Healthcare reforms, inertia polarization and group influence

(1)

Citation for this paper:

Contandriopoulos, D., Brousselle, A., Larouche, C., Breton, M., Rivard, M.,

Beaulieu, M., … Perroux, M. (2018). Healthcare reforms, inertia polarization and

group influence. Health Policy, 122, 1018-1027.

https://doi.org/10.1016/j.healthpol.2018.07.007

UVicSPACE: Research & Learning Repository

_____________________________________________________________

Faculty of Human and Social Development

Faculty Publications

_____________________________________________________________

Healthcare reforms, inertia polarization and group influence

Damien Contandriopoulos, Astrid Brousselle, Catherine Larouche, Mylaine Breton,

Michèle Rivard, Marie-Dominique Beaulieu, Jeannie Haggerty, Geneviève

Champagne, Mélanie Perroux

2018

© 2018 The Authors. Published by Elsevier B.V. This is an open access article under

the CC BY-NC-ND license (

http://creativecommons.org/licenses/by-nc-nd/4.0/

).

This article was originally published at:

(2)

ContentslistsavailableatScienceDirect

Health

Policy

jo u rn al h om ep a ge :w w w . e l s e v i e r . c o m / l o c a t e / h e a l t h p o l

Healthcare

reforms,

inertia

polarization

and

group

inﬂuence

Damien

Contandriopoulos

a,∗

_,

_Astrid

_Brousselle

b

_,

_Catherine

_Larouche

g

_,

_Mylaine

_Breton

c,d

_,

Michèle

Rivard

j

_,

_{Marie-Dominique}

_Beaulieu

h,i

_,

_Jeannie

_Haggerty

e,f

_,

Geneviève

Champagne

d

_,

_Mélanie

_Perroux

k

a_School_of_Nursing,_University_of_Victoria,_PO_Box_1700,_STN_CSC,_Victoria,_BC_V8W_2Y2,_Canada

b_School_of_public_{administration,}_University_of_Victoria,_PO_Box_1700,_STN_CSC,_Victoria,_BC_V8W_2Y2,_Canada

c_Department_of_Community_Health_Sciences,_University_of_Sherbrooke,_200-150,_place_{Charles-LeMoyne,}_Longueuil,_QC_J4K_0A8,_Canada d_Charles_LeMoyne_Hospital_Research_Center,_200-150,_place_{Charles-LeMoyne,}_Longueuil,_QC_J4K_0A8,_Canada

e_Department_of_Family_Medecine,_McGill_University,_5858,_chemin_de_la_{Côte-des-Neiges,}_Montreal,_QC_H3S_1Z1,_Canada f_St._Mary’s_Hospital_Research_Center,_3830,_avenue_Lacombe,_Montreal,_QC_H3T_1M5,_Canada

g_School_of_Public_Health,_University_of_Montréal,_7101,_avenue_du_Parc,_Montreal,_QC_H3C_3J7,_Canada

h_Department_of_Family_and_Emergency_Medicine,_University_of_Montréal,_2900,_boulevard_{Édouard-Montpetit,}_Montreal,_QC_H3T_1J4,_Canada i_Researcher_at_the_CHUM_Research_Center,_900,_rue_Saint-Denis,_Montreal,_QC_H2X_0A9,_Canada

j_Centre_for_Modern_Indian_Studies,_University_of_Göttingen,_Göttingen,_Waldweg,₂₆₃₇₀₇₃_Göttingen,_Germany k_Regroupement_des_aidants_naturels_du_Québec_(RANQ)₁₈₅₅_rue_Dézéry,_Montréal,_H1W_2S1,_Canada

a

r

t

i

c

l

e

i

n

f

o

Articlehistory:

Received14August2017

Receivedinrevisedform28June2018 Accepted9July2018

Keywords: Healthpolicy Politics Medicalunions Socialnetworkanalysis Quebec(Canada)

a

b

s

t

r

a

c

t

Healthcaresystemsperformanceisthefocusofintensepolicyandmediaattentioninmostcountries. Que-bec(Canada)isnoexception,wheresuccessivegovernmentshavestruggledfordecadeswithapparently intractableproblemsincareaccessibilityoverall,poorperformance,andrisingcosts.Thisarticleexplores theunderlyingcausesofthedisconnectionbetweenthehighsalienceofhealthcaresystemdysfunctions inbothmediaandpolicydebatesandthelackofpolicychangelikelytoremedythosedysfunctions.

Academically,publicpolicies’evolutionisusuallyconceptualizedastheproductofcomplex, long-terminteractionsamongdiversegroupswithspeciﬁcpowersourcesandpreferences.Inthiscontext,we

wantedtoexamineempiricallywhetherdivergencesinstakeholders’viewsconcerningvarious

health-carereformoptionscouldexplainwhycertainpolicychangesarenotimplementeddespiteconsensus

ontheirprogrammaticcoherence.

Theresearchdesignwasanexploratorysequentialdesign.Datawereanalyzednarrativelyaswellas graphicallyusingamethodderivedfromsocialnetworkanalysisandgraphtheory.

Resultsshowedstrikingintergroupconvergencearoundaprogrammaticallysoundpolicypackage cen-tredonthegeneralobjectiveofstrengtheningprimarycaredeliverycapacities.Thoseresults,interpreted inlightofpoliticalscienceelitistperspectivesonthepolicyprocess,suggestthattheincapacitytoreform thesystemmightbeexplainedbyoneortwogroups’havingadefactovetoinpolicy-making.

1. Introduction

Giventheamountofresourcesinvested,thesymbolic impor-tance attributed to the concept of health, and the healthcare system’sroleindetermininghealth,caredeliverysystemsarecore

∗ Correspondingauthor.

E-mailaddresses:damien1@uvic.ca(D.Contandriopoulos),

astrid@uvic.ca(A.Brousselle),catlarouche@hotmail.com(C.Larouche),

Mylaine.Bretron@usherbrooke.ca(M.Breton),Michele.Rivard@umontreal.ca

(M.Rivard),Marie-Dominique.Beaulieu@umontreal.ca(M.-D.Beaulieu),

Jeannie.Haggerty@mcgill.ca(J.Haggerty),Genevieve.Champagne3@usherbrooke.ca

(G.Champagne),Melanie.Perroux@umontreal.ca(M.Perroux).

componentsofmodernsocieties.Thismakesthemthefocalpoint ofintensepolicyandmediaattention.Suboptimalcareaccessibility orquality,aswellasinefﬁcienciesinresourceallocation,are gen-erallyperceivedinthemediaandpoliticalspheresaslegitimate policyinterventiontargets.Common wisdomdictatesthat gov-ernmentsandpublicinstitutionsareexpectedto ¨dosomething¨to correctexistingdeﬁciencies.Yettheproblem–solutiontrajectory isfarfromlinear,andinmostsystemsdeep-seatedperformance shortcomingspersistdespitea seeminglynever-endingcycleof reforms[1,2].

Academically, the evolution of public policies is usually explainedfroma perspectivethatis morepolitical than instru-mental. Thatis,policies aretheproduct of complex,long-term

https://doi.org/10.1016/j.healthpol.2018.07.007

(3)

interactions through institutional mechanisms among diverse groupswithspecificpowersourcesandpreferences[3–7].Insuch aview,publicpoliciesarenotorientedprimarilytowards imple-mentingprogrammaticallysoundsolutionstotackledocumented deficiencies, but rather towards findinga politicallyacceptable equilibriumforthecorestakeholdersinvolvedinthepolicyprocess [8–13].

Previous research by this team [93–96] suggests that most obstacles to translating evidence on care delivery into struc-turesandpractices arepolitical.Bythis wemeanthattheycan beexplained bypolicyoptions’ levelof divergencefrom domi-nantsocialnormsandpowerfulstakeholders’preferences.Inthis context,wedesignedanempiricalstudytoexaminewhether dis-agreementsinstakeholders’viewsonvarioushealthcarereform optionscouldexplainwhycertainpolicychangeswerenot imple-menteddespitebroadscientiﬁcconsensusontheirprogrammatic coherence.

However,theresultsofthisstudy,asreportedinthepresent article,pointinadifferentdirection.Despitearobustsequential mixed-methodapproachdesignedtoidentifyandmeasure inter-groupdivergencesinstakeholdergroups’preferences,wemostly foundconvergence.Thisarticlepresentsthemethodsandﬁndings ofourstudyandbuildsonpoliticalsciencetheoriestoprovidean alternativehypothesistotheproblem–solutiondivideinhealthcare policyinQuebec.

2. Context:Quebechealthcaresystemreforms

Data for this study come from Quebec, one of Canada’s 10 provinces.In theCanadianfederalsystem,healthcare provision comesunderprovincialjurisdiction.LikeallCanadianprovinces, Quebechasapublic,tax-funded,universalhealthcaresystemthat coversall«medicallynecessary»care.Whiletheﬁnancing com-ponentofthesystemperformswell,thedeliverycomponenthas struggledfordecadeswithsevereproblemsincareaccessibility, waitinglists,overcrowdedemergencyrooms,overallpoor perfor-mance,andrisingcosts[14–17].

All publicly appointed commissions since the beginning of Quebec’s publichealthcare system[18–20] have recommended policy options that are consistent with the characteristics of high-performinghealthcaresystemsasidentifiedinmostofthe available scientific literature [17,21–25]. Verybroadly, the rec-ommendationsstresstheneedtoimprovetimelyaccessthrough the development of an accountable, primary care-centred sys-temrelying oninterprofessional teams and strong information systems [26]. The same recommendations appear in various othernationaland provincial commissions’documents[27–30]. However,despite this apparentconvergence between scientific evidenceandprovincial-andnational-levelpubliccommissions, theanalysisofreformsactuallyimplementedinQuebecoverthe past20yearssuggeststhatmanycriticalelements(forexample, issuesrelatedtointerdisciplinarycareorphysiciancompensation models)weresystematicallyignored[17].

3. Conceptualframework

Since the middle of the 20th _century, _most _theories _on

policy-makinghaveacknowledgedthatstakeholders’opinionsand preferencesplayacoreroleinpolicy-makingandimplementation processes[31,32].

Someof the mostobvious theoretical strands that focus on stakeholders’orinterestgroups’roleinpolicy-makingprocesses are found in the literature on interest groups and lobbying [5,33–42].Yet,evenpoliticalsciencemodelsthatwerenot devel-opedfromthepostulatethatgroupsarethemaindeterminantof

policy-makingstillattributea majorrole togroups’preferences inpolicy-makingprocesses.Forexample,theAdvocacyCoalition Framework[3,4,43,44]andtheliteratureonpolicycommunities andnetworks[11,45–51]andonagenda-setting[52–57]allshare thecommonassumptionthat stakeholder groups’opinionsand preferencesstructurepolicy-making.

Theperceivedcausalmechanismsinvolvedintheprocessdiffer, however,dependingonthetradition.Stakeholderscanstructure policy-makingthroughtheircapacitytoinﬂuencepublicopinion [39,56–58],whichinturncanhaveanimpactonlegislatorsthrough potentialelectoralconsequences [11,12,35,36] or throughmore subtleideologicalstructuringprocesses[8,59,60].Stakeholderscan alsoinﬂuencelegislatorsmoredirectlythroughthecontrolof valu-ablecommodities[33–35],suchasmoney(throughpartyfunding) or,moreoften,information[38,40,41,61].

Likewise, most contemporary models of policy-making rest on the concepts of policy arenas [52,53,62,63] or subsystems [44,53,64–67]. Thoseare definedaslong-terminteractions bya setof relativelystableparticipantsarounda given policyissue, aimedatinfluencingtheadoptionandimplementationofpublic policies[51,53,65].Bydefiningmultipleagendas(usuallypolitical, media,andpublic),theearlyagenda-settingmodels[68,69]played acentralroleinspreadingtheideathatpolicy-makingisabroader processthaninitiallythought,intermsofwhoisinvolved(notonly legislatorsandinterestgroupsbutthemedia,researchers,etc.)and whereithappens(notonlyinformalgovernmentalinstitutionsbut alsoinprivatediscussions,publicopinion,etc.)[52–54].

Questionsrelatedtothelevelofinfluencethatstakeholdersand interestgroups actuallyhavein policy-makinghowever,largely debatedintheliterature[70].Pluralistmodelssuggestthata vari-ety ofgroups participate in policy-makingand implementation processes,whichenhanceseachgroup’spotentialforinfluencing policyagendaswhileensuringthatspecificgroupsdonotalways monopolizeagendas[5,36,71,72].However,a significantportion oftheliteratureoninterestgroupsalsoarguesthatspecific inter-estgroups,suchasthoserepresentingbusinessinterests,havea largersayinpolicyprocessesandhaveeven, inthemostelitist view,appropriatedgovernmentalprerogatives[73–76].The liter-atureon“irontriangles”hassuggested,forexample,thattightly knitgroups havingstable relationshipswithauthorities usually exertthemostinfluenceonpolicydecisions[10,12,47,77].Along thesamelines,certaincharacteristicsofthehealthpolicyfield,such asthetechnicalcomplexityofmostissuesorthepolitical clout ofmedicalorganizationscreatesconditionspropitiousforelitist processes[6,8,78].

Buildingonthepoliticalscienceliteraturesummarizedabove, we hypothesized that Quebec’s incapacitytoadopt and imple-mentcoherentpolicysolutionstocaredeliveryproblemsandpoor healthcaresystemperformance,despitethehighsalienceofthese issuesinboththemediaandpoliticalagendas,haspoliticalroots. Speciﬁcally,wepositedthatcompetinginterestsorviewsbetween differenttypesofstakeholders—eitherbetweenprofessional occu-pational groups suchas physicians, nurses,administrators, and pharmacists,orbetweenkeystakeholdersandthehealthcare pro-fessionals they represent—could explain why effective reforms remainelusive.

4. Methodsanddata

Thisprojectwasbasedonanexploratorysequentialdesign[79] dividedintotwophases,oneinvolvingin-depthinterviewswith keyrepresentativesofhealthcaresystemstakeholdergroupsand theother,asurveyofgroupsofprofessionals—physicians,nurses, and pharmacists—and administrators. The sequential approach hadtwoobjectives:ﬁrst,tointegratestakeholderrepresentatives’

(4)

viewsintothesurveyinstrumentand,second,toassessthe conver-gencebetweenthoserepresentatives’viewsandtheviewsheldby membersoftheprofessionalgrouptheyrepresented.Aswediscuss later,datafrombothcomponentswerethenanalyzedusingthe sameapproachaimedatidentifyingconsensusandpolarization. Theprojectreceivedethicalapprovalfromtheethicscommittee oftheResearchCentreoftheCentreHospitalierUniversitairede Sherbrooke(CHUS).

4.1. Qualitativeinterviewswithrepresentativesofstakeholder groups

Thefirstphaseofthestudyexploredtheviewsofkey stakehold-ersinQuebec’shealthcarepolicyarena.Weappliedaninclusive definitionoforganizedinterests“thatincludesboth membership-based organizations[...] and ‘memberless’institutions suchas hospitals,privatecompanies,andregionalboards”[60].Informants whocouldarticulatetheviewsofeachgroupwerethenidentified onthebasis thattheyoccupied, orhad occupied,key positions withinthe groups or organizations (for example, presidents of professionalorders,unions,andassociations,deansofhealth pro-fessionfaculties,CEOsofuniversityhospitalsorofregionalhealth andsocialservicesagencies,etc.).Thirty-twoorganizationswere identifiedandinvitedtoparticipate;ofthose,31accepted (Que-bec’sunionof specialistphysiciansdeclined). Weconducted 31 interviews,althoughthenumberofintervieweeswashigher,as manyorganizationsoptedforgroupinterviews.Interviewswere conducted between February 2013 and May 2013. The semi-structuredinterviews, which lasted from45minto two hours, werelooselystructuredaroundfourthemes:1)thestrengthsof thecurrenthealthcaredeliverysystem;2)themainchallengesand problemsfacingthesystem(notpresentedhere,astheiranalysis endedupbeingredundantwithidentifiedsolutions);3)thebest solutionstoimprovethehealthcaresystem’sperformance;and4) theactorsandinterestgroupsthatappearedtobethemost pow-erfulandinfluentialinshapinghealthcarepolicies.Allinterviews wererecordedandtranscribed.

Interviewswerecategorizedaccordingtotheinterviewees’ pro-fessionalgroups(11physicians,ninenurses,nineadministrators, andtwopharmacists).Thecategorizationwasbasedmoreon func-tionthantraining.Forexample,ahospitalCEOwhotrainedasan MDbutwasworkingexclusivelyinhospitalmanagementwas cat-egorizedasanadministrator.Alltranscriptswerethenreviewed bytwodifferentresearchers,whoindependentlysummarizedthe interviewees’maincommentsandideas.Thesesummarieswere thencompared,discussed,andcombinedintoasingletext.Atthis step,summarieswereanalyzednarrativelytoprovideanin-depth understandingofeachinterview[80,81].Thesummarieswerethen usedinductivelytobuildalistofcodesorschemata[82] represent-ingtheessenceofeachdistinctcomplexstatementorproposition putforwardbyeachinformant.Codesweredividedintothefour themescoveredbytheinterviewquestions:strengths,problems, solutions,andinfluentialgroups.Tworesearchersworked inde-pendentlyondevelopingschematainductivelywhileatthesame timecodingeachtranscript.Codingdifferenceswerereviewedby theteammembersinconsiderationoftheoriginalinterview tran-scriptsuntilalldiscrepanciesininterpretationwereresolved.At theendoftheprocessthecodingstructurewascomposedofnine strengths,41problems, 46solutions, and 10influential groups. Oncethecodes werefinalized,allsummarieswerereviewedto ensurecodingwasconsistentthroughoutthecorpus.

4.2. Quantitativesurveyofprofessionalgroups

Thesecondphaseofthestudyexaminedhealthcare profession-als’andadministrators’viewsonpotentialsolutionstoreformand

improvethehealthcaresystem.BetweenAugustandOctober2015, asurveywassentbyemailand/ormailto2491peoplewhowere eitherphysicians,nurses,pharmacists,oradministrators.Arandom sampleof750physicians,748nurses,and750pharmacistswas generatedfromlistingsofprofessionalregistrationbodies(Quebec CollegeofPhysicians,theNurses’Corporation,andthe Pharma-cists’Corporation).Foradministrators,wetargetedpersonsintop managementpositionsinhealthcareinstitutionsandusedacensus approachtargetingthewholepopulation(n=243).

Thesurveyincluded29Likert-scaleclosedquestionsandone open-endedquestionaskingrespondentstosuggest three solu-tionsthattheMinistryofHealthmightconsidertoimprovethe healthcare system. Detailed results from this survey are pub-lished elsewhere [97]. The analysis provided here focuses on answers to the open-ended question. A total of 919 respon-dentsprovidedusablesurveyanswers(37%overallresponserate, 40% for physicians, 26% for nurses, 45% for pharmacists, and 33% for administrators). The sample wasgenerally representa-tiveofeachpopulationgroup(sex,training,location),exceptfor nurses,whereuniversity-trainednursedweresigniﬁcantly over-represented.Responseratefortheopen-endedquestionwas31% overall,with774usableanswers(253physicians,160nurses,291 pharmacists,and70administrators.)

Whilethesurveyquestionaskedforthreepotentialsolutions, the number of solutions provided ranged from one to 10 per respondent, with a mean of 3.2. As in the first phase of the study,answerswereindependentlycodedbytworesearchers.The schemata-basedcodingstructurecreatedforthe ¨solutions¨themeof thefirststudyphasewasusedasastartingpoint.Thesameprocess adoptedinthefirstphase,oftworesearchersdeveloping induc-tiveschematainparallel,wasalsousedhere.Fromtheoriginal 46solutions¨schemata,threewereslightlyeditedforbetterfitwith theanswersprovided,onewasnotusedatall,threeweremerged, andeightnewcodeswerecreatedtocapturenewanswersgiven byrespondents,foratotalof51schemata-basedcodes.

4.3. Avisualanalysismethodforqualitativedata

Inthecontextofthisproject,wedevelopedanewmethodfor thevisualanalysisofqualitativedata[98].Thismethod,derived fromtheﬁeldofsocialnetwork analysis,isaimedatidentifying convergencesanddivergencesinrelationsbetweeninformantsand theirviews.

Aspresentedintheprevioussection,theendpointofthe cod-ingprocessofeachphasewasabinarymatrixlinkinginformants toschemata-basedcodescorrespondingtotheiranswers.Froma socialnetworkanalysisperspective,suchmatriceseachconstitute atwo-modenetwork[83,84],wherenodesrepresenteither infor-mantsortheir(schemata-summarized)opinions,andwhereties representthelinksbetweeneach informantandtheiropinions. Whatconnectsinformantsisthusnotdirectlinksbetween them-selves,butrathertheircommonanswers(codes)ininterviews.

The matrices were imported into Cytoscape 3.2.0 software andrepresentedasa setof force-directedsociogramsusingthe Fruchterman-Reingoldalgorithm [85].Force-directedalgorithms are based on the principle that nodes are mutually repulsive and ties constitute attractiveforces. Thismeansthat the more a schemata-basednodeis ina balanced positionbetween clus-tersofinformant-basednodes,themoreconsensualitisamong thoseinformants.Themostinterestingfeatureofthis visualiza-tionapproachisthatnodessharingmanylinkswillbepositioned together,far from thenodes with which they share few links. This clustering effect allows visual assessment of convergence betweeninformants,betweenschemata,andbetweeninformants andschemata.Tovalidatethisdataanalysisapproach,we trans-formedourinterview-basedtwo-modematrixintotwoone-mode

(5)

Fig.1. StrengthsofQuebec’shealthcaresystem.

matricesand computedJaccardsimilaritycoefﬁcientsthatwere plottedusingmetric multidimensionalscaling(MDS).Adetailed discussionofthisvalidationapproachispublishedelsewhere[98]. To facilitate visual analysis, we also mapped node colour basedon thetwo typesof nodesdisplayed onthesociograms: schemata/codenodes(lightgrey)andinformantnodes (adminis-trators:blue;pharmacists:yellow;nurses:pink;andphysicians: green). Lastly, the nodes’ sizeswere mapped according tothe numberoftiesconnectedtoonenode.Ingraphtheory,the num-beroftiesconnectedtoanodeiscalledthenodedegree.Larger schemata/codenodesonthegraphsthusrepresenthighsalience ideassharedbylargernumbersofinformants.Onallsociograms, thenodewiththehighestdegreewasplottedwithadiameter20 timesthesizeofthenodewiththelowestdegree.

Beforepresenting thesociograms,wewant sotostressthat thedatavisualizationapproachusedhereisspeciﬁcallyaimedat identifyinginter-groupsclusteringinopinionsandpreferences.It isn’tdesignedtofacilitatetheoverviewofallparticipants opin-ions(readersinterestedinsuchananalysisofthedatapresented hereshouldread [98]).Becauseof thegraphoptimization algo-rithmused,somenodesandtextmayoverlapandmakeithardto identifysmallorperiphericnodes.However,thisdoesnotaffect thereadabilityoftheclusteringbycolourortheanalysisofcentral nodes.

5. Results

As stated in the Introduction, this project was designed to examinewhetherpolarizationinstakeholders’viewsconcerning healthcarereformoptionscouldexplainwhysomeimportant

pol-icychangesarenotimplementeddespitebroadscientiﬁcconsensus ontheirprogrammaticcoherence.However,wefoundverylittle group-basedpolarization.Insteadwefoundimpressive levelsof consensus.Inthissection,webrieﬂypresentresultsfromthe qual-itativecomponentsummarizingkeystakeholderrepresentatives’ viewsonthestrengthsofQuebec’shealthcaresystem(Fig.1)and identifyingthemostpowerfulactorsinpolicy-makingprocesses (Fig.2).Wethengointomoredetailontheanalysisofthetwo com-ponents’resultsregarding policysolutionstoimproveQuebec’s healthcaresystemperformance(Figs.3and4).

5.1. MainstrengthsofQuebec’shealthcaresystem

Ofthe31keystakeholdergroupswhoparticipatedinthe qual-itative interviews, 28 listed at least one systemstrength. Two strength-basedschematawereparticularlysalientintheanswers: freeanduniversalprovisionofcare(degreeof17)andqualityof careprovided(degreeof16).

Theﬁgureshowsahighlevelofconsensusregardingthe sys-tem’scorestrengths.Theotherinterestingobservationisthatthere isnoobviousclusteringofinformantsaccordingtoprofessional groupafﬁliation(visualizedasnodecolour).Thissuggeststhere arenoobviousgroup-baseddivergencesinopinionsregardingthe system’scorestrengths.

5.2. Mostpowerfulactorsandinterestgroupsinhealth-related policy-making

Ofthe31keystakeholdergroupswhoparticipatedinthe qual-itative interviews,26 identiﬁedat leastoneinﬂuential actoror

(6)

Fig.2.MostinﬂuentialgroupsinQuebec’shealthcaresystem.

interestgroup.Onegroup-basedschemareallystandsouthere: medical unions1_—of _which _there _are _one _for _GPs _and _one _for

specialists—werelistedasthemostinﬂuentialgroupin20ofthe 26interviews.Overall,thecoreresultisthatthereisaverystrong consensusthatphysicians’unionsarethemostinﬂuentialgroups inQuebec’shealthcarepolicy-makingprocesses.

Fig.2showssomeclusteringofinformantsaccordingtotheir professionalgroup:physiciansaremostlylocatedatthebottom rightofthesociogram,whilenursesareatthetopleft.This clus-teringismostlyexplainedbythefactthatalmostallphysicians(10 outof11)identifiedphysicians’unionsbynamewhereasnurses identifiedbothphysicians’unionsand ¨physicians äsagenericgroup. Itisalsoworthmentioningthatnarrativeanalysisofthedata suggestsmostrespondentsbelieveitwouldbeimpossibleto imple-mentsignificantpolicychangewithoutthesupportofthemedical unions.

5.3. Solutionstoimprovethehealthcaresystem’sperformance

Asstatedearlier,wehavedatafrombothcomponentsofour studyregarding respondents’ opinionsonthe bestsolutions to improvetheperformanceofQuebec’shealthcaresystem.

1 _Throughout_Canada,_physicians_are_covered_by_the_Rand_formula_and_therefore requiredtopayuniondues,regardlessoftheirunionstatus.Inmostprovinces,the duesarepaidtoasinglebodythatcombinesthefunctionofaunionrepresentingall physiciansintheirnegotiationswiththeprovinceandthefunctionofaprofessional association.Quebec’sphysicians’representationstructuresaredifferentfromthis model.InQuebec,specialistphysiciansandfamilydoctorsareeachrepresentedby onedistinct ¨Federation ¨whichrepresentthemforcontractnegotiations.The profes-sionalassociation–AMQ–isdistinctfromtheFederationsandreliesonvoluntary membershipfromallphysicians’groups.Weusethegenericterm ¨union¨todescribe Quebec’stwomedicalfederations.

Respondentsinall31qualitativeinterviewsmentionedpolicy optionstheyperceivedasplausiblewaystoimprovethesystem’s performance.Informantsinthequalitativephaseeachmentioned betweensix and 24 solution-basedschemata(average 13).The fivemostsalientsolutionswere ¨Strengthenprimarycare capaci-ties¨(degreeof28); ¨Rethinkthewayphysiciansarepaid¨(degreeof 22); ¨Redefineprofessionalrolesinprimarycare¨(degreeof21); ¨More relianceonperformancemonitoring¨(degreeof19);and Ïncrease physicians’accountability¨(degreeof17).

Evenmoreinteresting,however,istheobservationthat,ifthe 10solution-basedschematacentredaroundthehighlysalientnode ¨Strengthenprimarycarecapacities¨(redcircleonFig.3)areranked bytheirsalience,itreadsasaquitecoherentpolicypackage:rethink thewayphysiciansarepaid;redefineprofessionalrolesinprimary care; increase thereliance onperformance monitoring mecha-nisms;implementasystem-wideelectronichealthrecord;involve morephysiciansandnursesinthesystem’sgovernance;extendthe hoursforprimarycaredeliverystructures;increasenurse practi-tionerandregisterednursestaffinginprimarycareteams;improve hospitalmanagementcapacities(thisonedoesnotreallyfitinthe list);relymoreonpivotnursesandcarenavigatorsembeddedin primarycareteams;makebetteruseofinformationtechnologies; and make thesystemmore patient-centric.Overall, while each respondentproposeddifferentsolutions,whenputtogether,the solutionsthatclusteratthecentreofFig.3correspondtothe essen-tialcharacteristicsofanefficienthealthcaresystem.

Anotherimportant element is thelow level of group-based polarization.Atﬁrstsight,thereissomegroup-basedclusteringin Fig.3:nursestendtoclustermostlyontheleftofthegraph, admin-istratorsmostlyontheright,thetwopharmacistsare together, andmostphysiciansareinthetopleft.However,acloseranalysis showsthatsuchclustersarelinkedtopreferencesforperipheral solution-basednodes.Forexample,nurserespondentsmentioned

(7)

Fig.3. Component1solutionstoimproveQuebec’shealthcaresystem.

mandatoryuniversityleveltrainingfornurses(degreeof2)or bet-terworkingconditionsfornurses(degreeof2).However,theredoes notseemtobeanysigniﬁcantgroup-basedpolarizationofpolicy preferencesforthecore,highlysalientelementsatthecentreofthe graph.

Thislackofpolarizationwasasurpriseforusandwarranted further consideration. As stated in the Methods section, 31 of the32groupsinvitedtoparticipateinthequalitativeinterviews accepted,withtherefusalcomingfromthespecialistphysicians’ union—the more powerful of Quebec’s two physicians’ unions. Giventheconsensusweobservedinourdataregardingthe cen-tralityofphysicians’unionsinpolicy-makingprocesses(seeFig.2) andthelackofobservedpolarization,wecarefullyre-analyzedthe positionoftheGPphysicians’unioninourdataset.Inthestructural analysis,itdidnotdivergemuchfromtheothergroups.However, thenarrativediscourseanalysisshowedthat,whencomparedto theotherinterviews,thesupportexpressedbytheGPs for pol-icyreformoptionsputforwardbyothergroupswasmorelimited inbothscopeandform.Moreover,astheconsensusamongother intervieweespointedtowardspolicyavenuestostrengthen pri-marycarecapacity,itwasnotsurprisingthatitgenerallyconverged withGP(mostlyprimarycare)physicians’unionpreferences.

In additiontothesolutions suggestedin the31 stakeholder interviews,774oftheprofessionals(physicians,pharmacists,and

nurses)andadministratorssurveyedprovidedusableanswersto theopen-endedsurveyquestionaboutsolutionstheywould sug-gest tothe Ministry of Health to improve Quebec’s healthcare system.Onaveragerespondentsprovided3.2solutions(maximum 10,minimum1,modeandmedian3).Onaverage,eachofthe51 solution-basedschematawasidentifiedby47.9respondents.The fivemostsalientsolutionswere: ¨Redefineprofessionalrolesin pri-marycare¨(degreeof317); Ëxtendprimarycareaccessibility¨(degree of199,thisschemaisabroadeningof Ëxtendthehoursforprimary care deliverystructures¨from phase 1); ¨Rethinktheway profes-sionalsarepaid¨(degreeof149,abroadeningof ¨Rethinktheway physiciansare paid¨); ¨Strengthencareappropriateness¨(degreeof 123),and Ïmplementasystemwideelectronichealthrecord¨(degree of 122). Four of thesolution-based schemata withthe highest saliencefromthephase1qualitativeinterviewswerealsointhe top-10highestsalienceinthesurveydata.

Onceagain,the10solution-basedschematawiththehighest degreesandthatsurroundthehighlysalientnode“Redeﬁne prfes-sionalrolesinprimarycare”(degree of317)togetherpresenta remarkablycoherentprimary-carecentredpolicyproposal:extend accessibilityandhoursforprimarycaredeliverystructures(degree of 199);rethink thewayprofessionalarepaid (degreeof 149); strengthencareappropriateness(degreeof123);implementa sys-temwideelectronichealthrecord(degreeof122);improvecare

(8)

Fig.4.Phase2solutionstoimproveQuebec’shealthcaresystem.

coordination(degreeof103);generallystrengthenprimarycare (degreeof94);improveaccesstohospital-basedtestsand tech-nicalplatformsforambulatorypatients(degreeof86); increase nursepractitionerand registerednursestafﬁngin primarycare teams;(degreeof78)andincreasepatients’involvementincare andmanagement(degreeof68).Itshouldberecalledthatthese datacomefromasurvey-basedopenquestionand,giventhisorigin, areimpressivelycoherent.

Anotherinterestingresultis theabsenceofanygroup-based clusteringin Fig. 4. Thismeans that, while individualopinions andpreferences mayvary,the factof beinga physician,nurse, senioradministrator,orpharmacistdoesnotseemtosignificantly structurethosepreferences.Asstatedearlier,ouroriginalintentin thisprojectwastomeasureandstudygrouppolarizationof pol-icypreferences.WhenthefirstiterationsofFig.4producedthese unexpectedresults,wetried alternativesettingsforgraph opti-mization,butnoreasonablesettingprovidedanyimprovementin theidentificationofgroup-basedpolarization.

6. Discussion

The results presented lead to two important observations. First, the performance improvement solutions put forward by informants in both phases of this study constitute an impres-sivelycoherentpolicyproposalcentredonstrengtheningprimary carecapacities. We describeit as coherentbecausethediverse propositionsmakingupthisproposalareinterdependentintheir contributiontoachievingthebroadpolicygoalofsystemic perfor-manceimprovement.

Giventhevisualizationtechniquesusedhere,thefactthereis aclearcoreofsolution-basedschematathatsimultaneouslyhave

highsalienceandoccupyacentralpositiononthegraphsuggests thereisstrongconsensusonthesolutionstobeimplemented.

Beyondsimplydescribingthecontentofgroups’policy prefer-ences,wealsoaimedtoexaminethehypothesisthatdisagreements instakeholders’viewsonvarioushealthcarereformoptionsexplain whycertainpolicychangesarenotimplementeddespitebroad sci-entificconsensusontheirprogrammaticcoherence.Asmentioned above,however,ourdataarenotconvergentwiththis hypothe-sis.Thisopensthefieldtoalternativeexplanations. Ifthereare bothstrongsupportforcoherentreformavenuesandsignificant inter-groupconsensusaroundpolicyoptions,whyisitapparently impossibletoreformtheQuebecsysteminawaythatsolves per-sistentproblems?

Wewanttodiscusstwointerconnectedpossibleexplanations here.First,institutionaldysfunctionsorlarge-scaleincompetence couldbe crippling government’s capacity for action tosuchan extentthatnodeliberateactionispossible.Second,theactual pol-icyinfluenceofcertaininterestgroupscouldbesuchthatoneor moregroupswithnoappetiteforsignificantreformhaveadefacto vetocapacityoverandaboveanyconsensusorcoalitioninthefield. The first source of explanation can take many forms. For example, some political science literature can be used to sug-gestgovernmentactioncouldbelimitedbyinstitutionalfailures whereinpolicy-makerseitherfailtoproperlyassessstakeholders’ preferencesorareunabletoenactthemintopolicy[86,87]. Alter-natively,organizationalscienceliteraturecanbeusedtostressthe practicalchallengesrelatedtoeffectingdeliberatechangeinlarge complexsystems[88,89].

Thesecondsourceofexplanationisanchoredinelitetheory per-spectives[8,74,90].Incontrasttopluralists,whoseepolicy-making asarelativelyopenprocess,elitistsconceiveofpolicy-makingas monopolizedbyinterestgroupsexpressingthepreferencesofthe

(9)

richestandmosteducated.Ourinterviewswithstakeholdergroup representatives—allofthemvery savvyregarding actual policy-makinginQuebec—revealedthat,intheirview,physicians’unions unambiguouslydominatethepolicyﬁeld(seeFig.2).Thisresult ishighlycompatiblewithelitetheoryperspectives.Thecentrality andpowerofphysicians’unionsisnosurprise,asittallieswith previousresearch[17].However,ourresultssuggestthesegroups’ powermayhavebeenunderestimated.

Is Quebec’s situation documented here an exception? The politicalcloutofQuebec’sphysicians’unionsislikelysomewhat idiosyncratictotheprovince.ApplyingtheRandformulato physi-cians (see first footnote) automatically creates an exceedingly well-fundedpoliticalactorabletotapintoheapsofsymbolic capi-tallinkedtothenatureofmedicineandillness.If,furthermore,that actor’smissionissolelyfocusedondefendingphysicians’ finan-cialandprofessionalinterests[91],itskewsthepoliticalfieldina givendirection.Revolvingdoors—wherethesamepersonsswitch back-and-forthfromuniondutiestohigh-leveldecision-making positionsingovernment—alsoplayaroleinexplainingthepolitical powerofphysicians’unionsinQuebec2_._We_believe_the

combina-tionofallthose factorsexplainsthepeculiar powerbalancewe observed(seeFig.2).However,webelievethatphysicians’unions andmedicalorganisationsinotherjurisdictionscanrelyonatleast somesimilarpowersources[8].

But the paper’s conclusion isn’t so much about the power wieldedby physicians’union thanaboutthepotential overem-phasisgiventointer-groupdivergencesinexplainingstalematesin healthcarepolicymaking.Asanalystswithalongstanding inter-estinQuebec’shealthpolicies,weweresurprisedbytheresults obtainedhere.Thissuggeststhat,evenwhenpayingattention,one canunderestimateconsensusoroverestimatethelevelof polar-ization. Theavailability cascade biasis likely atplay here.This cognitivebiasdescribestheartiﬁcialoverestimationofanidea’s credibilitythroughthefeedbackprocessofitsrepetitioninpublic discourses.ItrelateswiththeapocryphalquoteattributedtoJoseph Goebbelsdiscussingpropaganda:«Repeatalieoftenenoughandit becomesthetruth».Inparticularthefunctioningofmediareporting, whereissuestendtobeframeddialectically[92],islikelyto over-playpolarization.Butwhateverthecausesare,overestimatingthe possibilitythathealthpolicyreformstalematesareexplainedby divergencesinactors’preferencesispotentiallydamagingattwo levels.

First,itmightobscuretherealvetopointsandtensionsandthus limittheattentionpaidtothose.Second,itmightdownplaythe politicalacceptabilityofsomereformavenues.Ourresultssuggest theremightbeunderexploredachievablepolicytargetsforhealth policyreformwhereasigniﬁcantconsensusexistsfor evidence-basedinterventionsbutwhereoneorafewgroupswereableto obfuscatetheissueandblockprogress.Thiswouldbethepolicy equivalentofpluralisticignoranceinsocialpsychology.Asituation wherethecollectiverealizationthataconsensusisindeedpossible wouldtipthescaletowardsmakingchangepossible.

7. Conclusion

Drawing upon the political science literature, this study explored the hypothesis that Quebec’s apparent incapacity to adoptandimplementcoherentpolicysolutionstotacklepervasive healthcaredeliveryproblemswas,tosomeextent,explainedby competingstakeholderpreferences.Thedataweanalyzeddonot supportthishypothesis.Infact,ourresultsrevealstriking

inter-2_As_one_example,_the_person_who_was_president_of_the_specialist_{physicians’}_union duringtheinterviewdatacollectionphaseofthisstudywasthehealthminister duringthesurveydatacollectionandstillwasatthetimeofwritingthisarticle.

groupconvergencearoundwhatappearstobeaprogrammatically soundpolicypackagecentredonthegeneralobjectiveof strength-eningprimarycaredeliverycapacities.

Whentaken together,theseresultsare compatible withthe stronglyelitisthypothesisthattheapparentincapacityto imple-mentcoherentpolicyreformsdespitebroadscientiﬁcconsensuson theirprogrammaticcoherencecomesfromoneortwogroups’ hav-ingadefactovetoinpolicy-making.Thisconclusionisimportant atboththepracticalandresearchlevels.

Practicallyspeaking,thereislittledoubtleftthatthespecialist physicians’unioniskingofthepolicyjungleinQuebec.However, ifwhatwefoundistrue,andifindeedthereisastrongpolicy con-sensusamongmostorallotherstakeholders,ourresultscanplaya roleinimprovingthepoliticalrepresentationofthosepreferences andhelpingtoerodethisgroup’svetocapacity.

Attheconceptualandresearchlevels,describingastrong con-centrationofpoliticalpowerhardlysettlesmuch.Thesourcesof this power,howit is expressed,and how itinterconnects with institutionalrulesareonlysomeofthequestionsnowopened.

Acknowledgments

ThisstudywassupportedbyagrantfromtheCanadianInstitutes ofHealthResearch(CIHR)#272944.Theauthorsaregratefultothe FondsdeRechercheduQuébec–SantéandCIHR,whichfunded AstridBrousselle’sCanadaResearchChair,andCIHR,whichfunds DamienContandriopoulos’AppliedPublicHealthResearchChair.

References

[1]BusbyC,MuthukumaranR,JacobsA.Realitybites:howCanada’shealthcare systemcomparestoitsInternationalpeers.Ottawa:C.D.HoweInstitute;2018.

[2]JohnstonS,HogelM.Adecadelost:primaryhealthcareperformancereporting acrossCanadaundertheactionplanforhealthsystemrenewal.HealthcPolicy 2016;11:95–110.

[3]SabatierPA.Theneedforbettertheories.In:SabatierPA,editor.Theoriesofthe policyprocess:theoreticallensesonpublicpolicy.Boulder:WestviewPress; 1999.p.3–18.

[4]SabatierPA,Jenkins-SmithHC.Policychangeandlearning.Anadvocacy coali-tionapproach.Boulder(CO):WestviewPress;1993.

[5]BaumgartnerFR,LeechBL.Basicinterests.Theimportanceofgroupsinpolitics andinpoliticalscience.Princeton(NJ:PrincetonUniversityPress;1998.

[6]CarpenterD.Ishealthpoliticsdifferent?AnnualReviewofPoliticalScience 2012;15:287–311.

[7]ContandriopoulosD,DenisJ-L.Leadingtransformationinpublicdelivery sys-tems:apoliticalperspective.In:DentM,FerlieE,TeelkenC,editors.Leadership inthepublicSector:promisesandpitfalls.London:Routledge;2012.p.44–61.

[8]AlfordR.Healthcarepolitics:ideologicalandinterestgroupbarrierstoreform. Chicago:TheUniversityofChicagoPress;1975.

[9]ConsidineM.MakingUpthegovernment’smind:agendasettingina parlia-mentarysystem.Governance1998;11:297–317.

[10]GrantW.PressuregroupsandBritishpolitics.London:MacmillanPressLtd.; 2000.

[11]NagelP.Policygamesandvenue-shopping:workingthestakeholderinterface tobrokerpolicychangeinrehabilitationservices.AustralianJournalofPublic Administration2006;65:3–16.

[12]SmithMJ.Pressure,power&policy.Pittsburgh:UniversityofPittsburghPress; 1993.

[13]ContandriopoulosD,DenisJ-L,LangleyA.Whatstructureshealthcarereforms? AcomparativeanalysisofBritishandCanadianexperiences.In:TavakoliM, DaviesHTO,MalekM,editors.healthpolicyandeconomics:strategicissuesin healthcaremanagement.Aldershot(England:AshgatePublishingLtd.;2001. p.219–36.

[14]CSBE.Perceptionsetexpériencesdesmédecinsdepremiéreligne:leQuébec comparé.Résultatsdel’enquêteinternationalesurlespolitiquesdesantédu CommonwealthFundde2015.Quebec(Canada):Commissaireàlasanteetau bien-être,GouvernementduQuébec,2016.

[15]CSBE.Perceptionsetexpériencesdelapopulation:LeQuébeccomparé.Recueil desrésultatspourchacunedesquestionsdel’enquêteinternationalesurles politiquesdesantéduCommonwealthFundde2016.Quebec(Canada): Com-missaireàlasantéetaubien-eˆtre,GouvernementduQuébec,2017. [16]CSBE.Apprendredesmeilleurs:étudecomparativedesurgencesduQuébec.

Quebec(Canada):Commissaireàlasantéetaubien-être,Gouvernementdu Québec,2016.

[17]ContandriopoulosD,BrousselleA.Reliableintheirfailure:ananalysisof health-carereformpoliciesinpublicsystems.HealthPolicy2010;95:144–52.

(10)

[18]Castonguay-NepveuCommission.RapportdelaCommissiond’enquêtesurla santéetlebien-êtresocial.Québec:ÉditeurofﬁcielduQuébec,1967-1970. [19]ClairCommissionRapportetrecommandationsdelacommissiond’étudesur

lesservicesdesantéetlesservicessociaux:Lessolutionsémergentes.Québec: GouvernementduQuébec,2000.

[20]RochonCommission.Rapportdelacommissiond’enquêtesurlesservicesde santéetlesservicessociaux.Québec:PublicationsduQuébec,1988. [21]MechanicD.Thetruthabouthealthcare:whyreformIsnotworkinginAmerica

(criticalissuesinhealthandmedicine).Camden,UK:RutgersUniversityPress; 2008.

[22]ShortellSM,SchmittdielJ,WangMC,LiR,GilliesRR,CasalinoLP,etal.An empir-icalassessmentofHigh-performingmedicalgroups:resultsfromanational study.MedicalCareResearchandReview2005;62:407–34.

[23]BakerGR,MacIntosh-MurrayA,PorcellatoC,DionneL,StelmacovichK,Karen B.Highperforminghealthcaresystems:deliveringqualitybydesign.Toronto, ON:LongwoodsPublishingCorporation;2008.

[24]KatzA,GlazierRH,VijayaraghavanJ.Thehealthandeconomicconsequences ofachievingaHigh-qualityprimaryhealthcaresysteminCanada“Applying whatworksinCanada:closingthegap”.Ottawa,ON:CanadianHealthServices ResearchFoundation;2009.

[25]McMurchyD.WhatarethecriticalattributesandbeneﬁtsofaHigh-quality primaryhealthcaresystem?Ottawa,ON:CanadianHealthServicesResearch Foundation;2009.

[26]LewisS.Asysteminnameonly—access,variation,andreforminCanada’s provinces.NTheNewEnglandJournalofMedicine2015;372:497–500.

[27]OntarioHSRC.lookingBack,lookingforward,sevenpointsforaction.Toronto: OntarioHealthServicesRestructuringCommission;2000.

[28]RomanowRJ.Buildingonvalues:thefutureofhealthcareinCanada–ﬁnal report.Ottawa:CommissionontheFutureofHealthCareinCanada;2002.

[29]HealthCouncilofCanada.Strategicplan2008/2009–2012/2013:takingthe pulsetowardimprovedhealthandhealthcareinCanada.Toronto,ON:Health CouncilofCanada;2008.

[30]Fyke KJ. Caring for medicare: sustaining a quality system. Regina, SK: SaskatchewanCommissiononMedicare;2001.

[31]FreemanJL.Thepoliticalprocess.NewYork:Doubleday;1955.

[32]LindblomCE.Thescienceof ¨muddlingthrough.PublicAdministrationReview 1959;19:79–88.

[33]AinsworthSH.Evaluatinginterestgroupinﬂuence:theimportanceoflobbyists Atlanta.AnnualMeetingoftheAmericanPoliticalScienceAssociation1989.

[34]AinsworthSH,SenedI.Theroleoflobbyists:entrepreneurswithTwoaudiences. AmericanJournalofPoliticalScience1993;37:834–66.

[35]Austen-SmithD.Informationandinﬂuence:lobbyingforagendasandvotes. AmericanJournalofPoliticalScience1993;37:799–833.

[36]BerryJM.Theinterestgroupsociety.NewYork:Longman;1997.

[37]BirnbaumJH.Thelobbyists.HowinﬂuencepeddlersworktheirWayin Wash-ington.Toronto:RandomHouse;1993.

[38]DeFigueiredoJM.Lobbyingandinformationinpolitics.BusinessandPolitics 2002;4:125–9.

[39]KollmanK.Outsidelobbying.Publicopinionandinterestgroupstrategies. Princeton(NJ):PrincetonUniversityPress;1998.

[40]MilbrathLM.Lobbyingasacommunicationprocess.PublicOpinionQuarterly 1960;24:33–53.

[41]MilbrathLM.TheWashingtonlobbyist.Chicago:RandMcNally;1963.

[42]Nownes AJ,DeAlejandroKW.Lobbyinginthenewmillennium:evidence ofcontinuityandchangeinthreestates.StatePoliticsandPolicyQuarterly 2009;9:429.

[43]Jenkins-SmithHC,SabatierPA.Thedynamicsofpolicyorientedlearning.In: SabatierPA,Jenkins-SmithHC,editors.Policychangeandlearning:anadvocacy coalitionapproach.Boulder:WestviewPress;1993.p.41–56.

[44]SabatierPA,Jenkins-SmithHC.Theadvocacycoalitionframework.In:Sabatier PA,editor.Theoriesofthepolicyprocess:theoreticallensesonpublicpolicy. Boulder:WestviewPress;1999.p.117–66.

[45]CarpenterD,EsterlingKM,LazerDMJ.Thestrengthofweaktiesinlobbying networks:evidencefromhealth-carepoliticsintheUnitedStates.Journalof TheoreticalPolitics1998;10:417–44.

[46]CarpenterD,EsterlingKM,LazerDMJ.Friends,brokers,andtransitivity:who informswhominWashingtonpolitics?JournalofPolitics2004;66:224–46.

[47]JordanG,MaloneyWA.Accountingforsubgovernments:explainingthe per-sistenceofpolicycommunities.ADMSociety1997;29:557–83.

[48]KnoepfelP,Kissling-NafI.Sociallearninginpolicynetworks.Politics&Policy 1998;26:343–67.

[49]KnottJH.Themultipleandambiguousrolesofprofessionalsinpublic policy-making.Knowledge:Creation,Diffusion,Utilization1986;8:131–53.

[50]KonigT,BrauningerT.Theformationofpolicynetworks:preferences, insti-tutionsandactors’choiceofinformationandexchangerelations.Journalof TheoreticalPolitics1998;10:445–71.

[51]RhodesRAW.Policynetworks:aBritishperspective.JournalofTheoritical Pol-itics1990;2:293–317.

[52]DearingJW,RogersEM.Agenda-setting.ThousandOaks,CA:SagePublications; 1996.

[53]KingdonJW.Agendas,alternatives,andpublicpolicies.NewYork:Harper CollinsPublishers;1984.

[54]McCombsME,ShawDL.Theevolutionofagenda-settingresearch:twenty-Five yearsinthemarketplaceofideas.JournalofCommunication1993;43:58–67.

[55]Soroka SN.Issueattributesandagenda-settingbymedia,thepublic,and policymakers inCanada.InternationalJournalofPublicOpinionResearch 2002;14:264–328.

[56]SorokaSN,WlezienC.Opinionrepresentationandpolicyfeedback:Canadain comparativeperspective.CanadianJournalofPoliticalScience/Revue canadi-ennedesciencepolitique2004;37:531–59.

[57]SorokaSN,WlezienC.Opinion–Policydynamics:publicpreferencesand pub-licexpenditureintheUnitedKingdom.BritishJournalofPoliticalScience 2005;35:665–89.

[58]WestDM,HeithD,GoodwinC.HarryandLouiseGotoWashington:political advertisingandhealthcarereform.JournalofHealthPoliticsPolicyandLaw 1996;21:35–68.

[59]BachrachP,BaratzMS.Twofacesofpower.TheAmericanPoliticalScience Review1962;56:947–52.

[60]ContandriopoulosD.Onthenatureandstrategiesoforganizedinterestsin healthcarepolicymaking.ADMSociety2011;43:45–65.

[61]Smith P. Political communication inthe UK: a study of pressure group behaviour.Politics1999;19:21–7.

[62]BaumgartnerFR,JonesBD,LeechBL.Mediaattentionandcongressional agen-das.In:IyengarS,ReevesR,editors.Dothemediagovern?:Politicians,Voters, andreportersinAmerica.ThousandOaks,CA:SagePublications;1997.p. 349–63.

[63]BursteinP.Theimpactofpublicopiniononpublicpolicy:areviewandan agenda.PoliticalResearchQuarterly2003;56:29–40.

[64]StoneD.Policyparadox:theartofpoliticaldecisionmaking,reviseded. New-York:Norton&Company;2002.

[65]SabatierPA.Policychangeoveradecadeormore.In:SabatierPA,Jenkins-Smith HC,editors.Policychangeandlearninganadvocacycoalitionapproach.Boulder (CO):WestviewPress;1993.p.13–39.

[66]JonesBD,BaumgartnerFR.Thepoliticsofattention:howgovernment priori-tizesproblems.Chicago:TheUniversityofChicagoPress;2005.

[67]Baumgartner FR,Jones BD. Agendas andinstability in Americanpolitics. Chicago:TheUniversityofChicagoPress;2009.

[68]CobbRW,ElderCD.Thepoliticsofagenda-building:analternativeperspective formodernDemocratictheory.TheJournalofPolitics1971;33:892–915.

[69]McCombsME,ShawDL.Theagenda-settingfunctionofmassmedia.ThePublic OpinionQuarterly1972;36:176–87.

[70]DürA,DeBièvreD.Thequestionofinterestgroupinﬂuence.JournalofPublic Policy2007;27:1–12.

[71]HeaneyMT.Brokeringhealthpolicy:coalitions,parties,andinterestgroup inﬂuence.JournalofHealthPolitics,PolicyandLaw2006;31:887–944.

[72]LindblomCE,WoodhouseEJ.Thepolicy-makingprocess,3rd.ed.Englewood CliffsNJ:PrenticeHall;1993.

[73]YackeeJW,YackeeSW.Abiastowardsbusiness?Assessinginterestgroup inﬂuenceontheU.S.Bureaucracy.JournalofPolitics2006;68:128–39.

[74]LowiTJ.Theendofliberalism:ideology,policyandthecrisisofpublicauthority. NewYork:Norton;1969.

[75]GilensM,PageBI.TestingtheoriesofAmericanpolitics:elites,interestgroups, andaveragecitizens.Perspectivesonpolitics2014;12:564–81.

[76]SalisburyRH.Interestrepresentation:thedominanceofInstitutions.American PoliticalScienceReview1984;78:64–76.

[77]HecloH.Issuenetworksandtheexecutiveestablishment.In:KingA,editor.The NewAmericanpoliticalsystem.Washington:AmericanEnterpriseInstitute; 1978.

[78]DobrowMJ,GoelV,UpshurREG.Evidence-basedhealthpolicy:contextand utilisation.SocialScience&Medicine2004;58:207–17.

[79]Creswell JW, Plano Clark VL.Designing and conducting mixed methods research.ThousandOaks:SagePublications;2011.

[80]ChouliarakiL,FaircloughN.discourseinlatemodernity:rethinkingcritical discourseanalysis.Edinburgh:EdinburghUniversityPress;1999.

[81]HardyC,PalmerI,PhilipsN.Discourseasastrategicresource.HumanRelations 2000;53:1227–48.

[82]DiMaggio P. Culture and cognition. Annual Review of Sociology 1997;23:263–87.

[83]BorgattiSP,EverettMG.Networkanalysisof2-modedata.SocialNetworks 1997;19:243–69.

[84]BorgattiSP,EverettMG,JohnsonJC.Analyzingsocialnetworks.ThousandOaks, CA:Sage;2013.

[85]FruchtermanTM,ReingoldEM.Graphdrawingbyforce-directedplacement. Software:Practiceandexperience1991;21:1129–64.

[86]PetryF,MendelsohnM.PublicopinionandpolicymakinginCanada1994-2001. CanadianJournalofPoliticalScience2004;37:505–29.

[87]PowellGBJ.Politicalrepresentationincomparativepolitics.AnnualReviewof PoliticalScience2004;7:273–96.

[88]AlvessonM,SpicerA.Astupidity-basedtheoryoforganizations.Journalof ManagementStudies2012;49:1194–220.

[89]Brunsson N.The irrationality ofaction and action rationality: decisions. Ideologies and Organizational Actions Journal of Management Studies 1982;19:29–44.

[90]HewittC.Policy-makinginpostwarBritain:anation-leveltestofelitistand pluralisthypotheses.BritishJournalofPoliticalScience1974;IV:187–216.

[91]FMSQ.FMSQmissionstatement;2018.

[92]ContandriopoulosD,AbelsonJ,LamarcheP,BohémierK.Thevisiblepoliticsof theprivatizationdebateinQuebec.HealthcarePolicy2012;8:67–79.

[93]ContandriopoulosD.OntheNatureandStrategiesofOrganizedInterestsin HealthCarePolicyMaking,Administration&.Society2011;43(1):45–65.

(11)

[94]ContandriopoulosD,BrousselleA.Reliableintheirfailure:ananalysisof health-carereformpoliciesinpublicsystems.HealthPolicy2010;95(2-3):144–52.

[95]ContandriopoulosD,DenisJ-L.LeadingTransformationinPublicDelivery Sys-tems:APoliticalPerspective.In:DentM,FerlieE,TeelkenC,editors.Leadership inthePublicSector:PromisesandPitfalls.London:Routledge;2012.p.44–61.

[96]ContandriopoulosD,HudonR,MartinE,ThompsonD.Tensionsentrerationalité techniqueetintérêtspolitiques:l’exempledelamiseenœuvredelaLoisurles agencesdedéveloppementderéseauxlocauxdeservicesdesantéetdeservices sociauxau.QuébecCanadianPublicAdministration2007;50(2):193–217.

[97]AstridBrousselle,DamienContandriopoulos,JeannieHaggerty,Mylaine Bre-ton, Michèle Rivard, Marie-Dominique Beaulieu, Geneviève Champagne, MélaniePerroux.StakeholderViewsonSolutionstoImproveHealthSystem Performance,HealthcarePolicy.HealthcarePolicy2018.INPRESS.

[98]ContandriopoulosD,LaroucheC,BretonM,BrousselleA.Asociogramisworth athousandwords:proposingamethodforthevisualanalysisofnarrativedata. QualitativeResearch2017;18(1):70–87.