theattitudesofneonatologistsandneonatalnursesconcerningprenatalandneonatalend-of-lifedecision-making?Whatarethedifferencesbetweenphysiciansandnurses inattitudes towards thesedecisions?Which attitudes concerningprenatalandneonatalend-of-lifedecisionsandwhichdemographiccharacteristicsareassociatedwithpossibletreatmentoptionsthatareconsideredacceptableinahypotheticalcase?

4. Whichfactorsinvolvedinthedecision-makingprocesscan,accordingtoexperiencesfromneonatologistsandneonatalnurses,facilitateorimpedetheneonatalend-of-lifedecision-makingprocessinaFlemishneonatalintensivecareunit?

5. In what way are neonatologists and neonatal nurses supported by colleagues,psychologistsandthehospitalwardduringthedifficultprocessofend-of-lifedecisionsin

165

a Flemish neonatal intensive care unit? How sufficient is the current psychologicalsupportforcaregiversworkinginaFlemishneonatalintensivecareunit?

Methods

Toanswertheresearchquestionsandstudyobjectivesofthisdissertation,severaldata-collectionmethodsanddatasourceswereused,namelyamortality follow-backsurvey,anattitudeandpsychological support survey and a qualitative study with face-to-face semi-structuredinterviews.

Themortalityfollow-backsurvey

Themortalityfollow-backsurvey-methodfollowsthedesignofamortalityfollow-backsurveyonapopulation-levelbasedonalldeathcertificatesofstillbornsfrom22weeksofgestationorabirthweightof500gramonwards,andneonatesorinfantswhodiedbeforetheageofoneyear.All included stillbirthsordeathsoccurred inFlandersorBrussels and concerned foetusesorinfantswhosemotherwasaFlemishresidentatthetimeofdeath.ThedesignofthisstudywasidenticaltoasurveyconductedfromAugust1999toJuly2000,withtheexceptionofalongerinclusionperiodfromSeptember2016toDecember2017(12monthsin1999-2000versus16monthsin2016-2017).

Within fourmonthsafterdeath,everycertifyingphysicianreceiveda four-pagequestionnairethrough theFlemishAgency forCareandHealthwho is responsible forprocessing thedeathcertificates with an introductory letter containing patient identification characteristics. Toguaranteeanonymity,alawyerservedasanintermediarybetweentherespondingphysicians,theFlemishAgency forCare andHealth, and the researchers.The intermediary ensured thatcompletedquestionnairescouldneverbelinkedtospecificpatients,physiciansorhospitals.

Two separate questionnaires were used during the survey namely one questionnaire toaccompanydeathcertificatesthatcertifiedastillbirthandonequestionnairetoaccompanydeathcertificatesthatcertifiedthedeathofaninfantbeforetheageofoneyear.Thequestionnairesusedinthesurveyaimedtoinquireaboutpossibleprenatalandneonatalend-of-lifedecisionsthatprecededthedeathorstillbirthreportedonthedeathcertificate.Avalidatedquestionnaireusedtosurveyneonatalend-of-lifedecisionmakingdevelopedinthe1999-2000studywasusedas the basis for the current 2016-2017 questionnaires to ensure comparability of data. Bothquestionnairesfirstaskedwhetherthedeathoftheneonatehadbeensuddenandunexpected.Ifanswerednegatively,anend-of-lifedecisionwasconsideredpossibleandphysicianswereaskedwhetheranyend-of-lifedecisionsprecededthedeathorstillbirth.Theusedquestionnaires(inDutch)canbefoundinAppendix1and2.

Whenmorethanoneend-of-lifedecisionwasdenoted,thedecisionwiththemostexplicitlife-shorteningintentionwasdeemedmostimportant.Whenmorethanoneend-of-lifedecisionwiththe same life-shortening intentionwasnoted, administrationofdrugs (active)prevailedoverwithholding orwithdrawing treatment (passive). In case of an end-of-life decision, follow-upquestionswereaskedsuchas:byhowmuchtimewasthelifeoftheinfantshortened,whatwas

166

themostimportantreasonfordecidingontheend-of-lifedecision,andwhowasincludedinthedecision-makingprocess.Demographicinformationfromthedeathcertificateswasanonymouslylinkedwiththeirrespectivequestionnairedataafterdata-collectionwasfinished.

Theattitudeandpsychologicalsupportsurvey

Inorder toexamine theattitudesandperceivedpsychologicalsupportof involvedhealthcareprovidersinneonatalend-of-lifedecision-making,afullpopulationmailsurveywassetupinallneonatologistsandneonatalnursesworkinginaFlemishneonatalintensivecareunit.AllFlemishneonatalintensivecareunitsparticipatedinthisstudy.Theseneonatalintensivecareunitsweresituated in the following hospitals: Ghent University Hospital, Brussels University Hospital,LeuvenUniversityHospital,AntwerpUniversityHospital,AZSint-JanBrugge-Oostende,HospitalOost-LimburgGenk,HospitalGZAStAugustinusandZNAMiddelheim.

DatawascollectedbetweenMay1standMay31stof2017.Thegatekeepermethodwasused,wherearepresentativephysicianworkingineachneonatalintensivecareunithandedoutthequestionnaire to every neonatologist and every neonatal nurse in their respective ward.Physiciansandnurseswereinvitedtofilloutthequestionnaireandsenditbackbymeansofaprepaidenvelopetotheresearcherswithintheperiodofonemonth.

The questionnaire used in this survey was developed based on an existing Flemish attitudequestionnairefromtheyear2000onneonatalend-of-lifedecisions,andanAmericanstudyoncompassion fatigue, burnout and compassion satisfaction of neonatologists in a neonatalintensive care setting. A multidisciplinary team consisting of three sociologists, twopsychologists, three neonatologists and one gynaecologist developed the final questionnaire.Afterwards,thisquestionnairewascognitivelytestedonfiveneonatologistsfromfourseparatehospitals,threeneonatalnursesfromtwoseparatehospitalsandonegynaecologistinordertoensurevalidityoftheitems.Thequestionnaireconsistedofsevensocio-demographicquestionsand12itemsonperinatalend-of-lifedecisions.Sixoftheseattitudeitemsfocussedonneonatalend-of-lifedecisionsandsixitemsfocussedonprenatalend-of-lifedecisions(lateterminationofpregnancy). Attitudes were measured by indicating whether or not they agreed with thestatements,scoredonafive-pointLikertscale.Wealsopresentedahypotheticalcaseofafoetusbornat27weeksgestationwithadditionalcomplications;participantsweregivensevenpossibletreatmentoptionsandwereaskedtoindicatewhethertheywouldconsidereachoptiononafour-point Likert scale. Lastly, the questionnaire included statements about perceived stress,professional psychological support provided by the neonatal intensive care unit, andpsychological support provided by colleagues. We included a statement on the option ofexpressingprotest concerningan end-of-lifedecision,which couldbe an additional sourceofdistresswhenthisisdiscouraged.Thestatementswerescoredona5-pointLikertscale.Theusedquestionnaires(inDutch)canbefoundinAppendix3and4.

Face-to-facesemi-structuredinterviews

A qualitative study was conducted using semi-structured face-to-face interviews withneonatologistsandneonatalnursesworkinginaFlemishneonatalintensivecareunit.Wechose

167

aqualitativeresearchmethodologytocoverthecomplexity,subtletyandindividualspecificityofexperiencesintheend-of-lifedecision-makingprocessregardingneonatesthatwouldbemissedby a quantitative approach. Because of the sensitivity of the subjectwe opted for individualinterviews.

We recruited neonatologistsworking as resident physicians at one of four Flemish neonatalintensivecareunits (UniversityhospitalsofGhent,BrusselsandLeuven,andgeneralhospitalSint-JanBruges)betweenDecember2017andJuly2018whohadbeentheattending/treatingphysiciantoatleastonechildwhohaddiedatthewardwhereanend-of-lifedecisionwasmadeinthepastyear,andnurseswhohadbeenthemostinvolved.Noexclusioncriteriawereused.Aneonatologistofeachparticipatinghospitalinformedallneonatologistsandnurseswithintheirrespectivewardof thepurposeof the study, andprovided contactdetails of thosewilling toparticipate. Researchers contacted them and set up a date for the interview either at theirneonatalintensivecareunitorattheirhomeresidency.Purposefulsamplingwasusedtoselectparticipants.

A topic guide (see Appendix 5 and 6) was developed by a multidisciplinary team of nineexperiencedresearchersinthefieldsofend-of-lifecareandneonatology.Participantswereaskedwhatmadeiteasierormoredifficulttomakeend-of-lifedecisions.Beforetheinterview,ashortquestionnairewasadministeredtocollectsocio-demographicdata.Datawerecollecteduntilnonewbarriersandfacilitatorsemergedforbothneonatologistsandnursesseparately,anddatasaturationwasachieved.

Summaryofthemainfindings

Examining end-of-life decisions in stillbirths, neonates and infants in Flanders,Belgiumonapopulationlevel

Developingamethodologytostudytheprevalenceofend-of-lifedecisionsbeforeandafterbirth

Inchapter2wepresentedastudydesignaimedtoevaluateandmonitortheprevalenceofprenatalandneonatalend-of-lifedecisionsonapopulationlevelinFlanders,Belgium.Thisstudydesign involved thedevelopmentof a validated conceptual frameworkof end-of-lifedecisions across the entire foetal-infantile period and the development of a surveymethodology tostudy these foetal-infantileend-of-lifedecisions independentof thesettingwithinwhichthedeathorstillbirthtookplace.

Wecreatedanew,all-encompassingframeworktoconceptualizeend-of-lifedecisionsintheentirefoetal-infantileperiod,includingbothdeathsbeforebirthfromaviableageofthefoetusonwards(from22weeksofgestationorabirthweightof500gramormore)andlivebornneonates who died before the age of one year. Two dimensions were deemed important,namelythemedico-technicaldimensionthatclassifiedthemedicalactthatwasposed,andthemedico-ethical classification that classified the life-shortening intention of the physicianassociatedwith thatmedicalact. Intermsofmedicalacts,adistinctionwasmadebetween

168

non-treatmentdecisionssuchaswithholdingorwithdrawinglife-supportingtreatment,andadministering drugs or performing active medical interventions with a possible life-shorteningeffect.Thelife-shorteningintentionofthephysicianontheotherhandcouldeitherbe: 1) no intention to shorten life, yet a potentially life-shortening effect was taken intoaccount,2)apotentiallylife-shorteningeffectwaspartlyintended,yetnotthemainaimofthemedicalact,and3)the life-shortening intentionwasexplicit.Basedon this framework,wedevelopedtwoseparatebutsimilarquestionnairesinordertoexamineend-of-lifedecisionsinstillbornsandneonatesrespectively.

Ouraimwastostudyend-of-lifedecisionsinlive-borninfantsandstillbornsfrom22weeksofgestationonwardsonapopulationlevel.Basedonpreviousexperienceinneonates,childrenandadults,usingdeathcertificatesasthebasisforsendingoutquestionnaireswasdeemedideal. For stillbirths between 22 and 26 weeks of gestation, this procedure proved to bechallengingasadeathcertificateisnotmandatoryunder26weeksofgestationandthusoursamplingframeworkbymeansofdeathcertificatescouldpotentiallybeincomplete.However,usingtheonlyotherregistryofallstillbirths,namelythebirthregistry(livebornandstillborn)oftheStudycentreforPerinatalEpidemiology(SPE),woulddrasticallydecreasethereliabilityof our responses since delays in processing these documents could take up to one year.Therefore, we chose to rely on the robust mortality follow back survey-method for bothdeceased neonates and stillbirths, with some minor adjustments to improve coverage ofstillbirthsinthecrucialperiodbetween22and26weeksofgestation.TheFlemishAgencyofCareandHealth,whichprocessesalldeathcertificates, startedencouragingregistrationofstillbirthsfrom22weeksonwardsforepidemiologicalpurposesduringthedata-collectionofour study. In addition to this method, we provided our questionnaires to the ten largestmaternitywards inFlanders so thatphysicianswere encouraged to fill themout for eachstillbirth from22weeksof gestationonwards, inaddition to fillingout the accompanyingdeathcertificate.

Physiciansfilledoutthemainpartofadeathcertificateforeveryneonataldeathorstillbirth,whichincludeddemographicandmedicalinformation.Afterwards,thecentraladministrationauthorities, in our case the Flemish Agency of Care and Health, received the filled-outcertificates.TheAgencywasresponsibleforsendingoutquestionnairesandaccompanyingletterswithpatientinformationtophysiciansforeachdeathcertificatedenotingthedeathofaneonateorastillbirth from22weeksofgestationonwards.Thephysician identified theinfant, according to the information on the accompanying letter, and filled out thequestionnaire. All filled-out questionnaires were sent to a lawyer, who was bound byconfidentiality and thus safeguarded the anonymity of the physician, patient, parents andhospitals.Afterdatacollectionwasfinished,thelawyerlinkeddatafromthequestionnaireswithinformationonthedeathcertificates.

Thedevelopedresearchprotocol is the firsttostudyend-of-lifedecisions instillbornsanddeceasedneonatesand infantsunder theageofoneyearonapopulation levelwithinonestudydesign.Weareconvincedthatregularrepetitionofthisstudyinthefutureisneededinordertomonitorandevaluatechangesinend-of-lifepracticesundereverchangingsocietal,legalandclinicalinfluencesinavulnerablegroupoffoetusesandinfantswhoareunabletospeak for themselves. By basing inclusion of all deaths and all stillbirths on the death

169

certificates,thisresearchmethodcanbeusedinothercountries,irrespectiveofdifferentlegalframeworks regarding perinatal end-of-life decision-making, making internationalcomparativestudiespossible.Providingtheseprevalenceestimates,notonlyinFlanders,caneventuallyaidthedevelopmentofobstetric,neonatalandpaediatricguidelinestosupportaverydifficultethicalend-of-lifedecision-makingprocessindailypractice.

Theprevalenceofend-of-lifedecisionsintheneonatalperiod

Chapter3ofthisdissertationfocusedonprovidingpopulationestimatesoftheprevalenceofend-of-lifedecisionsinneonatesandinfantsinFlandersovertwostudyperiods(1999-2000and2016-2017).Theseestimateswereexaminedbymeansofthedevelopedpopulation-levelmortalityfollow-backsurveywedescribedinthepreviousparagraphs.

Atotalnumberof276neonatesandinfantsdiedbetweenSeptember1st2016andDecember31st2017(229filled-outquestionnairesreceived,83%responserate);and292neonatesandinfants died between August 1st 1999 and July 31st 2000 (253 filled-out questionnairesreceived,87%responserate).Studyresultsshowedthattheprevalenceofneonatalend-of-lifedecisionshasstayedrelativelystableacrossbothtime-pointsatabout60%ofneonatalandinfantdeathsbeingprecededbyanend-of-lifedecision.Non-treatmentdecisionsarestillthemostprevalentat34%ofallneonatalandinfantdeathsin1999-2000,comparedto37%in2016-2017.Withholdingtreatmentoccurred in13%ofallneonataland infantdeathsin1999-2000and12%in2016-2017,whilewithdrawing treatmentwasprevalent in21%ofcasesin1999-2000and25%in2016-2017.Administeringmedicationwithapotentiallylife-shorteningeffectstayedrelativelystableat16%in1999-2000comparedto14%in2016-2017, while the prevalence of administering medication with an explicit life-shorteningintentionoccurred inasimilargroupof7%in1999-2000and10%in2016-2017.Despitestableprevalenceratesoverall,importantshiftsinthetypeofend-of-lifedecisionbeingmadeindifferentagegroupswerenoted.End-of-lifedecisionswerenowsignificantlymoreoftentakenafterthefirstweekoflife(74%ofdeathsbetween7and27daysoldwasprecededbyanend-of-lifedecisionin2016-2017comparedto50%in1999-2000,p=0.03;64%ofdeathsafter 27 days of life in 2016-2017 compared to 38% in 1999-2000, p=0.003). In deathsoccurring in the firstweekof life, prevalenceof end-of-lifedecisions significantlydropped(55%ofdeathsin2016-2017comparedto72%in1999-2000,p=0.01).Afterthefirstweekoflife, end-of-life practice in Flanders considerably changed compared to 17 years ago, asdecisionstowithdrawlife-sustainingtreatmentoradministermedicationwithanexplicitlife-shorteningintentionbecomenoticeablymoreprevalent.In1999-20009%ofdeathsbetween7and27dayswasprecededbyadecisiontowithdrawtreatmentandtherewerenocaseswheremedicationwithanexplicitlife-shorteningintentionwasadministered,whilein2016-2017 in the same age group, withdrawing treatment and administering medication withexplicitlife-shorteningintentionwereeachprevalentin26%ofcases.Afterthefirst27daysoflife,theprevalenceofwithdrawingtreatmentrosefrom16%in1999-2000to31%in2016-2017,andtheprevalenceofadministeringmedicationwithexplicitlife-shorteningintentionrosefrom2%to10%.

This chapter shows that end-of-life decisions continue to be an integral part of medicalpracticeinextremelyillneonatesandinfants,withthreeinfivedeathsbeingprecededbysuch

170

decisions,whichindicatestheneedtodiscusstheirpermissibilityandrequirementsforgoodclinicalpracticeamongsthealthcareproviders.

Attitudes,viewsandexperiencesofhealthcareprovidersinvolvedinneonatalend-of-lifedecision-making

Attitudesofneonatologistsandneonatalnursesconcerningperinatalend-of-lifedecisions

In chapter4we present the attitudes of neonatologists and neonatal nursesworking in aneonatalintensivecareunittowardsperinatalend-of-lifedecisions,examinedbymeansofafull-populationmailsurvey.

We found that overall, considerable support for both prenatal and neonatal end-of-lifedecisionscouldbenotedamongstFlemishneonatalhealthcareproviders.Intermsofprenatalend-of-life decisions, between 80 and 98% of neonatologists and nurses consideredtermination of pregnancy at a viable term acceptable in case of severe or lethal foetalanomalies.Whenthefoetusishealthy,yetthelifeofthemotherisindanger,morethan60%of neonatologists and nurses found termination of pregnancy at a viable term acceptable.However,when the foetus is healthy but themother has a severe psychological problem,acceptance rates drop to 15% in both physicians and nurses. In extremely ill livebornneonates,between80and100%ofallparticipatinghealthcareprovidersfoundnon-treatmentdecisionssuchaswithholdingorwithdrawingtreatmentacceptable,regardlessofwhetherthelife-shorteningeffectwassolelytakenintoaccountorexplicitlyintended.Asidefromgeneralconsensus between neonatologists and neonatal nurses on the abovementioned types ofprenatal and neonatal end-of-life decisions, some differences in attitudes between bothhealthcare providers could be noted. Administering medication with a potentially life-shorteningeffectwasconsideredacceptablebythemajorityofbothhealthcareproviders,yetneonatologistsweresignificantlymorelikelytoagreetothispractice(96%)thannursesdid(84%,p=0.02).Conversely,thoughmorethanhalfofbothhealthcareprovidersfoundactivelyadministeringmedicationwithanexplicit life-shortening intentionacceptable, thepracticewasmoreoftenconsideredacceptablebynurses(74%)thanbyneonatologists(60%,p=0.02).Thisdespite the fact thatactivelyhasteningdeathbymeansofmedicationiscurrentlynotlegallytoleratedwithintheBelgianlegalframework.

Ourstudythusfoundalargeacceptanceofbothprenatalandneonatalend-of-lifedecisionsinneonatologistsandneonatalnurses,evenfordecisionsthatcurrentlyfalloutsidetheBelgianlegal framework. However, physicians and nurses differed slightly in their acceptance ofdifferenttypesofend-of-lifedecisions,whichcouldpossiblyberelatedtonursesnotcarryingthefinallegalresponsibilityofthesemedicaldecisions.Thesefindingsindicatetheimportanceofincludingbothperspectivesinthesedifficultdecisionsattheendofaninfant’slife.

171

Barriersandfacilitatorsforneonatologistsandneonatalnursesregardingneonatalend-of-lifedecision-making

Inchapter5,weexploredbarriersandfacilitatingfactorsexperiencedbyneonatologistsandneonatalnursesduringtheend-of-lifedecision-makingprocessinaneonatalintensivecareunit.Hereby,weaimedtoprovideinsightinthecomplexityofneonatalend-of-lifedecisionsindailypractice,andtheindividualnatureofpersonalexperiencesonthistopic.

Somebarriersandfacilitatorsarelinkedwiththecharacteristicsofthespecificcase.Thesefactors relate to either the ill neonate, the parents or the involved healthcare providers.Decisions seemedeasierwhen thebadprognosiswas evident fairlyquicklyas opposed towhenthereisalotofprognosticinsecurity,andexploringallcurativetreatmentoptionsfirsttoensurethattheend-of-lifedecisionistheonlyavailableoptionlefttoreducesufferingofthechildhelpedmakedecisionseasier.Healthcareprovidersindicateaneasierdecision-makingprocess when parents have the same culture and language as the physicians and nursesinvolved.Previousexperienceofhealthcareproviderswithend-of-lifedecisionsisconsideredacrucialinfluencingfactor,astheyarebetterabletoanticipatethechild’sfuturecondition.

On a process level,we consider factors that are related to characteristics of the decision-makingprocessitself.Intensecommunicationbetweenhealthcareprovidersandparentsisimperative for an easier end-of-lifedecision-makingprocess. Furthermore, communicationamongsthealthcareprovidersisessential,forexamplebyinstallingregularmultidisciplinaryconsultationsordebriefings.Additionally, decidingon an end-of-lifedecision canbemadeeasierbyconsideringalldirectionsthechild’sconditioncantakeinadvanceduringoneormore advance care planning conversations between parents and healthcare providers.Hereby, decisions regarding themedical responses in each situation canbemadewithoutbeingrushedbyanacutedeteriorationofthechild.

Afinallevelincludesfactorsrelatingtotheoverarchingstructureoftheward,thehospitalandthebroader society that could influencedecision-making.Emotionalandpracticalsupportfrom colleagues at the ward, or lack thereof, is crucial in end-of-life decision-making inneonates. Additionally, the lack of separate rooms to ensure privacy during bad-newsconversations,andtheshortageofavailabletrainedpersonnelinend-of-lifecarewereclearlyidentifiedasbarriers forend-of-lifedecision-making.Lastly, thecurrentBelgianlegislationwasmentionedasaninfluencingfactor.Whenmentioned,neonatologistsandnursesstatedthat theyexperience the lackofalegal frameworktoallowforactivelyendingthe lifeofaneonate in extreme cases to be an important barrier, especially in contrast to during thepregnancy, where the option to terminate as soon as a life-limiting foetal abnormality isdiagnosedisavailable.

Ourqualitativeinterviewstudyrevealedbarriersandfacilitatorsduringneonatalend-of-lifedecision-makingwhich could lead to recommendations for improving thisprocess indailypractice.Theserecommendationsincludeestablishingregularmultidisciplinarymeetingstoincludeallhealthcareprovidersandreduceunnecessaryuncertainty,routinelyimplementingadvance care planning in severely ill neonates to make important decisions beforehand,

172

creatingprivacy forbad-newsconversationswithparentsandreviewing thecomplex legalframeworkofperinatalend-of-lifedecision-making.

Psychologicalsupportinend-of-lifedecision-makingforneonatologistsandneonatalnurses

Chapter 6 of this dissertation focussed on the perceived stress that neonatologists andneonatalnursesexperienceduringanend-of-lifedecision-makingprocessintheirneonatalintensivecareunit,andtheirperceivedpsychologicalsupportbothfromcolleaguesandfromprofessionals.Thiswasexaminedbymeansofafull-populationmailsurvey.

The majority of neonatologists and nurses agreed that making an end-of-life decision(neonatologists)orbeingconfrontedbyone(nurses)causedmorestressthanusual(73%and70%respectively).Duringthedecision-makingprocessfortheseend-of-lifedecisions,mostphysicians(86%)indicatedthattheyfeltsupportedbytheircolleagues.However,fewerthanhalfoftheneonatalnurses(45%)agreedthatthephysicianslistenedtotheiropinionswhenthesedecisionswerebeingmade.Whilemostneonatologists(88%)agreedthattheirneonatalintensive careunit provides sufficient opportunity to express any reservations theymighthave about certain end-of-life decisions, only 32% of nurses agreed with this statement.Almostalloftheparticipatingneonatologistsandneonatalnursesagreedthattheycantalktotheircolleagueswhensomething isbotheringthemregardinganend-of-lifedecision(94%and92%respectively).Furthermore,whentheydidnotagreewithanend-of-lifedecisionthathadbeenmade,halfofneonatologists(53%)and65%ofnursesagreedthattheycouldopttonolongerbeinvolvedinthatparticularcase.Despitethefactthatbothgroupsofhealthcareprovidersindicatedthattheycouldtalktotheircolleagueswhensomethingregardingend-of-life decision-making bothered them, 57% of neonatologists and 60% of neonatal nursesindicated that they would prefer their neonatal intensive care unit to provide morepsychologicalsupport forstaffmemberswhentheywerebeingconfrontedwithend-of-lifedecisions. Furthermore, only 41% of neonatologists and 50% of nurses agreed that theyreceivedsufficientpsychologicalsupportfromtheirneonatalintensivecareunitafterapatientoftheirsdied.

Our findings seem to suggest that neonatal intensive care units need professional ad hoccounselling or standard debriefings, as we believe they could substantially improve theperceivedlackofsupportindicatedbycliniciansworkingattheneonatalintensivecareunit.Furthermore, we believe that including nurses in interdisciplinary end-of-life discussionscouldnotonlyincreasethequalityofthesedecisions,butcouldpossiblyalsobenefitthenursesthemselvesbyreducingmoraldistresscausedbybeingexcludedfromthedecision-making.

173

Discussionofthemainfindings

Comparing Flemish neonatal end-of-life decision-making with internationallyavailableevidence

Neonatalmortalityvarieswidelyacrosscountries.Asidefromdifferencesinneonatalmortality,differencesinethicalperspectiveexistbetweencountriesintheacceptabilityanduseofmedicaldecisions at the end of an infant’s life. It is therefore necessary to compare the informationgatheredwithin this dissertation on neonatal end-of-life decisions in Flanders, Belgiumwithinternationally available data in order to unveil country-specific factors influencingdecision-making.

Internationalcomparisonofthepracticeofnon-treatmentdecisions

Generally,inEurope,non-treatmentdecisionssuchaswithholdingorwithdrawingtreatmentarewellaccepted,andthemajorityofphysiciansworkinginneonatalintensivecarereporthaving been involved in at least one case in which limits to intensive care were set.Internationally, the likelihood of limiting intensive treatment in neonates is known to bedependentonthepositiveornegativeattitudeofphysicianstowardsthesetypesofend-of-lifedecisions. The positive attitude of Flemish neonatal healthcare providers towards non-treatmentdecisionsandthecorrespondinghighprevalenceofthesetypesofdecisionsintheentirepopulationofneonataldeathsbeforetheageofoneyearreportedinthisdissertationcorroborate these findings.We could therefore hypothesize that our prevalence estimatesconcerning non-treatment decisions could possibly be comparable to those of Europeancountrieswithasimilarlypositivestanceontheseend-of-lifedecisionssuchastheUKandtheNetherlands. Physiciansworking inneonatal healthcare inEuropean countriessuch as theBalticstates,Italy,SpainandGermanyhaveastrongerpro-lifeattitude.WhileFlandersandcountriessuchastheUKandtheNetherlandscouldthusbeconsideredtohaveapermissiveattitude towards non-treatment decisions with a potentially life-shortening effect, otherEuropeancountriesmightbemorerestrictive.

Theprevalenceofnon-treatmentdecisionsinFlandersin2016-2017was37%ofallneonataldeaths(chapter3).ThisisslightlyhigherthanthatoftheNetherlands,whichwasestimatedat31%in2010.ReportsfromneonatalintensivecarecentresintheUnitedStates,theUnitedKingdom,AustraliaandEuropeshowthatbetween40and93%ofneonataldeathsoccurafterwithholding or withdrawing artificial ventilation or other life-sustaining treatments. Thedifferencebetweenthesepopulation-basedestimates(37%inourstudy)andtheprevalenceestimatesofthenumberofdeathsinspecializedneonatalintensivecareunitsinternationallybeingprecededbyanon-treatmentdecision(between40and93%)canberelatedtoseveralfactorsincludingsomemethodologicaldifferencesinassessingprevalence,andadifferencebetween assessing prevalence estimates within highly specialized single centre studiescomparedwithbroadpopulationdataacrosssettings.Aspopulation-levelstudiesarescarce,actual international comparisons between our population estimates and estimates fromcountriesoutsideoftheNetherlandswereimpossible,whichshoulddefinitelyberemediedin

174

futurestudiesinordertoexaminecountry-specificinfluencesonclinicalpractice.Thestudydesigndescribedinchapter2ofthisdissertationwouldbeidealforthispurpose.

Internationalcomparisonsoftheadministrationofmedicationwithanimplicitorexplicitlife-shorteningintention

The findingswithin thisdissertation indicate thatFlandershasa fairlypermissive climatetowardsmoreactivetypesofend-of-lifedecisionssuchasadministeringmedicationwithapotentialorexplicitlife-shorteningintention,evenwhenthesedecisionscurrentlyfalloutsideoftheBelgianlegalframework.

Weseethatthelife-shorteningintentionofadministeringmedicationbeingeitherimplicitorexplicit makes a crucial difference in whether the Flemish permissive attitude could becorroborated internationally. In Switzerland, 95% of physicians and nurses working in aneonatalintensivecareunitfoundadministeringsedativesoranalgesicsacceptable,evenifthis might cause respiratory depression and death. However, when the life-shorteningintention of administering medication becomes explicit, acceptance rates of Swissneonatologists and nursesdrop to 24%. InCanada, a survey on the attitudes of Canadianpaediatricians revealed a collective unease towards non-voluntary euthanasia in never-competentchildren,suggestingthatCanadianpaediatriciansandneonatologistsmightbealotlesspermissivethanthoseinFlanders.InFrance,amultidisciplinaryworkinggrouponethicalissuesinperinatalmedicineevenstatedthatactstodeliberatelyhastenapatient’sdeathareboth legally and morally forbidden, indicating an even more restrictive attitude. Flemishneonatalhealthcareprovidersmightthusbemuchmorepermissivetowardsadministeringmedicationwithapotentialorexplicitlife-shorteningintentioninextremelyillneonatesandinfantsthanhealthcareprovidersinothercountries.

Asidefromareflectiononinternationalattitudestowardstheadministrationofmedicationwith a potential or explicit life-shortening intention, international prevalence estimatesshould be considered. A multi-national study (EURONIC) in eight European countries(Belgium not included) revealed that between 32% and 89% of physicians working in aneonatalintensivecareunithadpreviouslyadministeredpainandsymptomrelief,despitetherisk of respiratory depression and even death. These numbers varied greatly betweencountries, with France, the Netherlands and Sweden reporting the highest number ofphysicianswithpreviousexperienceinadministeringsedativesandanalgesicsevenattheriskofhasteningdeath(86-89%),andItalybeingtheonlyreportedcountrywithratesunder50%(namely 32%). Furthermore, this study revealed that administering medication with thepurposeofendinglifeinneonatesoccursveryrarelyinthemajorityofreportedEuropeancountries.Only2-4%ofphysiciansworkinginaneonatalintensivecareunitinItaly,Spain,Sweden,GermanyandtheUKreportedeverhavingtakenthesetypesofdecisions.Inarecentfollow-upstudyoftheEURONICstudyin2016inGermany,SwitzerlandandAustria,97%ofphysiciansreportedinanonlinesurveythattheyhaveadministeredsedativesandanalgesicseven at the risk of potentially hastening death at least once. When considering theadministrationofdrugswithanexplicitlife-shorteningintentionhowever,theproportionofphysicianseverhavingmadethisdecisioninGermany,SwitzerlandandAustriadropsto4%.IntheEURONICstudies,physicianswereaskedwhetherornottheyhadeverpreviouslytaken

175

specific typesofneonatalend-of-lifedecisionsand thereforewehaveno indicationofhowfrequent thesedecisionsactuallyare.Providingaclearcomparisonbetweendata fromtheEURONICstudiesandourprevalenceestimatesofadministeringmedicationwithapotentialor explicit life-shortening intentionwithin an entire population of deceased neonates andinfantsduringasetperiodoftimethereforeprovestobedifficult.

The only available prevalence estimates regarding the use of medication to implicitly orexplicitlyhastendeathinneonatesisfromtheNetherlands,indicatingthatmedicalpracticemight not be as comparable to Flanders as previously suspected. Where our prevalenceestimates indicate that 14% of all neonatal deaths in 2016-2017 were preceded by theadministration of medication with a potentially life-shortening effect, the Netherlandsreportedaprevalenceofjust4%in2010.Whenlookingattheprevalenceofadministeringmedication with an explicit life-shortening intention, the difference between the 10%prevalenceestimatesofFlandersin2016-2017withtheDutchestimateof1%in2010isevenmorestriking,particularlyconsidering that theNetherlandshavea legal frameworkwhichpermits such decisions in rare cases of extremely ill infants where Flanders does not.Availabilityofalegalframeworkthusnotnecessarilyleadstoanincreaseintheprevalenceofthispractice.However,thelackofinternationalpopulationestimatesmakesitimpossibletodraw robust and valid conclusions on the impact of these different legislative choicesregardingthepermissibilityandrulesforgoodclinicalpracticeofvariousneonatalend-of-lifedecisionsontheiractualprevalence.

ThepossibleimpactoftheBelgianlegalcontextonneonatalend-of-lifedecisions

The permissive climate towards neonatal end-of-life decisions amongst Belgian neonatalhealthcareprovidersmustbeviewedwithinthebroadercontextoftheBelgianlegalandmedicalculture. As Belgium has both a fairly liberal law on termination of pregnancy and a law oneuthanasiainadultsandcompetentminors(seechapter1),itcouldbedebatedthattheBelgiummedicalandlegalcultureasawholecouldbeconsideredasmoreacceptingofcertaindecisionsattheendofaperson’sliferegardlessoftheiragethanisinternationallythecase.

Asneonatesarenotcompetenttorequest,andreceive,euthanasia,theycouldthusbeclassifiedas a vulnerable group that falls outside the jurisdiction of the euthanasia law, but yet couldexperience an influence of its implementation. If this is the case, the implementation of theeuthanasialaw foradultsandcompetentminorsshould lead to1)an increaseofdeliberatelyendingthelifeofneonateswithsevereconditions,and2)thisincreaseshouldbeattributedtophysiciansfeelingmoreateasewiththepracticeexplicitlyduetotheexistenceorextensionoftheeuthanasialaw.Prevalenceestimatesprovidedinthisdissertationcanonlyprovideinsightintothefirstclaim,namelythattheprevalenceofadministeringmedicationwithanexplicitlife-shortening intention stayed relatively constantat7% twoyearsbefore; and10%of the totalpopulationofdeceasedneonatesandinfants15yearsafterimplementationoftheeuthanasialawinadults(threeyearsafter theadditionofcompetentminors).Furthermore,acausalrelationbetweenthe implementationof the euthanasia law foradultsandcompetentminors,andtheconsiderable prevalence of administration of medication to intentionally hasten death inneonates,asmentionedinpoint2,cannotbeprovenbythedatapresentedinthisdissertation.

176

As healthcare providers specifically mention the influence of the existence of the law ontermination of pregnancy on neonatal end-of-life decision-making, andmore specifically therestrictionsanduneasiness theysometimes feelwhentheyareunable to intentionallyhastendeathinsufferingneonatesandinfantswhenapregnancycouldbeterminatedforexactlythesamediagnosisinanunbornfoetus,itmightbepossiblethattheexistenceofaliberalterminationofpregnancylawhasaninfluenceondecision-makingafterbirth.However,astheBelgianlawontermination of pregnancywas instated in 1990, prevalence estimates on neonatal end-of-lifedecisionsfrom1999-2000and2016-2017discussedinthisdissertationwillthusnotbeabletopoint out any changes following the implementation of this law. Furthermore, whether thepermissiveattitudeofhealthcareproviders inperinatalcare followsrather thanprecedes theimplementationoftheterminationofpregnancylawcanbedebated.

When comparing prevalence estimates of administering medication with an explicit life-shortening intention in Flanders - a regionwhere this practice is currently not regulated bymeans of a protocol or a law - with the Netherlands - who chose to provide guidelines andregulations forbestpractice–,wesee that thispracticeoccursmoreoftendespite thelackofregulation. Our prevalence estimates on a practice that is currently not legally tolerated,combined with the knowledge that attitudes of Flemish neonatal healthcare professionalstowardsadministeringmedicationwithexplicitlife-shorteningintentionarepermissible,raisesthequestionofwhetherguidelines,protocolsorlawsareneededtomonitorthesedecisionsinsuch a vulnerable patient group. However, the existence of apermissive attitude of involvedhealthcareproviderstowardsthesedecisions,andtheexistenceofempiricalevidenceindicatingthat thesedecisionsare actuallymade indaily clinical practicedonot automaticallywarrantsupporttowardstheselegislativechanges.Whileourinterviewstudyshowedthatneonatologistsandnursesfindthelackofalawallowingforactivelyhasteningdeathbymeansofmedicationinsevere cases to be a barrier in decision-making, they also indicated to be wary of possiblestandardisationbymeansofalaw.Whiletheexistenceofaprotocoloralawtolegallyallowthesedecisionsmightaiddecision-making,andcouldpossiblyprovideguidelinestowardswhatwouldbeconsideredbestpracticeinthesecases,cautioniswarranted.Thisextremelysensitiveissueneedsfurtherinterdisciplinarydebate,includingphysicians,ethicistsandpolicymakers.

Supportforhealthcareprovidersduringtheneonatalend-of-lifedecision-makingprocess

Akeyfindingofthisdissertationisthatpsychologicalandpsychosocialsupportforhealthcareprovidersworkinginneonatalend-of-lifecareiscurrentlylacking(chapter5and6).Bothbeingpartofanend-of-lifedecision-makingprocessandexperiencingthedeathofaneonateintheircare causes a considerable amount of stress for involved physicians and nurses. Duringinterviews,neonatologistsandneonatalnursescontinuouslystressedhowdealingwithseverelyill newborns canweigh on their emotionalwellbeing, especiallywhen the infant looks like ahealthy, full termbaby, orwhenhealthcareprovidershaveyoung infantsof their ownwhichcausesthemtoprojectthehardshipstheyviewandexperienceonthejobontheirownhouseholdsituation.Additionally,theyindicatedthatbeingpartofaneonatalend-of-lifedecision-makingprocessisnevereasy,andthatdiagnosticandprognosticinsecuritycanheavilyweighontheirstateofmind.

177

Tocopewiththeelevatedamountsofstressduetobeingconfrontedwithend-of-lifedecisionsandinfantdeathonaregularbasis,ourstudiesshowedthathealthcareprovidersturntotheirpeer colleagues for support. Though the positive impact of collegial support from peers onwellbeingofthehealthcareprovidersshouldnotbeoverlooked,itisnotsufficienttocopewiththe stressors associated with end-of-life decision-making and infant-death in the Flemishneonatalintensivecareunits.Counsellingforbereavedparentsafterperinatallosstohelpthemcope ismuchmore readilyavailable than it is forhealthcareproviders, as they areoftennotrecognizedasabereavedpersonbysocietyortheirworkenvironment.Asaresultofthis,mostrecommendationsandguidelinesonpsychosocialsupportduringdeathandend-of-lifecareinneonates focusses on providing physicians, nurses and other healthcare professionals withconcretetoolstooptimallyattendtoparentsintheirdecision-makingprocessandgrief.Caringforthehealthcareprovidersinthiscasebecomessecondaryorevennon-existent,eventhoughthe addedemotionaldistressofdealingwith these extremelydifficult decisions regularly canprove to bemore than they can copewith. Healthcare providerswho suffer from emotionaldistressandevenburn-outarefurthermoreknowntohaveadiminishedcapacitytocarefor,andshowempathytowardstheillneonatesintheircareandtheirparents.Caringforthehealthcareproviders might thus not only benefit their wellbeing on a personal level, but it might alsoconsiderablyimprovetheirabilitytocarefortheinfantsandsupportthefamilies.Thelackofprofessionalsupportforhealthcareprovidersworkinginaneonatalintensivecareunit,asshownbyseveralstudiesinthisdissertation,shouldthusobviouslybeaddressedandresolved.

Theroleofpalliativecareinneonatalend-of-lifedecisionmaking

In caring for extremely ill neonates and infants, deciding to either reorient care from cure-orientedandlife-extendingtocomfortandpalliativecare,ortoprovidebothcure-orientedandcomfort care concurrently, ispart of daily clinicalpractice. Inprenatal andneonatalpractice,palliative care is a relatively new field. The role of end-of-life decisions, and the possibleimplications of data provided within this dissertation, within such a perinatal or neonatalpalliativecareapproachisthusstillunclear.

When parents receive a life-limiting diagnosis for their child, it is extremely important thathealthcareprovidersprovidethemwithanempathetic,understandableandbalancedoverviewofalltreatmentoptions,includingactiveandcure-orientedinterventions,end-of-lifedecisions,andpalliativecare.As thepracticeofwithholdingorwithdrawingunnecessarily invasive life-supporting treatment inmodernneonatal intensivecareunitsglobally iswellsupported,andresultsofthisdissertationcorroboratedtheircentralroleinneonatalend-of-lifecare(chapter3),we can expect these conversations between healthcare providers and parents to includediscussing non-treatment decisions. Furthermore, we can expect decisions to withhold orwithdrawunnecessarilyinvasivetreatmenttobeanintegralpartofashiftfromcurativecaretopalliative care. A prime example of this is the practice of compassionate extubation, whereassistedventilationwhichisoftenvitalforsurvivalofthechildiswithdrawntoincreasecomfort.Asidefromotherimportantcomponentssuchasadvancecareplanningandsupportforparents,providingadequatepainreliefandcomfortareacrucialcomponentofperinatalandneonatal(palliative) care. Withholding or withdrawing treatment is therefore usually followed byincreasinganalgesicsandsedativestotreatthedyingneonates’painandsuffering.Thehighdosesofpainmedicationneededtoprovidereliefforsufferingneonateswithinthecontextofproviding

178

goodpalliativecarecould,andareoften(chapter3)givenevenwhenalife-shorteningeffectwasforeseenorevenintended.

Embeddingtheend-of-lifedecision-makingprocesswithinaneonatalorperinatalpalliativecareapproachcouldbeusefultoaddressthecomplexfamilyneedsinanemotionallyturbulenttimebyprovidingafamily-centredapproachwithafocusonparental(spiritualandcultural)values,memory making, and compassionate communication between parents and providers.Additionally, such a palliative care approach could benefit healthcare providers, as existingperinatalpalliativecareprotocolsoftenincludesectionsonpsychosocialstaffsupport,usedtoimprove quality of care and counteract moral distress, burnout and compassion fatigue. Aneonatalorperinatalpalliativecareapproach thus includesnotonlyadequatepainreliefandcomfort for the child, but also has a strong emphasis on compassionate communication andpsychosocialsupportforparents,familymembersandinvolvedhealthcareproviders.

Implicationsandrecommendations

Implicationsandrecommendationsforpractice

- Attentionshouldbegiventocreatingaprivateroomforbad-newsconversationsintheneonatal intensive careunitand inotherhospitalwardswere such conversations areprevalentandnecessary.

- Installingaroutineuseofadvancecareplanningwithparentsinneonateswithasevereprognosiscouldaiddifficultdecisions.

- Prognostic uncertainty can be reduced by installing regular multidisciplinary teammeetings and debriefings, and routinely asking for a second opinion from otherphysicians.

- Difficultiesinworkingwithparentswhohaveadifferentculturalbackgroundorspeakadifferent language than that of the involved healthcare providers could possibly bereducedbyconsultinganeonatalorperinatalpalliativecareteam.Neonatalandperinatalpalliativecareteamsputampleemphasisonconversationaltrainingandcompassionatecommunicationbetweenparentsandhealthcareproviders,makingthemideallyplacedtomediateduringdifficultend-of-lifedecision-makingprocesses.

- Neonatologists,neonatalnursesandotherhealthcareprofessionalsworkinginaneonatalintensivecareunitshoulddevelopgeneralistpalliativecareskills. InBelgium, there iscurrently no formal training on neonatal palliative care available to aid healthcareprovidersinattainingtheseneonatalpalliativecareskills.Includingamoduleonneonataldeath and end-of-life decision-making in standard curricula for neonatologists andneonatalnursesincreasesclinicalexperienceandend-of-lifecommunicationskillsearlyonintraining,whichleadstoenhancedconfidenceandfewernegativeexperienceswithend-of-lifecareintheneonatalintensivecareunits.

- We suggest the implementation of regular formal debriefings with the entire teamresponsibleforcaringforaneonatewhodiedwithintheunit.Hereby,opportunitiesarecreatedtoreviewanddiscusswhatcouldhavebeenimproved,whichcouldaidinfutureend-of-lifecases.

179

- Werecommendcounsellingsessionsforhealthcareproviderswhowereinvolvedinend-of-life cases during regular work hours, as opposed to them attending counsellingsessionsonavoluntarybasisorduringunpaidtime.

Implicationsandrecommendationsforpolicy

- Withoutreliablepopulation-levelprevalenceestimates,ethicalandlegaldiscussions,andeven legislativedecision-making, arebasedon experiences andviewpoints of a selectnumberofconsultedexperts,whilepopulation-datacanprovideanactualempiricalbasisonifandhowoftenvariousend-of-lifepracticesoccurinthepopulation.Withinanever-changingsocietywherethereisacontinuousriseinmedicalpossibilitiestosavethelifeofneonateswithseverehealthconcerns,systematicmonitoringofend-of-lifedecisionsonapopulation-levelonaregularbasisisparamount.Policymakersshouldsupportthisrecurrentperiodicmonitoring inorder tobeawareofanysignificantchanges indailypracticewhichmightwarrantlegislativeorpolicy-changes.

- As current national and international guidelines on the acceptability and adequateperformanceoftheseneonatalend-of-lifedecisionsarelacking,theremightstillbealotof uncertainty among physicians and nurses regarding their permissibility andrequirements for good clinical practice.Ourprevalence estimates canprovide expertswithastartingpointtodiscussthepossibleformulationoftheseguidelinesorlegislativealternatives further. Additionally, the prevalence estimates and possible barriers andfacilitators healthcare providers experience during a neonatal end-of-life decision-making process discussed within this dissertation might be an ideal starting pointtowards formulating aids and guidelines towardswhat is considered best practice inthesecases.

Implicationsandrecommendationsforfutureresearch

- Inputfrom(bereaved)parentswasmissingfromthenarrativeofthisdissertation,yetparentalviewsarecrucialtoprovideacomprehensivepictureofaneonatalend-of-lifedecision-makingprocess.

- End-of-lifedecision-makinginneonatesisirrevocablyconnectedtoprenatalend-of-lifedecision-making, as a lot of congenital disorders or anomalies can be diagnosedprenatally. Future research should continue to include both prenatal and neonataldecisions,usingtheframeworkprovidedinchapter2.Additionally,asshowninchapter4,whenexaminingattitudes,opinionsandexperiencesofinvolvedhealthcareproviders,bothprenatalandneonatalshouldbeconsidered.

- Futurestudiesshouldfocusoncollectingpopulationdatainternationally.Internationalcomparativeresearchcan identifycountry-specificor evenregion-specific factors thatmight influence the occurrence of end-of-life decisions and end-of-life practice.Furthermore,itcouldprovideevidenceofdifferingmedicalculturesconcerningneonatalend-of-lifecare.

- Neonatal end-of-lifedecisionsare embedded inneonatalandevenperinatalpalliativecare.Althoughcrucialelementsofapalliativecareapproacharealreadyimplementedinregular perinatal practice, the existence of actual perinatal palliative teamsinternationally is rare. As this is a relatively new and emerging research field that

180

addressesmuchneededsupportforextremelyillinfantsbeforeandafterbirthaswellasfor their families and involved healthcare providers, future research should focus onevaluatingthebestmodelofcarewithinthissetting.

181

-

182

Nederlandstaligesamenvatting

183

Inleiding

Indelaatstedecenniazijnhetaantalmedischeentechnischeinterventiesvoorhetbehandelenvanextreemziekepasgeborenenenneonatensterkgestegen.DesondankssterftinVlaanderenongeveer8.7perduizendkinderengedurendedefoeto-infantieleperiode:vanafdegeboortevaneenfoetusmeteengeboortegewichtvan>500gramofeenzwangerschapsduurvan22wekentotenmetdeleeftijdvanéénjaar.Hetgrootstedeelvandezeoverlijdensvindtplaatsineendienstneonatale intensieve zorgen en wordt voorafgegaan door een levenseindebeslissingmet eenmogelijkslevensverkortendeffect.Hierondervallenzowelniet-behandelbeslissingenzoalshetnietinstellenofstakenvaneenmogelijkslevensverlengendebehandeling,enhettoedienenvanmedicatie,beidemeteenmogelijkeofuitdrukkelijkelevensverkortendeintentie.Ookprenataalzienwedelaatstejareneenbelangrijkestijgingindekwaliteitvandiagnostischetechniekenzoalsgenetischescreeningenprenatalebeeldvorming,waardooreensteedsgroteraantalcongenitaleafwijkingenprenataalinplaatsvanneonataalkunnenwordenvastgesteld.Wanneerdithetgevaliskunnenlevenseindebeslissingenprenataalwordengemaakt,zoalshetnietinstellenvaneenactieveprenatalebehandelingofhetvroegtijdigafbrekenvandezwangerschap.

Hetethischdilemmainsommigevandezegevallentussenhetreddenvanhetlevenvandefoetusofpasgeboreneenonzekerheidoverdeprognoseop latere leeftijd,vereisteendoordachteenprofessionele afweging van zowel de ouders als de betrokken zorgverleners. Hoewel eenindividueleafwegingvandezeethischedilemma’sbijelkespecifiekecasenoodzakelijkis,kunnenpopulatiegegevensoverwat zich in vergelijkbaremedische situaties voordoetwaardevol zijnvoorbetrokkenzorgverleners,zekeringevalvanonzekerheidoverdeprognoseofhetbestaanvaneenmeningsverschiltussenbetrokkenpartijen.BeschikbaaronderzoekzowelbinnenBelgiëals in het buitenland is onvolledig of verouderd, en daarom ongeschikt als leidraad voor dehuidigepraktijk.Binnendestudiesvanditproefschrifthebbenweonsdaaromgerichtophetonderzoekenvandehuidigeklinischepraktijkrondhetmakenvanlevenseindebeslissingenineenkwetsbarepopulatievankinderenvanafeenlevensvatbarezwangerschapsduurtotenmetde leeftijd van één jaar. Het doel van dit proefschrift was tweeledig: 1) inzicht geven in deprevalentievanlevenseindebeslissingenbijdoodgeborenen,pasgeborenenenzuigelingen,en2)dieper ingaan op wat het betekent om als zorgverlener deel uit te maken van eenlevenseindebeslissingsprocesindedagelijksepraktijk.

Een overzicht van de beschikbare informatie inzake foeto-infantielelevenseindebeslissingen

Deprevalentievanfoeto-infantielelevenseindebeslissingen

Populatie-studieswaarbijalle sterfgevallenbinneneenbepaaldeperiodewordenbekeken,zijnideaalomaccurateprevalentieschattingenteleverenaangezienzeeentotaalbeeldgevenvandepraktijkonafhankelijkvandeziekenhuissettingofdediagnose.Dergelijkestudiesrondlevenseindebeslissingenzijnzeldzaambijpasgeborenenenzuigelingen,enzelfsonbestaandwanneerhetgaatomlevenseindebeslissingenbijdoodgeborenen.Bijpasgeborenenwordenprevalentieschattingen vaak gebaseerd op dossierstudies binnen één of enkele neonataleintensievezorgafdelingen.Bijditsoortstudieszienwedat40%tot93%vandesterfgevallen

184

binnenzulkedienstvolgtnahetstopzettenvaneenlevensverlengendebehandeling.Deenigebeschikbarepopulatie-studieszijnafkomstiguitNederlandin2014enBelgiëin2000.Dezestudiestoondenaandatlevenseindebeslissingengemaaktwordeninongeveer60%vanalleoverlijdensonderdeleeftijdvanéénjaar.Bijdoodgeborenenbestudeerdenbestaandestudiesin 2003 en 2000-2005 tot nu toe enkel de prevalentie van een laattijdigezwangerschapsafbreking inhet tweedeofderde trimester vande zwangerschap.Er isdusweinigbeschikbareinformatieoverlevenseindebeslissingenbijdoodgeborenenandersdanhet afbreken van de zwangerschap, over het voorafgaande beslissingsproces, en overpatiëntkarakteristieken naast zwangerschapsduur en de aan- of afwezigheid van foetaleafwijkingen.

Voorafgaandaandeontwikkelingvandestudiesbinnenditproefschriftwerddenoodvoorrecente, populatie-gebaseerde gegevens rond de prevalentie van levenseindebeslissingenaangegeven door Vlaamse zorgverleners uit zowel de prenatale als neonatale zorg.Aanbevelingen voor de klinische praktijk kunnenmomenteel enkel gebaseerd worden opverouderde en mogelijks niet langer relevante populatiegegevens door de constanteveranderingen op maatschappelijk, juridisch en klinisch vlak. Voorbeelden van dezemaatschappelijke veranderingen zijn onder andere de implementatie van wetten rondpatiëntenrechten, palliatieve zorg en euthanasie bij volwassenen in 2002; en dewet overeuthanasiebijkinderenin2014.Dezewettenzijnniettoepasbaarbinnenonzepopulatievanpasgeborenen en zuigelingen, aangezien ze beperkt zijn tot volwassenen enwilsbekwameminderjarigen.Tochkunnenzemogelijkeen invloedhebbenopdeprenataleenneonatalepraktijk.InternationaalwerdinNederlandhetGroningen-protocolgeïmplementeerd,wathettoedienenvanmedicatiemeteenexplicietelevensverkortendeintentiebijpasgeborenenenzuigelingenwettelijkmogelijkmaakt inextremegevallen.Dezenationaleen internationalemaatschappelijkeveranderingenhebbenmogelijkeen invloedgehadopdeprevalentievanlevenseindebeslissingeninBelgië.Hiernaastzouookdetoenameinmedischeentechnischeinterventiespre-enpostnataaleeninvloedkunnenhebbenopdeklinischepraktijk.Actueleenbetrouwbareprevalentieschattingenzijndusbroodnodigomzorgverleners,beleidsmakersen onderzoekers een beeld te geven van de huidige klinische praktijk vanlevenseindebeslissingenbinneneenzeerkwetsbarepopulatievankinderenonderdeleeftijdvanéénjaar.

Attitudesvanzorgverlenersomtrentfoeto-infantielelevenseindebeslissingen

Eerderonderzoektoondeeengrotevariabiliteitaaninwelketypeslevenseindebeslissingenwordengemaakt,zelfbijpasgeborenenmetdezelfdepathologie.Ditomdatbeslissingenaanhet levenseinde beïnvloed worden door een groot aantal contextuele factoren zoals debeschikbare middelen binnen de ziekenhuissetting; en de sociale, culturele en religieuzeovertuigingenvanoudersendebetrokkenzorgverleners.Naastdezecontextuelevariabelenspelen attitudes van zorgverleners een cruciale rol indebesluitvorming. Zelfs binnenéénzorgteam werden belangrijke verschillen gevonden tussen de houding van artsen enverpleegkundigentenaanzienvanlevenseindebeslissingen.Persoonlijkekenmerkenvandezorgverleners kunnen dus een cruciale rol spelen bij de besluitvorming rondlevenseindebeslissingenbijpasgeborenenenzuigelingen.

185

Uitvoorgaandonderzoeknaardeattitudesvanartsenin10Europeselandenin2000bleekdat de waarschijnlijkheid van het beperken van levensondersteunende behandelingen bijpasgeborenen sterk afhankelijk is van het land; de religie, het geslacht en het al dan niethebbenvankinderenvandebetrokkenarts;endeprevalentievankinderenmeteenextreemlaaggeboortegewicht opdedesbetreffendedienst.Bovendienbleekuit een zelfrapportagevragenlijst en een retrospectieve review van ziekenhuisdossiers dat een onbedoeldlevensverkortendeffect vanhet toedienenvanopioïdenacceptabelwordtgeacht voor eengrootdeelvanartsenwerkendopeenneonataleofpediatrischedienstintensievezorgen.Dezestudies geven een beperkt beeld van de praktijk aangezien ze geen rekening hieldenmetattitudestenopzichtevanbeslissingendievoordegeboortemogelijkwaren.Weachtenditbelangrijk omdat attitudes en beslissingen voor en nade geboorte zeer sterk gelinkt zijn.Neonatologen worden namelijk vaak geraadpleegd bij prenatale beslissingen rond hetlevenseinde, wat aantoont dat hun attitudes ten aanzien van prenatale en neonatalelevenseindebeslissingendusbestonderéénoverkoepelendonderzoekwordenbekeken,zodateenvergelijkingtussenbeidepraktijkenmogelijkwordt.Vanwegederelevantievanattitudesvanzorgverlenersbinnenhetdebatrondlevenseindebeslissingenpre-enpostnataal,waseenafzonderlijkdeelvanditproefschriftgewijdaanhetonderzoeknaarattitudesmetbetrekkingtot levenseindebeslissingen in de foeto-infantiele periode van neonatologen en neonataleverpleegkundigen.

Barrièresenfaciliterendefactoreninhetbeslissingsprocesvanneonatalelevenseindebeslissingenvoorbetrokkenzorgverleners

Ondanks de grote impact van het maken van levenseindebeslissingen op betrokkenzorgverleners hebben weinig studies tot nu toe gefocust op welke factoren door hen alsbehulpzaamofhinderendwordenervaren.KwalitatievestudiesbijzorgverlenerswerkendineendienstneonataleintensievezorgeninNoorwegentoondenaandatbeslissingenoverhetaldannietvoortzettenvanlevensondersteunendebehandelingenbemoeilijktwordendooreengebrek aan zekerheidoverdeprognose enhet toekomstig lijdenvanhet kind.VerdertoondendezestudiesinNoorwegenaandatdeambivalentietussenhetwillenbetrekkenvanouders bij het beslissingsproces en hen besparen van onnodig lijden kan leiden totbesluiteloosheidoverdehoeveelheidinformatiediezorgverlenersverstrekkenaandeouders.

Dezebestaandestudiesoverbarrièresenfaciliterendefactoreninhetbeslissingsprocesvanneonatale levenseindebeslissingen voor betrokken zorgverleners hebben een aantalbelangrijke tekortkomingen. Ze richten zich voornamelijk op specifieke types vanlevenseindebeslissingen,zoalshetstakenofnietinstellenvaneenbehandeling,inplaatsvanhetvolledigespectrumvanmogelijkebeslissingeninachttenemen.Verderrichtendemeestestudies zich op de ervaring van ouders, waardoor het standpunt van de betrokkenzorgverlenersvaakwordtvergeten.Éénvandehoofdstukkenvanditproefschriftfocustezichdaaromopdefactorendieneonatologenenneonataleverpleegkundigenalsbehulpzaamofmoeilijkervarenbijhetnemenvanbeslissingenaanheteindevanhetlevenvanpasgeborenenopeendienstneonataleintensievezorgen.Kennisoverdeinvloedvandezefactorenophetbesluitvormingsproces kan dienen als een startpunt voor het formuleren van concreteaanbevelingenomdetoekomstigepraktijkteverbeteren.

186

Psychologischeondersteuningvanzorgverlenerstijdenshetmakenvanlevenseindebeslissingen

Neonatologenenneonataleverpleegkundigendieopeendienstneonataleintensievezorgenwerkenervarenvaakstressenmoreledrukvanwegedehogeeisenvanhunberoep.Vooralwanneereenkindbinnenhunzorgnietlangervoordelenervaartvandeagressieveofzelfsnuttelozebehandelingenen een levenseindebeslissingmoet/kanwordengenomen, kandestress hoog oplopen. Net zoals bij zorgverleners werkend op een dienst pediatrischeintensievezorgenervarenzeverdriet,hulpeloosheidenfrustratiewanneerzenietinstaatzijnom het leven van een kind te redden. Vanwege dit leed zijn neonatologen en neonataleverpleegkundigenergvatbaarvoorhetontwikkelenvaneenverminderdemogelijkheidommedeleventetonenmethunpatiëntenennaasten,enhebbenzeeenverhoogdrisicoophetontwikkelenvaneenburn-out.Indezeinstantieszijnzorgverlenersnietlangercapabelomdeemotionele druk van hun job het hoofd te bieden. Psychosociale ondersteuning vanzorgverlenerswerkend op een dienst neonatale intensieve zorgenwerdmomenteel reedsopgenomeninrichtlijnenvoordedagelijksepraktijk,maardemeestevandezeaanbevelingenmet betrekking tot psychosociale ondersteuning zijn eerder gericht op het voorzien vanconcrete tools voor neonatologen en neonatale verpleegkundigen om ouders optimaal teondersteunen bij besluitvorming en rouw. Onderzoek naar de ervaren psychosocialeondersteuningbijzorgverlenersisbijgevolgookonbestaande.

Voorzoveronsbekendis,bevatteslechtséénonderzoekspecifiekeaanbevelingenuitsluitendgerichtophetondersteunenvanzorgverlenersbinneneenneonatalepalliatievezorgsetting.CatlinenCarteradviseerdenverplichte,formelevergaderingenofadviessessiesalsonderdeelvan reguliere werkuren van zorgverleners, in plaats van het zoeken naar professionelepsychosociale ondersteuning buiten de werkuren of op vrijwillige basis. Verder raden zeziekenhuisafdelingenaanomhetmogelijktemakenvoorzorgverlenersomteweigerendeeluit temakenvaneen specifieke stervensbegeleidingwanneer zedeze last emotioneelnietkunnendragen.Omwillevanhetbelangvanondersteuningvoordezorgverlenersbinneneenlevenseindebeslissingsproces, focuste een laatste studie binnen dit proefschrift op hetonderzoekenvandeervarenpsychologischeondersteuningvanneonatologenenneonataleverpleegkundigenwerkendopeendienstneonataleintensievezorgenbinnendecontextvanlevenseindebeslissingen. Ondersteuning van de betrokken zorgverleners kan cruciaal zijnvoorhetoptimaliserenvandeaangebodenzorgenempathievoorzowelpatiëntjesalsouders.

Doelstellingenvanditdoctoraat

De focus van dit proefschrift is besluitvorming aan het levenseinde van doodgeborenen,pasgeborenen en zuigelingen op populatieniveau over verschillende ziekenhuissettings,patiënten en artsen. De volgende twee doelstellingen, met hun eigen specifiekeonderzoeksvragen,leidenditproefschrift:

Deeerstedoelstellingwashetonderzoekenvanlevenseindebeslissingspraktijkenenbeslissingenbij doodgeborenen, pasgeborene en zuigelingen inVlaanderen,Belgiëoppopulatieniveau.Devolgendeonderzoeksvragenwerdenbeantwoord:

187

1. Welkstudiedesignkangebruiktwordenomdeprevalentievanverschillendebeslissingenaan het levenseinde voor en na de geboorte betrouwbaar te bestuderen? Welkedatabanken op populatieniveau kunnen gebruikt worden om zowel prenatale alsneonatale levenseindebeslissingen te bestuderen, en hoe kunnenwe anoniem contactopnemenmetdeartsdiebetrokkenisbijgevallenvandoodgeboorteofoverlijden?

2. Watisdeprevalentievanverschillendelevenseindebeslissingenindeneonataleperiode?Is de prevalentie in de loop van de tijd veranderd in vergelijking met de vorigegegevensverzameling in Vlaanderen in 1999-2000? Wat zijn de klinische endemografische kenmerken van pasgeborenen en zuigelingen wiens overlijdenvoorafgegaanwerddoordezeverschillende typesvan levenseindebeslissingen?Welkeomstandighedenworden geassocieerdmet verschillende soorten beslissingenaan hetlevenseindevanpasgeborenenenzuigelingen?

Het tweede doel van dit proefschriftwas het in kaart brengen van attitudes, opvattingen enervaringenvanbetrokkenzorgverleners,namelijkneonatologenenneonataleverpleegkundigen,metbetrekkingtotprenataleenneonatalelevenseindebeslissingen.Binnenditdoelwerdendevolgendeonderzoeksvragenbeantwoord:

3. Watzijndeattitudesvanneonatologenenneonataleverpleegkundigentenaanzienvanprenatale en neonatale levenseindebeslissingen? Wat zijn de verschillen in attitudetussen artsen en verpleegkundigen? Welke attitudes ten aanzien van prenatale enneonatalelevenseindebeslissingenvanzorgverleners,enwelkevanhundemografischekenmerken, worden geassocieerd met het al dan niet aanvaardbaar vinden vanverschillendebehandeloptiesineenhypothetischecase?

4. Welke factorenvanhet levenseindebeslissingsproceskunnen,volgensdeervaringvanneonatologen en neonatale verpleegkundigen, het besluitvormingsproces aan hetlevenseindevanpasgeborenenenzuigelingenbelemmerenofvergemakkelijken?

5. Op welke manier worden neonatologen en neonatale verpleegkundigen ondersteunddoorcollega’s,psychologenendeziekenhuisafdelingtijdenshetbeslissingsprocesbijhetmaken van levenseindebeslissingen binnen een dienst neonatale intensieve zorgen?Wordt de ervaren psychologische ondersteuning door zorgverleners als voldoendegeacht?

Methoden

Omde onderzoeksvragen en onderzoeksdoelstellingen van dit proefschrift te beantwoorden,werden verschillende methoden voor gegevensverzameling gebruikt, namelijk eenvragenlijststudie bij artsen gebruik makend van overlijdensattesten, een attitude enpsychologische ondersteuningssurvey bij artsen en verpleegkundigen, en een kwalitatiefonderzoekmetface-to-facesemigestructureerdeinterviewsbijartsenenverpleegkundigen.

Devragenlijststudieopbasisvanoverlijdensattesten

Omonderzoeksvraag2en3tebeantwoordenmaaktenwegebruikvaneenvragenlijststudieopbasisvandeoverlijdensattestenvanalledoodgeborenenmeteenzwangerschapsduurvanmeer

188

dan22wekenofeengeboortegewichtvanmeerdan500gram,enalleoverlijdensvoordeleeftijdvanéénjaar.AllegeïncludeerdedoodgeboortesenoverlijdensvondenplaatsinVlaanderenofBrussel,enbijelkefoetus,pasgeboreneofzuigelingwasdemoedereeninwonervanVlaanderenophetmomentvandoodgeboorteofoverlijden.Hetdesignvandezestudiewasidentiekaaneenvoorgaandonderzoekopbasisvanoverlijdensattestenuitgevoerdtussenaugustus1999enjuli2000,metuitzonderingvaneenlangereinclusieperiodevanseptember2016totdecember2017(12maandenin1999-2000tenopzichtevan16maandenin2016-2017).

Binneneenperiodevanviermaandennahetoverlijdenontvingelkeartsverantwoordelijkvoorhet ondertekenen van de overlijdensattesten een vragenlijst en een begeleidende brief metidentificatiegegevens over het specifieke patiëntje via het Vlaams Agentschap voor Zorg enGezondheid,dieverantwoordelijkisvoordeverwerkingvanoverlijdensattesten.Omanonimiteitte garanderen, diende een advocaat als intermediair orgaan tussen de artsen, het VlaamsAgentschapvoorZorgenGezondheidendeonderzoekersverantwoordelijkvoorhetverwerkenvandegegevens.Dezetussenpersoonzorgdeervoordatingevuldevragenlijstennooitgekoppeldkondenwordenaaneenspecifiekepatiënt,artsofziekenhuis.

Erwerdgebruikgemaaktvantweeafzonderlijkevragenlijstentijdensditonderzoek,namelijkeen vragenlijst bij overlijdensattesten omtrent een doodgeboorte en een vragenlijst bijoverlijdensattestenomtrenteenoverlijdenvaneenlevendgeborenkindonderdeleeftijdvanéénjaar. De vragenlijsten die tijdens dit onderzoek werden gebruikt hadden als doel om teinformerennaarmogelijkeprenataleenneonatalelevenseindebeslissingenvoorafgaandaandedoodgeboorte of het overlijden van het patiëntje.Wemaakten gebruik van een gevalideerdevragenlijst die vroeger reeds gebruiktwerdom levenseindebeslissingenvoorafgaandaaneenoverlijdenvaneenpasgeboreneofzuigelingonderdeleeftijdvanéénjaartebestuderen,zodatvergelijkbaarheidvangegevensovertijdgewaarborgdwerd.Bijdestartvanbeidevragenlijstenwerd gevraagd of de doodgeboorte of het overlijden plotseling en volledig onverwachtplaatsvond. Bij een negatief antwoord werd een levenseindebeslissing voorafgaand aan dedoodgeboorteofhetoverlijdenmogelijkgeacht,waarnadevragenlijstindetailhetaldannietvoorkomen van deze levenseindebeslissingen naging. De gebruikte vragenlijsten kan jeterugvindeninAppendix1en2.

Wanneerersprakewasvanmeerderelevenseindebeslissingenbijeenspecifiekedoodgeboorteof overlijden, werd de beslissing met de meest expliciete levensverkortende intentie alsbelangrijkstebeslissingweerhouden.Wanneermeerdanéénlevenseindebeslissingmetdezelfdelevensverkortende intentie werd aangegeven, werd de toediening van medicatie (actief)verkozen boven het staken of niet instellen van een behandeling (passief). Wanneer eenlevenseindebeslissingwerdaangegeven inde vragenlijstwerdeneen reeksbijvragengesteld,zoalsbijvoorbeelddetijdsduurwaarmeehetlevenwerdverkort,debelangrijksteredenvoorhetnemenvandebeslissing,enwiebetrokkenwasbijhetbesluitvormingsproces.Demografischeinformatieaangegevenopdeoverlijdensattestenwerdanoniemgekoppeldaandevragenlijstdatanavoltooiingvandedataverzameling.

189

Deattitudeenpsychologischeondersteuningssurvey

Om de attitudes en ervaren psychologische ondersteuning van betrokken zorgverleners bijneonatalebeslissingenaanhetlevenseindeteonderzoeken,werdeenvragenlijststudieopgesteldbijalleneonatologenenneonataleverpleegkundigenwerkendineendienstneonataleintensievezorgen inVlaanderen.AlleVlaamsedienstenneonatale intensieve zorgennamendeelaanditonderzoek en waren gevestigd in de volgende ziekenhuizen: Universitair ziekenhuis Gent,Universitair ziekenhuis Brussel, Universitair ziekenhuis Leuven, Universitair ziekenhuisAntwerpen,AZSint-JanBrugge-Oostende,ziekenhuisOost-LimburgGenk,ziekenhuisGZASint-AugustinusenziekenhuisZNAMiddelheim.

Gegevenswerdenverzameldtussen1en31mei2017.Degatekeepermethodewerdgebruikt,waarbijeenartswerkendopelkvandeachtdeelnemendeziekenhuisdienstenverantwoordelijkwasvoorhetuitdelenvandevragenlijstenaanalleartsenenverpleegkundigenwerkendophundienst.Artsenenverpleegkundigenwerdengevraagdomdevragenlijstintevullenenterugtesturennaardeonderzoekersdoormiddelvaneenprepaidenvelop.

DevragenlijstdiegebruiktwerdtijdensdezesurveywerdontwikkeldopbasisvaneenbestaandeVlaamseattitude-vragenlijstuithetjaar2000overlevenseindebeslissingenbijpasgeborenenenzuigelingen,eneenAmerikaansevragenlijstnaarhetonderzoekenvanburn-outbijneonatologenbinnen een intensieve zorgsetting. Een multidisciplinair team van drie sociologen, tweepsychologen, drieneonatologenen eengynaecoloogontwikkeldede finale vragenlijst.Nadienwerd deze vragenlijst cognitief getest bij vijf neonatologen werkend op vier afzonderlijkeziekenhuisafdelingen, drie neonatale verpleegkundigen werkend op twee afzonderlijkeziekenhuisafdelingen, en één gynaecoloog om de validiteit van de items te garanderen. Devragenlijst bestonduit zeven socio-demografischevragenen12 itemsoverhunattitudes tenopzichtevanperinatalelevenseindebeslissingen.Zesvandezeattitudeitemswarengerichtoplevenseindebeslissingen in de neonatale periode, de overige zes attitude items focusten opprenatalebeslissingen(zwangerschapsafbrekingbijeenlevensvatbarefoetus).Attitudeswerdengemetendoormiddelvaneenvijf-puntLikert-schaal,waaropdedeelnemerskondenaanduidenof ze het al dan niet eens warenmet de aangeboden stellingen.We presenteerden ook eenhypothetische casestudie van een foetus geboren na 27 weken zwangerschap, waarbijadditionelecomplicatiesoptradennadegeboorte.Participantenkondenhierbijopeenvier-puntLikert-schaalaanduidenofzezevenmogelijkebehandeloptiesaldannietzoudenoverwegen.Tenslottebevattedevragenlijstookitemsrondervarenstress,deaanwezigheidvanprofessionelepsychosocialeondersteuningvoorzorgverleners,endeervarensteunvancollega’stijdenshetmakenvanlevenseindebeslissingenophundienst.Deelnemerskondenopbasisvaneenvijf-puntLikert-schaal aanduiden inwelkemate ze akkoord gingenmet deze stellingen. De gebruiktevragenlijstenkanjeterugvindeninAppendix3en4.

Deface-to-facesemigestructureerdeinterview-studie

Als laatste werd een kwalitatieve interview-studie uitgevoerd waarbij semigestructureerdeinterviewswerdenafgenomenbijneonatologenenneonataleverpleegkundigenwerkendopeenVlaamse dienst neonatale intensieve zorgen. We kozen voor een kwalitatief design om

190

deelnemersdemogelijkheidtebiedenomopeenopenmanierhunmeningtedelen,waardoorervoldoende aandacht kon worden geschonken aan hun individuele ervaringen. Binnen dezeinterviews focusten we op de ervaren barrières en faciliterende factoren van zorgverlenerstijdenshetmakenvanlevenseindebeslissingen.

We rekruteerden neonatologen die werkzaam waren op één van de volgende vier dienstenneonatale intensieve zorgen: Universitair ziekenhuis Gent, Universitair ziekenhuis Brussel,Universitair ziekenhuis Leuven of AZ Sint-Jan Brugge-Oostende. De interviews vondenplaatstussendecember2017en juli2018.Deneonatologenwerdengeachtbetrokkentezijnbij tenminste één kind waarvan het overlijden voorafgegaan werd door een levenseindebeslissingbinnen het afgelopen jaar. Verder rekruteerden we op dezelfde afdelingen ook neonataleverpleegkundigendiebinnenhetafgelopenjaarverantwoordelijkwarenvoordezorgvantenminsteéénkindwaarvanhetoverlijdenvoorafgegaanwerddooreenlevenseindebeslissing.Eenverantwoordelijkeartsopelkeafdelinginformeerdealleneonatologenenverpleegkundigenopzijn/haardienstoverhetdoelvanhetonderzoek.Contactgegevensvangeïnteresseerdeartsenenverpleegkundigenwerdendaarnaovergemaaktaandeonderzoekers.Alleinterviewsvondenplaatsineenafgeslotenruimteopdeziekenhuisafdeling,ofbijdedeelnemersthuis.

Degebruiktetopicguide(zieAppendix5en6)werdontwikkelddooreenmultidisciplinairteamvannegenonderzoekersmetervaringbinnenhetonderwerpvanpalliatievezorg,levenseindeenneonatologie.Aandedeelnemerswerdgevraagdwathetgemakkelijkerofnetmoeilijkermaaktevoorhenom levenseindebeslissingentenemenbijpasgeborenenenzuigelingen.Voorafgaandaanhet interviewwerdeenkortevragenlijstafgenomenomsocio-demografischegegevens teverzamelen.Dataverzamelingwerdafgerondwanneerergeennieuwebarrièresoffaciliterendefactorennaarvoorkwamentijdensinterviews.

Belangrijkstebevindingen

Het onderzoeken van levenseindebeslissingspraktijken en beslissingen bijdoodgeborenen, pasgeborenen en zuigelingen in Vlaanderen, België oppopulatieniveau

Hetontwikkelenvaneenmethodeomdeprevalentievanlevenseindebeslissingenvoorennadegeboorteteonderzoeken

In hoofdstuk 2 presenteerdenwe een studie design om de prevalentie van prenatale enneonatale beslissingen rond het levenseinde op populatieniveau in Vlaanderen, België, teevalueren en te monitoren. Dit design omvatte de ontwikkeling van een gevalideerd,conceptueel kader van levenseindebeslissingen gedurende de gehele foetaal-infantieleperiode,endeontwikkelingvaneenonderzoeksmethodeomdezebeslissingentebestuderen,onafhankelijkvandesettingwaarinhetoverlijdenofdedoodgeboorteplaatsvond.

Wecreëerdeneennieuw,allesomvattendkaderombeslissingenoverhetlevenseindeindegehelefoetaleinfantieleperiodeteclassificeren.Binnenditconceptuelekaderincluderenwebeslissingenbijdoodgeborenenmeteenzwangerschapsduurvanmeerdan22wekenofeen

191

geboortegewichtvanmeerdan500gram,enbeslissingenvoorafgaandaaneenoverlijdenvoorde leeftijd van één jaar. Twee dimensies werden belangrijk geacht, namelijk demedisch-technische dimensie die de medische handeling classificeerde die werd gesteld, en demedisch-ethischeclassificatiediedelevensverkortendeintentievandeartsinverbandmetdiemedischehandelingomvatte.Watmedischehandelingenbetreft,werdeenonderscheidgemaakt tussen niet-behandelbeslissingen, zoals het niet instellen of staken van eenbehandeling, en het toedienen van medicatie of het uitvoeren van een actieve medischeinterventie met een mogelijks levensverkortend effect. De medisch-ethische classificatieomvattediemogelijkelevensverkortendeintentiesnamelijk:1)geenintentieomhetleventeverkorten, maar hetmogelijks levensverkortend effect werd in rekening gebracht, 2) eenmogelijks levensverkortend effectwas aanwezig,maar hetwas niet het hoofddoel van demedischehandeling,en3)deintentieomhetleventeverkortenwasexpliciet.Opbasisvanditconceptuele kader werden twee afzonderlijke, vergelijkbare vragenlijsten ontwikkeld omlevenseindebeslissingenbijrespectievelijkdoodgeborenenenoverledenkinderenonderdeleeftijdvanéénjaarteonderzoeken.

Het hoofddoel van de studie was het onderzoeken van levenseindebeslissingen bijdoodgeborenenvanaf22wekenzwangerschapenbijoverledenkinderenonderdeleeftijdvanéén jaar op populatieniveau. Gebaseerd op eerdere ervaringen bij pasgeborenen,minderjarigenenvolwassenenwerdgeopteerdvoorhetgebruikvanoverlijdensattestenalsbasisvoorhetverzendenvanonzevragenlijsten.Voordoodgeborenentussen22en26wekenzwangerschapbleekdezemethodenietideaal,aangezienhetaangevenvaneendoodgeboortebinnen deze periode door middel van een overlijdensattest niet verplicht is. Onssteekproefkader bleek dus mogelijks onvolledig. Desondanks verkozen we deze methodeboven het versturen van vragenlijsten gebaseerd op het geboorteregister van hetStudiecentrumvoorPerinataleEpidemiologie(SPE),aangezieneenvertragingbijverwerkingvan deze gegevens tot één jaar in beslag kon nemen, wat de betrouwbaarheid van deantwoordenoponzevragenlijstdrastischzouverlagen.Wekozendaaromvooreenmortalityfollow-back survey methode voor zowel overleden pasgeborenen en zuigelingen als voordoodgeborenen,mits kleineaanpassingenaandemethodologie voordoodgeborenen indecrucialeperiodetussen22en26wekenzwangerschap.HetVlaamsAgentschapvoorZorgenGezondheid, dat alle overlijdenscertificaten verwerkt, moedigde de registratie vandoodgeboortesvanaf22wekenzwangerschapactiefaantijdensdeperiodevandata-collectie.VerderwerdenvragenlijstenverstrektaandetiengrootstematerniteiteninVlaanderen,zodatartsendemogelijkheidhaddenomdezeintevullennaelkedoodgeboorte,naasthetinvullenvandebijhorendeoverlijdenattesten.

Artsen vulden bij elk overlijden of elke doodgeboorte een overlijdensattest in, inclusiefdemografische en medische gegevens van het kind. Nadien ontvingen de centraleoverheidsinstanties, in ons geval het Vlaams Agentschap voor Zorg en Gezondheid, deingevuldeoverlijdensattesten.HetAgentschapwasverantwoordelijkvoorhetverzendenvandevragenlijstenvoorelkontvangenoverlijdensattestnaardeattesterendeartsvermeldophetattest,inclusiefeenbegeleidendebriefmetpatiëntkenmerkennodigvoordeidentificatievanhetkind.Deartsidentificeerdehetkindopbasisvandevoorzienegegevens,envuldedevragenlijstin.Allevragenlijstenwerdenverzondennaareenadvocaatdieverantwoordelijkwas voor het anonimiseren van de ontvangen gegevens, waardoor anonimiteit van de

192

betrokken arts, het patiëntje en het ziekenhuis gewaarborgd werd. Na afsluiting van dedataverzamelingwerdenallegegevensvandevragenlijstengekoppeldaandedemografischeenklinischegegevensvanhetoverlijdensattest.

Ditontwikkelendeonderzoeksprotocolisheteerstestudiedesignwaarbeslissingenrondhetlevenseindebijdoodgeborenen,pasgeborenenenzuigelingenonderzochtkunnenwordenoppopulatieniveau binnen één onderzoeksopzet. We zijn ervan overtuigd dat dit soortonderzoekregelmatigmoetwordenherhaaldomeventueleveranderingenindepraktijkvanlevenseindebeslissingen in kaart te brengen, onder een voortdurend veranderendemaatschappelijke, juridische en klinische invloed. Omwille van het gebruik vanoverlijdenscertificatenvoordeverzendingvanonzevragenlijstenkandezemethodeookinandere landengebruiktworden, ongeachtmogelijke verschillen inhetwettelijk kadermetbetrekking tot levenseindebeslissingen. Hierdoor zijn internationale vergelijkingen tussenlandenmogelijk.Hetverstrekkenvanprevalentieschattingen,nietenkelinVlaanderenmaarook internationaal, kan uiteindelijk helpen bij de ontwikkeling van richtlijnen terondersteuning van zorgverleners tijdens het maken van deze ethische beslissingen in dedagelijksepraktijk.

Deprevalentievanlevenseindebeslissingenbijpasgeborenenenzuigelingen

Hoofdstuk3vanditproefschriftwasgerichtophetleverenvanprevalentieschattingenvanverschillendetypeslevenseindebeslissingenbijpasgeborenenenzuigelingeninVlaanderenover twee verschillende studieperiodes (1999-2000 en 2016-2017). Dezeprevalentieschattingen werden onderzocht door middel van de ontwikkeldeonderzoeksmethodebeschrevenbinnendevorigeparagrafen.

Een totaal aantal van 276 neonaten en zuigelingen stierf tussen 1 september 2016 en 31december 2017 (229 ingevulde vragenlijsten ontvangen, 83% respons rate); en 292pasgeborenenenzuigelingenstierventussen1augustus1999en31juli2000(253ingevuldevragenlijsten ontvangen, 87% respons rate). Onderzoeksresultaten toonden aan dat deprevalentie van levenseindebeslissingen bij pasgeborenen relatief stabiel is gebleven overbeidestudieperiodes.Ongeveer60%vanalleneonataleenkindersterfteonderdeleeftijdvanéénjaarinVlaanderenwerdvoorafgegaandoorzulkebeslissingen.Nietbehandelbeslissingenkomennogsteedshetmeestvoor,namelijkbij34%vanalleoverlijdensin1999-2000en37%in2016-2017.Hetnietinstellenvaneenbehandelingkwamvoorbij13%vanalleneonataleen zuigelingensterfte in 1999-2000 en 12% in 2016-2017, terwijl het staken van eenbehandelingvoorkwambij21%vandegevallen in1999-2000en25%vandegevallen in2016-2017. Het toedienen vanmedicatiemet eenpotentieel levensverkortend effect bleefstabiel op 16% in 1999-2000 vergelekenmet 14% in 2016-2017. De prevalentie van hettoedienenvanmedicatiemeteenexplicietlevensverkortendeffectbleefrelatiefconstantmet7% in 1999-2000 en 10% in 2016-2017. Ondanks het feit dat de prevalentie vanlevenseindebeslissingen over beide studieperiodes relatief stabiel is gebleven, merkenwetoch grote veranderingen binnen leeftijdsgroepen. Levenseindebeslissingen werden nusignificant vaker genomen na de eerste levensweek dan in 1999-2000 (74% van desterfgevallentussen7en27dagenoudwerdvoorafgegaandooreenlevenseindebeslissingin2016-2017vergelekenmet50%in1999-2000,p=0.03;en64%vandesterfgevallenna27

193

dagenoudin2016-2017vergelekenmet38%in1999-2000,p=0.003).Bijsterfgevallendieplaats vonden binnen de eerste levensweek daalde de prevalentie vanlevenseindebeslissingenaanzienlijk(55%vandesterfgevallenin2016-2017vergelekenmet72% in 1999-2000, p=0.01). Na de eerste levensweek veranderde de praktijk vanlevenseindebeslissingen in Vlaanderen in vergelijking met 17 jaar geleden, aangezienbeslissingenomlevensverlengendebehandelingentestakenofmedicatiemeteenexplicietelevensverkortendeintentietoetedienenaanzienlijkstegen.In1999-2000werd9%vandesterfgevallen tussen 7 en 27 dagen oud voorafgegaan door een beslissing om een reedsingesteldebehandelingtestaken,enwerdenergeengevallenaangegevenwaarmedicatiemeteenexplicietelevensverkortendeintentiewerdtoegediend.In2016-2017daarentegenwerdbinnendezelfdeleeftijdsgroepin26%vandegevalleneenbehandelinggestaaktenin26%vandegevallenmedicatie toegediendmeteenexplicietelevensverkortende intentie.Nadeeerste27dagensteegdeprevalentievanhetstakenvaneenbehandelingvan16%in1999-2000naar31%in2016-2017,endeprevalentievanhet toedienenvanmedicatiemeteenexplicietelevensverkortendeintentiesteegvan2%naar10%.

Binnen hoofdstuk 3 toondenwe aan dat levenseindebeslissingen een integraal onderdeelblijvenvandemedischepraktijkbij hetbehandelenvanextreemziekekinderenonderdeleeftijdvanéénjaar,aangeziendrieopdevijfsterfgevallenbinnendezegroepvoorafgegaanwerddoordergelijkebeslissingen.Ditgeeftaandatdetoelaatbaarheidenvereistenvooreengoedeklinischepraktijkbesprokenmoetenwordenonderbetrokkenzorgverleners.

Attitudes,opvattingenenervaringenvanbetrokkenzorgverlenersmetbetrekkingtotprenataleenneonatalelevenseindebeslissingen

Attitudesvanneonatologenenneonataleverpleegkundigentenopzichtevanperinatalelevenseindebeslissingen

Inhoofdstuk4besprokenwedeattitudesvanneonatologenenneonataleverpleegkundigenwerkend op een dienst neonatale intensieve zorgen ten opzichte van perinatalelevenseindebeslissingen. Deze werden onderzocht door middel van een post survey oppopulatieniveau.

Wevondendateralgemeenaanzienlijkesteunkonwordengevondenvoorzowelprenataleenneonatale levenseindebeslissingen bij neonatale zorgverleners. In geval van prenatalelevenseindebeslissingenachtte80tot98%vanalleneonatologenenverpleegkundigendateenlaattijdige zwangerschapsafbreking na een levensvatbare termijn aanvaardbaar was bijdiagnosevanernstigeoflethalefoetaleafwijkingen.Wanneerdefoetusgezondismaarhetlevenvandemoederdoordezwangerschapingevaarwordtgebracht,vondmeerdan60%van de neonatologen en verpleegkundigen een zwangerschapsafbreking op levensvatbaretermijnaanvaardbaar.Wanneerdefoetusechtergezondismaardezwangerschapeengevaarvormt voor de psychologische gezondheid, daalt de aanvaardbaarheidsgraad tot 15% bijzowelartsenalsverpleegkundigen.Bijextreemziekepasgeborenenofzuigelingenvond80tot100%vanalledeelnemendezorgverlenersniet-behandelbeslissingenzoalshetstakenofnietinstellen van een behandeling aanvaardbaar, ongeacht of deze beslissing genomen werd

194

rekening houdend met een mogelijk levensverkortend effect of met een explicietelevensverkortendeintentie.Naastdealgemeneconsensustussenartsenenverpleegkundigenop de bovenvermelde levenseindebeslissingen konden ook een aantal verschillen tussenzorgverleners worden waargenomen. Het toedienen van medicatie met een potentieellevensverkortendeffectwerddoordemeerderheid vanbeide zorgverleners aanvaardbaargeacht, maar neonatologen waren significant meer geneigd om in te stemmen met dezepraktijk (96%)danverpleegkundigen(84%,p=0.02).Anderzijdsvondenverpleegkundigenhet toedienen van medicatie met een expliciete levensverkortende intentie vakeraanvaardbaar (74%) dan neonatologen (60%, p=0.02). Ondanks het feit datdeze praktijkmomenteel nietwettelijk toegestaan is in België, lagde aanvaardbaarheidsgraad bij beidezorgverlenersbovende50%.

Onze studie stelde een grote aanvaardbaarheid vast van zowel prenatale als neonatalelevenseindebeslissingen bij neonatologen en neonatale verpleegkundigen, zelfs voorbeslissingen die momenteel buiten het Belgische wettelijke kader vallen. Artsen enverpleegkundigen verschilden echter enigszins in hun aanvaardbaarheid ten opzichte vanspecifieke types levenseindebeslissingen. Deze verschillen tussen zorgverleners houdenmogelijkverbandmethetfeitdatverpleegkundigenvaaknietdeeindverantwoordelijkheidvoor de medische beslissingen dragen. Deze bevindingen wijzen op het belang van hetincluderenvanhetperspectiefvanzowelartsenalsverpleegkundigentijdenshetmakenvanlevenseindebeslissingenindeperinataleperiode.

Barrièresenfaciliterendefactorenbijhetmakenvanlevenseindebeslissingenbijpasgeborenenenzuigelingen

In hoofdstuk 5 onderzochten we de barrières en faciliterende factoren die zorgverlenerservarentijdenshetmakenvanlevenseindebeslissingenbijpasgeborenenenzuigelingenopeen dienst neonatale intensieve zorgen. Hiermee wilden we inzicht verschaffen in decomplexiteitvanhetlevenseindebeslissingsprocesbinnendedagelijksepraktijkdoorinzichttegevenindeindividueleenpersoonlijkeervaringenvanzorgverleners.

Sommigevandezebarrièresenfaciliterendefactorenzijngelinktaandespecifiekekenmerkenvandecaseinkwestie.Dezefactorenhebbenbetrekkingophetziekekind,deoudersofdebetrokken zorgverleners. Het beslissingsproces werd gemakkelijker geacht wanneer eenslechte prognose vrij snel duidelijk was in het ziektetraject dan wanneer prognostischeonzekerheid aanhield, ofwanneer alle curatievemogelijkheden eerst verkendwerden omiedereenervanteverzekerendatdelevenseindebeslissingdeenigemogelijkeoptiewasomhet lijden van het kind te verzachten. Zorgverleners gaven aan dat eenlevenseindebeslissingsprocesgemakkelijkerverliepwanneerdeoudersdezelfdeculturele-entaalachtergrondhaddenalsdebetrokkenartsenenverpleegkundigen.Ervaringinhetmakenvanlevenseindebeslissingenwerdookbeschouwdalseencrucialefactor,ditomdatervarenzorgverlenersbeterkunnenanticiperenopdetoekomstigemedischetoestandvanhetkind.

Op procesniveau beschouwen we factoren die verband houden met kenmerken van hetspecifieke besluitvormingsproces zelf. Intensieve communicatie tussen zorgverleners enouders is cruciaal om het besluitvormingsproces makkelijker te laten verlopen. Ook

195

communicatietussenallebetrokkenzorgverlenerszelfisvangrootbelang,bijvoorbeelddoorregelmatig multidisciplinair overleg of debriefings. Bovendien kan het nemen vanlevenseindebeslissingengemakkelijkerwordengemaaktdoorhetroutinematigopstellenvanvoorafgaandezorgplanninggesprekkentussenallebetrokkenen.Tijdensdezegesprekkenkangeanticipeerd worden op alle mogelijke medische uitkomsten van het kind, waardoorbeslissingengemaakt kunnenworden in alle rust inplaats van tijdensperiodes van acuteachteruitgang.

Een laatsteniveauomvat factorenmetbetrekking totdeoverkoepelendestructuurvandeafdeling,hetziekenhuisendebrederesamenlevingdiemogelijkeeninvloedkunnenhebbenop besluitvorming aan het levenseinde van pasgeborenen en zuigelingen. Emotionele enpraktische steun van collega’s op de afdeling, of het gebrek hiervan, is cruciaal voorzorgverlenerstijdenshetmakenvanlevenseindebeslissingen.Bovendienwerdhetgebrekaaneen afzonderlijke ruimte voor slecht-nieuws gesprekken op de dienst en het tekort aanervaren personeel getraind in levenseinde- en palliatieve zorg geïdentificeerd als eenbelangrijke barrière in het levenseindebeslissingsproces. Als laatste werd ook de huidigeBelgische wetgeving genoemd als een beïnvloedende factor. Neonatologen enverpleegkundigengavenaandathetontbrekenvaneenwettelijkkaderominuitzonderlijkegevallenactiefintegrijpenenhetlijdenvanhetkindtebeëindigendoormiddelvanmedicatiemet een expliciet levensverkortende intentie een belangrijke barrière vormt in hetbeslissingsproces. Vooral omdat deze beslissingen tijdens de zwangerschap wel kunnenwordengemaaktonderdewetrondzwangerschapsafbrekingen.

Onzekwalitatieveinterviewstudiebrachtverschillendebarrièresenfaciliterendefactorenaanhet licht rond het levenseindebeslissingsproces bij pasgeborenen en zuigelingen. Degeïdentificeerde factoren kunnen leiden tot een aantal specifieke aanbevelingen om hetbeslissingsprocesindedagelijksepraktijkteverbeteren,zoalshetopzettenvanregelmatigemultidisciplinaire overlegmomenten en debriefings met alle betrokken zorgverleners omprognostischeonzekerheidtereduceren,hetroutinematigimplementerenvanvoorafgaandezorgplanning gesprekken met ouders en zorgverleners bij extreem zieke kinderen zodatbelangrijkebeslissingenvoorafgemaaktkunnenworden,hetcreërenvanprivacyvoorslecht-nieuws gesprekken, en eenmogelijke herziening van het complexe juridische kader rondperinatalelevenseindebeslissingen.

Psychologischeondersteuningvoorzorgverlenersbijhetmakenvanlevenseindebeslissingenbijpasgeborenenenzuigelingen

Hoofdstuk6vanditproefschriftrichttezichopdestressdiezorgverlenerservarentijdenshetmakenvanlevenseindebeslissingenopeendienstneonataleintensievezorgen,endeervarenpsychosociale ondersteuning van zowel collega’s als professionele instanties. Dit werdonderzochtdoormiddelvaneenpostsurveyoppopulatieniveau.

Demeerderheidvandeondervraagdeneonatologenenneonataleverpleegkundigengafaandat het nemen van levenseindebeslissingen (artsen), of het geconfronteerd worden metlevenseindebeslissingen (verpleegkundigen) meer stress veroorzaakt dan normaal(respectievelijk73%en70%).Tijdenshetbesluitvormingsprocesvoordezebeslissingengaf

196

eenmeerderheidvandeartsenaandatzezichondersteundvoeldendoorhuncollegaartsen(86%).Minderdandehelftvandeverpleegkundigendaarentegengafaandatartsennaarhunmeningluisterdentoenlevenseindebeslissingengenomenwerden(45%).Hoeweldemeesteneonatologen(88%)heteenswarendathunneonatale intensievezorgafdelingvoldoendeopportuniteitenaanboodomeventuelebedenkingenofbezwarenteuitenomtrentgemaaktebeslissingen,werdditbijverpleegkundigenslechtsbevestigddoor32%vandedeelnemers.Bijna alle deelnemende artsen en verpleegkundigen waren het eens dat ze bij hunrechtstreeksecollega’sterechtkondenvooreengesprekwanneerhenietsdwarszatomtrentgenomenlevenseindebeslissingen(respectievelijk94%en92%).Bovendiengafdehelftvandeneonatologen(53%)en65%vandeverpleegkundigenaandatzeervoorkondenkiezenomnietlangerbetrokkentewordenbijeenkindjewaarbijzehetnieteenswarenmetdegenomenbeslissingen.Ondankshet feitdatbeidezorgverlenersaangavendatzebij collega’sterechtkondenvooreengesprek,gaf57%vandeneonatologenen60%vandeverpleegkundigenaandatzewildendathunafdelingmeerpsychologischesteunaanboodwanneerzorgverlenersgeconfronteerdwerdenmethetmakenvanlevenseindebeslissingen.Bovendienwasslechts41%vandeneonatologenen50%vandeverpleegkundigenhetermeeeensdatdeafdelingvoldoendepsychologischeondersteuningboodnahetoverlijdenvanpatiënten.

Deze bevindingen wijzen erop dat een dienst neonatale intensieve zorgen meer ad hocprofessionele psychologische ondersteuning moet bieden aan zorgverleners diegeconfronteerd worden met het overlijden van patiëntjes of het maken vanlevenseindebeslissingen. Daarnaast radenwe standaarddebriefings aan om zorgverlenersextrateondersteunenbijhetmakenvandezebeslissingen.Verderzijnwevanmeningdatverpleegkundigenmeerbetrokkenmoetenwordenbij interdisciplinaireoverlegmomententijdenshetlevenseindebeslissingsproces.Ditzounietenkeldekwaliteitvandezebeslissingenkunnenverbeteren,maarzoumogelijkookeeninvloedkunnenhebbenopdeervarenstressvanverpleegkundigenveroorzaaktdoorhetfeitdatzenogtevaaknietbetrokkenwordenbijhetmakenvanbeslissingenvoorhunpatiëntjes.

Besprekingvandebelangrijkstebevindingen

Eeninternationalevergelijkingvanneonatalelevenseindebeslissingen

Neonatale sterfte varieert sterk tussenverschillende landen.Naastdezevariatie inneonatalesterftecijfers bestaan er internationale verschillen in ethisch perspectief wat deaanvaardbaarheid en deprevalentie van verschillende levenseindebeslissingen betreft.Het isdaaromnoodzakelijk omdeverzamelde informatiebinnendit proefschrift te vergelijkenmetinternationaal beschikbare gegevens om zo landsfactoren te identificeren die mogelijk eeninvloedhebbenophetmakenvanlevenseindebeslissingenbijpasgeborenenenzuigelingen.

Eeninternationalevergelijkingvanhetmakenvanniet-behandelbeslissingen

Overhetalgemeenwordenniet-behandelbeslissingenzoalshetstakenofniet instellenvaneen behandeling in Europa goed geaccepteerd, en geeft een meerderheid van de artsenwerkzaamopeendienstneonataleintensievezorgenaanreedsbetrokkengeweesttezijnbij

197

tenminsteééngevalwaarintensievezorgengelimiteerdwerden.Internationaalzienwedatdekansophetlimiterenvanintensievezorgenbijpasgeborenensterkafhankelijkisvandepositieve of negatieve houding van de betrokken artsen ten opzichte van dit soortlevenseindebeslissingen.DepositievehoudingvanVlaamse zorgverleners ten aanzienvanniet-behandelbeslissingenbijpasgeborenenendeovereenkomstigehogeprevalentievanditsoortbeslissingenbinnendegehelepopulatiesterfgevallenvoordeleeftijdvanéénjaar,zoalsgerapporteerdinditproefschrift,bevestigendezebevindingen.Wezoudendaaromkunnenveronderstellendatonzeprevalentieschattingenmetbetrekkingtothetvoorkomenvanniet-behandelbeslissingenmogelijkvergelijkbaarzijnmetanderelandenwaarzorgverlenerseendergelijke positieve houding ten opzichte van deze niet-behandelbeslissingen rapporterenzoals het Verenigd Koninkrijk en Nederland. Artsen werkzaam op een dienst neonataleintensieve zorgen inEuropese landenzoalsdeBaltische staten, Italië, Spanje enDuitslandhebben daarentegen een sterkere pro-life attitude. Hoewel Vlaanderen, het VerenigdKoninkrijkenNederlandbeschouwdkunnenwordenalslandenmeteentolerantehoudingten opzichte van niet-behandelbeslissingen met een potentieel levensverkortend effect,kunnenandereEuropeselandendusmogelijkrestrictieverzijn.

Deprevalentievanniet-behandelbeslissingeninVlaanderenin2016-2017bedroeg37%vanallesterfgevallenonderdeleeftijdvanéénjaar(hoofdstuk3).DezeschattingenzijnietshogerdancijfersuitNederland,waardeprevalentiein2010op31%werdgeschat.Uitrapportenvanneonatale intensieve zorg afdelingen inde Verenigde Staten, hetVerenigd Koninkrijk,Australië en Europa bleek dat 40% tot 93% van de neonatale sterftes op deze dienstenvoorafgegaan werd door het staken of niet instellen van een mogelijks levensreddendebehandeling.Hetverschiltussendeprevalentieschattingenvanonzepopulatiestudie(37%)endeprevalentieschattingenvanhetaantalsterfgevallenbinnengespecialiseerdedienstenneonatale intensieve zorgenvoorafgegaandoor eenniet-behandelbeslissing (40-93%)kanverschillendeoorzakenhebben.Eerstenvooralbestaanermethodologischeverschillenindebeoordelingvandeprevalentievanlevenseindebeslissingen,maarverderbestaaterookeengrootverschiltussenhetvoorkomenvandezebeslissingenbinneneengespecialiseerdeunitenhetvoorkomenvanniet-behandelbeslissingenbinnendetotalepopulatievansterfgevallenonderdeleeftijdvanéénjaaroververschillendesettings.Omdatstudiesoppopulatieniveauschaarszijn,warenvalide internationalevergelijkingentussenonzeprevalentieschattingeneninternationaalbeschikbarecijfers,metuitzonderingvandezeuitNederland,onmogelijk.Inde toekomst zijndeze internationale vergelijkingenoppopulatieniveaubroodnodigomdespecifiekeinvloedvanverschillendelandsfactorenopdeklinischepraktijkteonderzoeken.Hetonderzoeksdesignbeschreveninhoofdstuk2vanditproefschriftisideaalomditdoeltebereiken.

Eeninternationalevergelijkingvanhettoedienenvanmedicatiemeteenimplicieteofexplicietelevensverkortendeintentie

DeresultatenbeschrevenbinnenditproefschriftgevenaandatVlaandereneenvrijtolerantklimaatheeftvoormeeractievevormenvanlevenseindebeslissingenzoalshettoedienenvanmedicatiemet een potentiële of expliciete levensverkortende intentie, zelfswanneer dezebeslissingenmomenteelbuitenhethuidigeBelgischewettelijkekadervallen.

198

We zien dat de levensverkortende intentie ommedicatie toe te dienen, hetzij impliciet ofexpliciet,eencruciaalverschilmaakt indevraagofdeVlaamseaccepterendehoudingvanzorgverlenersinternationaalkanwordenbevestigd.InZwitserlandgaf95%vandeartsenenverpleegkundigenwerkendopeendienstneonataleintensievezorgenaanhettoedienenvansedatievaofanalgeticaacceptabeltevinden,zelfsalsditkonleidentotademhalingsnoodenvroegtijdigoverlijden.Wanneerhettoedienenvandezemedicatiedaarentegeneenexplicietelevensverkortende intentie had daalde de aanvaardbaarheidsgraad naar 24% van deondervraagdezorgverleners.InCanadatoondeeenonderzoekbijpediatersaandatzezichcollectief weerhoudend opstelden ten opzichte van niet-vrijwillige euthanasie bij niet-competentekinderen,watsuggereertdatCanadesekinderartsenenneonatologenmogelijkseenstukmindertolerantzijntenopzichtevandezelevenseindebeslissingendanhunVlaamsecollega’s.InFrankrijkverklaardeeenmultidisciplinairewerkgroeprondethischekwestiesinperinatale geneeskunde dat handelingen om opzettelijk de dood van een patiënt tebespoedigenzowelwettelijkalsmoreelverbodenzijn,watduidelijkeenrestrictievehoudingaangeeftvanFransezorgverleners.Vlaamsezorgverlenerswerkendinneonataleintensievezorgenhebbendusmogelijkeenveeltoleranterehoudingtenopzichtevanhettoedienenvanmedicatie met een potentieel of expliciet levensverkortend effect bij pasgeborenen enzuigelingendanhuninternationalecollega’s.

Naast een vergelijking van de internationale verschillen in attitudes van zorgverlenersomtrent het toedienen van medicatie met een potentieel of expliciete levensverkortendeintentie,iseeninternationalevergelijkingvaneffectieveprevalentieschattingennodig.Uiteenmultinationalestudie(EURONIC)inachtEuropeselanden(Belgiënietinbegrepen)bleekdat32-89%vandeartsenwerkendopeendienstneonataleintensievezorgenaangafdatzeooitpijn-ensymptoommedicatiehaddentoegediend,ondankshetrisicoopademhalingsnoodenvroegtijdigoverlijden.Dezecijfersvarieerdensterktussenlanden:inFrankrijk,NederlandenZwedengaf86-89%vandeartsenaanervaringtehebbenmethettoedienenvansedatievaofanalgetica,zelfswanneereenrisicoopademhalingsnoodofvroegtijdigedoodvanhetkindmogelijk was; terwijl de prevalentie van artsen met eerdere ervaring in dezelevenseindebeslissingeninItaliëslechts32%bedroeg.BovendienbleekuitonderzoekdathettoedienenvanmedicatiemeteenexplicietdoelomhetlevenseindetebespoedigenzeerzeldenvoorkwamindemeesteonderzochteEuropeselanden.Slechts2-4%vandeartseninItalië,Spanje, Zweden, Duitsland en het Verenigd Koninkrijk gaf aan dat ze ooit eerder zulkebeslissingen hadden gemaakt. In een recent vervolgonderzoek van EURONIC in 2016 inDuitsland,ZwitserlandenOostenrijkmeldde97%vandeartsendoormiddelvaneenonlineenquête dat ze ten minste éénmalig medicatie hadden toegediend met het risico op eenvroegtijdig overlijden. Wanneer de toediening van deze medicatie een explicietlevensverkortende intentie had, daalde het percentage van artsen met ervaring in dezebeslissingentot4%.BinnendezeEURONIC-studieswerdartsengevraagdofzeooiteerdereenbepaalde levenseindebeslissinghaddengemaakt inde volledige loopvanhun carrière.Ditgeeftonsgeenindicatieoverhoeprevalentdezebeslissingeneffectiefzijnbinnendedagelijksepraktijk.HetisdanookmoeilijkomeenduidelijkevergelijkingtemakentussendegegevensvandeEURONIC-studiesenonzeprevalentieschattingenomtrenthettoedienenvanmedicatiemet eenpotentieel of expliciete levensverkortende intentiebinnende totalepopulatie vanoverledenkinderenonderdeleeftijdvanéénjaarbinneneenbepaaldeperiode.

199

Deenigebeschikbareprevalentieschattingenomtrenthettoedienenvanmedicatiemeteenpotentieel of expliciete levensverkortende intentie bij pasgeborenen en zuigelingen zijnafkomstiguitNederland.WanneerwedezeprevalentieschattingenvergelijkenmetVlaamsecijfers,zienwedatdemedischepraktijkmisschientochnietzovergelijkbaarisalseerderwerdvermoed.Waaronzecijfersaangevendat14%vanallesterfgevallenonderdeleeftijdvanéénjaar in2016-2017werdvoorafgegaanaanhet toedienenvanmedicatiemet eenmogelijkslevensverkortend effect, rapporteerdeNederlandeen prevalentie van slechts 4% in 2010.Wanneerwekijkennaardeprevalentievanhettoedienenvanmedicatiemeteenexplicietelevensverkortende intentie is het verschil tussen de prevalentieschatting van 10% inVlaanderenin2016-2017meteenprevalentievan1%inNederlandin2010nogopvallender.Wat het laatste verschil zo opvallendmaakt is het feit danNederland eenwettelijk kadervoorzietdathetmakenvanditsoortbeslissingeninextremegevallenmogelijkmaakt,terwijldezebeslissingeninVlaanderenmomenteelnietwettelijkzijntoegestaan.Debeschikbaarheidvan een ondersteunend wettelijk kader leidt dus niet noodzakelijk tot een toename inprevalentievanhettoedienenvanmedicatiemeteenexplicietelevensverkortendeintentie.Het ontbreken van internationale populatieschattingen maakt het echter onmogelijk omrobuusteengeldigeconclusiestetrekkenoverdeimpactvanverschillendekeuzesomtrenthetwettelijkomkaderenvandezebeslissingenopdeklinischepraktijk.

Eenreflectieoverincidentie,attitudesenervaringenvanzorgverlenersinrelatiemethethuidigwettelijkkader

Het tolerante klimaat voor het maken van levenseindebeslissingen bij Vlaamse neonatalezorgverleners moet worden gekaderd binnen de bredere Belgische juridische en medischecultuur. België heeft zowel een vrij liberale wetgeving rond laattijdigezwangerschapsafbrekingen als een euthanasiewetgeving bij volwassenen en competenteminderjarigen (zie hoofdstuk 1). Daarom zoudenwe kunnen argumenteren dat de Belgischejuridischeenmedischecultuurinzijngeheelbeschouwdkanwordenalstolerantertenopzichtevanlevenseindebeslissingenongeachtdeleeftijddaninternationaalhetgevalis.

Omdatpasgeborenenzelfniet instaatzijnomeuthanasieaan tevragenendus teontvangen,kunnen we ze definiëren als een kwetsbare groep die momenteel buiten de bestaandeeuthanasiewetgeving valt, maar waar een mogelijke invloed van de implementatie van dezewetgevingmogelijkis.Omeeninvloedvandewetgevingaantetonen,zoudeimplementatievande euthanasiewetgeving voor volwassenen en competente minderjarigen leiden tot 1) eentoenameindeprevalentievanhetopzettelijkbeëindigenvanhetlevenvanpasgeborenenmeternstigeaandoeningen,en2)moetdezetoenametoegeschrevenwordenaanhetfeitdatartsensneller geneigd zijn om dit soort beslissingen te maken omwille van het bestaan van dezewetgevingondanksdatdezenietophunpatiëntenvantoepassingis.Deprevalentieschattingengegevenbinnenditproefschriftkunnenenkel inzichtgeven inpunt1vandezevoorwaarden:namelijk dat de prevalentie van het toedienen van medicatie met een explicietelevensverkortendeintentierelatiefconstantblijftop7%vandegehelepopulatievanoverlijdensonder de leeftijd van één jaar voorafgegaan door dergelijke beslissing twee jaar voor deimplementatie, en 10% van de populatie overlijdens 15 jaar na de implementatie van deeuthanasiewetgeving(driejaarnadetoevoegingvancompetenteminderjarigen).Eenoorzakelijkverband tussende implementatie vande euthanasiewetgeving enhet feit dat een aanzienlijk

200

aantal overlijdensonderde leeftijd van één jaar voorafgegaanwordtdoorhet toedienenvanmedicatiemeteenexplicietelevensverkortendeintentieindepraktijk,zoalsvermeldinpunt2,kanalsdusdanignietbewezenwordenmetgegevensbinnenditproefschrift.

Zorgverlenersgavendesondanksweldeinvloedvaneenanderewetgevingaanbinnenhetkadervanonzeinterviewstudie,namelijkdebestaandewetgevingrondzwangerschapsafbrekingen.Zovermeldenzehetcontrasttussendeonmogelijkheidomwettelijkhetlevenvaneenpasgeboreneteverkortendoormiddelvanmedicatiemeteenexplicietelevensverkortendeintentieterwijleenzwangerschapkanwordenbeëindigdvoorexactdezelfdediagnosebijeenongeborenfoetusalseen belangrijke barrière. Het bestaan van een liberale wetgeving rond laattijdigezwangerschapsafbreking voor ernstige en lethale foetale afwijkingen kan dus mogelijk eeninvloedhebbenopbesluitvormingaanhetlevenseindevanpasgeborenenenzuigelingen.Omdatde Belgische wetgeving omtrent zwangerschapsafbreking in 1990 geïmplementeerd werd,kunnendeprevalentieschattingenrondneonatalelevenseindebeslissingentussen1999-2000en2016-2017binnenditproefschriftgeen informatieverschaffenoverdemogelijke invloedvandezewetgeving.Bovendienkangedebatteerdwordenoverdevraagofeentolerantehoudingvanbetrokken zorgverleners in perinatale zorg de implementatie van dergelijke wetgevingvoorafgaat,eerderdandathetvolgtopdeimplementatievandezewetgeving.

Wanneer we de prevalentieschattingen voor het toedienen vanmedicatiemet een explicietelevensverkortendeintentieinVlaanderen–eenregiowaardezepraktijkmomenteelnietwordtgereguleerddoormiddelvaneenprotocolofwetgeving–vergelijkenmetdezeinNederland–waar richtlijnen en voorschriften voor deze praktijk verstrektwerden binnen het Groningenprotocol-,zienwedatdezepraktijkvakervoorkomtondankshetgebrekaanregelgeving.Onzeprevalentieschattingenomtrenteenpraktijkdiemomenteelnietwettelijkwordtgetolereerd,incombinatiemetdetolerantehoudingvanVlaamsezorgverlenerstenopzichtevanhettoedienenvan medicatie met een expliciet levensverkortend effect, roept de vraag op of richtlijnen,protocollenofwetgevingennodigzijnomdezebeslissingenindedagelijksepraktijkbinnenzoeenkwetsbaregroepoptevolgen.Desondanksleidthetbestaanvanzoeentolerantehoudingvanbetrokkenzorgverlenersenhetbestaanvanempirischbewijsvanheteffectiefvoorkomenvan deze praktijk binnen de Vlaamse zorgverlening niet automatisch tot brede steun binnenperinatale geneeskunde voor het ontwikkelen van dergelijke wetswijzigingen. Hoewel onzeinterviewstudie aantoonde dat neonatologen en verpleegkundigen het ontbreken van eenwettelijkkaderommedicatietoedieningmetexpliciete levensverkortende intentiemogelijk temakeninspecifiekegevallenalseenbelangrijkebarrièrezien,gavenzeookaanergophunhoedete zijn voor de mogelijke gevolgen van pogingen om zulke individuele beslissingen testandaardiseren en reguleren binnen een restrictieve wetgeving. Het ontwikkelen van eenprotocol of wetgeving om deze beslissingen wettelijk te reguleren en toe te staan kan dusmogelijkhelpenbijbesluitvormingindeklinischepraktijk.Tochwillenwehierinvoorzichtigzijn.Deze uiterst gevoelige ethische kwestie behoeft zeker verder interdisciplinair debat, waarbijbetrokkenzorgverleners,ethicienbeleidmakersbetrokkenmoetenworden.

201

Psychosocialeondersteuningvoorzorgverlenerstijdenshetnemenvanneonatalelevenseindebeslissingen

Een belangrijke bevinding van dit proefschrift is de vaststelling van het ontbreken vanpsychologische en psychosociale ondersteuning voor zorgverleners werkzaam op een dienstneonataleintensievezorgenbijhetmakenvanlevenseindebeslissingen(ziehoofdstuk5en6).Zoweldeeluitmakenvaneenlevenseindebeslissingsprocesalsbetrokkenheidbijhetoverlijdenvaneenpasgeboreneveroorzaakteenaanzienlijkehoeveelheidstressbijbetrokkenartsenenverpleegkundigen. Tijdens de interviews benadrukten neonatologen en neonataleverpleegkundigenvoortdurendhoehetomgaanmeternstigziekekinderenkandoorwegenophun emotionelewelzijn, vooralwanneerde babyer gezond en voldragen uitziet, ofwanneerzorgverlenerszelfkinderenhebbenwaardoordemoeilijkesituatiesdiezeervarenbinnenhunjobalsnelgeprojecteerdwordenophuneigengezinssituatie.Bovendienishetbelangrijkomtevermelden dat deel uitmaken van een levenseindebeslissingsproces bij pasgeborenen nooitgemakkelijkis,endatdediagnostischeenprognostischeonzekerheidzwaarkandoorwegenopdegemoedstoestand.

Om deze verhoogde hoeveelheid stress het hoofd te bieden wanneer zorgverlenersgeconfronteerdwordenmetlevenseindebeslissingenwendenartsenenverpleegkundigenzichvaaktothuncollega’sterondersteuning.Hoewelhetpositieveeffectvancollegialesteunophetwelzijn van zorgverleners zeker niet vergeten mag worden, is het niet voldoende om destressoren gelinkt met levenseindebeslissingen en kindersterfte het hoofd te bieden.Psychosociale ondersteuning voor ouders na het verlies van hun kind is veel frequenterbeschikbaar dan ondersteuning voor betrokken zorgverleners. De meeste aanbevelingen enrichtlijnenomtrentpsychosocialeondersteuningbijzorgaanhetlevenseindevanpasgeborenenrichtzichdaaromookophetondersteunenvanartsenenverpleegkundigenbijrouwbegeleidingvan ouders. Psychosociale ondersteuning van de zorgverleners zelf wordt in dit geval alssecundairofzelfsonbelangrijkgeacht.Desondankskanhetervarenemotioneleleedvandezezorgverlenersleidentotburn-out,depressieeneenverminderdvermogenomempathietetonennaarpatiëntenenouderstoe.Hetvoorzienvanzorgvoorzorgverlenerskandusnietalleenhunpersoonlijkwelzijnbevorderen,hetkanookdezorgdiezeverstrekkenaanpatiëntenenoudersverbeteren. Het gebrek aan professionele ondersteuning voor artsen en verpleegkundigenwerkendopeendienstneonatale intensievezorgen,zoalsaangegevenbinnenditproefschrift,moetdusduidelijkgeadresseerdenopgevolgdworden.

Derolvanpalliatievezorgbijhetmakenvanneonatalelevenseindebeslissingen

Hetheroriënterenvanzorgvancuratiefnaarpalliatief,ofhetgelijktijdigaanbiedenvanbeide,isonderdeel van de dagelijkse klinische praktijk bij het behandelen van extreem ziekepasgeborenenenzuigelingen.Indeprenataleenneonatalepraktijkispalliatievezorgtotnutoeeenrelatiefnieuwenonontgonnenveld.Derolvanhetmakenvanlevenseindebeslissingen,endemogelijke implicaties vande cijfers die besprokenworden in dit proefschrift, binnen eendergelijkeperinataleofneonatalepalliatievezorgbenaderingismomenteelnogergonduidelijk.

202

Wanneeroudersgeconfronteerdwordenmeteenernstigeoflethalediagnosevoorhunkindishet uiterst belangrijk dat zorgverleners hen een empathisch, begrijpelijk en evenwichtigoverzicht geven van alle behandelingsopties, inclusief actieve en op genezing gerichteinterventies,beslissingenrondhetlevenseindeenpalliatievezorg.Hetstakenofnietinstellenvaneenmogelijk levensreddendebehandeling iseen internationaalondersteundepraktijkbijdezeextreemziekekinderen,endedatabinnenditproefschriftonderschrijvenhuncentralerolbinnendelevenseindezorgbijpasgeborenenenzuigelingen(ziehoofdstuk3).Daaromkunnenweverwachtendat zorgverlenersbij dergelijke slecht-nieuwsgesprekkenmetoudersookdemogelijkheidtothetbeperkenvandezorgdoormiddelvaneenniet-behandelbeslissingzullenaanhalen.Verderkunnenweookveronderstellendatniet-behandelbeslissingeneencrucialestapzijn inhet overschakelenvan curatieve en levensreddendebehandelingennaarpalliatieve enlevenseindezorg.Eenideaalvoorbeeldishetextuberenvanhetkindomcomfortteverhogen,ookalisdekunstmatigebeademingcruciaalvooroverleving.Naastcomponentenzoalsvoorafgaandezorgplanning enpsychosocialeondersteuningvanoudersbestaat eenneonataleof perinatalepalliatieve zorgbenadering bij deze kinderen uit het aanbieden van adequate pijn- ensymptoombestrijding.Hetstakenofnietinstellenvanbehandelingenwordtdaaromvaakgevolgddoorhettoedienenvanpijnstillersensedativaomhetlijdenvanhetkindteverzachten.Dehogedosissen pijn- en symptoombestrijding die nodig zijn om adequatepalliatieve zorg te biedenkunnen, en worden (zie hoofdstuk 3), vaak toegediend zelfs wanneer dit een potentieel ofexplicietlevensverkortendeffectteweegbrengt.

Wanneerhet levenseindebeslissingsproces ingebedwordtbinneneenperinatale of neonatalepalliatievezorgbenaderingkanertijdenseenemotioneelturbulenteperiodevoldoendeaandachtgegeven worden aan de complexe behoeften van het gezin. Dit omdat zulke palliatievezorgbenaderingeneengezinsgerichtezorgaanbiedenmeteengrotefocusopde(spiritueleenculturele)waarden van de ouders, hetmaken van blijvende herinneringenmet hun kind, enempathische en duidelijke communicatie tussen ouders en zorgverleners. Dit soortondersteuning is cruciaal bij beslissingen aan het einde van het leven van pasgeborenen enzuigelingen.Verderbiedteenpalliatievezorgbenaderingookvoldoendeondersteuningvoordebetrokkenzorgverleners.Bestaandeperinatalepalliatievezorgprotocollengevennamelijkmeeraandacht aanpsychosociale ondersteuningvanhetpersoneel, omzodekwaliteit van zorg teverbeterenenstress,burn-outencompassionelevermoeidheidtegentegaan.Eenneonataleofperinatale palliatieve zorgbenadering omvat dus niet alleen voldoende pijn- ensymptoombestrijdinggerichtopcomfortvoorhetkind,maarhetlegtookeensterkenadrukopempathische communicatie en psychosociale ondersteuning voor ouders, familieleden enbetrokkenzorgverleners.

Implicatiesenaanbevelingen

Implicatiesenaanbevelingenvoordepraktijk

- Ermoetvoldoendeaandachtwordenbesteedaanhetcreërenvaneenprivéruimtewaarslecht-nieuws gesprekken in alle rust kunnen plaatsvinden. Dit zowel op dienstenneonatale intensieve zorgen als op alle andere ziekenhuisafdelingen waar dergelijkegesprekkenregelmatigplaatsvinden.

203

- Het routinematig implementeren van voorafgaande zorgplanningsgesprekken metoudersingevalvaneenslechteprognosevanhetkindkaneenfaciliterendefactorzijnbijmoeilijkelevenseindebeslissingen.

- Prognostischeonzekerheidkanwordenverminderddoorhetinstallerenvanregelmatigemultidisciplinaireoverlegmomentenendebriefings, enhet stelselmatigbetrekkenvan(externe)expertsomeensecondopiniontebekomen.

- Moeilijke gesprekken met ouders omwille van een cultuur- of taalbarrière kunnenwordenverholpenofverminderddoorhetraadplegenvangespecialiseerdeneonataleofperinatale palliatieve zorgteams. Deze teams worden verondersteld om voldoendeervaringentraininggenotentehebbenomtrenthethebbenvanslecht-nieuwsgesprekkenenempathischecommunicatietussenoudersenzorgverleners.Daaromzijnzeinstaatom te bemiddelen tijdens extreem moeilijke levenseindebeslissingsprocessen tussenoudersenzorgverleners.

- Neonatologen,neonataleverpleegkundigenenanderebetrokkenzorgverlenerswerkendop een dienst neonatale intensieve zorgen zoudenmoetenworden verondersteld omgeneralistischepalliatievezorgvaardighedenteontwikkelen.InBelgiëisermomenteelgeenformeletrainingbeschikbaarrondhetaanbiedenvanadequateneonatalepalliatievezorgmethetdoelomzorgverlenersdezevaardighedenaanteleren.Hetopnemenvaneen module over neonatale sterfte en besluitvorming rond het levenseinde instandaardcurricula voor neonatologen en neonatale verpleegkundigen verhoogt deindividuele klinische ervaring van zorgverleners. Verder verwerven ze tijdens dezemodulescrucialecommunicatievaardigheden.Dezeopleidingenzoudenhetvertrouwenvanzorgverlenersinhuneigenvaardighedenomtrenthetbiedenvanadequatepalliatieveenlevenseindezorgkunnenverhogen,enhunnegatievekijkoplevenseindezorgkunnenverminderen.

- We stellen voor om regelmatig formele debriefings te houden met het betrokkenzorgteamnahetoverlijdenvaneenkindopdeafdeling.Hierdoorcreëerjemogelijkhedenomtebesprekenwatfoutging,watvlotverlopenisenwatzouhelpenbijtoekomstigelevenseindebeslissingen.

- We raden diensten neonatale intensieve zorgen aan om standaard psychosocialeondersteuningtevoorzientijdensdewerkurenvoorzorgverlenersdiebetrokkenwarenbijeenlevenseinde-ofpalliatievezorgproces.

Implicatiesenaanbevelingenvoorhetbeleid

- Zonder betrouwbare prevalentieschattingen op populatieniveau zijn ethische enjuridische discussies en zelfs wettelijke besluitvorming gebaseerd op ervaringen enstandpuntenvaneenselectaantalgeraadpleegdeexperts.Populatiegegevensvormeneenfeitelijke,empirischebasisvoordezediscussies,aangezienzekunnenaangevenofenhoevaak verschillende levenseindebeslissingen voorkomen binnen een bepaalde groeppatiënten. Onze samenleving ondergaat continue veranderingen en evoluties, en demedische behandelopties voor extreem zieke kinderen neemt steeds toe. Daarom issystematische monitoring van levenseindebeslissingen op populatieniveau van hetgrootste belang. Beleidsmakers moeten deze systematische periodieke monitoringondersteunen.Zokunnenbeslissingenenwijzigingenopwettelijkofbeleidsvlaksteeds

204

gebaseerdwordenopdemeestrecentecijfersenmogelijkewijzigingenvandedagelijksepraktijk.

- Aangezienermomenteeleengebrekisaannationaleeninternationalerichtlijnenoverdeaanvaardbaarheidenadequateuitvoeringvanlevenseindebeslissingenbijpasgeborenenenzuigelingen,isernogsteedsveelonzekerheidoverhuntoelaatbaarheidenvereistenvoor een goede klinische praktijk. Onze prevalentieschattingen kunnen experts eenstartpuntbiedenomdeontwikkelingvandezerichtlijnenofmogelijkewettelijkekadersverder te bespreken. Bovendien kunnen de prevalentieschattingen en de besprokenbarrièresenfaciliterendefactorentijdenshetbesluitvormingsproces,zoalsaangegevenbinnen dit proefschrift, een ideaal uitgangspunt zijn voor het formuleren vanaanbevelingenenrichtlijnenoverwatinmoeilijkegevallenalsbestepraktijkkanwordenbeschouwd.

Implicatiesenaanbevelingenvoorverderonderzoek

- Input van ouders ontbrak binnen het narratief van dit proefschrift. Toch zijn deervaringenenopvattingenvanouderscruciaalomeenvolledigbeeldtegevenvanhetbesluitvormingsproces bij levenseindebeslissingen onder de leeftijd van één jaar.Toekomstigonderzoekzouzichdaarommoetenfocussenophetonderzoekenvanhunervaringenenbelevingen.

- Neonatale levenseindebeslissingen zijn onherroepelijk verbonden met mogelijkebeslissingentijdensdeprenataleperiode.Ditomdateengrootdeelvandeaangeborenaandoeningen of afwijkingen prenataal kunnen worden vastgesteld. Toekomstigonderzoekmoetzowelprenatalealsneonatalelevenseindebeslissingenblijvenomvattenbinnen eenzelfde onderzoekssetting, eventueel gebruikmakend van het conceptueelkaderbeschreveninhoofdstuk2.Bovendienmoet,zoalsinhoofdstuk4werdgetoond,bijhet onderzoeken vanattitudes,meningen en ervaringen van betrokken zorgverlenerszoweldemeningvanprenatalealsneonataleexpertenoverwogenworden.

- Toekomstige studies moeten gericht zijn op het verzamelen van internationalepopulatiedata omtrent de prevalentie van prenatale en neonatalelevenseindebeslissingen. Internationaal vergelijkend onderzoek kan landspecifieke ofzelfs regiospecifieke factoren identificeren die de prevalentie van dezelevenseindebeslissingen en de bijhorende klinische praktijk kunnen beïnvloeden.Bovendien zou internationale data inzicht kunnen bieden in verschillende medischeculturenmetbetrekkingtotneonatalezorgaanhetlevenseinde.

- Levenseindebeslissingenbijpasgeborenenenzuigelingenzijnduidelijksterkingebedinneonatale en zelfs perinatale palliatieve zorg. Hoewel cruciale elementen van eenpalliatieve zorgbenadering reeds worden geïmplementeerd in de reguliere perinatalepraktijk, is het bestaan van gespecialiseerde perinatale palliatieve zorgteamsinternationaalzeldzaam.Aangezienditeenrelatiefnieuwenopkomendonderzoeksveldisdat instaatvoordecrucialeondersteuningvoorextreemziekebaby'svoorennadegeboorte, evenals voor hun families en betrokken zorgverleners, zou toekomstigonderzoekzichmoetenrichtenophetevaluerenvanhetbestezorgmodelbinnendezesetting.

205

CurriculumVitaeandlistofpublicationsofLaureDombrecht

206

Curriculumvitae

LaureDombrecht, bornMay 11th 1992 (Tielt, Belgium), graduatedwith honours fromGhentUniversityin2015asamasterofscienceinexperimentalpsychology.InOctober2015shejoinedtheEnd-of-LifeCareResearchgroupofGhentUniversityandVrijeUniversiteitBrussel(VUB)asa doctoral researcher. Her doctoral research project focussed on end-of-life decisions instillbirths,neonatesandinfantsinFlanders.ShewassupervisedbyProf.dr.LucDeliens,Prof.dr.JoachimCohen,Prof.dr.KennethChambaere,anddr.KimBeernaert.Recently,incollaborationwithProf.dr.KennethChambaere,dr.KimBeernaert,Prof.dr.FilipCoolsandProf.dr.KristienRoelens,sheobtainedafour-yeargrantfromResearchFoundationFlanders(FWO)todevelopand pilot test the first Belgian perinatal palliative care program within standard perinatalhealthcare(aphase0-2study).

207

Listofpublications

Gruber,J.,Strauss,G.P.,Dombrecht,L.,&Mittal,V.A.Neuroleptic-freeyouthatultrahighriskforpsychosis evidence diminished emotion reactivity that is predicted by depression andanxiety.SchizophreniaResearch,2018March.

Dombrecht,L.,Beernaert,K.,Roets,E.,Chambaere,K.,Cools,F.,Goossens,L.,Naulaers,G.,DeCatte,L.,Cohen,J.,Deliens,L.andonbehalfoftheNICUconsortium.Aposts-mortempopulationsurveyonfoetal-infantileend-of-lifedecisions:aresearchprotocol.BMCPediatrics,2018August.

DombrechtL.,DeliensL.,ChambaereK.,BaesS.,CoolsF.,GoossensL.,NaulaersG.,RoetsE.,PietteV.,CohenJ.,BeernaertK.(andtheNICUconsortium).Neonatologistsandneonatalnurseshavepositiveattitudestowardsperinatalend-of-lifedecisions,anationwidesurvey.ActaPaediatrica,2019April.

DombrechtL.,PietteV.,DeliensL.,CoolsF.,ChambaereK.,GoossensL.,NaulaersG.,CornetteL.,BeernaertK.,CohenJ.,andonbehalfoftheNICUconsortium.Barrierstoandfacilitatorsofend-of-lifedecision-makingbyneonatologistsandneonatalnursesinneonates:aqualitativestudy.JournalofPainandSymptomManagement,2019October.

DombrechtL.,CohenJ.,CoolsF.,DeliensL.,GoossensL.,NaulaersG.,BeernaertK.,ChambaereK.on behalf of the NICU consortium. Psychological support in end-of-life decision-making inneonatalintensivecareunits:fullpopulationsurveyamongneonatologistsandneonatalnurses.PalliativeMedicine,2019November.

DombrechtL.,BeernaertK.,ChambaereK.,CoolsF.,GoossensL.,NaulaersG.,CornetteL.,LarocheS., Theyskens C., Vandeputte C., Van de BroekH., Cohen J., Deliens L. End-of-life decisions inneonatesandinfants:arepeatedpopulation-levelmortalityfollow-backstudy(Submitted).

RoetsE.,DierickxS.,DeliensL.,ChambaereK.,DombrechtL.,BeernaertK.,RoelensK.Healthcareprofessionals’attitudestowardsterminationofpregnancyatviablestage(Submitted).

208

Listofpresentationsgivenat(inter)nationalconferencesandseminars

Theresearchprotocolofapost-mortemsurveyonend-of-lifedecisionsinstillbirths,neonatesandinfants,SecondinternationalconferenceonEndoflifelaw,ethics,policyandpractice,13-15September2017,Halifax,NovaScotia,Canada(oralpresentation).

Het onderzoeksprotocol van een post-mortem studie naar levenseindebeslissingen bijdoodgeborenen, pasgeborenen en zuigelingen in Vlaanderen, België., Nederlands-Vlaamsewetenschapsdagen Palliatieve zorg, 30 November – 1 December 2017, Amsterdam, TheNetherlands(posterpresentation).

Watzijndeattitudesvanneonatologenenverpleegkundigentegenoverlevenseindebelissingenbij pasgeborenen met een ernstige afwijking? Een vragenlijststudie., Nederlands-Vlaamsewetenschapsdagen Palliatieve zorg, 30 November – 1 December 2017, Amsterdam, TheNetherlands(posterpresentation).

Attitudes of neonatologists and neonatal nurses on neonatal end-of-life decisions, a fullpopulationsurvey.,10thworldresearchcongressoftheEuropeanAssociationforPalliativeCare,24-26May2018,Bern,Switzerland(posterpresentation).

Neonatologists’ and neonatal nurses’ attitudes towards perinatal end-of-life decisions., ThirdinternationalconferenceonEndoflifelaw,ethics,policyandpractice,7-9March2019,Ghent,Belgium(oralpresentation).

ExperiencedSupportwhenConfrontedwithEnd-of-LifeDecisionsintheNeonatalIntensiveCareUnit:ANationwidePopulationSurveyamongNeonatologistsandNeonatalNurses.,16thworldcongress of the European Association for Palliative Care, 23-25May 2019, Berlin, Germany(poster presentation,winner of the award of one of the three best abstracts in the categoryPalliativeCareinChildrenandAdolescents).

Inzicht in eindelevens beslissingen, congres controversen in de perinatale geneeskunde, 28September2019,Leuven,Belgium(invitedspeaker).

Barrièresenfaciliterendefactorenbijhetmakenvanlevenseindebeslissingeninpasgeborenenbijneonatologenenneonataleverpleegkundigen:eenkwalitatievestudie.,Nederlands-VlaamseWetenschapsdagen Palliatieve Zorg, 21-22 November 2019, Antwerp, Belgium (oralpresentation).

Psychologischeondersteuningtijdenshetmakenvanlevenseindebeslissingenineenneonataleintensieve zorgen afdeling: een full population survey bij neonatologen en neonataleverpleegkundigen.,Nederlands-VlaamseWetenschapsdagenPalliatieveZorg, 21-22November2019,Antwerp,Belgium(oralpresentation).

209

Appendix

210

211

212

213

214

215

216

217

218

219

220

221

222

223

224

225

226

227

228

229

230

231

232

233

234

235

Vragen(interviewer) Prompts(Manieromverderintegaanopwatdeartsofverpleegkundigevertelt)

Introductie- Naamvragen- Bedankenvoorhunaanwezigheid- Jezelf(interviewer)voorstellen- Doelvanhetonderzoekenhetgesprektoelichten

- Wijzenopvertrouwelijkheid- Wijzenophetfeitdatzehetgesprekteallentijdekunnenstopzettenindiengewenst

- Uitlegvandeinformedconsent- VragenomhunGSMuittezetten

Nagaanofdeinformedconsentondertekendwerd

Introductievraag- Ikwilhetinditinterviewgraaghebbenoverdeproblematiekvanlevenseindebeslissingen.Welkelevenseindebeslissingenwordenhieropdedienstsomsgemaakt?

- Wijzenopanderevormenvanlevenseindebeslissingenindiennodig

Kernvragen- Watmaakthetvoorjoumoeilijkerwanneererzo’nlevenseindebeslissingenwordengemaakt?

- Voorjezelfalsarts/verpleegkundige(eigenrolbenadrukken)

- Bijvragenstellen:Þ Watbedoeljehierpreciesmee?Þ Kanjehierwatmeeropingaan?Þ Kanjehiereenconcreetvoorbeeldbijgeven

Þ Enwatmaaktehethierdanmoeilijker?

- DuidelijkwetenwelkeELDhetis.Indienalgemeen:vragenofhetbijandereELDsookzogaat

- Voorbeeldenvanmomentendiehetmoeilijkermaken

- Watmaakthetvoorjoumakkelijkerwanneererzo’nlevenseindebeslissingenwordengemaakt?

- Watzouhetvoorjougemakkelijkerkunnenmaken?

- Voorjezelfalsarts/verpleegkundige(eigenrolbenadrukken)

- Bijvragenstellen:Þ Watbedoeljehierpreciesmee?Þ Kanjehierwatmeeropingaan?Þ Kanjehiereenconcreetvoorbeeldbijgeven

Þ Enwatmaaktehethierdangemakkelijker?

- DuidelijkwetenwelkeELDhetis.Indienalgemeen:vragenofhetbijandereELDs

236

ookzogaat- Voorbeeldenvanmomentendiehetgemakkelijkermaken

- Voeltuzichondersteunddoorcollega’sbijhetnemenvanlevenseindebeslissingen?

- Voeltuzichondersteunddooroudersbijhetnemenvanlevenseindebeslissingen?

- Eventueelverschiltussen:Þ Ondersteundtijdenshetbeslissingsproces

Þ Ondersteundachteraf(psychologischeondersteuningbv)

- Welkgevoelhebtudannahetnemenenuitvoerenvanzo’nlevenseindebeslissingen?

- Enhebtudanhetgevoeldatdejuistebeslissinggemaaktwerd?

Indiennietvermeld:- EthischecommissiebetrokkenbijELDs?- Zijnersomsmomentenindecommunicatiemetouders/collega’sdiehetbeslissingsprocesmakkelijkerofmoeilijkermaken?

Eindvragen- Deinterviewermaakteenkortesamenvattingvanhetgesprek.- Vindtuditeengoedesamenvattingvanhetgesprek?- Zijnernogzakendienietaanbodgekomenzijnwaaruhetgraagnogoverwillenhebben?SlotvraagErisvandaagheelwatbesprokengeweest.Voorweafronden,hebtunogvragen?Heelergbedanktvooruwtijdenbijdrageaanonzestudie.Indienunogbijkomendevragenofopmerkingenhebtoverditinterview,dedata,ofhetonderzoekinhetalgemeen,aarzeldannietomcontactoptenemenmetdeuitvoerendeonderzoeker(geefgegevensmee).

237

238

239

Vragen(interviewer) Prompts(Manieromverderintegaanopwatdeartsofverpleegkundigevertelt)

Introductie- Naamvragen- Bedankenvoorhunaanwezigheid- Jezelf(interviewer)voorstellen- Doelvanhetonderzoekenhetgesprektoelichten

- Wijzenopvertrouwelijkheid- Wijzenophetfeitdatzehetgesprekteallentijdekunnenstopzettenindiengewenst

- Uitlegvandeinformedconsent- VragenomhunGSMuittezetten

Nagaanofdeinformedconsentondertekendwerd

Introductievraag- Ikwilhetinditinterviewgraaghebbenoverdeproblematiekvanlevenseindebeslissingen.Welkelevenseindebeslissingenwordenhieropdedienstsomsgemaakt?

- Wijzenopanderevormenvanlevenseindebeslissingenindiennodig

Transitievragen- Inwelkematewordjij/ubetrokkenbijhetnemenvanzo’nlevenseindebeslissingen?

- Niettelangblijvenstilstaanbijdezevraag

- Doel=wordenzebetrokkenjaofnee:Þ Ja=kernvragenbehoudenÞ Nee=doorvragenwaarzedanwel

bijbetrokkenworden(bvuitvoeringvandebeslissing,ondersteuningvanouders…)+dekernvragenandersformuleren:alserzo’nlevenseindebeslissingengemaaktworden,watmaaktdiesituatiedanmakkelijker/moeilijkervoorjou?

Kernvragen- Watmaakthetvoorjoumoeilijkerwanneererzo’nlevenseindebeslissingenwordengemaakt?

- Voorjezelfalsarts/verpleegkundige(eigenrolbenadrukken)

- Bijvragenstellen:Þ Watbedoeljehierpreciesmee?Þ Kanjehierwatmeeropingaan?Þ Kanjehiereenconcreetvoorbeeld

bijgevenÞ Enwatmaaktehethierdan

moeilijker?- DuidelijkwetenwelkeELDhetis.Indienalgemeen:vragenofhetbijandereELDs

240

ookzogaat- Voorbeeldenvanmomentendiehetmoeilijkermaken

- Watmaakthetvoorjoumakkelijkerwanneererzo’nlevenseindebeslissingenwordengemaakt?

- Watzouhetvoorjougemakkelijkerkunnenmaken?

- Voorjezelfalsarts/verpleegkundige(eigenrolbenadrukken)

- Bijvragenstellen:Þ Watbedoeljehierpreciesmee?Þ Kanjehierwatmeeropingaan?Þ Kanjehiereenconcreetvoorbeeld

bijgevenÞ Enwatmaaktehethierdan

gemakkelijker?- DuidelijkwetenwelkeELDhetis.Indienalgemeen:vragenofhetbijandereELDsookzogaat

- Voorbeeldenvanmomentendiehetgemakkelijkermaken

- Voeltuzichondersteunddoorcollega’sbijhetnemenvanlevenseindebeslissingen?

- Voeltuzichondersteunddooroudersbijhetnemenvanlevenseindebeslissingen?

- Eventueelverschiltussen:Þ Ondersteundtijdenshet

beslissingsprocesÞ Ondersteundachteraf

(psychologischeondersteuningbv)- Welkgevoelhebtudannahetnemenenuitvoerenvanzo’nlevenseindebeslissingen?

- Enhebtudanhetgevoeldatdejuistebeslissinggemaaktwerd?

Indiennietvermeld:- EthischecommissiebetrokkenbijELDs?- Zijnersomsmomentenindecommunicatiemetouders/collega’sdiehetbeslissingsprocesmakkelijkerofmoeilijkermaken?

Eindvragen- Deinterviewermaakteenkortesamenvattingvanhetgesprek.- Vindtuditeengoedesamenvattingvanhetgesprek?- Zijnernogzakendienietaanbodgekomenzijnwaaruhetgraagnogoverwillenhebben?SlotvraagErisvandaagheelwatbesprokengeweest.Voorweafronden,hebtunogvragen?Heelergbedanktvooruwtijdenbijdrageaanonzestudie.Indienunogbijkomendevragenofopmerkingenhebtoverditinterview,dedata,ofhetonderzoekinhetalgemeen,aarzeldannietomcontactoptenemenmetdeuitvoerendeonderzoeker(geefgegevensmee).