Towards process-oriented care delivery in hospitals - [PDF Document] (2024)

Towards process-oriented care delivery in hospitals

Citation for published version (APA):

Vos, L. (2010). Towards process-oriented care delivery in hospitals. [Doctoral Thesis, MaastrichtUniversity]. NIVEL. https://doi.org/10.26481/dis.20101008lv

Document status and date:Published: 01/01/2010

DOI:10.26481/dis.20101008lv

Document Version:Publisher's PDF, also known as Version of record

Please check the document version of this publication:

• A submitted manuscript is the version of the article upon submission and before peer-review. There canbe important differences between the submitted version and the official published version of record.People interested in the research are advised to contact the author for the final version of the publication,or visit the DOI to the publisher's website.• The final author version and the galley proof are versions of the publication after peer review.• The final published version features the final layout of the paper including the volume, issue and pagenumbers.Link to publication

General rightsCopyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyrightowners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with theserights.

• Users may download and print one copy of any publication from the public portal for the purpose of private study or research.• You may not further distribute the material or use it for any profit-making activity or commercial gain• You may freely distribute the URL identifying the publication in the public portal.

If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license above,please follow below link for the End User Agreement:

www.umlib.nl/taverne-license

Take down policyIf you believe that this document breaches copyright please contact us at:

[emailprotected]

providing details and we will investigate your claim.

Download date: 16 Sep. 2022

Towardsprocess­orientedcaredelivery inhospitals LetiVos

thesis_Vos_17x24_v01.pdf

ISBN978‐94‐6122‐022‐6http://www.nivel.nl[emailprotected]Telefoon0302729700Fax0302729729©2010NIVEL,Postbus1568,3500BNUTRECHTCoverdesign: MarcObbens Wordprocessing/layout: ChristelvanWellPrinting: Datawyse/UniversitairePersMaastricht Allrightsreserved.Nopartofthispublicationmaybereproduced,storedinaretrievalsystemortransmitted, in any form or by any means, electronic, mechanical, photocopying, recording orotherwise,withoutthepriorwrittenpermissionofNIVEL.Exceptionsareallowedinrespectofanyfairdealingforthepurposeofresearch,privatestudyorreview.

thesis_Vos_17x24_v01.pdf

Towardsprocess­orientedcaredeliveryinhospitals

PROEFSCHRIFT terverkrijgingvandegraadvandoctoraandeUniversiteitMaastricht, opgezagvandeRectorMagnificus,Prof.mr.G.P.M.F.Mols, volgenshetbesluitvanhetCollegevanDecanen, inhetopenbaarteverdedigen opvrijdag 8oktober2010om12.00uur door LetiVos

thesis_Vos_17x24_v01.pdf

Promotores:Prof.dr.G.G.vanMerodeProf.dr.C.Wagner(VrijeUniversiteitAmsterdam)Prof.dr.P.P.Groenewegen(UniversiteitUtrecht)Beoordelingscommissie:Prof.dr.C.Spreeuwenberg(voorzitter)Prof.dr.H.J.J.M.Berden(UniversiteitvanTilburg)Prof.dr.J.vanEngelshovenProf.dr.J.A.M.MaarseProf.dr.J.deVries(RijksuniversiteitGroningen)ThestudiespresentedinthisthesiswereconductedattheFacultyofHealth,Medicineand Life sciences, Maastricht University Medical Centre+ and NIVEL, the NetherlandsInstitute forHealth ServicesResearch inUtrecht. Part of the researchwas performedwith financial support of MESOS Medical Centre and ZonMw, the NetherlandsOrganisation for Health Research and Development. Printing of the book has beensupportedfinanciallybyNIVEL,theNetherlandsInstituteforHealthServicesResearch.

thesis_Vos_17x24_v01.pdf

‘Itsoundedanexcellentplan,nodoubt,andveryneatlyandsimplyarranged;the

onlydifficultywas,thatshehadnotthesmallestideahowtosetaboutit;....’

Alice’sAdventuresinWonderlandbyLewisCarroll

thesis_Vos_17x24_v01.pdf

thesis_Vos_17x24_v01.pdf

│7

Contents1 Generalintroduction 92 Towardsanorganisation‐wideprocess‐orientedorganisation ofcare:aliteraturereview 313 Evaluatinghospitaldesignfromanoperationsmanagement perspective 574 Doescase‐mixbasedreimbursem*ntstimulatethedevelopmentof process‐orientedcaredelivery? 775 Howtoimplementprocess‐orientedcare:acasestudyonthe implementationofprocess‐orientedin‐hospitalstrokecare 936 Applyingthequalityimprovementcollaborativemethodto processredesign:amultiplecasestudy 1117 Generaldiscussionandconclusion 135Summary 159Samenvatting 167Dankwoord 175CurriculumVitae 179Listofpublications 183

thesis_Vos_17x24_v01.pdf

thesis_Vos_17x24_v01.pdf

│9

Chapter1

Generalintroduction

thesis_Vos_17x24_v01.pdf

thesis_Vos_17x24_v01.pdf

Chapter1 │11

1.1 IntroductionIn April 2000, Susan M. went for her preventive breast cancer screeningmammogram.Themammogramshowedanabnormality,andSusanhadtoseeasurgeon.Thefirstavailableappointmentdatewas fiveweeks later.Thesurgeonthen decided he wanted the radiologist to perform a needle biopsy. The firstavailableappointmentdate for the radiologistwasanother fourweeks later. ‘Icompletelyfreakedout,’Susansays. ‘Icouldn't imaginespendinganothermonthwiththishangingovermyhead.’Ultimately,aweekaftertheneedlebiopsySusanwastoldbythesurgeonthattheabnormalitywasnothingtoworryabout.PatientslikeSusanM.intheaboveexampleoftenhavetowaitforlongperiodsbefore they are seen by medical specialists, and subsequently need to waitagain for diagnostic examinations or for treatment while experiencing greatuncertainty about their illness. In 2007, 19.7% of all Dutch hospitaldepartments (medical specialties) indicated that patients had to wait longerthan fourweeks for their first visit to theoutpatients’ clinic, and6.4%of thedepartments indicated that after their first visit to the outpatients’ clinic,patients had to wait longer than seven weeks for their clinical treatment1.Although these waiting times still exceeded the desired period, they wereconsiderably less than they were in 2005. In 2005, respectively 22.8% and9.1% of all Dutch hospital departments indicated that patients had to waitlongerthanfourweeksfortheirfirstvisittotheoutpatients’clinic,andlongerthansevenweeksfortheirclinicaltreatment1.Nationalandinternationalattentiontoqualityproblemslikewaitingtimesandthe poor coordination of care activities has made hospital management andmedical specialists aware that the way care is delivered needs to bereorganised.However,littleisknownabouthowtoimprovetheorganisationofcaredeliveryintermsofqualityandefficiency.Moreover,it isnotknownhownewwaysoforganisingcaredeliveryshouldbeimplementedintoahospital’sorganisation.Until now,many of the ideas for improvements of coordinationandprocesscontrolhavebeenderivedfromresearchinindustry,includingthesuccessfulconceptreferredtoas‘businessprocessorientation’.Thisconcept’sbreakthroughoccurredatthebeginningofthe1990sunderthename‘BusinessProcess Reengineering’2. Successful examples of the application of theprinciplesofbusinessprocessorientationfortheorganisationofcaredeliveryto specific patient groups are known: for instance, involving women withsuspected breast cancer (see Box 1.1). It is unknown, however, whetherbusiness process orientation or process‐oriented care delivery can also beappliedsuccessfullyathospitallevel.

thesis_Vos_17x24_v01.pdf

12│ Towardsprocess­orientedcaredeliveryinhospitals

Box1.1 Process‐orientedcaredeliveryforwomenwithsuspectedbreastcancer

Forwomenwith suspected breast cancer, the organisation of care has improved considerablyduringthelast10years.Inthepast,womenwereforcedtowaitformanydays,orevenweeks,tosee the medical specialist; then they had to return to the hospital many times for diagnosticexaminations, and were given the results at the earliest one week after the last examination.However, the introductionof an ‘outpatient clinic formamma care’ changed this.Womenwithsuspected breast cancer now have access to a special outpatient clinic within five days of theinitialmammogram. In this clinic, patients are attended toby a surgeononamultidisciplinaryteamthatisdevotedtodelivermammacare.Onthesamedayasseeingthesurgeon,thepatientreceives all prescribed diagnostic examinations (palpation by surgeon or nurse practitioner,diagnostic imaging [mammogram, echography], and cytology or needle biopsy). In addition,appointments aremade for furtherdiagnostic examination suchasMRIwhen this is indicated.The patient is given the results of the examinations within five days of this first visit to thehospital.

Therefore, this thesisaimstoextendtheknowledge inthefieldofhealthcareoperationsmanagement by testing the hypothesis that the implementationofprocess­oriented caredelivery leads tobetteroutcomes in termsofqualityandefficiency at hospital level. The scope of this thesis is limited to the process‐orientedorganisationofpatient flow.Hence,organisationof the flowofotherresourcesthroughthehospitalsuchascapital, labour,andmaterials(suppliesneededforoperations,medicalfiles,andsoon)arenottakenintoaccount.In this introductory chapter, attention is first paid to the organisationalcharacteristics ofDutchhospitals, the setting of this study. Next,wedescribehowcaredeliveryhasbeenorganisedtraditionally,andtrytoexplainwhyandhow hospitals shouldwork towards a process‐orientedway of care delivery.Subsequently, certain conflicts that can evolve during the introduction ofprocess‐oriented care delivery in a hospital are discussed. This chapter endswiththegeneralaim,researchquestions,andanoutlineofthisthesis.1.2 Dutchhospitals:organisationalcharacteristicsAs a result of the tendency towards specialisation in themedical profession,Dutchhospitalstraditionallyhavefunctionalorganisationstructures3.Withinafunctionalstructure,organisationaldepartmentsarebuiltaroundtheskillsandknowledgeoftheirprofessionals,likeinternalmedicine,radiology,andsoon3‐5.In this organisational structure, responsibilities are coupled to the differentorganisationaldepartmentsandnottocaredeliveryprocesses.Belowthelevelofthehospitalboardaremanagerswhoareresponsiblefortheperformanceof

thesis_Vos_17x24_v01.pdf

Chapter1 │13

divisions or clusters. Within these divisions or clusters, operational units ordepartments functionwith their ownresponsiblemanagers. In theseunitsordepartments,professionalsaregroupedaccordingtotheirspecialty.In the Netherlands, most medical specialists (65%) are contracted but notsalariedbythehospital6,7.Asaresult,medicalspecialistsarepoorlyintegratedinto the hospital organisation. However, specialists are responsible for thepatient’smedicalcareprocessanditsoutcome,andthereforeplayacrucialrolein the realisation of the hospital organisation’s objective. During the last 10years, the hospital organisation and its contracted medical specialists havedemonstrated increasing integration by the implementation of dualmanagementstructures,whichmeansthatmedicalspecialistsparticipateinthemanagementof hospitals.Thus,medical specialists are co‐responsible for thepolicies,management, andorganisationofcaredelivery. Inpractice,hospitalsgivemeaningtothisdualmanagementstructureindifferentways:forexample,bymaking amedical specialist amember of thehospital boardor by the co‐management of clusters of departments8. However, the divide betweenmanagementandmedicalspecialistsisnotdiminished.Therealityisastrategiccontrolbattlebetweenhospitalmanagementandmedicalstaff7,9.Duetotheabove‐mentionedcharacteristics,Dutchhospitalscanbeconsidereddurable virtual organisations: the patient may have the impression thatservices are provided by one and the same organisation, but in fact manydifferent autonomous organisations, the medical specialties, are operatingbehindthesharedfacade,andareinvolvedinthedeliveryofservices6.Withinthisvirtualorganisation,eachdepartmentoperatesasaseparate‘silo’withitsown goals3, 4, because responsibilities are coupled to the differentorganisational departments and not to care delivery processes. As aconsequence,themanagementofeachorganisationaldepartmentisconcernedwith maximising its own efficiency rather than with the efficiency of entirepatient care processes10. This means two sets of goals co‐exist and perhapscompete: those at the strategic hospital level, where hospital policies andstrategictargetsaresetandinvestmentsrequiredaredetermined,andthoseatthe operational department level, where freedom exists in terms of how toorganiseoperations6.AdditionalinformationaboutDutchhospitalcareispresentedinBox1.2.

thesis_Vos_17x24_v01.pdf

14│ Towardsprocess­orientedcaredeliveryinhospitals

Box1.2 Dutchhospitalcare

Hospitalcategories:HospitalcareintheNetherlandsisdeliveredalmostexclusively inprivatenot‐for‐profit organisations. In 2009, therewere eighty‐five general hospitals, eight universityhospitals,andanumberofspecialisedhospitals11.Generalhospitalsconcentrateontreatmentandnursing.Twenty‐sixofthegeneralhospitalsalsoprovidemedicaltrainingandhighlyspecialisedcare that requires expensive and specialised instruments (e.g. heart andneurosurgery, in vitrofertilisation). University hospitals also conduct scientific research and education for medicalprofessionals, and develop new medical technologies and techniques. Specialised hospitalsconcentrateonspecificformsofcareoronaspecificillness(e.g.orthopaedics,asthma,epilepsy,ordialysis).Medicalspecialists(orclinicians):Mostmedicalspecialistsworkinonehospital.Themajorityisself‐employed,workinginapartnershipwithotherspecialistsinthesamefield.In2007,therewere 15,360 medical specialists in the Netherlands for approximately 16,4 million Dutchinhabitants.Thisrepresentsagrowthof20%comparedtothenumberofmedicalspecialists in200012. However, the growth in full‐time equivalents ofmedical specialists is smaller: namely,16%.MedicalspecialistsarerepresentedintheMedicalBoardofahospitalandareorganisedinnationalprofessionalgroupspermedicalspecialty.Access to hospital care: In the Netherlands, general practitioners (GPs) are gatekeepers forhospital care. Except for cases of emergency, patients only consult a medical specialist afterreferralbyaGP13.Between2000and2007,thepercentageofDutchpeoplewhovisitedamedicalspecialistatleastonceayearincreasedfrom38to41%14.Somepatientsonlyvisitahospitalfordiagnosis and/or therapy and then leave (outpatients), while others are admitted and stayovernight(inpatients).

1.3 TheorganisationofcaredeliverywithinhospitalsOwing to institutionalpressure, topmanagement inDutchhospitalscurrentlyfeels the need to pay closer attention to patients’ interests and to search forways to tailor theorganisationof caredelivery accordingly.This section firstexplainstheprinciplesoftheorganisationofcaredeliveryandthewayinwhichcare traditionally is organised; it then describes why and how hospitals cantailortheorganisationofcaredeliverytopatients’needsormakecaredeliverymoreprocess‐oriented.TheorganisationofcaredeliveryinhospitalsbelongstothefieldofOperationsManagement(OM).OMinhospitalsconcernsthedesign,planning,andcontrolof coordination mechanisms between patient flow and diagnostic andtherapeutic activities to maximise output/throughput given availableresources, taking into account different requirements for delivery (elective,semi‐urgent, urgent), acceptable standards for delivery reliability (waiting

thesis_Vos_17x24_v01.pdf

Chapter1 │15

times), and acceptable medical outcomes15. OM in hospitals has two aims:increasingtheefficiencyandqualityofcaredelivery.Thedesign,planning,andcontrolofcoordinationmechanismsaredeterminedbythestructureofthehandledlogisticalconcept.Thisconceptconsistsoffourmainelements16:1. Thebasicformofanorganisationorthemainorganisationstructure;2. Theoperatingsystem;3. The information system (where in the organisation is what kind of

informationneeded);and4. The organisation of personnel (qualifications, competences and

responsibilitiespertask).The strategic goals of an organisation (including logistic goals)make up thestartingpointforthedesignofthesefourcoherentelements.Thealignmentofthefourelementsultimatelydetermineshowcaredeliveryisorganised.1.3.1Traditional­functional­logisticalconceptforcaredeliveryAs described in Section 1.2, hospitals traditionally have a functionalorganisationstructure.Thestructureofthelogisticalconceptforthecontrolofpatientflowwithinthistraditionalorganisationcanbedescribedasfollows:Thebasic form of thehospital looks likeanetwork (seeFigure1.1).Within afunctionalstructure,eachdepartmentoperatesasaseparate‘silo’withitsowngoals3, 4. Due to current increasing specialisation, the patient’s care processfrequently consists of a chain of care activities delivered by differentdepartments.Figure1.1Functionalbasicstructure(basedon:VanMerodeetal.[2004]17)

In the operating system of a hospital with a functional basic structure,departmentsarecontrolledindependentlyofeachother(seeFigure1.2),whichmeans that theydonot exchange information for theplanningandcontrolofpatientflows(‘stand‐alone’).Asaconsequenceofthefunctionalstructure,each

thesis_Vos_17x24_v01.pdf

16│ Towardsprocess­orientedcaredeliveryinhospitals

careactivitywithinachainisplannedonthebasisofavailabilityofresources,after the previous care activity is completed. This leads to considerablecoordinationand to a number of transfer points perpatient. The result is anindeterministicorfuzzypatternofpatientflowthroughthehospital.Afterall,there aremanyways inwhich each patient canmove through the hospital’snecessarychainofcareactivities.Inaddition, theavailabilityofdepartmentalcapacities is frequentlynotbasedonactualdemand,butisdeterminedwithoutinformationaboutthevolumeandthe type of patients needing to be served (‘push’ planning). For example, theplanningofanoperating theatre–whichdetermineswhichmedical specialtymayusetheoperatingtheatreatwhichtimeoftheweek–issetdownyearly.This planning is thus done before patients are scheduled for surgery, andthereforebeforetheneededoperatingcapacityperdaypermedicalspecialtyisknown.Figure1.2 Functionaloperationalcontrolsystem–‘standalone’

The way in which processes are controlled determines what information isneeded bywhom to take the appropriate decision at the rightmoment. In afunctional logistical concept, every department controls its own patient flowwith theaimtooptimise itscapacityuse,whichmakes thedegreeofcapacityutilisationimportantinformationwithregardtocontrollingthedepartments.Within the functional logistical concept, personnel is organised according tospecialisation.Asdescribedinthischapter’s introduction,theuseofthisfunctional logisticalconcepthasledtowaitingtimesanddelaysinpatients’careprocess.

thesis_Vos_17x24_v01.pdf

Chapter1 │17

1.3.2Aprocess­orientedlogisticalconceptforcaredeliveryBeforeexplainingwaystotailortheorganisationofcaredeliverywithrespectto patients’ needs, the reasons for developing a process‐oriented logisticalconceptasamannertotakepatients’requirementsintoaccountarediscussed.ReasonsforchangingthefunctionallogisticalconceptofhospitalsHospitals are currently under external pressure to improve the quality andefficiency of care delivery, andmore specifically to improve the coordinationbetween specialists’ subtasks to provide all components of care needed bypatients.Thisisaresultofseveraldevelopments,whichinclude:‐ Anincreasedcomplexityofprocessesasaresultofchangesinhealthcare

needs, the use of new technologies, and increased specialisation. Theincreasing number of patients with more multidisciplinary health careneeds calls for additional coordination. At the same time, the increasingspecialisation of medical professionals and the development of newtechnologiesincreasinglydividescaredeliveryintospecialistsubtasksthatrequireevenmorecoordination18;

‐ Theneedforefficientutilisationofresourcesandreductionincostsduetopoliticalpressure tocontrolhealthcareexpenditures. In theNetherlands,healthcareexpenditures increasedfrom26,9billioneuro in2000to42,5billioneuroin200614;

‐ Increased pressure on the part of stakeholders – patients and healthinsurance companies – to improve the quality and efficiency of services,amongothersbydecreasingwaitinglistsandin‐processwaitingtimes6.

These pressures to improve the quality and efficiency of care delivery areformalised by governmental actions. Since 1996, Dutch hospitals have, forexample,beenboundbylawtoprovideeffective,efficient,andpatient‐centredcare delivery19. In addition, the Dutch government has introduced regulatedcompetition in the hospital sector to force hospitals to rethink theirperformance (price‐quality ratio, focus on patient’s needs). Further, thegovernment has implemented a new case‐mix‐based reimbursem*nt systemthataimstostimulatecoordinationofcareactivitiesincareprocessesandtheefficiency of care delivery. This new reimbursem*nt system will graduallyreplacethefee‐for‐servicepaymentofhospitalcare,whichrewardsdeliveryofindividualprocedures(i.e.hospitaladmissions,numberof in‐patientdays)uptoapredefinedmaximum20.Inthisnewsystem,hospitalbudgetsarebasedonthe number of delivered Diagnosis Treatment Combinations (DTC). A DTCconsistsofalldiagnosis‐and treatment‐relatedcosts incurredby thehospitaland the medical specialists. Thus, it covers the whole chain of delivered

thesis_Vos_17x24_v01.pdf

18│ Towardsprocess­orientedcaredeliveryinhospitals

services: from an initial consultation or examination to the final check‐up21.Consequently, instead of on the number of individual procedures, hospitalbudgetsbecomedependentontheefficientandeffectivedeliveryofDTCs,andthus the total path of diagnosis and treatment of a patient,which consists ofseveral subtasks. However, also within this new system hospital budgets arenot open ended. Hospitals negotiate with health care insurers regarding themaximumnumberofDTCstheymaydeliverwithin1year.However, it is difficult tomeet these requirements for caredeliverywithin afunctional organisation because of the limited possibilities of managers tocontrolthemanycoordinationandtransferpointsinthecareprocesses5.Thisisdueto:‐ The ill‐defined ownership of care processes for a defined category of

patients versus the clearly defined responsibilities of the medicalspecialtiescontributingservices.Specialistsplayakeyrole,astheydecideonthepatient’stransfertoanextstageintheprocess,butthecoordinationofcareactivitieswithintheprocessasawholedoesnotlieintheirhands6;

‐ Thelimitedsharingbetweenmedicalspecialistsandmanagerswithregardto available information about the chain of operations that forms thepatientprocess;

‐ The need to coordinate different parts (departments) of the hospitalorganisation,withoutbeingabletocontrolallpartsoftheorganisation6.

Therefore,theneedtoimprovethequalityandefficiencyofcaredeliverycallsforanewlogisticalconcept.Process­orientedlogisticalconceptsTooptimisepatients’carepathsthroughthehospital,thecomplexityofserviceintegrationwithitsmanycoordinationandtransferpointsinthecareprocessmust be reduced17. According to Galbraith, this can be accomplished in twodifferentways:A. Changingtheoperatingsystemofthelogisticalconcept;orB. Changingthebasicformoftheorganisation22.Bothways lead theoretically to amore process‐oriented organisation of caredelivery.Ad.A.Changingtheoperatingsystemtoaprocess­orientedoperationalcontrolsystemWithin this approach, top management decides to keep the functional basicstructure and to change the operating system towards a process‐orientedoperationalcontrolsystemforpatientswhoneedstandardcare:namely,thoseforwhomtherequiredcomponentsofcaredeliveryareknownaheadof time

thesis_Vos_17x24_v01.pdf

Chapter1 │19

because they are embedded in guidelines, task routines or protocols23 (seeFigure 1.3). This system implies that individual patients’ demands areaggregated into logistically hom*ogeneous groups and that departmentsexchange information to coordinate the flow of these patient groups acrossdepartmentboundaries. Informationexchange is thenused tobridgebarrierserected by the functional organisation structure by bringing together allmedicalspecialiststhatareinvolvedinthedeliveryofcaretoaspecificpatientgroup24. As well as this, the information exchange gives departments theopportunitytoadjusttheircapacityplanningtotheexpecteddemandofthesepatientgroups,whichreducesthechanceofwaitingtimesanddelaysforthosepatients(‘pull’planning).Figure1.3 Process‐oriented operational control system within functional organisational

structure–‘integrationbycommunication’

Thus,within this approach, hospitals organise carepathwaysorprogrammesfor specific patient groups, which establish the sequence of care activities(diagnostics, consultations, treatment) along with the responsibilities ofprofessionalsinvolvedinthediagnosisandtreatmentofhom*ogeneouspatientgroups, as in the example of ‘process‐oriented care delivery forwomenwithsuspectedbreastcancer’describedinBox1.1.Ad.B.Changingthebasic formoftheorganisationtoaprocess­orientedbasicstructureInthisapproach,topmanagementdecidestorestructureitsbasicstructureintonewself‐containedmultidisciplinarydepartmentsthatarebasedontheneedsofthepatients(seeFigure1.4).Thesedepartmentsarethencomposedinsuchawaythattheycanhandleacareprocessascomprehensivelyaspossible,haverelatively few interdependencies with other departments, and anindeterministicpatientflowisreduced4,25,26.

thesis_Vos_17x24_v01.pdf

20│ Towardsprocess­orientedcaredeliveryinhospitals

Figure1.4 Process‐orientedbasicstructure

Asaresultofthischangeinbasicstructure,achangeintheoperationalcontrolsystemisnotneededbecausedepartmentsnowhandleacareprocessintegrally(see Figure 1.5). Within the organisational units, the tasks are performedautonomously and integratively by cross‐functional teams27. In comparisonwith functional logistical concepts, employees are confronted with morediversified and multifaceted tasks. In the case of complex processes, thefunctional division of work within the team can be maintained; thecoordination is still facilitated by the elimination of departmental bordersbetweenteammembers4.Figure1.5 Process‐oriented operational system within a process‐oriented basic structure

(‘integralcontrol’)

This change in thebasic structure isnot only convenient forpatientswho fitwithin one multidisciplinary department (shorter waiting time, even fewerpatienttransfersbetweendepartments)butitalsosimplifiesthecoordinationof patient flow in the organisation, which in its turn makes more preciseplanning possible28. Patients who have to be treated within more than onemultidisciplinary department also experience advantages because fewertransfers are needed. The feasibility of a process‐oriented basic structure islimited by the pronounced functional specialisation ofmedical professionals,which is indispensable because of the know‐how intensity of many hospitalservices4. Moreover, it is not clear whether the creation of multidisciplinarydepartments will be feasible for strategically less important services (e.g. tosmallmultidisciplinarypatientgroups).

thesis_Vos_17x24_v01.pdf

Chapter1 │21

Anexampleofahospitalthathandlessuchaprocess‐orientedbasicstructureistheMayoClinicinRochester(US).Thishospitalisstronglydifferentiatedwithrespecttopatientgroupsortypeofillness.Foreachpatientgroup,forexampleheart diseases, an integrated system exists from prevention to revalidation.Medical specialists and allied health professionals are specialised in thetreatment of a certain patient group. Each patient experiences a hospitalspecialisedforaspecificillness.These process‐oriented logistical concepts described in this section ([A] theimplementation of a process‐orientedoperational control systemand [B] theimplementation of a process‐oriented basic structure) require differentinformation systems and another organisation of personnel. In this kind oforganisation, theperformanceofcareprocessesand in‐processwaiting timesbecome more important than the capacity utilisation of departments. Thismeans that ICT systems have to be redesigned to deliver the correctinformationto thepersonwhohas to takethe timelydecisiontocontrolcareprocesses.Inaddition,divisionoftasks,competences,andresponsibilitieshaveto be reconsidered. It requires thatmedical specialists and allied health careprofessionals specialise in specific diseases, and that management isresponsible for the efficient performance of care processes insteadof for thecapacityutilisationoftheirdepartment.To achieve optimal efficiency, the physical layout of a building ideally has tosupport the new process‐oriented logistical concept29. The layout influenceswork processes, patient flow, information, and decisions across theorganisation.However, asexpertshavenoted,a supportingphysical layout isextremelydifficulttorealiseinapre‐existingbuilding30.1.4 Implementation of new logistical concepts in a hospital

organisation Intheprevioussection,wedescribedtwopossiblewaystochangethelogisticalconcept insuchawaythatpatients’needsare integrated into thewaycare isdelivered. These changes need to be introduced into the hospital system. Asdescribed in Section 1.2, a traditional hospital system exists of topmanagement, divisions, and functional departments, in which medicalprofessionals are grouped according to their specialty; this system needs tooperate within a hospital building in order to reach organisational goals.Conflictsarisewhengoalsanddecisionsof thecomponentsof this traditionalhospital system (top management, divisions/departments, medical profes‐

thesis_Vos_17x24_v01.pdf

22│ Towardsprocess­orientedcaredeliveryinhospitals

sionals)andpreconditionssuchasthehospitalbuildinglayoutarenotalignedwith each other when a new logistical system is introduced. Figure 1.6visualises three types of conflict. In this section, these conflicts are discussedandpossiblesolutionsaresuggested.Figure1.6 Threetypesofconflict

Conflict1:Conflictbetweenthephysicallay­outofahospitalbuildingandaprocess­orientedlogisticalconceptTo achieve optimal efficiency, the physical layout of a building ideally has tosupport the logistical concept being applied. The basic form of the logisticalconceptisfrequentlyalsothebasisforthelayoutofthehospitalbuilding,suchas the organisational separation of outpatients’ and inpatients’ clinics in afunctional basic form31.However, the life cycleof ahospital building ismuchlonger than that of a logistical concept. During the life of a building,environmental changes occur, such as ageing populations, more demandingpatients,new insights concerning treatment, and technological developments.Thesemayleadtochangesinthelogisticalconcept, likethechangetowardsaprocess‐oriented logistical concept.As a consequence, therewill be a conflictbetweenthelayoutofahospitalbuilding(whichisbasedonthebasicformofapreviousfunctionallogisticalconcept)andthenewlyappliedlogisticalconcept,which can result in the implementation of the process‐oriented logisticalconcepthavinglimitedeffects.

Topmanagement

DivisionsandDepartments

Medicalprofessionals

Hospitalbuildinglayout

Strategicgoalsanddecisionsincludinglogisticalconcept

Goalsanddecisionsofdivisionsanddepartments

Goalsanddecisionsofmedicalprofessionals

HospitalsystemConflict1

Conflict2

Professionalgroupspermedicalspecialty

Conflict3

thesis_Vos_17x24_v01.pdf

Chapter1 │23

Possible solution: since a hospital building layout is static and can only bealtered at considerable expense, it is essential to create flexibility in thebuildingtoovercomethisconflictbetweenthelayoutofthebuildingandnewlydeveloped logistical concepts. This flexibility can, for example, be createdthrough the standardisationof consultation rooms32.Medical professionals ofdifferent specialties are then no longer bound to a specific section of thehospitaltoperformtheirsubtaskinacareprocess.Toensurethattheflexibilityof the designed hospital building layout fits a desired logistical concept, thelayoutshouldbeassessedbeforethehospitalisbuilt,basedonpossiblefuturedevelopmentsinlogisticalconceptsandchangesinpatientmix.Conflict 2: Conflict between top management’s wish to implement aprocess­oriented logistical concept, the responsibilities of departments,andinterestsofmedicalspecialistsAn effective implementation of a process‐oriented logistical concept requiresthat the efficiency of care processes is given priority above the capacityutilisation of a department, both by department managers and medicalspecialists. After all, there is a trade‐off between short waiting times andcapacity utilisation: a higher capacity utilisation (‘full agendas’) reduces thepossibilityof immediatecare forpatients,and leads towaiting times28. In thetraditional Dutch hospital organisations in which middle managers areresponsiblefortheperformanceoftheirdepartmentsordivisions,theneedforcoordination at process level can consequently lead to friction withmanagement responsibilities at division or department level. As departmentsloseautonomy,medicalspecialistsfeeltheyarelessabletoinfluenceworkingconditions and to have control over their production. This hampers theimplementationofaprocess‐orientedlogisticalconcept.Possiblesolution:Onewaytoovercomethisconflictbetweentopmanagement’swishto implementaprocess‐oriented logisticalconceptontheonehand,andtheresponsibilitiesofdepartmentsandthe interestsofmedicalspecialistsontheotherhand,isfortopmanagementtobasetheallocationoffinancialmeans,inwhole or in part, on the total care of a patient during an acute episode ofillness33.Thisisinlinewiththetheoryoftargetengineering,whichstatesthatexternalpressuresrelatedtopaymentsystemscanbeusedtobridge internalconflicts of interests hampering efficiency and quality5, 34. As a result of theprocess‐based payments, chain of caremembers have towork together fromthestart toorganisecaredelivery insuchawaythat itwill deliver fullvalue.Theresultshouldbethatthecaredeliverybecomesbetteradaptedtotheneedsof patients, but also that quality and efficiency improves as chain of caremembersanalyse,redesign,andmonitortheirprocessesinmoredetail.

thesis_Vos_17x24_v01.pdf

24│ Towardsprocess­orientedcaredeliveryinhospitals

Conflict 3: Conflict between the autonomy ofmedical specialistswhiledelivering care inaccordancewith thevaluesof theirownprofessionalgroupsandtherequirementsofaprocess­orientedlogisticalconceptThe introduction of a process‐oriented logistical concept requires frommanagersandmedicalspecialistsanotherviewofthereality.Theyhavetoshifttheirfocusfromfunctiontoprocess.However,thischangeofperspectivebringsprofessionalsintoanareaofconflict:medicalspecialistswanttopractisetheirprofession autonomously in accordance with the high professional values oftheir own group, but are also expected to contribute to the process‐orientedcaredelivery35.Medicalprofessionalswhohavegrownaccustomedtoworkinginkeepingwithparticularproceduresduringyearsof training andeducationhave to change their routines36. They feel that their autonomyat the level ofcaredeliverytoanindividualpatientisbeinglimitedbecausetheiractionhastofitthepathchosen37.Thisfeelstothemasthoughtheyhavetogiveuppartoftheir professional autonomy, as the implementation of process‐orientedlogistical concepts requires that the components of care delivery to specificpatientgroupsbestandardised.Consequently,thesenewroutinesor‘rulesforworkprocesses’donotdevelopautomaticallywhen topmanagementdecidestoimplementaprocess‐orientedlogisticalconcept.Possiblesolution:Therearemanykindsofprocessimprovementorinnovationmethodstodeveloproutinesthatfitaprocess‐orientedlogisticalconcept.Itis,for instance, assumed that the use of process improvement or innovationmethods that focus on changing the patterns of interaction betweenprofessionals and the redesign of the overall work flow stimulatesprofessionalstodevelopandimplementnewroutines.However,apreconditionisthatprofessionalsaremotivatedtoparticipateinprojectsthataimtosetnewroutines. Professionals can then execute their autonomywith other involvedprofessionalsindecidingwhatformastandardcareprogrammeshouldtake. 1.5 StudyaimandresearchquestionsCentralhypothesis:‘Theimplementationofprocess­orientedcaredeliveryleadstobetteroutcomesintermsofqualityandefficiencyathospitallevel’.In order to test the central hypothesis of this thesis, the overall researchquestionis:‘Is the implementation of a process­oriented logistical concept effective forimprovingqualityandefficiencyofcaredeliveryathospitallevel?’

thesis_Vos_17x24_v01.pdf

Chapter1 │25

Thegeneralresearchquestionisdividedintoanumberofsub‐questionsandisexploredinseveralstudies.Thedesignsandmethodsusedinthesestudiesaredescribed in the chapters of this thesis. First, it is important to knowwhichprocess‐oriented logistical concepts topmanagement wants to implement toimprove thequality and efficiencyof caredelivery at hospital level: does topmanagement tend to change the operating system and/or the basic form tostrive towards process‐oriented care delivery? And how effective are theselogisticalconceptsinimprovingthequalityandefficiencyofcaredelivery?Thisleadstothefirstresearchquestion:1. ‘Whichprocess­orientedlogisticalconceptsdohospitalsimplementtodeliver

process­oriented care,andhoweffectiveare these in improving thequalityandefficiencyofcaredelivery?’

Asdescribed inSection1.4, itmaybeexpected that thesuccessofaprocess‐orientedlogisticalconceptisinfluencedbyanumberofconflictsthatmayarisewhenanewlogisticalconcept is introduced inahospitalorganisation. In thisthesis,thepossibleeffectsofanevaluationoftheflexibilityofhospitalbuildinglayout,aprocess‐basedpaymentsystem,andimplementationmethodsusedtoovercometheaccompanyingconflictsareassessedinordertoexplorewhetherthe use of these ‘tools’ may positively affect the outcomes of theimplementation of a process‐oriented logistical concept. This leads to thefollowingresearchquestions.2. ‘Doesevaluationoftheflexibilityofhospitalbuildinglayoutcontributetothe

implementationofanewlogisticalconcept?’3. ‘Doprocess­basedpayments contribute to the implementationofaprocess­

orientedlogisticalconcept?’4. ‘Do improvementand innovationmethodscontribute to the implementation

ofaprocess­orientedlogisticalconcept?’1.6 OutlineofthesisTheresearchquestionsareaddressedinChapters2‐7ofthisthesis(seeTable1.1). Chapter 2 dealswith the first research question. It presents a literaturereview on experiences of hospitals that have implemented process‐orientedlogistical concepts of caredelivery.Next, in Chapter 3, the contributionof anevaluationmethod to examine the flexibility of hospital building layout with

thesis_Vos_17x24_v01.pdf

26│ Towardsprocess­orientedcaredeliveryinhospitals

regard to the implementation of new logistical concepts is assessed usingcomputer simulation techniques. Chapter 4 is concerned with the thirdresearch question, analysing the effect of process‐based payment systems ontheimplementationofaprocess‐orientedlogisticalconcept.Subsequently, thecontributions of a process innovation and improvement method on theimplementation of a process‐oriented logistical concept are assessed inChapters5and6.Chapter5focusesontheeffectivenessofaprocessinnovationmethodthatradically implementstheprocess‐oriented logisticalconcept,andChapter6centresontheeffectofaprocessimprovementmethodthataimstoimplementincrementallyaprocess‐orientedlogisticalconceptwiththehelpofstandardisedchangeideas.Finally,Chapter7presentsthemainfindingsofthisthesis on the sub‐questions, along with a discussion addressing someconsiderations of the research, and conclusions regarding themain researchquestion. In addition, several recommendations for future research andhospitalpracticearepresented.Table1.1 Overviewofthethesis

Researchquestion Chapter

1. Implementationofprocess‐orientedlogisticalconceptsandtheir

effectivenessinhospitals

Chapter2Towardsanorganisation‐wideprocess‐orientedorganisationofcare:aliteraturereview

2. Contributionofanex‐anteevaluationmethodforhospitalbuildinglayout

totheimplementationofnewlogisticalconcepts

Chapter3Evaluatinghospitaldesignfromanoperationsmanagementperspective

3. Contributionofprocess‐basedpaymentsystemstotheimplementationofaprocess‐orientedlogisticalconcept

Chapter4Doescase‐mix‐basedreimbursem*ntstimulatethedevelopmentofprocess‐orientedcaredelivery?

Chapter5Howtoimplementprocess‐orientedcare:acasestudyontheimplementationofprocess‐orientedin‐hospitalstrokecare

4. Contributionofprocessimprovement/innovationmethodstotheimplementationofaprocess‐orientedlogisticalconcept Chapter6Applyingthequalityimprovement

collaborativemethodtoprocessredesign:amultiplecasestudy

thesis_Vos_17x24_v01.pdf

Chapter1 │27

References1. Westert GP, Van den Berg MJ, Koolman X, Verkleij H: Dutch health care

performance report. Bilthoven: National Institute for Public Health andEnvironment;2008.

2. HammerM,Champy J:Reengineering thecorporation:amanifesto forbusinessrevolution.London:NicholasBrealey;1993.

3. Lega F, DePietro C: Converging patterns in hospital organization: beyond theprofessionalbureaucracy.HealthPolicy2005,74(3):261‐281.

4. Vera A, Kuntz L: Process‐based organization design and hospital efficiency.HealthCareManagementReview2007,32(1):55‐65.

5. Van Merode GG: A prelude of 2004 Antwerp quality conference: targets andtarget values ‐ integrating quality management and costing. Accreditation andQualityAssurance2004,9:168‐171.

6. De Vries G, Bertrand JWM, Vissers JMH: Design requirements for health careproductioncontrolsystems.ProductionPlanning&Control1999,10(6):559‐569.

7. PlochgT,LombartsK,WitmanY,KlazingaNS,KruijthofK:Doctorsandmanagers.ProblemsinDutchhospitalsresemblethoseinBritishhospitals.BritishMedicalJournal2003,326(7390):656.

8. Van Lindert H, Hutten J, Groenewegen P: Specialist and hospital policy: theclassical organsiation disappears (in Dutch). Medisch Contact 2003,58(30/31):1164‐1166.

9. Groenewegen PP, Hansen J, Ter Bekke S: Professies en de toekomst:veranderende verhoudingen in de gezondheidszorg. Utrecht: VVAA/Springer;2007.

10. Ludwig M, Van Merode F, Groot W: Principal agent relationships and theefficiencyofhospitals.TheEuropeanJournalofHealthEconomics2010,11:291‐304.

11. RIVM: Volksgezondheid toekomst verkenning, nationale atlas volksgezondheid(http://www.zorgatlas.nl),versie3.18.Bilthoven:RIVM.

12. Capaciteitsorgaan: Capaciteitsplan 2008 voor de medische, tandheelkundige,klinischtechnologischeenaanverwantevervolgopleidingen.Utrecht;2008.

13. SchäferW,KronemanM,BoermaW,VandenBergM,WestertG,DevilléW,VanGinnekenE:TheNetherlands:healthsystemreview.Healthsystemsintransition2010,12(2):1‐240.

14. CBS,Statistischjaarboek.Voorburg/Heerlen:CentraalBureauvoordeStatistiek;2008.

15. Vissers JMH, Beech R: Health operationsmanagement: patient flow logistics inhealthcare.London:Routledge;2005.

16. VisserHM,VanGoorAR:Workingwith logistics (inDutch).Groningen/Houten:Wolters‐Noordhoff;2004.

thesis_Vos_17x24_v01.pdf

28│ Towardsprocess­orientedcaredeliveryinhospitals

17. Van Merode F, Molema H, Goldschmidt H: GUM and six sigma approachespositionedasdeterministictoolsinqualitytargetengineering.AccreditationandQualityAssurance2004,10:32‐36.

18. LeeKH:Thehospitalmovement‐Acomplexadaptiveresponsetofragmentationofcareinhospitals.AnnalsAcademyofMedicineSingapore2008,37(2).

19. MinistryofHealth,WelfareandSport:Careinstitutionsqualityact(inDutch).TheHague;1996.

20. FolmerK,MotE:Diagnosisandtreatmentcombinations inDutchhospitals.TheHague:CPBNetherlandsBureauforEconomicPolicyAnalysis;2003.

21. Custers T, Arah OA, Klazinga NS: Is there a business case for quality in theNetherlands?Acriticalanalysisoftherecentreformsofthehealthcaresystem.HealthPolicy2007,82(2):226‐239.

22. Galbraith JR: Organization design: an information processing view. Interfaces1974,4(3):28‐36.

23. BohmerRMJ:Medicine's service challenge:blending customand standard care.HealthCareManagementReview2005,30(4):322‐330.

24. Gemmel P, Vandaele D, Tambeur W: Hospital process orientation (HPO): thedevelopmentofameasurementtool.Gent:GhentUniversity,Belgium;2007.

25. Vanhaverbeke W, Torremans H: Organizational structure in process‐basedorganizations.KnowledgeandProcessManagement1999,6(1):41‐52.

26. LovePED,GunasekaranA,LiH:Puttinganengineintore‐engineering:towardaprocess‐orientedorganisation.InternationalJournalofOperations&ProductionManagement1998,18(9/10):937‐949.

27. MajchrzakA,WangQ:Breakingthefunctionalmind‐setinprocessorganizations.HarvardBusinessReview1996,74(5):93‐99.

28. Van Merode GG: Planning and reaction in care logistics. Oration (in Dutch).UniversityMaastricht;2002.

29. Liker JK:TheToyotaway:14managementprinciples fromtheworld'sgreatestmanufacturer.NewYork[etc.]:McGraw‐Hill;2004.

30. Bevan H, Glenn R, Bate P, Maher L,Wells J: Using a design approach to assistlarge‐scale organizational change: ‘10 high impact changes’ to improve thenational health service in England. The Journal of Applied Behavioral Science2007,43(1):135‐152.

31. Vissers JMH, De Vries G: Working on care processes. Oration (in Dutch).Rotterdam:ErasmusUniversity;2005.

32. Van Merode GG, Verreusel R, Vrieze KOJ, Zeemering S: Constructing newhospitals: the effects of innovative care organization. In Health SciencesSimulationConferenceProceedings:2004;SanDiego;2004:25‐28.

33. DavisK:Payingforcareepisodesandcarecoordination.NewEnglandJournalofMedicine2007,356(11):1166‐1168.

thesis_Vos_17x24_v01.pdf

Chapter1 │29

34. Kaplan RS, Cooper R: Cost and effect: using integrated cost systems to driveprofitabilityandperformance.Boston:HarvardBusinessSchoolPress;1997.

35. DeVriesG, VanTuijlH (Eds.):Health care under pressure: vital tension in themiddleoforganisations,abusinessapproach.Houten:BohnStafleuvanLoghum;2006.

36. VanRaakA,Groothuis S, VanderAaR, LimburgM,Vos L: Shifting stroke carefromthehospitaltothenursinghome:explainingtheoutcomesofaDutchcase.JournalofEvaluationinClinicalPractice(acceptedforpublication).

37. Berg M, Schellekens W, Bergen C: Bridging the quality chasm: integratingprofessional andorganizational approaches toquality. International Journal forQualityinHealthCare2005,17(1):75‐82.

thesis_Vos_17x24_v01.pdf

thesis_Vos_17x24_v01.pdf

│31

Chapter2

Towardsanorganisation­wideprocess­orientedorganisationofcare

aliteraturereview

Arevisedversionofthischapterwassubmittedas:LetiVos,SarahE.Chalmers,MichelL.A.Dückers,PeterP.Groenewegen,CordulaWagner,GodefridusG.vanMerode.Towardsanorganisation‐wideprocess‐orientedorganisationofcare:aliteraturereview.

thesis_Vos_17x24_v01.pdf

thesis_Vos_17x24_v01.pdf

Chapter2 │33

2.1 BackgroundDuringthelastdecade,hospitalshavetriedtomovefromfunctionaltoprocess‐orientedorganisationforms.Inaprocess‐orientedhospital,thefocusisontheprocess of care instead of on functional departments such as radiology andinternalmedicine.The central ideaofprocess‐orientedorganisationdesign isthatorganisingahospitalaroundcareprocessesleadstomorepatientcentredcare, cost reductions and quality improvements1. The breakthrough of theprocess‐orientationconcepttookplaceatthebeginningofthe1990sunderthename ‘business process reengineering’1. Since then, many hospitals haveundertakenactionstomakecaredeliverymoreprocess‐oriented, forexamplebytheimplementationofcareprogrammes,clinicalpathwaysorcarepathwaysforspecificpatientgroups.However,manyhospitalsstrugglewiththequestionof how to deal with process‐orientation at hospital level. The realisation ofprocess‐orientationwithintheentirehospitalorganisationdemandsmoreofanorganisation than performing single projects. Hospitals need to balance theoptimisation of care processes with efficiency in use of resources in thefunctional departments, for example the use of scarce resources by severalpatientgroups2.2.1.1 TheoryProcess­orientedorganisationdesignTraditionally, hospitals have a functional organisation structure. Within thisorganisationaldesign individualswithasimilarareaofexpertisearegroupedinto departments1, 3‐6. In this functional organisation design all departmentsstrivetooptimisetheirownfunctioning,butareoftenunabletointegratetheirservicestomeettheneedsofpatients7.Acomplexsetofpatientflowsemergeswhere the care of the patient, their records, and the resources necessary forcare have to be transferred across department boundaries in ways that arehardtopredict7.Bottlenecksoccurwhereonedepartmentpushespatientsintoanotherdepartmentthat isnotreadytotakecareofthem.Due tothis lackofcoordinationbetweendepartments,afunctionalorganisationusuallystruggleswithadaptationandefficiencyproblemsincareprocesses7.According to Vera et al. (2007)1, it is necessary to overcome the functionaldivision of labour when an organisation wants to become process‐oriented.Withinaprocess‐orientedorganisationaldesignthefocusisontheprocessofcare insteadof organisationaldepartments.Theorganisational design is thendominated by cross‐functional patient flow. Further, the optimisation of the

thesis_Vos_17x24_v01.pdf

34│ Towardsprocess­orientedcaredeliveryinhospitals

performance of eachdepartment becomes subordinate to the optimisation ofthe patient flow across the hospital2, 8. Additionally, it is important that allmembersof thedifferentdisciplines involved in thecaredeliveryofapatientwork together as a groupand strive to achieve commongoals4, 9. Ideally, thephysicallayoutisalsoadaptedtothecareprocesses7,10.However,expertshavenotedthatthis isextremelyhardtoachieve inapre‐existing building11.Table2.1outlinesthedistinctionsbetweenfunctionalandprocess‐orientedorganisa‐tionaldesign.Table2.1 Characteristicsoffunctionalorganisationandprocess‐orientedorganisations

Functionalorganisation Process­orientedorganisation

Organisationdesign

Similarcapacitiesaregroupedinadepartment(accordingtotheirspecialisation,educationandtraining)1,3,productlayout39

Similarcapacitiesaregroupedinadepartment(accordingtotheirspeciali‐sation,educationandtraining)1,3,productlayout39withadditionalcoordinatingstructures(e.g.careprogrammes)4‐or‐Multidisciplinaryorganisationaldepartmentswhichareorganisedaroundandbasedoncareprocesses1,13,processlayout7,10,39,layoutfollowsprocess13

OrganisationalOrientation

Verticalorientation40,objectivesforanorganisationaldepartmentcanonlybelinkedindirectlytovalueforthepatient13

Patient‐oriented13;horizontalorientationthatcutsacrosstheorganisationaldepartments4,13,activitiescandirectlybelinkedtovalueforthepatients10,40

Managementfocus

Managingdepartments(piecesoftheprocess)40,optimisingdepartmentperformance(capacityuse)7

Managingprocesses(holisticview)14,40,optimisingpatientflow

Decisionmaking

Centralised14 Devolvedtomultidisciplinaryteams13

Responsibilityforcareprocesses

Nooneisinchargeoftheprocesses,becauseworkisorganisedaroundtasks13

Processownershavethefullresponsibilityfortheeffectiveandefficientrunningofacareprocess13

­table2.1continues­

thesis_Vos_17x24_v01.pdf

Chapter2 │35

­table2.1continued­

Functionalorganisation Process­orientedorganisation

Coordinationbetweendepartments

Adhoc,frequenthandoversofpatientsbetweendepartmentswhichremainlargelyuncoordinated40,41

Systematiccoordinationofhandoversandcoworkingasrule41throughadditionalstructuralcoordinationdimensionsatthetopofthefunctionalstructure13‐or‐Departmentshaverelativelyfewinterdependenciesbecauseeveryonerelevanttotheprocessbelongstothesamedepartment,coordinationacrossdepartmentsiskeptataminimum1,13

Patientflow Unstructured,unforeseeableandill‐defined7,40,andthereforealotofvariationincareactivitiesforthesamepatientgroups

Defined40andthereforepredictable7,exceptforclinicalexceptionstostandardisedcareprocesses

Inefficiencycostsincareprocesses

Lotsofwasteandtransferpointsresultingininefficiencycostsinthecareprocesses7

Lowerinefficiencycostsincareprocessestheninfunctionalorganisation,becausewasteandtransferpointsarereduced7

Implementationofhospitalwideprocess­orientationVeraetal.(2007)1andGemmeletal.(2008)4describedtwomainapproachestoredesignfunctionallyorientedcaredeliverytoamoreprocess‐orientedcaredelivery:A. Byimplementingcoordinationmechanisms;B. Byorganisationalrestructuring.Ad.A.ByimplementingcoordinationmechanismsIn the coordination mechanism approach the functional organisation is notchanged, but coordinating structures, like care programmes or clinicalpathways, are put on top of the existing organisation structure for therealisation of a smooth patient flow across boundaries of hospitaldepartments4.Thesecoordinatingstructures,intheformoflateralconnections,are used to bridge barriers erected by an organisation’s structure. Theyestablishthesequenceofcareactivities(diagnostics,consultations,treatment)and the responsibilities of professionals involved in the diagnosis andtreatment of hom*ogeneous patient groups. As a consequence, everybody

thesis_Vos_17x24_v01.pdf

36│ Towardsprocess­orientedcaredeliveryinhospitals

involved in the care process should know what to expect in the next, andprevious,steps.Ad.B.ByorganisationalrestructuringIn the organisational restructuring approach, an organisational restructuringhas to take place in which the needs of a patient form the basis of the neworganisationstructure4.Thismeansthatthehealthcarechain,whichcontainsmultipleservicesanddisciplines,hastobeoptimallyorganisedandintegratedwith reference to the actual needs of the patient. In the extreme case, everypatient can be regarded as a ‘project’ for which specific resources aretemporarily united4. An aggregation of similar projects has been called a’product line’12. These product lines are then organised in separatemultidisciplinary departments, which bring together clinicians, nurses, alliedhealth professionals, management staff, and clerical staff. Thesemultidisciplinarydepartmentscanhandleacareprocessascomprehensivelyaspossible and have relatively few interdependencies with other multi‐disciplinarydepartments1,13,14.The adoption of either of these approaches does not automatically imply anincreaseinprocess‐orientation4.Toeffectuateprocess‐orientation,achangeofwork processes is needed as well. Clinicians, for example, have grownaccustomed to working according to particular procedures during years oftrainingandeducation15.These routinesare repetitive, recognisable,patternsof actions.Routines are confirmedandboundby formal, informal,writtenorunwrittenrules16,17likeorganisationalprocedures,protocolsandguidelinesforcare delivery, contracts, agreements and job descriptions15. Adoption of anapproachtomovetowardsaprocess‐orientedorganisationisanadoptionofacollection of rules as well, which, like other rules, are intended to structure,guide, constitute, allow,obligeorprohibit particular actionsand interactions.However, thesenewrulesarenotalways followed17and it isunknownwhich(combinationof)rulesareeffective.2.1.2 StudyaimIn an effort to extend the knowledge about transitions towards process‐orientationathospital level,weperformeda literaturereview.Theaimoftheliterature review is to reportanddiscussapproaches thathospitalsadopt forthe development to process‐oriented organisations and the accompanyingfactorsforsuccessandfailureinordertoderivelessonsforfuturetransitions.

thesis_Vos_17x24_v01.pdf

Chapter2 │37

2.2 Methods2.2.1 SearchstrategyWe searched the Pubmed, Embase and Business Source Premier (BSP)databasesforrelevantEnglishlanguagearticleswithanabstractfromJanuary1998 through May 2009. First, we searched the Medical Subject Headingsdatabasetofindusefulkeywords(MeSHheadings)andselectedsixpotentiallyrelevant terms: Efficiency, Organisational; Patient‐Centred Care; ProcessAssessment (health care); Organisational Innovation; Product LineManagement;HospitalRestructuring.Next,weperformedaMajorTopicsearchin Pubmed using theseMeSH terms in combinationwith theMeSH headingsHospitals and Hospital Administration. These two terms were added to thesearch command because every study had to involve a hospital redesignregarding the management of the internal organisation of the hospital. InEmbaseweusedtheselectedMeSHsubheadingsaskeywordsinoursearch.ForthesearchinBSP,thelistwithallavailablestandardkeywords(subjects)inthedatabase was scanned to find useful subjects. We selected 15 potentiallyrelevant terms (‘advanced planning & optimisation’, ‘advanced planning &scheduling’, ‘business logistics’, ‘business logistics management’, ‘corporatereorganisations’, ‘health care reform’, ‘organisational change‘, ‘organisationalstructure’, ‘process optimisation’, ‘product lines’, ‘product orientation‘,‘production engineering’, ‘reengineering [Management]’, ‘work design’,‘workflow’). We searched the BSP database with these keywords incombinationwiththeterm‘hospital’.2.2.2 StudyselectionanddataextractionAfter performing our search with the selected MeSH headings, articles werereviewed on the basis of the title and abstract. The studies had to assesshospitalredesignthataimedtoimprovethecontrolofat leasttwointerferingcareprocessesintermsofprocess‐relatedtopics.Thestudiedredesignsshouldnot(mainly)beaimedatchangingthespecificsofclinicalpractice,butshouldconcern improvements in the flow of patients through thehospital. InclusionandexclusioncriteriaaresummarisedinTable2.2.Wedecidednottospecifyinclusion criteria on outcome measures too strictly beforehand. Process‐orientation is a broad concept, covering a variety of structure, process andoutcomeparameters. Furthermore,wedidnot set criteria for theused studydesigns for the evaluation of the redesigns towards process‐orientedorganisations. In order to understand and evaluate this kind of interventions

thesis_Vos_17x24_v01.pdf

38│ Towardsprocess­orientedcaredeliveryinhospitals

research methods need to shed light on the interaction between theinterventionanditscontext18.Therefore,studiesusingobservationalresearchmethodsarealsoincludedinthisstudynexttoquantitativestudies.Table2.2 Inclusionandexclusioncriterialiteraturereview

Inclusioncriteria Exclusioncriteria

Articleshould:‐ Containanabstract;‐ BewritteninEnglish;‐ Focusonhospitalorganisations;‐ Addressarestructureorredesignofpatientflowatorganisationallevel,oratleastfortwointerferingcareprocesses;

‐ Containadescriptionofthetransformationprocess/actualintervention;

‐ Beastudyandnotaneditorial,lettertotheeditor,oropinionpiece;

‐ HavebeenpublishedafterJanuary1st1998andbeforeMay1st2009.

Articlefocuseson:‐ Staffsatisfactionand/orchangeonlyconcernsjobredesignorresponsibilitychanges;

‐ Changingtheorganisationalstructureorredesigningatorganisationallevelwithoutaimingimprovementofpatientflow;

‐ Changingthehealthstructuresatnationallevels;‐ Changinghospitalownershiporaffiliation;‐ Projectswithmainpurposeoffinancialimprovement,exceptwherethisisusedtoformbasisoforganisationalchangeorincentives;

‐ Changingtheorganisationofasinglefunctionalunitorasinglecarepathway;

‐ Changeinsoftwareand/orhardwareandITwithnointendedeffectonpatientsflows;

‐ Descriptionofmethods,modelandtheorieswithoutempiricaldata;

‐ Themanagementofredesignandchangeprojects;‐ Redesignofbuildings.

Tworeviewers(LVandSC)independentlyscannedtitlesandabstractstoselectstudies for consideration. Initial disagreements on study selection wereresolved by reaching consensus. Publications were selected for furtherassessmentofthefulltextifinclusioncriteriaweremetorifitwasimpossibleto determine this based on the abstract. We used a standardised extractionchecklist to obtain data on the main characteristics of the redesigns, studydesign, approachesused, relevant results, and factors for success and failure.Further, we looked in particular whether hospitals undertook specificmeasures to promote the adoption of new rules of the process‐orientedorganisationdesignwithinworkingprocedures.Additionally,weperformedanextra searchon the internetusingGoogle® tofind additional information about the redesigns that were described in theincludedarticlesofoursearch.Forthissearchweusedthenameofthehospitalandthekeywords‘redesign’and‘reengineering’.

thesis_Vos_17x24_v01.pdf

Chapter2 │39

2.3 ResultsFigure2.1showstheflowofpapersthroughthereview.Overall,325abstractsof articles published between January 1998 and May 2009 were identified.During abstract screening, 282 articles were excluded because they did notmeet the inclusion criteria. A total of 43 articles was selected for detailedreview, 33 additional articles were excluded subsequently for not meetinginclusion criteria. Three of the ten remaining articles described differentaspectsoftheredesignofpoliclinicoA.Gemelli(PG)19‐21,andtwootherarticlesdescribed different aspects of the redesign of the Leicester Royal Infirmary(LRI)22, 23. The remaining four articles described redesigns of Denver Health(DH),FlindersMedicalCenter(FMC)andUniversityofWisconsinHospitalsandClinics (UWHC). As a result, a total of five redesigns are described in thisreview.Oursearchonthe internetusingGoogle®providedextra informationabouttheredesignsofDH24,FMC25‐27andLRI28,29.The study designs, approaches used, applied supporting measures for theadoptionoftheapproach,reportedoutcomesandfactorsforsuccessandfacedchallengesofthefiveincludedredesignsaresummarisedinTable2.3basedontheretrievedliterature.

thesis_Vos_17x24_v01.pdf

40│ Towardsprocess­orientedcaredeliveryinhospitals

Figure2.1 Selectionprocessforstudiesincludedinanalysis

329Potentiallyrelevantarticlesidentifiedandscreenedforretrieval

Pubmed(n=200)BusinessSourcepremier(n=113)Embase(n=16)

325Uniquearticlesidentified

4duplicatearticlesexcluded

282articlesexcludedonscreeningtitlesandabstracts

NofocusonhospitalorganisationsNorestructureorredesignatorganisationallevel,oratleastfortwointerferring careprocessesEditorials,letterstotheeditor,commentariesoropinionpiece

43Potentiallyappropriatearticlesidentifiedforfurtherreview

Pubmed(n=37)BusinessSourcePremier(n=6)Embase(n=0)

33articlesexcludedafterfull­textreviewNofocusonhospitalorganisationsNorestructureorredesignatorganisationallevel,oratleastfortwointerfering careprocessesFocusisonstaffsatisfaction/jobredesign,healthstructuresatnationallevel,changeofhospitalownership/affliation,financialimprovement,changeofasingledepartment/carepathway,changeinICT,redesignofsupplysystems,redesignofbuildingsEditorials,letterstotheeditor,commentariesoropinionpiece

10articlesincludedinfinalreview

thesis_Vos_17x24_v01.pdf

Table2.3

Overview

ofincludedredesigns

DenverHealth(DH)

FlindersMedical

Center(FMC)

LeicesterRoyal

Infirm

ary(LRI)

PoliclinicoA.Gem

elli

(PG)

UniversityofW

isconsin

HospitalsandClinics

(UWHC)

Setting

A398‐bedhospitalin

Denver,UnitedStates

A500‐bedteaching

generalhospitalin

Adelaide,Australia

A(>)1000‐beduniversity

hospitalinLeicester,

UnitedKingdom

A1500‐bedteaching

hospitalinRoma,Italy

A489‐bedtertiarycare

centreinMadison,United

States

Aimredesign

Toim

provepatient

safetyandsatisfaction,

efficienciesandcost

reductions,andjob

satisfaction

Toim

provepatientflow

throughtheem

ergency

department(ED),

medicalandsurgical

patients

Toim

provehospital

performanceinallareas

(includinghospitalcosts,

patientprocesstimes,

lengthofin‐hospital

stay)dramatically

Tointroduceanew

patient‐oriented

mentality;toreduce

costs

Toim

proveefficiencyand

patientsatisfaction,and

stabilisinginstitutional

financialhealthwhile

keepingqualityhigh

Studydesign

Uncontrolledbefore‐

afterstudy,includingan

analysisofpositiveand

negativeantecedent

conditions

Uncontrolledbefore‐

afterstudy

Uncontrolledbefore‐

afterstudyandaprocess

evaluation

Uncontrolledbefore‐

afterstudy

Uncontrolledbefore‐after

study

Evaluation

period

2003–2008

2003–2007

1995‐1998

1995–1998

2000‐2004

Redesigned

services

Clinicaland

administrativeprocesses

Clinicalcare(first

emergencycare,then

surgicalcare,m

edical

care)

Allpatientservices

(outpatients’andclinical

care)

Allpatientservices

(outpatients’andclinical

care)

Heartandvascularcare,

oncologyandpaediatric

care

Applied

approach

Coordinationmechanism

approach

Coordinationmechanism

approach

Coordinationmechanism

approach

Coordinationmechanism

approach

Organisational

restructuringapproach

thesis_Vos_17x24_v01.pdf

Chapter2 │ 41

DenverHealth(DH)

FlindersMedical

Center(FMC)

LeicesterRoyal

Infirm

ary(LRI)

PoliclinicoA.Gem

elli

(PG)

UniversityofW

isconsin

HospitalsandClinics

(UWHC)

Measuresto

changeworking

procedures

Notreported

Notreported

Processm

anagem

ent

Notreported

Incentivesforproduct

linesanddepartments

Results

Reductionsinoperating

room

expenses;fewer

droppedpatientcalls;

costsavings

Positiveresultsfor

redesignatthe

emergencydepartment

(lesscongestion;reduced

throughputtime);No

outcom

esreportedfor

theelectivesurgicalcare

processredesign

Theimpactofredesign

onhospitalservices,

costsandorganisation

wasnotasdramaticas

initiallyanticipated

(initialtargetsw

ere

ambitious);Theoverall

efficiencywasnot

transformed(asassessed

throughaquantitative

evaluationofits

performance)

Positiveresultsforthe

introductionoftheDC

andreorganisationof

surgicalwards;Results

ofthemedicalwardsare

positivebuthavetobe

furtherimprovedto

reachgoalsofthe

redesign

Financial:eachserviceline

demonstratedimproved

percentm

argin,im

proved

netrevenues,and

increasesinlocaland

regionalmarketshare;

Operational:operational

efficiency,measuredby

patientvolum

echange,

inpatientlengthofstay

data,improvedfrom

pre

servicelinemetrics;

Patientsatisfaction:

improvedpatients

satisfactionsurveysw

ere

documentedforeach

serviceline

thesis_Vos_17x24_v01.pdf

42│ Towardsprocess­orientedcaredeliveryinhospitals

DenverHealth(DH)

FlindersMedical

Center(FMC)

LeicesterRoyal

Infirm

ary(LRI)

PoliclinicoA.Gem

elli

(PG)

UniversityofW

isconsin

HospitalsandClinics

(UWHC)

Factorsfor

success

Thechangestrategywas

builtonideasthatw

ere

developedandtestedin

precedingprojects;

Leaderoftransformation

wasaclinician,who

drew

onherprofessional

statusandfamiliarity

withclinicalpractice;

Politicalandfinancial

supportofthecity;

Trainingofnurses,

cliniciansandmiddle

managersinLean

improvem

ent

techniques;Previous

(positive)experience

withchange

managem

ent

Leadershipbysenior

executives;Clinical

leadership;Team‐based

problemsolving;Afocus

onpatientjourney;

Accesstodata;

Ambitioustargets;

Externalfacilitatorsto

breakdownthe‘silo’

mentalityandfacilitating

multidisciplinary

team

work;Organisation‐

alreadiness;Selectionof

projects‐startthe

redesignprocesswitha

problemthatobviously

needstobefixed;Strong

performancemanage‐

ment;Aprocessfor

maintainingim

prove‐

ment;Communicating

themethodologyand

resultsinmanydifferent

ways,i.e.Leanthinking

days

Notreported

Notreported

Enthusiasticparticipation

ofcliniciansandtheir

willingnesstochange

practicepatternsto

achievecareefficiencies;

Administrativesupport

whichmadeitpossibleto

reorganiseandrelocate

careunitswithinthe

hospitaltocentraliseareas

ofspecialtycareandto

adoptuniversalnursing

practicesonunitswhere

patientshadsimilar

requirem

ents

thesis_Vos_17x24_v01.pdf

Chapter2 │43

DenverHealth(DH)

FlindersMedical

Center(FMC)

LeicesterRoyal

Infirm

ary(LRI)

PoliclinicoA.Gem

elli

(PG)

UniversityofW

isconsin

HospitalsandClinics

(UWHC)

Challenges

Tomanagesystem

‐wide

changesw

hilehorizontal

communicationacross

occupations,

departmentsandsitesis

impeded;Toprom

otethe

useofindustrial

techniquestoclinicians

whiletheylack

experienceworkingwith

theseproblemsolving

andqualityim

provem

ent

techniques;Tomanage

shortcom

ingsinIT

infrastructurein

providingdataforR

IEs;

Tomobilise(financial)

resourcesneededforthe

redesignwhilethe

hospitalhassafetynet

obligations(cannot

deleteservices)

Tomanagethetension

betweenthebottom

‐up

approachofRedesigning

Careandthemoreusual

‘commandandcontrol’

(top‐dow

n)process

adoptedbyhealthcare

managersw

ho,oncethe

problemisidentified,see

theirroleascomingup

withasolutionthat

front‐linestaffthenhave

toim

plem

ent

Tomobiliseenough

commitm

entto

reengineerwhileclinical

involvem

entin

laboratorieswaslow;To

ignoretheneedfor

tailoringofinterventions

toclinicalsituations;To

managedivergentviews

aboutnatureand

purposeofservices

betweenreengineersand

clinicians;Tomanage

changesthatcrossed

specialtyanddirectorate

boundaries;Tohavethe

rightambition(results

maynotbeatexpenseof

learningorgenerate

cynicism

insteadof

interestandenthusiasm)

Tomanagechangesthat

involvemorehospital

departments.For

exam

ple,insurgical

wards,theactivityasa

wholeisconditionedby

theoperatingroom

s,whileinmedicalwards,

functioningisvery

complexandinteracts

withtheentirehospital

Togetagreementfor

collaborationofstaff

cliniciansandtheir

willingnesstochange

practicepatterns

thesis_Vos_17x24_v01.pdf

44│ Towardsprocess­orientedcaredeliveryinhospitals

Chapter2 │45

2.3.1MaincharacteristicsofredesignsfoundThe articles reported on redesigns in Australia (FMC), Italy (PG), UnitedKingdom(LRI)andUnitedStates(DH,UWHC)19‐23,30‐32.Twooftheseredesignsaimedtoimplementprocess‐orientationforallhospitalservices(PG,LRI)19‐23.The other redesigns were limited to clinical care (DH, FMC)30, 31 and threeservicelines(heartandvascularcare,oncologyandpaediatriccare)(UWHC)32.Allredesignsaimedtoimprovethepatientflowthroughthehospital,andsomehadadditionalgoals:costreductions/efficiencyimprovements19‐23,31,32,patientsafety31,patientsatisfaction31,32andjobsatisfaction31.2.3.2 StudydesignsAllredesignswereevaluatedinuncontrolledbefore‐afterstudydesigns.FromtheassessmentofthePG,DHandFMCredesignspreciseinformationonstudydesign, data gathering strategies and outcome measures were lacking. TheevaluationoftheLRIredesigncontainedanassessmentofchangesinquantityand costs of the health care delivered using routine hospital and healthauthority data sources and specific monitoring data of the redesignprogramme29.Besides,aprocessevaluationthataimedtodescribeantecedents,context, implementationand impactof theLRIredesignandtoderive lessonsregardingmanagementofchange,wasperformed29.Forthisprocessevaluationadditional qualitative data were gathered by documentation research,interviews,andnotesfrominformalconversationsandobservationaldatafrommeetings.The evaluationof theUWHC redesign included service‐linemetricsonfinancialperformance,operationalefficiencyandpatientsatisfaction,usinghospitaldataandpatientsurveys32.2.3.3 Approachesusedtomovetowardsaprocess­orientedorganisationCoordinationmechanismapproachFour of the five redesigns (DH, FMC, LRI and PG) followed the coordinationmechanismapproach for the implementationofprocess‐orientation.Threeofthese redesigns (DH,LRIandPG) identified first commonprocessingsteps inmedical treatment processes of patients, e.g. triage, diagnosis and treatment.They subsequently analysed and optimised these processing steps byimplementingcoordinationmeasures.

thesis_Vos_17x24_v01.pdf

46│ Towardsprocess­orientedcaredeliveryinhospitals

Denver Health’s (DH) selected five overarching processing steps, ‘access’,‘inpatientflow’, ‘outpatientflow’, ‘operatingroomflow’and ‘billing’,astargetsfor the redesign of clinical and administrative processes24, 31. For eachprocessingstep,adetailedmapwascreatedtodiagramitscurrentstate,idealstate, and likely future state. DH then initiated a series of week‐long ‘Rapid‐Improvement Events (RIEs)’ five of which were conducted each month toimprove individual processes within each processing step. In these RIEsprocesses were mapped and unnecessary activities removed. A RIE for theprocessing step ‘access’ was for example to improve the telephone‐callabandonment rate.Next to theoptimisationof commonprocessing steps,DHfocused on development of its infrastructure for information technology andworkforce (identifying the ‘right people’ through personnel selectiontechniques).Leicester Royal Infirmary (LRI) identified four hospital processing steps,‘patient visit’, ‘patient test’, ‘emergency entry’, ‘hospital stay’, and planned toredesign these processing steps within specially created ‘laboratories’22, 23.Originally, they planned to redesign the ‘patient test’ and ‘patient visit’(diagnosticservicesandoutpatientclinics)firstbeforeredesigning‘emergencyaccess’ and ‘patient visit’ (clinical care processes). However, this phasedapproachwasreplacedbyplanstoredesigntheprocessingstepsconcurrentlytoreducechancesofcreatingapartiallyredesignedorganisationandtomanagethe interaction between hospital processes and challenging existingdepartmental and functional boundaries. Nevertheless, reengineering becamemore local thancorporatebecause itwasshapedandmanagedat the levelofgroupings of functional departments. The ‘laboratories’ were dismantled andthe responsibility and accountability for redesign projectswere shifted fromreengineers in laboratories to functional departments to better suit theredesignoftheprocessingstepstolocalinterestsandagendas.ThePoliclinicoA.Gemelli (PG) identified fiveprocessing stepsof themedicaltreatment process of patients as targets for their redesign: ‘emergency care’,‘outpatient care’, ‘diagnostic service and laboratories’, ‘operating rooms’ and‘medical/surgicalcare’19,20.Subsequently,PGidentifiedpatientgroupsthatareprocessedequallywithintheseprocessingsteps,e.g.outpatientsor inpatientsthatarebookinganoutpatient (follow‐up)appointment.Next, theyoptimisedthese processing steps, starting at the pre‐hospitalisation process and thescheduling foroutpatientsappointments.Thepre‐hospitalisationprocesswasfor example optimised by planning all preoperative care activities (routinetests,initialpatientevaluation)on1day.

thesis_Vos_17x24_v01.pdf

Chapter2 │47

Incontrasttothethreeredesignsdescribed,theFlindersMedicalCenter(FMC)didnotfocusitsredesignattheoptimisationof individualprocessingstepsofcare processes (e.g. scheduling outpatients’ appointments), but on theoptimisationof thepatient flowbetweenandwithinprocessing stepsof careprocesses25, 30.TheFMCfirstdividedtheclinicalcareprocessesinemergency,surgical and medical care. Within these three groups, FMC identified highvolume patient flows by searching for patient groups that had a number ofprocessing steps in common (‘patient‐care families’), for example for ‘shortemergencycare’(likelytobedischarged)andfor‘longemergencycare’(likelyto be admitted). Next, they looked at the processing steps of the identifiedpatient‐care families to improve the sequencing of the processes involved byeliminating ‘waste’: steps in a care process that do not add value to a careprocess(e.g.waitingtimes,unnecessarymovementofpersonnelandpatients).Thisinvolvedmappingoutthedailyprocessesforclinicalteams,thenobtainingagreement on new sequences. Once an efficient and effective way ofundertakingaprocesshadbeendevelopedandagreedon,itbecameastandardprocedure.Thishappenedforinstanceforthewaymedicalstafforganisetheirdayacrossthehospital25,30.Whileusingthismethod,theFMCworkedgraduallytowards process‐orientation of their clinical care processes: first theyredesignedallemergencycareprocesses,followedbythesurgicalandmedicalcareprocesses.OrganisationalrestructuringapproachUniversity of Wisconsin Hospitals and Clinics (UWHC) followed theorganisational restructuring approach. The UWHC gradually worked to aproductlinematrixstructure,inwhichdisease‐andpatientbasedprocessesarestreamlinedinfocusedclinicalunits.Aninternalandexternalmarketanalysisled to the selection of the first three clinical areas (heart and vascular care,oncologyandpaediatriccare)forservicelinedevelopment32.Thesethreeareashad the necessary leadership in place, institutional strength, and there wasregionalneedfortheseservices.Theserviceswerecentralisedtogeographicalareasofthehospitaldedicatedtocareandmanagementofthesepatientgroups.Thisincludedrelocationandredesignofhospitalunitsanddiagnosticfacilitiesforheartandvascularpatients,theoncologyservicelineand theconstructionofafreestandingadjacentchildren’shospitaltower32.In2006,theUWHCwasplanningtoexpandfromthreetosixservicelines.Thenewestadditionsweretransplantation,neuroscienceandorthopaedics.

thesis_Vos_17x24_v01.pdf

48│ Towardsprocess­orientedcaredeliveryinhospitals

2.3.4 SupportingmeasurestochangeworkingproceduresIt appeared that two hospitals took supporting measures to promotecompliance to the rules of the process‐oriented organisation design on thework floor.Within the redesignofLRI, hospitalmanagement tried to enforcecompliance by changing authority and power structures. LRI introducedprocess management as an attempt to strengthen managerial accountabilityand responsibility for patient processes at the level of the functionaldepartments;andtoimprovemanagerialcommunicationanddecisionmakingacrossfunctionaldepartments22,23.UWHCdevelopedanincentivisationprocesswhichallowedbothdepartmentsandproduct lines tohave financial rewardsforsuccess,inordertoenforcecompliancetothenewworkingmethodsaswellastosustainthequalityofallservicesthatwerenotyetredesigned32.2.3.5 ReportedoutcomesoftheredesignsThereare largedifferencesbetween the typesofoutcomesdescribed.Of fourredesigns(FMC,PG,LRIandUWHC)datafrombeforeandafterimplementingchangestobecomeprocess‐orientedwerereported20,30,32.Thereportedresultsof the FMC and PG redesigns were limited to a number of positive processrelated outcomes of patient groups or specific departments (e.g. throughputtimes, waiting times, length of in‐hospital stay)20, 30. LRI’s redesign led toimprovements but thesewere not as big as initially anticipated. Besides thisLRI’s systemredesigndidnot lead tomoreoverall efficiency22, 23.LRIdidnotsucceedinsignificantlyreconfiguringpreviouspatternsoforganisation:clinicaldirectoratesandspecialtiessurvivedasorganisationalforms23.TheredesignofUWHC resulted in improved operational efficiency, patient satisfaction andfinancial performance32. Of the remaining redesign, DH, only qualitativedescriptionsof the resultswere reported in the retrieved literature: ‘It led toreductions in operating roomexpenses, fewer droppedpatient calls and costsavings’31.2.3.6 FactorsforsuccessandchallengesfacedIn three redesigns (FMC,DH andUWHC)we found factors for success in theretrieved literature, including: senior management support27; clinicalleadership and involvement27, 31, 32; team‐based problem solving27; adequateICT support27, 31; administrative support32; ambitious targets27; externalfacilitators27;organisationalreadiness27;selectionandexecutionofprojectsin

thesis_Vos_17x24_v01.pdf

Chapter2 │49

order of urgency27; using a change strategy that already proved to besuccessful31; good communication and training in the quality improvementtechniques27,31.In the retrieved literatureaboutall fiveredesignschallenges to the redesignswerereported(seealsoTable2.3).Themainchallengesthatwerereportedbythe hospitals that followed the organisational restructuring approach wererelated to the improvement techniques used within the redesigns, theorganisational structure, and the nature of care delivery. Three of the fourhospitals (FMC, DH and LRI)mentioned that the technique used for processimprovementwassometimeschallenging.Twoofthesehospitalsmadeuseof‘Lean’ as core technique, which originates from industry. The aim of thistechniqueistooptimisecareprocessesorprocessingstepsbytheeliminationof activities that do not add value to the patients, like waiting times ormovementsofstaffandpatients.InDH,theapplicationof‘Lean’wassometimesdifficult because clinicians lack experience with this kind of improvementtechnique31. InFMC,the ‘Lean’ techniqueposedachallengetothemiddle andseniormanagers30. They had to change roles from the traditional, top down,problemsolvingresponsibilitiestowardsamorebottom‐upapproach,inwhichthey first had to understand how thework is done aswell, aswhat the rootcausesofdelaysareandotherimpedimentstoflow,beforetheycouldcomeupwithasolution.InLRI, theredesignwasbasedonBusinessProcessRedesign,whichaimsatradicalimprovements.Consistentwiththelogicthatpeopleneedtothinkbigandradicallytorealisebig improvements,LRIsetambitiousaimsfor its redesign, but these turnedout to be too ambitious,which cameat theexpense of learning and generated cynicism instead of interest andenthusiasm28.Furthermore, two hospitals (LRI and DH) reported that the nature of caredelivery prevented them to fully apply the selected approach to come to aprocess‐orientedorganisation.DH,didnotfeelfree,likemostfirmsinindustry,to delete services and focus on strategically important services31. ThishamperedDH to free financial resourcesneeded for the redesign. In LRI, thenatureofcaredeliveryhampered ‘rollingout’aredesignedprocess inarapidand mechanistic fashion. The need to tailor the redesigned processes todifferentclinicalsituationstooktime.In addition, three hospitals (PG, DH and LRI) reported that the existingdepartmentaland functionalboundarieshamperedthe implementationof theredesign. PG experienced that making changes was much more difficult indepartmentswhich interactwith theentirehospital than indepartments, e.g.

thesis_Vos_17x24_v01.pdf

50│ Towardsprocess­orientedcaredeliveryinhospitals

surgical wards, that are conditioned by one department, like the operatingroom20.DHperceiveda lackofhorizontal communicationacrossoccupations,departments and sites31. LRI experienced that making changes across theinterfaces of existing specialties and clinical directorates was a slow anddifficult process. The introduction of process management to improvemanagerialcommunicationanddecisionmakingacrossspecialtiesandclinicaldirectorates could not significantly change this pattern23. In contrast, thehospitalthatadoptedtheorganisationalrestructuringapproachdidnotreportanyoftheabovementioneddifficulties.Instead,theUWHCreporteddifficultiesincliniciancollaboration32.2.4 DiscussionWorldwide,hospitalorganisationsarechangingtheirfunctionalstructureintostructureswhich focusonpatientcareprocesses. In this literaturereviewweassessed five examples of hospitals that pursued a process‐orientedorganisational form and the accompanying factors affecting their success offailure in the redesign process. The study points out that four out of fivehospitals tried to move to a process‐oriented organisation of care by theimplementation of coordination mechanisms. Only one of them followed theorganisational restructuring approach. From the results of these hospitals itseems that the adoption of either approach can possibly lead to the desiredprocess‐orientation.TheUWHCredesigndemonstratedthattheadoptionoftheorganisational restructuring approach can be relatively successful: patientsatisfaction, financial outcomes and operational outcomes of the redesignedservices were improved. However, the UWHC adopted the organisationalrestructuringapproachforonlythree,andlateronsix,strategicallyimportantpatient groups. This leaves the question whether the organisationalrestructuringapproachwouldalsobesuccessfulforstrategicallylessimportantservicesorfortheorganisationofcaredeliveryforpatientswithneedsnotthatdonotfitwithinexistingproductlines.Veraetal.(2007)1alreadypointedoutthatthiscouldbedifficult,becausepoliticalandethicalobligationsofhospitalsprevented them from withdrawing services to focus only on strategicallyimportantones.Three of the four other hospitals (DH, FMC and PG) demonstrated that thecoordinationmechanismapproachcanleadtopositiveresults,buttheydidnotreportontheresultsveryextensively.FMCandPGonlyreportedsomegeneralresults on process measures, and DH only reported qualitative descriptions.LRI,ontheotherhand,evaluateditsredesignextensively,buttheresultswere

thesis_Vos_17x24_v01.pdf

Chapter2 │51

disappointing: financial outcomes and practice patterns showed noimprovement.Patientsatisfactionwasnotmeasured.Fromthereportedfactorsfor failure it appeared that the adoption of the coordination mechanismapproachwasconstrained,particularlybythefunctionalorganisationaldesignofhospitals.Improvementofcontrolatprocesslevelrequiresthatdepartmentssubordinate theperformanceof theirdepartment (utilisationof resources) totheservicelevelprovidedtopatients(e.g.shortaccesstimes,waitingtimesfordiagnosticexaminations,and throughput times)33.AnattemptofLRI tobreakthe previous pattern of the functional organisation by the implementation ofnon committal process management, did not work23, 28. This underlines theimportance of measures that enforce compliance. Vera et al. (2007)1recommend,forexample,theestablishmentofincentivesthatarebasedontheperformanceofcareprocesses.Further, itseemsthatwithinthisapproachaninitialfocusonlogisticallyhom*ogeneouspatientgroups,assistedbybottom‐upknowledgeofhealthcareprofessionals,couldhelptoovercome thefunctionaldivisionof labour.From thedescriptionof the redesigns it appeared that theredesignsthathadan initial focusonoverarchingprocessingstepsofmedicaltreatment of patients (DH, LRI and PG), in the end mainly implementedimprovements indepartmentalprocesses insteadof improvements thatmadepatient flow more smoothly. For example, PG optimised the scheduling ofoutpatients,DHoptimised theoutpatient flowandLRI thepatientvisit at theoutpatientclinic.Optimisationofthese‘isolated’processingstepsdoesnotleadto more collaboration between departments and more process‐orientationwithin the whole care trajectory of patients. Besides, the optimisation oflinkages between the processing steps will in this way escape from theattentionofthehospital.However, itappearedfromtheFMCredesignthataninitialfocusonlogisticallyhom*ogenouspatientgroupsencouragedhealthcareprofessionals to work together as a group to optimise linkages betweenprocessingsteps,andtodeleteallstepsinacareprocessthatdidnotaddvalue.Nexttothesespecificpointsofinterestforthedifferentapproachestobecomeprocess‐oriented, we could derive some more general lessons for futureredesignsfromtheresultsoftheliteraturereview.First,tailoringisneeded.LRItriedtoroll‐outgeneralredesignsofprocessingstepstoeveryclinicalsituation,butthisappearedtobeimpossibleduetothemultitudeofdifferentclinicalanddisease patterns. Second, clinical engagement, and additional support for theuseofqualityimprovementtechniques,iscrucialtothesuccessoftheredesign.TheevaluationsoftheDH,FMC,LRIandtheUWHCredesignspointedoutthatchanges toclinicalservicescannotsucceedwithout the inputofclinicians.Ontheotherhand,itappearedintworedesigns(FMCandDH)thatclinicianslackexperience in applying improvement techniques. Besides, professionals

thesis_Vos_17x24_v01.pdf

52│ Towardsprocess­orientedcaredeliveryinhospitals

workinginhospitalsalreadyfacehugedemandsontheirtime,and,justifiably,maynotalwaysbewilling toprioritise timeconsumingservice redesignoverspendingtimewithpatients.Therefore,itisveryimportanttoinvolvecliniciansinredesigningservicesandtoalsoofferthemadministrativesupport.Unfortunately,wearenotable to judgewhichof the twoapproachesdeliversthebestresultsinwhichcirc*mstances.Forsuchanassessmentmorestudiesareneeded.Suchstudieshavetoincludeevidenceonstudydesign,objectives,approach, patient population and results. Limitations of the review methodfollowedshouldbeconsideredininterpretingtheresults.Asinanyreviewwemayhavemissedrelevantstudies.Webelievethatgiventheworldwideamountof activity of hospitals to becomeprocess‐oriented, a very limited number ofstudieshasbeenpublishedaddressingapproachestomovetowardsaprocess‐oriented organisation design. This is probably due to the nature of thephenomenon studied. Like other types of planned change or innovation,(successful)organisationwideredesignmovessequentiallyfromawarenessofgaps, to identification of solutions, implementation of selected solutions, andinstitutionalisation of solutions12. This hampers evaluation andpublicationofthese kinds of interventions. Another explanation of the limited number ofstudieswe found could be the fact that process‐orientation in hospitals doesnot succeed and that studies about failures are not published. The cause offailure could be the strong institutionalised functional division of tasks inhealth care systems, which is also apparent within the education of medicalprofessionals.To extend theories and knowledge about the best approaches to becomeprocess‐oriented,andhowtoovercomebarriers tosuccess, it is important toassess each effort to implement a redesign and to assess it properly. Sinceredesigns are complex interventions that are introduced into complex anddiverse ‘socialworlds’34, 35, it is important that the typesof researchmethodsusedtounderstandandevaluatetheseredesignsshedlightontheinteractionbetween the characteristics of the redesign and its context. Quantitativeresearchmethods tend to focus on general trends and are unlikely to clarifyinteractionprocessesbetween the redesignand its context35‐38.Thereforewerecommendtheuseofdiversequalitativeresearchmethods(e.g.observation,semi‐structured interviews) as well as quantitative methods. Qualitativeresearchmethodsprovidean informativeanalysis ‘discerningwhatworks forwhom, in what circ*mstances, in what respects and how’18. This analysis isneededtoprovidea‘thickdescription’oftheredesignprocess,i.e.ameaningfuldescription of the redesign process in its context. The quantitativemeasurements have to include a wide range of measurements. Besides

thesis_Vos_17x24_v01.pdf

Chapter2 │53

logisticaloutcomes,itisalsoimportanttomeasurefinancialoutcomes,patientsatisfaction and process‐orientation of health care professionals. These lastmeasurements are very important: after all, top management may changestructures, including reporting responsibilities of middle and lowermanagement,butthisdoesnotautomaticallyleadtomoreprocess‐orientationin work processes, which is needed to effectuate the process‐orientation athospital level. Frequently, additional measures are needed to change oldworkinghabitsandroutines.2.6 ConclusionDue to the limitations of the evidence, it is not known which approach,implementation of coordination measures or organisational restructuring,produces the best results in which situation. Therefore, more research isneeded.Withinthisresearch,theuseofdiversequalitativemethodsinadditionto quantitative measures is recommended to be able to understand theinteractionbetweenthecharacteristicsoftheredesignsandtheircontext.Hospitalsareadvisedtotakethefactorsforfailuredescribedintoaccountandto takesuitableactions tocounteract theseobstacleson theirway tobecomeprocess‐orientedorganisations.

thesis_Vos_17x24_v01.pdf

54│ Towardsprocess­orientedcaredeliveryinhospitals

References1. Vera A, Kuntz L: Process‐based organization design and hospital efficiency.

HealthCareManagementReview2007,32(1):55‐65.2. Vissers JMH, Beech R: Health operationsmanagement: patient flow logistics in

healthcare.London:Routledge;2005.3. Braithwaite J: Strategic management and organisational structure:

transformationalprocessesatworkinhospitals.AustralianHealthReview1993,16:383‐404.

4. Gemmel P, Vandeale D, Tambeur W: Hospital process orientation (HPO): thedevelopment of a measurement tool. Total Quality Management & BusinessExcellence2008,19(12):1207‐1217.

5. Lega F, DePietro C: Converging patterns in hospital organization: beyond theprofessionalbureaucracy.HealthPolicy2005,74(3):261‐281.

6. Van Merode GG: A prelude of 2004 Antwerp quality conference: targets andtarget values ‐ integrating quality management and costing. Accreditation andQualityAssurance2004,9:168‐171.

7. Van Merode F, Molema H, Goldschmidt H: GUM and six sigma approachespositionedasdeterministictoolsinqualitytargetengineering.AccreditationandQualityAssurance2004,10:32‐36.

8. Berg M, Schellekens W, Bergen C: Bridging the quality chasm: integratingprofessional andorganizational approaches toquality. International Journal forQualityinHealthCare2005,17(1):75‐82.

9. Ludwig M, Van Merode F, Groot W: Principal agent relationships and theefficiencyofhospitals.TheEuropeanJournalofHealthEconomics2010,11:291‐304.

10. Liker JK:TheToyotaway:14managementprinciples fromtheworld'sgreatestmanufacturer.NewYork[etc.]:McGraw‐Hill;2004.

11. Bevan H, Glenn R, Bate P, Maher L,Wells J: Using a design approach to assistlarge‐scale organizational change: '10 high impact changes' to improve thenational health service in England. The Journal of Applied Behavioral Science2007,43(1):135‐152.

12. Shortell SM, Kaluzny AD (Eds.): Health care management: organization designandbehaviour.4thedition.CliftonPark:ThomsonDelmar;2000.

13. Vanhaverbeke W, Torremans H: Organizational structure in process‐basedorganizations.KnowledgeandProcessManagement1999,6(1):41‐52.

14. LovePED,GunasekaranA,LiH:Puttinganengineintore‐engineering:towardaprocess‐orientedorganisation.InternationalJournalofOperations&ProductionManagement1998,18(9/10):937‐949.

thesis_Vos_17x24_v01.pdf

Chapter2 │55

15. VanRaakA,Groothuis S, VanderAaR, LimburgM,Vos L: Shifting stroke carefromthehospitaltothenursinghome:explainingtheoutcomesofaDutchcase.JournalofEvaluationinClinicalPractice(acceptedforpublication).

16. Scott RW: Institutions and organizations (2sd edition). Thousand Oaks: Sage;2001.

17. VanRaakA,PaulusA,CuijpersR,VeldeC:Problemsofintegratedpalliativecare:aDutchcasestudyandcooperationintheregionofArnhem.Health&Place2008,14(4):768‐778.

18. PawsonR,GreenhalghT,HarveyG,WalsheK:RealistReview‐anewmethodofsystematic reviewdesigned for complex policy interventions. Journal ofHealthServicesResearchandPolicy2005,10(S1):21‐34.

19. Catananti C, Celani F, Cambieri A, De Angelis C: Phase 1: identifying criticalsuccessfactors.Rays1998,23(2):270‐287.

20. Catananti C, Celani F, Marchetti M: Management process reengineering of the‘PoliclinicoA.Gemelli’:outcomesoftesting.Rays1998,23(2):295‐309.

21. Marano P, Fioretti P, Bellomo S, Ceruti R, Masera P, Zaccheo L: Hospitalreengineeringprojectatthe‘PoliclinicoA.Gemelli’.Rays1998,23(2):263‐269.

22. McNulty T: Reengineering as knowledge management: a case of change in UKhealthcare.ManagementLearning2002,33(4):439.

23. McNultyT,FerlieE:Processtransformation:limitationstoradicalorganizationalchange within public service organizations. Organization Studies (01708406)2004,25(8):1389‐1412.

24. NuzumR,McCarthy D, Gauthier A, Beck C: Denver health: a high‐performancepublichealthsystem.NewYork:TheCommonwealthFund;2007.

25. Ben Tovim DI, Bassham JE, Bennett DM, Dougherty ML, Martin MA, O'Neill SJ,Sinco*ck JL, Szwarcbord MG: Redesigning care at the Flinders medical centre:clinicalprocessredesignusing‘leanthinking’.MedicalJournalofAustralia2008,188(6Suppl):S27‐31.

26. Ben TovimDI, DoughertyML, O'Connell TJ,McGrath KM: Patient journeys: theprocessofclinicalredesign.MedicalJournalofAustralia2008,188(6Suppl):S14‐17.

27. McGrath KM, Bennett DM, Ben Tovim DI, Boyages SC, Lyons NJ, O'Connell TJ:Implementing and sustaining transformational change in health care: lessonslearntabout clinicalprocess redesign.Medical JournalofAustralia2008,188(6Suppl):S32‐35.

28. McNulty T, Ferlie E: Reengineering health care. The complexities oforganizationaltransformation.Oxford:OxfordUniversityPress;2004.

29. BrownsIR,McNultyT:Re‐engineeringLeicesterRoyalInfirmary:anindependentevaluationofimplementationandimpact.Sheffield:SchoolofHealthandRelatedResearch;1999.

thesis_Vos_17x24_v01.pdf

56│ Towardsprocess­orientedcaredeliveryinhospitals

30. Ben Tovim DI, Bassham JE, Bolch D, Martin MA, Dougherty M, SzwarcbordM:Leanthinkingacrossahospital:redesigningcareattheFlindersmedicalcentre.AustralianHealthReview2007,31(1):10‐15.

31. HarrisonMI,Kimani J: Building capacity for a transformation initiative: systemredesignatDenverHealth.HealthCareManagementReview2009,34(1):42‐53.

32. TurnipseedWD,LundDP,SollenbergerD:Product linedevelopment:astrategyforclinicalsuccessinacademiccenters.AnnalsofSurgery2007,246(4):585‐590.

33. Vissers JMH: The development of operationsmanagement in hospitals and theroleofmiddlemanagers(inDutch).InHealthcareunderpressure:vitaltensionin themiddle of organisations, a business approach. Edited by DeVriesG, VanTuijlH.Houten:BohnStafleuvanLoghum;2006.

34. Bate P,Medel P, Robert G: Organizing for quality: the improvement of leadinghospitalsinEuropeandtheUnitedStates.Oxford:Radcliffe;2008.

35. Powell A, Rushmer R, Davies H: A systematic narrative review of qualityimprovement models in health care. Edinburgh: NHS Quality ImprovementScotland;2009.

36. BerwickDM:Thescienceofimprovement.JAMA2008,299(10):1182‐1184.37. Øvretveit J, Staines A: Sustained improvement? Findings from an independent

casestudyoftheJonkopingqualityprogram.QualityManagementinHealthCare2007,16(1):68‐83.

38. ØvretveitJ:Aframeworkforqualityimprovementtranslation:understandingtheconditionality of interventions. Joint Commission Journal onQuality and Safety2004(GlobalSupplement):15‐24.

39. Ozcan YA: Quantitative methods in health care management: techniques andapplications.SanFrancisco:Jossey‐Bass;2005.

40. McCormack KP, Johnson WC: Business process orientation: gaining the e‐businesscompetitiveadvantage.BocaRaton:CRCPressLLC;2001.

41. FeachemRG,SekhriNK,WhiteKL:Gettingmorefortheirdollar:acomparisonoftheNHSwithCalifornia'sKaiserPermanente.BMJ2002,324(7330):135‐141.

thesis_Vos_17x24_v01.pdf

│57

Chapter3

Evaluatinghospitaldesignfromanoperationsmanagementperspective

Thischapterwaspublishedas:Leti Vos, Siebren Groothuis, Godefridus G. van Merode. Evaluating hospital redesignfrom an operationsmanagement perspective. Health CareManagement Science 2007:10:357‐364.

thesis_Vos_17x24_v01.pdf

thesis_Vos_17x24_v01.pdf

Chapter3 │59

3.1 IntroductionInmanywesterncountriesweobservetendenciestowardsprocessdrivencare.Ideally,thelayoutofthebuildingshouldbeadjustedtothislogisticalconceptinordertogroupthefacilitiesalongthelineofthephasesofthecareprocess.Thisfacilitatestheintegrationofvariousoperationsinoneflowataratedeterminedbytheneedsofapatientandwithleastamountofdelayandwaiting1.Buildinganewhospitalgivesopportunitiestomatchthelayoutofthehospitalbuildingwiththedesired logisticalconcept. Inthatcasethe layoutmustbesuitabletodealwithall the flowsofpatientsandgoodsnowandinmanyyearstocome.After all, the layout cannot be adapted very easily and only at high cost.Designingalayoutisachallengingjobbecauseoftheuncertaintiesconcerningfuturenumbersofpatients,patientmixandnewmedicaltechnologies.Howcanwe accomplish that the layout of a hospital,which creates conditions for thedynamicoperationalcontrol,staysappropriateforcaredeliveryinthefuture?Itisimportanttodesignahospitallayoutwhichsupportshospitalstrategy,butalsotakesthefuturecirc*mstancesandtheuncertaintiesintoaccount.Inorderto assure that a design supports the efficient and effective operating of careprocesses now and in the future, it would be useful to have an evaluationmethodfortheassessmentoftheflexibilityandfitofthebuildingdesign.Thepurposeofthisresearchistotestsuchamethodandillustrateitbyacase.Theremainderofthispaperisdividedintofourmajorsections.Firstweexplaintherelationship between the operations management and hospital floor plandesign.Inthefollowingsectionamethodforevaluationoftheflexibilityandfitofthebuildingdesignwillbepresented.Nextthismethodwillbeappliedinacase study. In the last part we discuss the case and the contribution of theevaluationmethodforthefieldofoperationsmanagement.3.2 Relationship between operationsmanagement and hospital

designFrom an operations management perspective hospitals need to maximiseoutput/ throughput with available resources, taking into account differentrequirements for delivery flexibility (elective/appointment, semi‐urgent,urgent), acceptable standards for delivery reliability (waiting list, waiting‐times) and acceptable medical outcomes by the design, planning,implementationandcontrolofcoordinationmechanismsbetweenpatientflowsanddiagnosticandtherapeuticactivities2.Inordertorealiseshortthroughput

thesis_Vos_17x24_v01.pdf

60│ Towardsprocess­orientedcaredeliveryinhospitals

times and short displacement distances in the building the different flowsbetweenthevariouslocationsmustbetakenintoaccount.Fromanarchitecturalviewhospitalsconsistofstaticfloorplans,withelementslike corridors, consultation and examination rooms, reception desks, waitingfacilities etcetera. From the view of operations management hospitals aresystemsinwhichallbuildingelementsarerelatedtoeachother.Persons,goodsand informationmove from one building element to another. This results inflowsofpeople,goodsandinformation.A logisticsystemhas tocontrol theseflowsandsupportthehospital’soperations.There is a close link between hospital strategy, the choice of the system foroperationalplanningandcontrolandthehospital’s layout.Ahospitalstrategyisanactionplanforfuturedevelopmentandincludeschoiceofpatientmixintermsofdiseaseandtreatmentandmarketsegments,medicaltechnologyandserviceconcept.Arealistichorizonforthehospitalstrategy isalways limited:between5 and10years.However the life cycleof abuilding ismuch longer.Duringthelifeofthebuildingenvironmentalchangesmayoccur,likeageingofpopulation,moredemandingpatients,andnewinsightsabouttreatments.Thismayleadtochangesinthehospital’sstrategyandthusthepatientmix,medicaltechnology and the operations management system. This means that the fitbetweenhospital’slayout,patientmixandtheoperationsmanagementsystemisadifficultissuebecausethetimehorizonsdonotmatch.Thecostofchangingthe layoutof an existinghospital building canbe enormous. Flexibility of thebuilding to adapt to changes in patientmix, operationsmanagement conceptandtechnologyisthereforeessential.In this paper we assume that the floor plan design is ‘flexible and fit’ if itsupportsandfacilitatestheoperationalcontrolnowandisappropriatetodealwithfuturedevelopments.Thedegreeofcirculationoftheflowcanbeusedtomeasure this flexibility and fit.After all it is congestionof flow that indicatesthat there is a local capacity problem and thus a non optimal fit betweenbuildingdesignandoperational control. In thatcase thedesign isnot flexibleenoughtodealwith(variationsin)theflow.Duetothisdisturbanceinflowinone place, capacity in other places cannot be used in an efficient way3.Blockagesintheflowcanincreasewaitingandthroughputtimeswhichhaveanegativeeffectonthequalityofservicedeliveryandcapacityuse.Thenumberofpeoplethatiswalkingaroundandhowmanygoodsaretransportedthroughthehallwaysatthesametimeatthesameplaceinthebuildingisameasureofthedegreeofcirculationorcongestionoftheflow.

thesis_Vos_17x24_v01.pdf

Chapter3 │61

Figure 3.1 represents the relationship between floor plans and operationalplanningandcontrolsystemasdescribedabove.Onthebasisofthismodelwedesignedanevaluationmethodfortheflexibilityandfitofthebuildingdesignforoperationsmanagement,whichwillbedescribedinthefollowingsection.Figure3.1 Relationshipbetweenoperationsmanagementandhospitaldesign/floorplan

3.3 EvaluationmethodInthissectionamethodtoevaluatetheflowandthustheflexibilityandfitofbuildingdesignwillbedescribed(seealsoTable3.1).Oneofthekeyfeaturesofthismethodisdiscreteeventsimulation.Todeterminethedegreeofcirculationoftheflow,thehospitalsystemcanbedefined as a flow system in which different segments of the building aredistinguished (step 1). Each segment represents a part of the corridor of thebuilding throughwhichpeopleandgoods flow to theirdestination.Corridorscan be divided in segments on basis of function of segments: part of the

Futuredevelopments

FloorplanOperationalplanningand

control

Strategy

Outcome:Flow

QualityandreliabilityofservicesThroughputtimesWaitingtimes

Efficientcapacityuse

‘Flexibilityandfit’ofbuildingdesign

thesis_Vos_17x24_v01.pdf

62│ Towardsprocess­orientedcaredeliveryinhospitals

corridoralongreceptionsdesks,alongentrancesofclinics,alongwaitingroomsorpartsofthecorridorwhichdonothaveaspecificfunction.Table3.1 Stepsoftheevaluationmethod

Stepnumber

1 Distinguishsegmentsofthefloorplan2 Determineflowsofpersonsandgoods3 Designexperiments4 Implementthemodel5 Runthesimulationmodel,experimentandanalysetheresults

An example is given in Figure 3.2. The different segments in this figure arenumbered.Figure3.2 Exampleofdivisioninsegments

The time spent by people and goods in each segment is determined by(walking) speed and the dimensions of each segment.Walking speed can beestimatedon85meter/min4.Afterdividing the flow system in segments, theflows, based on floor plans and data on (present) numbers of patients

Toilets

thesis_Vos_17x24_v01.pdf

Chapter3 │63

(appointments) and of personnel and the movement of goods, including thelogisticalplanningconceptforcontrollingtheseflows(pushorpull)havetobedefined(step2).Measurementofthedegreeofcirculationbasedontheseflowsgives an indication of the functionality of the system. Next an experimentaldesignisdevelopedindicatingrangesofchangesofflowintensityanddirection(resultofchangeoflogisticalconcept),andonchangesorincreaseofintensityoftheflow(resultofchangedpatientmixandnewtechnologies)(step3).Thenextstepisassessingthenumberofpeopleandgoodswhicharepresentonthesameplaceatthesametimeinthedifferentexperimentalscenarios(step4).To accomplish the above steps a discrete event simulation model has beenbuilt.Discreteeventsimulationisthemodellingofasystemasitevolvesovertime by representation in which state variables changes instantaneously atseparate points in time5. The floor plan of the hospital is represented assegmentsinthesimulationmodel.Furthermorethesimulationmodelcontainstheexpectedamountofflowsofpeopleandgoods.Thesegmentsarelocationswithlimitedcapacity.The number of people in a segment is represented by the performanceindicatorofmaximumvalueof thenumberofpeoplewhoarepresenton thesame time in the same segment. The floor plandesign has to be able to dealwiththismaximumnumberofpeople.Thereforeit is importanttoassessthismaximumvalue.Besidesthemaximumvalueitisevenmoreimportanttoknowhowoftenthismaximumvalueoccurswithinsingletimeslotsofonehour.Step5 exists of running the simulationmodel (to test functionality). In this phaseseveralexperimentsregardingflowintensityanddirectioncanbedone(totestthe ability to adapt to developments). These experiments are determinedaccording to an experimental design6. After simulation results have to beinterpreted. The steps described above are summarised in Table 3.1. Theusefulnessofthismethodisillustratedinacasestudy.3.4 CasestudyInthissectionweapplythismethodfortheevaluationoftheflexibilityandfitofthedesignforoperationsmanagementtoacase.ThecaseselectedconcernsanewDutchhospitalwhichistheresultofthemergeroftwohospitalsandwillbebuiltonanewlocation.Thisnewhospitalwantstointroducea‘21stcenturyairport’operationsmanagementconceptforthedesignofanoutpatientclinic.Thepurposeofthe‘21stcenturyairport’conceptistheefficientuseofspaceofthehospitalbuildingbycentralisingthewaitingareas.Theassessmentanswers

thesis_Vos_17x24_v01.pdf

64│ Towardsprocess­orientedcaredeliveryinhospitals

thequestionwhetherthehospitaldesignwillallowafreeflowofpatientsandgoods.3.4.1 SettingTheoutpatientclinichasonecentralwaitingareaandalimitedwaitingcapacityinthedifferentambulatorydepartments.Eachambulatorydepartmenthasitsownreceptiondesk,whichislocatedalongthemaincorridors.Tworoomsforconsultation and/or examining patients are available for each medicalspecialist. According to the ‘21st century airport’ concept patients and theircompanionswill wait in the central waiting area for a call to leave for theirdestination: the ambulatory department of their specialist. Every ambulatorydepartment is situated at amain corridor. Patients take themain corridor towalk to their destination. The ‘21st century airport’ concept assumes that nopatients and companions have to wait at the ambulatory departments. Thismeans that the consultation or examination room is directly available at thetime the patient and his companion arrive at the ambulatory department. Apatient will only move from one room, e.g. a consultation room to anotherroom,e.g.anexaminationroomatthedepartmentwhentheexaminationroomisavailable.This logisticalconceptcanonlybeappliedtohighlystandardisedpatients groups. The flow of these patients is certain. The successive careactivities(consultations,examinations,diagnostics,andtreatments)arealreadyplannedbefore thepatient arrives at thehospital.Non standardpatientswillwaitin(highlyequipped)roomsbeforetheycangotoanotherroom.Aftertheirvisitsomepatientsmakeanewappointmentatthereceptiondesk.Thenumberof patients that will make an appointment differs per specialism. Alsopersonnelwalksthroughthecorridorstotheirdestinationatthebeginningandendoftheirworkingday.Dataconcerningtheroutingoftheexistingpatientmixandstaffwereprovidedby the hospital. The hospital provided also the data concerning the workinghoursofthestaff.Thesedatawereconsideredasfactsandwerenotsubjectofthisresearch.Dataonthemovementofgoodswerenotavailableandthereforenotincludedinthisstudy.3.4.2MethodThis paragraph describes how the evaluation method is used in the casedescribedinparagraph3.4.1.

thesis_Vos_17x24_v01.pdf

Chapter3 │65

Step1:DistinguishsegmentsofthefloorplanTo measure the degree of circulation the corridors of the outpatient’s clinichavebeendividedinsegmentsonbasisofavailablefloorplans.Eachsegmentrepresents a part of the corridor. The segments can be classified in fivecategories: segments with one reception desk at the side of the corridor,segmentswithtworeceptiondesks,segmentswhicharepartofanintersectionof twocorridors,segmentswhicharepartofamainway toanotherbuilding,remainingsegments (Figure3.2).Thewidthofeachsegmentwas 2.55meter,thelengthsvariedfrom2to5.4meters.Step2:DetermineflowsofpersonsandgoodsFor this case data were available on (present) numbers of patients(appointments) and of personnel and the movement of goods in the twohospitals and were extrapolated by the hospital to the new hospital. Thehospitalassumesthatfiftypercentofthepatientswillbeaccompaniedbyoneperson and the other fifty percent by two persons. On basis of these dataexpectedflowsaredefinedtoassessthefunctionalityofthedesign.Tocontroltheflowsontheoutpatientclinicthehospitalusesapullsysteminlinewiththeproposedlogisticalconcept.Inordertohavetherightpatientontherighttimeon the rightplace, thehospital tries to reduce the variation in arrivals at thedepartmentsbythecreationofacentralwaitingarea.Fromthisareapatientsareaskedataspecificmomenttomovetoaspecificlocationatthedepartment.Thismoment is determined by the availability of theneeded resources (staffand/orequipment).Step3:DesignexperimentsAccordingtothepresentedmodelinFigure3.1uncertaintiesinflowintensitycanbe theresultof threefactors:used logisticalconcept,anddevelopmentofpatientmixandtechnologies.Theabilityofthedesignoftheoutpatients’clinictomeettheseuncertaintiesisassessedinanexperimentaldesign.Thereforewedetermined three parameters: arrival pattern (logistical concept, intensity oftheflow),numberofpatients(patientmixandtechnologies)andthedurationofmakinganappointment(logisticalconcept,flowdirection).Threevaluesareusedforsimulatingdifferentarrivalpatterns(parameter1)attheambulatorydepartments of the outpatients’ clinic in the experimental design: arrivalschedule A (arrivals per hour determined by means of a probabilitydistribution), arrival schedule B (patients arrive every 10 minutes with anuncertaintyof5minutes),arrivalscheduleC(patientsarriveevery10minuteswith an uncertainty of 2 minutes). Parameter 2, number of patients, whichconcerns the flow intensity, has only two values: the present number ofpatientsandasituation inwhichthenumberofpatients increasedwith25%.

thesis_Vos_17x24_v01.pdf

66│ Towardsprocess­orientedcaredeliveryinhospitals

Parameter3, thedurationof theprocessofmakingan appointment, has alsotwovalues:T(1,2,3) (Triangulardistribution [minimum,modus,maximum])andT(1.5,2.5,3.5).Afullexperimentaldesignwasappliedwhichresultedin12experiments (see Table 3.2). This study assessed one fixed logistical conceptandthereforethiswasnotafactorintheexperimentaldesign.Table3.2 Experiments

Arrivalpattern Numberofpatients Durationofmakingan

appointment

Experiment A B C Present

number

+25% T

(1,2,3)

T

(1.5,2.5,3.5)

1. X X X 2. X X X 3. X X X 4. X X X5. X X X6. X X X7. X X X 8. X X X 9. X X X 10. X X X11. X X X12. X X X

Step4:ImplementthemodelIn our case themodel is implemented inMedModel7.MedModel is a discreteeventsimulationenvironmentwithagraphicalinterface.Step5:Runthesimulationmodel,experimentandanalysetheresultsWesimulated100days(100replications).Wealsosimulatedsingle timeslots(seeevaluationmethodsection).3.5 ResultsThe results concern the flow of patients, companions, staff and visitors. Dataweregatheredforeachofthehundredreplications.FunctionalityFortheresultsofthesimulationoftheexpectedflowsseeTable3.3,scheduleA.Theresultsforsevensegmentsofthemaincorridorsoftheoutpatients’clinicarepresented.Thesesevensegmentsrepresentthefivecategoriesasdescribed

thesis_Vos_17x24_v01.pdf

Chapter3 │67

in step 1. These results show that there are more people in corridors withreceptiondesksatthesideofit.Table3.3 Resultsofexperiment1,2and3.Maximumvalueofnumberofpeople inasegment

(100replications)andtheservicetimeatadeskisT(1,2,3)distributedfordifferentschedules

Morestrictlyorganisedarrivalpattern

ScheduleA ScheduleB ScheduleC

Segment(category)*

Dimensions(meters)

Min Max Mdn Min Max Mdn Min Max Mdn

B11(2)2receptiondesks

5.4*2.55 6 14 9 6 10 8 8 12 9

B13(1)1receptiondesk

5.4*2.55 9 18 11 6 11 8 7 12 9

B14(3)Intersection

2*2.55 4 9 6 4 9 6 5 9 6

D11(5)Nospecificfunction

5.4*2.55 7 15 9 7 14 9 7 13 9

D13(5)Nospecificfunction

5.4*2.55 4 10 6 4 9 6 4 10 6

F5(1)Onereceptiondesk

5.4*2.55 9 18 11 8 14 9 9 15 11

F16(4)Corridortopsychiatriccentre

5.4*2.55 8 18 12 8 20 12 6 24 12

*Thesegmentscanbeclassifiedinfivecategories:segmentswithonereceptiondeskatthesideofit (1), segmentswith two reception desks (2), segmentswhich are part of a intersection of twocorridors (3), segmentswhicharepart of amainway to anotherbuilding (4), segmentswithnospecificfunction(remaining)(5).Min=minimumvalueofthemaximumnumberofpatients;Max=maximumvalueofthemaximumnumberofpatients;Mdn=medianvalueofthemaximumnumberofpatients.AbilitytomeetfuturedevelopmentsToassesstheabilityofthesystemtomeetfuturedevelopmentsweperformedthe12experimentsoftheexperimentaldesign.Table3.3shows theresultsofexperiment 1, 2 and 3: the simulation of several arrival patterns in arrivalschedules A, B and C. The values measured per schedule do not differ

thesis_Vos_17x24_v01.pdf

68│ Towardsprocess­orientedcaredeliveryinhospitals

significantly.Aminimumvalueofthemaximumnumberofpatientsrepresentsthelowestmaximumnumberofpatientsthat ispresentatthesametimeinasegment inhundreddays.Amaximumvalueof18means that therewere18peopleatthesametimepresentinthesamesegmentonatleastoneday(ofthehundreddayssimulated).Thesamevaluesweremeasuredwhenwecollecteddataduringspecifictimeslots,e.g.from9till10a.m.Onlyduringlunchbreakofpersonnelthemaximumvaluesdecreased.When the patients’ stay at the reception desk increased with half a minuteT(1.5,2.5,3.5)butthenumberofpeopleremainedthesameinexperiment7,8and9,themaximumvalueswerenotaffected(seeTable3.4).The results of experiment 7, 8 and 9 which assess increase of flow volume,showthatthemaximumvaluesincreasedwith1or2personswhenthenumberofpatientsincreasedwith25%(seeTable3.5).Inexperiment10,11and12boththenumberofpatientsandthedurationoftheprocessareincreased.TheresultsarepresentedinTable3.6.

thesis_Vos_17x24_v01.pdf

Chapter3 │69

Table3.4 Resultsofexperiment4,5and6.Maximumvalueofnumberofpeople inasegment(100 replications) and the service time at a desk is T(1.5, 2.5, 3.5) distributed fordifferentschedules

Morestrictlyorganisedarrivalpattern

ScheduleA ScheduleB ScheduleC

Segment

(category)

Dimensions

(meters)

Min Max Mdn Min Max Mdn Min Max Mdn

B11(2)

2receptiondesks

5.4*2.55 8 17 10 7 11 8 8 13 9

B13(1)

1receptiondesk

5.4*2.55 9 18 12 7 11 8 7 12 9

B14(3)

Intersection2*2.55 5 9 6 4 7 6 4 8 6

D11(5)

Nospecificfunction

5.4*2.55 7 14 9 7 13 9 7 15 10

D13(5)

Nospecificfunction

5.4*2.55 4 8 6 4 9 6 4 10 6

F5(1)

Onereceptiondesk

5.4*2.55 9 20 13 8 16 11 9 17 11

F16(4)

Corridortopsychiatriccentre

5.4*2.55 8 18 12 6 16 12 8 18 12

Min=minimumvalueofthemaximumnumberofpatients;Max=maximumvalueofthemaximumnumberofpatients;Mdn=medianvalueofthemaximumnumberofpatients.

thesis_Vos_17x24_v01.pdf

70│ Towardsprocess­orientedcaredeliveryinhospitals

Table3.5 Resultsofexperiment7,8and9.Maximumvalueofnumberofpeople inasegment(100replications)andtheservicetimeatadeskisT(1,2,3)distributedfordifferentscheduleswithanincreasednumberofpatients

Morestrictlyorganisedarrivalpattern

ScheduleA ScheduleB ScheduleC

Segment

(category)

Dimensions

(meters)

Min Max Mdn Min Max Mdn Min Max Mdn

B11(2)

2receptiondesks

5.4*2.55 8 15 11 7 11 9 8 12 10

B13(1)

1receptiondesk

5.4*2.55 9 20 13 8 12 10 10 16 11

B14(3)

Intersection2*2.55 5 9 6 5 9 6 5 11 6

D11(5)

Nospecificfunction

5.4*2.55 4 8 6 4 12 6 4 14 6

D13(5)

Nospecificfunction

5.4*2.55 9 22 14 9 15 10 10 17 11

F5(1)

Onereceptiondesk

5.4*2.55 9 22 14 9 15 10 10 17 10

F16(4)

Corridortopsychiatriccentre

5.4*2.55 8 18 12 8 22 12 10 20 12

Min=minimumvalueofthemaximumnumberofpatients;Max=maximumvalueofthemaximumnumberofpatients;Mdn=medianvalueofthemaximumnumberofpatients.

thesis_Vos_17x24_v01.pdf

Chapter3 │71

Table3.6 Results of experiment 10, 11 and 12. Maximum value of number of people in asegment(100replications)andtheservicetimeatadeskisT(1.5,2.5,3.5)distributedfordifferentscheduleswithanincreasednumberofpatients

Morestrictlyorganisedarrivalpattern

ScheduleA ScheduleB ScheduleC

Segment

(category)

Dimensions

(meters)

Min Max Mdn Min Max Mdn Min Max Mdn

B11(2)

2receptiondesks

5.4*2.55 9 16 11 7 10 9 8 13 10

B13(1)

1receptiondesk

5.4*2.55 11 23 14 9 16 10 9 16 11

B14(3)

Intersection2*2.55 5 9 6 5 10 6 5 9 6

D11(5)

Nospecificfunction

5.4*2.55 8 14 10 7 13 9 8 16 10

D13(5)

Nospecificfunction

5.4*2.55 4 10 6 4 8 6 4 9 6

F5(1)

Onereceptiondesk

5.4*2.55 12 27 16 9 22 12 7 19 13

F16(4)

Corridortopsychiatriccentre

5.4*2.55 8 20 12 8 18 12 10 18 12

Min=minimumvalueofthemaximumnumberofpatients;Max=maximumvalueofthemaximumnumberofpatients;Mdn=medianvalueofthemaximumnumberofpatients.

thesis_Vos_17x24_v01.pdf

72│ Towardsprocess­orientedcaredeliveryinhospitals

3.6 ImplicationsandconclusionFirst the results of the case will be discussed. Then the contribution of theevaluationmethodfortheassessmentoftheflexibilityandfitofbuildingdesignwillbeoutlined.3.6.1 DiscussionandconclusionofthecasestudyThe evaluation method showed that the floor plan design of the outpatientclinic supports the functionality of the system and facilitates expected flows.Theresultsofarrivalschedule1experiment1showedtheexpectedmaximumnumber of persons in the different locations of a hospital building given theproposed logistical concept. Any higher value for the number of people in acertain segment indicates a higher possibility of congestion of the flow. Thesegmentsalongreceptiondesk(s)seemthelocationswhicharemostsensitivefor congestion. But no real problems did occur during the simulationexperiments.Sothefloorplandesigncanbeconsideredtobefunctionalgiventheexpectedsituation.Theabilitytomeet futuredevelopments isassessed inthe twelve experiments. These experiments showed that there is a smalldifference between the median of the maximum values and the overallmaximumvalueinthethreearrivalschedules(seeTable3.3).Theresultsfromexperiment 4‐6 and 10‐12 indicate that the placement of reception desksdirectlyalongthecorridorsalsoinfluencestheflow.Whenthedurationoftheprocessesatreceptiondesks(whicharedirectlyplacedalongthecorridors)isextended, themaximum values of the number of people in the segments arehigher.Experiment7‐12showus thatwhenthenumberofpatients increasesthelimitationsofandrequirementsonlayoutaremoreevident.Theseresultsshowthe importanceof theevaluationofafloor plandesign. Inthis case the building seems to support the functionality of the system andfacilitates expected flows. However, the case study shows that if the flowschange in intensity, direction or volume themaximumnumber of patients inthe segments increases and thiswill lead to congestion. The design does notseem to be flexible enough to deal with variations in the factors mentionedabove: used logistical concept, patient mix and technologies. The intendedoperations management system assumes that patient routings can bestandardisedandcompletelycontrolledoncestarted.Accordingtothehospitalmanagement variation in the flow will decrease after the implementationperiodof ‘21stcenturyairport’concept.Thehospitalmanagementistakinganenormous risk by constructing a hospital building that allows a limited

thesis_Vos_17x24_v01.pdf

Chapter3 │73

variation in flow. We recommend to increase the flexibility and fit of thebuilding design through the creation of decentralised waiting rooms at theambulatory departments. Decentralised waiting rooms can counteract theeffect of fluctuations in the flow. Besides this,more flexibility can be createdthrough standardisation of rooms, especially consultation rooms. Both theaddition of decentralised waiting rooms and standardisation of consultationrooms will reduce the dependency on the used logistical concept and thusincreasetheflexibilityandimprovethefitofthebuildingdesign.In thiscasestudynowaitingtimesat theambulatorydepartmentwere takenintoaccountbecauseoftheassumptionmadebythehospitalmanagementthatthepatientswillbedirectlyseenbythemedicalspecialistaftertheirarrivalatthe ambulatory department. This assumption is an essential part of theoperations management concept where patients are paged from a centralwaiting area to a specific care unit when the medical specialist becomesavailable(‘pull’logistics).Thepresenceofprivacylinesbeforereceptiondesks,whichhasareducingeffectonavailablecapacityinsegmentscouldalsonotbesimulated.Theseprivacylineswouldalmostcertainlyhaveadisturbingeffecton the flows.Further,except lunchbreak,nomovementsofpersonnelduringofficehoursaresimulatedbecauseofunavailabilityofdata.Thesemovementsare additional to the movements modelled in the simulation model andthereforewillincreasethedeterminedmaximumvalues.3.6.2 DiscussionofthesimulationmodelThesimulationmodelprovidedinformationaboutthenumberofpersonsinacertainsegment. In thesimulationmodelallpersonshaveaconstantwalkingbehaviour.We can imagine that people in particular cases slow downorwillstop to look around which direction they have to go e.g. when they areapproachingacorridororcrossing.Thiskindofbehaviourisnotincludedinthemodel.Alsoeffectsof theuseofwheelchairs,rollators,stretchersandbuggieshavenotbeentaken intoaccount.Thesemovementaidswill takemorespaceandprogressslowerthanasinglepersondoes,whichwilldisturbtheflowwithgreat probability. Enlarging the scope of the simulation model will allow anassessment ingreaterdetail.Toextend thesimulationmodelasks for furtherresearch.

thesis_Vos_17x24_v01.pdf

74│ Towardsprocess­orientedcaredeliveryinhospitals

3.6.3 ContributionoftheevaluationmethodThis research tried to fill the gap between theworld of architecture and theworldofoperationalcontrol.Flexibilityandfitofabuildingdesignimpliesthatthe static floor plan, made by the architect, meets the requirements of thedynamics of operational control. The case study shows that discrete eventsimulationisausefultechniquefortheevaluationoftheflexibilityandfitofabuildingforoperationsmanagement.Bothfunctionalityandtheabilitytomeetfuturedevelopmentsof thedesign canbe assessed.When the capacitymeetsthe requirement to deal with the flows in combination with the chosenlogisticalconcept,thebuildingcanbeconsideredtobefunctional.Thiscanbeassessed through a simulation study of the expected situation. The ability tomeetfuturedevelopmentscanbedeterminedthroughsimulationofvariationsin future developments such as used logistical concept, patient mix andtechnology.Thetechniquesusedinthisstudyarenotonlyusefulforevaluationof the new designs. Simulation can also be used for the evaluation of newlogisticconcepts forcareprocesses inexistingbuildings.Specificexperimentscanbedesignedtoassesstheimplicationsofthelogisticalconcept.Evaluationofnewlogisticconceptsbeforeintroducingthemcanpreventtheoccurrenceofproblemsintheoverallflowsystem.Assessing the flexibility and fit of a building in the design phase is in ouropinionofgreatimportanceforoperationsmanagement.Inthisphaseitisstillpossibletoadjustthelayoutofthebuilding.Weillustratedtheusefulnessoftheevaluation method with a case. In this example the building design showsshortcomings regarding the ability to meet future developments. Because ofthisevaluation, thedesignof theoutpatientcliniccanbeadjusted inorder toreachahigherdegreeofflexibilityandfitforoperationsmanagementandthusahigherdurability.

thesis_Vos_17x24_v01.pdf

Chapter3 │75

References1. Liker JK:TheToyotaway:14managementprinciples fromtheworld'sgreatest

manufacturer.NewYork[etc.]:McGraw‐Hill;2004.2. De Vries G, Bertrand JWM, Vissers JMH: Design requirements for health care

productioncontrolsystems.ProductionPlanning&Control1999,10(6):559‐569.3. Van Merode GG, Groothuis S: Hospitals as complexes of queuing systems. In

HealthSciencesSimulation2003:January19‐23,20032003;OrlandoFL.Editedby Anderson JG, Katzper M. Society for Modeling and Simulation International(SCS);2003:39‐44.

4. AbernethyB,HanrahanSJ,KippersV,MackinnonLT,PandyMG:Thebiophysicalfoundations of human movement. 2nd edition. Champaign, IL etc.: HumanKinetics;2005.

5. Law AM, Kelton WD: Simulation modelling and analysis. Second edition. NewYork,USA:McGraw‐Hill;1991.

6. Banks J: Handbook of simulation : principles, methodology, advances,applications,andpractice.NewYork,NYetc.:Wiley;1998.

7. MedModelhomepage[www.promodel.com/products/medmodel].

thesis_Vos_17x24_v01.pdf

thesis_Vos_17x24_v01.pdf

│77

Chapter4

Doescase­mixbasedreimbursem*nt

stimulatethedevelopmentofprocess­orientedcaredelivery

Thischapterwasacceptedforpublicationas:LetiVos,MichelL.A.Dückers,CordulaWagner,GodefridusG.vanMerode.Doescase‐mixbased reimbursem*nt stimulate the development of process‐oriented care delivery?HealthPolicy.

thesis_Vos_17x24_v01.pdf

thesis_Vos_17x24_v01.pdf

Chapter4 │79

4.1 IntroductionDuring the last decade, it is believed that fundamental shortcomings in theorganisation of hospital care can cause quality problems like long waitingtimes,wasteofresourcesandpoorcoordinationofcaredelivery1.Traditionally,hospitals are functionally organised intomedical (e.g. internalmedicine) andfacilitatingdepartments(e.g.laboratory),whichoperateasseparate‘silos’withtheir owngoals2, 3. Each silo is concernedwithmaximising its ownefficiencyratherthanthatofcareprocesses4.Asaconsequence,eachcareactivitywithinacareprocessofapatientwillbeplannedonbasisofavailabilityofresources,after the previous care activity is completed. This means that activitiesbelongingtoapatient’scareprocessaretreatedasbeingindependentinsteadofasbeingconnectedtoachainofinterdependentcareactivities.This poor coordination of care delivery is partially due to the use of fee‐for‐service systems of provider payment4. A fee‐for‐service system rewards thedeliveryofindividualprocedures(i.e.hospitaladmissions,numberofin‐patientdays)andstimulatestheoveruseofservices,duplicationofservicesanduseofcostlyspecialisedservices.Onewaytostimulatecoordinationandefficiencyofcaredelivery,istobasereimbursem*nt,inwholeorinpart,onthetotalcareofapatientduringanacuteepisodeof illness5.This is in linewith the theoryoftarget engineering, which states that external pressures related toreimbursem*nt systems can be used to bridge internal conflicts of interestshampering efficiency and quality6, 7. Based on this premise, many Westerncountries introduced Diagnosis Related Groups (DRGs) or similar case‐mixbasedgroupingsystemsforhospitalreimbursem*nt.AlsointheNetherlandsanew reimbursem*nt system was developed and initiated in 2005, that willgradually replace the fee‐for‐service payment of hospital care8. In this newsystem, hospital budgets are based on the number of delivered DiagnosisTreatment Combinations (DTCs). A DTC consists of all diagnosis‐ andtreatment‐related costs incurred by the hospital and the clinician. Thus, itcovers the pathway from an initial consultation or examination to the finalcheck‐up9. Consequently, hospital budgets become dependent upon theefficient and effective delivery of DTCs instead of on individual procedures.However, within this new system hospital budgets are not open‐ended.HospitalsnegotiatewithhealthcareinsurersregardingthemaximumnumberofDTCstheymaydeliverwithin1year.Toactuallyachieveabettercoordinatedandmoreefficientcaredeliveryacrosshospital department boundaries, hospitals may introduce a more process‐oriented way of working by the implementation of coordinating structures,

thesis_Vos_17x24_v01.pdf

80│ Towardsprocess­orientedcaredeliveryinhospitals

called care pathways or care programmes (see Figure 4.1)3, 10. Coordinatingstructures establish the sequence of activities (diagnostics, consultations,treatment) and the professionals’ responsibilities in the diagnosis andtreatment of hom*ogeneous patient groups. This potentially improves thecoordinationofcaredeliveryandreduceswaitingtimesanddelayswithincareprocesses forpatients.Hereafter,werefer to thesecoordinatingstructuresascareprogrammes.Figure4.1 Caredeliveryinafunctionalorganisationorprocess‐orientedorganisation

Althoughtheimplementationofcareprogrammesisconsideredamajorstep,itdoes not automatically imply that health care professionals adopt a process‐oriented way of working11. To actually change work processes, theimplementationofcareprogrammesmustbeaccompaniedby:- Alesshierarchicalorganisation,inwhichpeoplehavemoreresponsibility,

increased decision making capabilities, and act more autonomously andflexible12;

- Lessfragmentationofresponsibilitiesbyappointingprocessowners10,13;- Protocols, that ensuresmooth coordination, continuity, and lessvariation

betweencareprocessesperpatient1,14;and- Aprocess‐orientedviewheldbyallemployees13.

thesis_Vos_17x24_v01.pdf

Chapter4 │81

Davis (2007)5 argues that, given the dispersion of care across clinicians andpractices,thefragmentationofcaredeliveryandlackofcontinuityinclinician‐patient relationships, extensive evaluations of new reimbursem*nt methodsareneeded.Anumberofstudiesalreadyevaluatedtheeffectofcase‐mixbasedreimbursem*nt on output measures, like the length of stay15‐17. From thesestudies, itappearsthatadmissions,theaveragelengthofstayandintensityofcare,andthushealthcarecostshave,atleastinitially,decreasedasintended18.Inadditiontotheseintendedeffects,severaldysfunctionaleffects,suchasDRG‐creeping(‘adeliberateandsystematicshiftinahospital’sreportedcasemixinorder to improve reimbursem*nt’) and DRG‐dumping (‘an attempt to avoidtreatingdifficult cases’),havebeenreported18.Thesestudiesdidnotexaminetheeffectson theorganisationofcaredelivery, e.g. thecoordinationbetweencare activities. It is, however, very important to get insight into the effect oncare coordination because it is believed that hospital administrators adoptcase‐mixreimbursem*ntinhospitalbudgetingsystemsasameansofshowingconformitywith institutionalised rules and expectations19, but decouple theirbudgeting systems from the internal operations of the organisation. If that isthecase,abettercarecoordination,whichisveryimportantfromthepatient’spointofview,willnotbeachievedandimprovementsinefficiencymaybeonlythe result ofDRG‐creeping anddumping. Therefore,we explore in this studywhether case‐mix based reimbursem*nt stimulates activities to develop careprogrammes,andeventuallyaprocess‐orientedwayofworking.4.1.1 HypotheticalmodelOnthebasisofliterature,weconstructedahypotheticalmodel(seeFigure4.2)to demonstrate how case‐mix reimbursem*nt should contribute to theestablishmentofcareprogrammes,withtheultimategoaltobecomeprocess‐oriented. To meet the new requirements of case‐mix reimbursem*nt, Dutchhospitalshave tobase theallocationofbudgetsonprocesses (thenumberofdeliveredDTCs)ratherthanonindividualprocedures.Inthiswayhospitalscanalign the different departmental interests with the organisational goal ofmaximising the total hospital budget according to the theory of targetengineering. Thus, the model starts with the adoption of the newreimbursem*ntsystemwithinthehospitals’budgetingprocess(process‐basedbudgeting)andconsistsoffourhypothesisedrelations.

thesis_Vos_17x24_v01.pdf

82│ Towardsprocess­orientedcaredeliveryinhospitals

Figure4.2 Hypothesisedrelations

Hypothesis 1: A more process­based budgeting system stimulatesactivitiestoestablishcareprogrammes.When financial resources are allocatedonbasis of careprocessperformance,optimisation of care processes in keepingwith certain targets – all of whichinvolveatimefactor(e.g.shortthroughputtime)–ismoreimportantthanhighutilisation of departmental resources20. Hence, it is assumed that differentdepartmentsworkingforthesamepatientgroupwillundertakemoreactionstoimplementcareprogrammesandmeetprocesstargets.Hypothesis2:Amoreprocess­basedbudgetingsystemcoincideswiththeuseofmoreprocess­orientedindicatorsbyclinicians.For clinicians to assume responsibility for the establishment of ‘their’ careprocess, balanced steering information is a prerequisite21. Therefore, wehypothesise that process‐based budgeting coincides with the use of moreprocess‐orientedperformanceindicatorsbyclinicians.Hypothesis3:Theuseofmoreprocess­orientedindicatorsbycliniciansisrelated to a higher number of activities geared to establishing careprogrammes.Theinformationgatheredusingmeasurementofprocess‐orientedperformanceindicatorshelpstoidentifyareasforimprovement22, 23.Outputindicators(e.g.throughput times) tell hospital management and clinicians whether they arereaching targets24.Additionally, processmeasures, like thenumberofpatienthandovers between clinicians and physical transfers of patients betweendepartments,gaugethedeliveredcare25.Thenumberofphysicaltransfers,forexample, informs about the (non)integration of various operations in a careprocess. An optimal integration of various operations means a betteradjustment to patients’ needs and minimal delays and waiting times26. Wethereforeassume thatuseofprocess‐oriented indicators stimulates clinicianstoundertakeactionstodevelopandimprovecareprogrammes.

Processbasedfinancialallocation

Processorientedindicators

Activitiestoestablishcareprogrammes

%careorganisedincareprogrammes2 4

thesis_Vos_17x24_v01.pdf

Chapter4 │83

Hypothesis 4: More activities to establish care programmes lead to ahigherproportionofcaredeliverythatisorganisedincareprogrammes.Finally,itseemsplausiblethathospitalsconductingmoreactivitiestoestablishcareprogrammeshavehigherproportions of caredelivery organised in suchprogrammes than hospitals that undertake fewer or no activities. However,when hospitals are at the end of their transition towards a process‐orientedwayofworking,itislikelythattheycarryoutfewerornoactivitiestoestablishcareprogrammesbutcanhaveahighproportionofcaredeliveryorganisedincareprogrammes.4.2 Materialsandmethods4.2.1 DatacollectionTo assess the effect of case‐mix reimbursem*nt on the development towardsprocess‐oriented organisations, we conducted a survey during the winter of2007/2008.AquestionnairewassenttoallDutchhospitals(N=96),includinguniversity(N=8),tertiaryteaching(N=26),general(N=59),andspecialisedhospitals (N = 3), followed by a reminder after approximately three weeks.Recipients were hospitals’ chief executive officers, who in most instancescompletedthequestionnairetogetherwithqualitymanagers.Questions included: the extent to which process‐based budgeting is applied(interval: completely disagree 0.0 – completely agree 5.0); the use of fourdifferent process‐oriented performance indicators (dichotomously scaled:yes/no);activities toestablishcareprogrammes(ordinallyscaled:none, forasinglepatientgroup,forseveralpatientgroups,hospitalwide);thepercentageof care delivery that is/can be organised in care programmes (interval: 0‐100%); indicators for process‐orientation (decision making [interval:completelycentralised0.0–completelydecentralised5.0];theestablishmentofprocessowners[interval:completelydisagree0.0–completelyagree5.0];theemployees’viewofcaredeliveryasaprocess[interval:completelydisagree0.0– completely agree 5.0]; (clinical) protocols for specific diseases[dichotomously scaled: yes/no]; (organisational) protocols for routing ofpatients [dichotomous scaled: yes/no]); and hospital characteristics (type ofhospital, the number of fulltime equivalents (FTEs) employed personnel, andthenumberofpeoplelivingintheareathehospitalsserve).This questionnaire is in fact part of a larger survey carried out on a regularbasis(at leastevery5years)andaimstoassessthedevelopmentalstageofa

thesis_Vos_17x24_v01.pdf

84│ Towardsprocess­orientedcaredeliveryinhospitals

hospital’s qualitymanagement system27.However, followingdevelopments inthe hospital sector, the questions involving the application of process‐basedbudgetingandthepercentageofcaredeliverythatis/canbeorganisedincareprogrammesareintroducedinthesurveyof2007.Althoughseveralprocess‐orientedindicatorsexist,welimitedourselvesinthisquestionnaire to: ‘length of in‐hospital stay’, ‘throughput times’, ‘number ofphysical transfers between departments’, and ‘number of patient handoversbetweenclinicians’.Theseindicatorsareclosestrelatedtooneoftheprinciplesof theestablishmentof careprogrammes,which is to simplify theprocessbyreducingthenumberofinteractionsandhandovers28.Besides,theseindicatorsare the known indicators among hospitals and are frequently used withinprojectsthataimtoestablishcareprogrammes24.4.2.2 DataanalysisDescriptive statistics were used to assess the extent to which process‐basedbudgeting is applied, the use of four different process‐oriented performanceindicators, thenumber of activities geared to establish careprogrammes, thepercentageofcaredeliverythat is/canbeorganised incareprogrammes,andthe indicators of process‐orientation in hospitals. Next, structural equationmodelling(SEM),astatisticaltechniquefortestingandestimatingrelationshipsusingacombinationofstatisticaldataandqualitativecausalassumptions,wasusedtoassessthehypothesisedrelationsasvisualisedinFigure4.2.ToapplySEM, all observed variables in themodel need to have a normal distribution.Therefore, variables were first assessed for univariate and multivariatenormality. Univariate normality was assessed using skewness statistic √b1,kurtosis statistic b2, and the D’Agostino & Pearson K2 omnibus test fornormality29,30.Inaddition,Srivistava’sandSmall’stestofmultivariatekurtosisand skew, and an omnibus test of multivariate normality based on Small’sstatistics,wereusedtocheckmultivariatenormality30,31.Asignificantvalueofoneofthesestatistics(p<0.05)indicatesadeviationfromnormality.Subsequently, thehypothesisedrelationshipswereassessed intwosteps:onethat tested the hypotheticalmodel and a second testwithout non‐significantrelations.Modelfitdependsonanumberoffeaturesthatneedtobeexaminedinadditiontothesignificanceofparameterestimates.Welimitedourselvestocommon fit statistics like Chi‐square, the Tucker‐Lewis Index (TLI), theConfirmativeFitIndex(CFI),andtheRootMeanSquareErrorofApproximation(RMSEA). A significant Chi‐square represents ‘badness of fit’. This test is

thesis_Vos_17x24_v01.pdf

Chapter4 │85

suitableformodelswithasamplesizeupto100.TLI‐andCFIvaluesbelow.90indicate that the model can be improved. Values between .90 and .95 areacceptable,andvaluesabove .95aregood32.Goodmodels,moreover,haveanRMSEA value of equal to or lower than .05, values between .05 and .08 areconsidered acceptable, values higher than .10 indicate a poor fit 33, 34. SEManalysesinthisstudywereconductedusingAMOS16.0.Finally,weassessedwhetherimplementationofcareprogrammescontributedtothetransitiontowardsprocess‐orientation.Therefore,Spearmancorrelationcoefficients were computed between the percentage of care in careprogrammesandtheindicatorsofprocess‐orientationthatweremeasuredonan interval scale. Independent sample T‐tests were performed for theindicatorsmeasuredonadichotomousscale.4.3 ResultsSixty‐twohospitalscompletedandreturnedthequestionnaire(aresponserateof65%),includingfiveuniversity,fourteentertiaryteaching,fortygeneral,andthree specialised hospitals. On average, university hospitals have the mostpersonnel(infulltimeequivalents)andthehighestnumberofpeopletoserve,followedbytertiaryteaching,general,andspecialisedhospitals.Table4.1givesanoverviewofthedatagatheredinthehospitalsurvey.

thesis_Vos_17x24_v01.pdf

86│ Towardsprocess­orientedcaredeliveryinhospitals

Table4.1 Overviewsurveydata

Mean(sd) Range(min­max)

√b1*(p) b2^(p) D’Agostino&PearsonKsq#(p)

Process­basedbudgetingwithinthehospitalProcess‐basedfinancialallocation 2.1(1.0) 0‐4 ‐0.22(0.49) ‐1.39(0.16) 2.41(0.30)Process­orientedperformanceindicatorsUseofprocess‐orientedperformanceindicators:

1.7(0.8) 0‐3 0.38(0.24) ‐0.16(0.88) 1.39(0.50)

- Indicatorforthroughputtimes(%,[n])

(47.5[29])

- Indicatorforlengthofin‐hospitalstay(%,[n])

(63.9[39])

- Indicatorfornumberofhandovers(%,[n])

(11.5[7])

- Indicatorfornumberoftransfers(%,[n])

(9.8[6])

ActivitiestoestablishcareprogrammesActivitiestoestablishcareprogrammes:

2.2(0.7) 0‐3 ‐0.10(0.75) ‐1.42(0.16) 2.12(0.35)

- None(%,[n]) (1.6[1])- Asingleproject(%,[n]) (17.7[11])- Severalprojects(%,[n]) (58.1[36])- Hospitalwide(%,[n]) (22.6[14])Percentageofcaredeliverythatisorcanbeorganisedincareprogrammes%ofcaredeliveryincareprogrammes

33.1(20.1) 10‐100 1.23(0.00) 1.89(0.06) 13.70(0.00)

%ofcaredeliverypotentiallyincareprogrammes

71.0(14.2) 30‐100 ‐0.37(0.25) 0.77(0.44) 1.92(0.38)

Indicatorsofprocess­orientationDecentraliseddecisionmaking 2.3(1.0) 0‐5 ‐0.04(0.90) ‐0.05(0.96) 0.02(0.99)Processviewemployees 2.5(1.0) 1‐5 0.07(0.83) ‐0.46(0.65) 0.26(0.88)Processownership 3.1(1.0) 1‐5 ‐0.64(0.06) 0.09(0.93) 3.63(0.16)Protocolsforroutingofpatients(%,[n])

(74.2[46])

Protocolsforspecificpatientgroups(%,[n])

(93.5[58])

*Skewnessstatistic√b1,kurtosisstatisticb2.#D’Agostino&PearsonK2omnibustestfornormality.

thesis_Vos_17x24_v01.pdf

Chapter4 │87

Thesedatashowthat:- 33.3%(n=19)ofthehospitalsallocatefinancialresourcesonthebasisof

careprocessestosomeextent(score>2.5);- Cliniciansinhospitalsgenerallyusetwoprocess‐orientedperformance

indicators,namely‘lengthofin‐hospitalstay’and‘throughputtimes’,tomonitorandadjusttheorganisationofcareprocesses.Theindicatorsforthenumberoftransfersbetweendepartmentsandpatienthandoversbetweencliniciansareusedbyapproximately10%ofthehospitals;

- Mosthospitals(n=50,81%)undertakeseveralprojectstoestablishcareprogrammes.Onlyonehospitalisnotactivelyinvolvedinestablishingcareprogrammes;

- Onaverage,33.1%(sd=20.1)ofcaredeliverywasorganisedincareprogrammesinthewinterof2007/2008and71.0%(sd=14.2)ofcaredeliverycanbeorganisedincareprogrammesaccordingtotherespondents.

Dataonprocess‐orientationindicatorsshowthat:- Onaverage,decisionmakingismorecentralisedthandecentralised,aswas

indicatedbyone‐thirdofthehospitals(33.9%,n=21);- 75.4%(n=46)ofthehospitalsappointprocessownerstocareprocesses

(score>2.5);- Caredeliveryisregardedasaseriesoflinkedcareactivitiesintheminority

(39.3%,n=24)ofthehospitals(score>2.5);- 93.5%ofthehospitalshas(clinical)protocolsforspecificdiseasesand

almost75%has(organisational)protocolsforroutingofpatients.Beforetestingthehypothesisedrelationships,wefirstassessedtheunivariateandmultivariatenormalityof thevariablesof thehypotheticalmodel (Figure4.2).Testsforunivariatenormalityindicatethatthevariablescanberegardedasnormallydistributedexceptforthevariable‘%patientsincareprogramme’(see Table 4.1). To normalise this positively skewed variable, we applied asquareroottransformation.Afterthistransformation,normalitystatisticsweregood(√b1=0.52[p=0.12],b2=0.15[p=0.88],D’Agostino&PearsonK2=2.50[p=0.29]).Thetransformedvariablewasusedinfurtheranalysis.Inaddition,statistics indicate that multivariate normality of the data can be assumed(multivariatekurtosis[Srivistava’s test:χ=3.85,df=4,p=0.43;Small’stest:Q1=4.69,df=4,p=0.43];multivariateskewness[Srivistava’stest:χ=2.51,df=4,p=0.16;avariantofSmall’stest:VQ2=4.04,df=4,p=0.40];omnibustestofmultivariate normality based on Small’s statistics [VQ3 = 8.74, df = 8, p=0.37]).

thesis_Vos_17x24_v01.pdf

88│ Towardsprocess­orientedcaredeliveryinhospitals

Next,thehypotheticalmodelwastestedusingSEM.Itwasshownthatprocess‐basedbudgetingdoesnothaveasignificanteffectonactivitiestoestablishcareprogrammes(hypothesis1:b=0.02;p=0.80),butdoescoincidesignificantlywith the use of process‐oriented performance indicators by clinicians(hypothesis2:b=0.30;p=0.02).Theuseofahighernumberofperformancemeasures has in turn a positive relation with activities to establish careprogrammes(hypothesis3:b=0.21;p=0.01).Further,anincreaseinactivitiestoestablishcareprogrammesispositivelyrelatedtothepercentageofpatientsbeing treated in care programmes (hypothesis 4: b = 0.40; p = 0.00). Fitmeasuresforthismodelaregood:Chi‐square‘badnessoffit’ isnotsignificant(0.96;p=0.62),theCFIandTLIarealmost1.00andtheRMSEAisclosetozero.Inasecondtest,thenon‐significantrelationbetweenprocess‐basedbudgetingandactivitiestoestablishcareprogrammeswasremoved(Hypothesis1).Testresultsdemonstratethatrelations2,3,and4arestillsignificantandestimatesareunchanged.Chi‐square isnotsignificant (1.02;p=0.80).TheTLIandCFIarealmost1.00;theRMSEAisclosetozero.Thismeansthatthemodelfitsofthefirstandsecondmodelaresimilar.However,becausethesecondmodel ismorerestrictive,itismoreinformativeandshouldthusbepreferred.Statistics show that thepresence ofmore careprogrammes is not associatedwithhigherscoresonindicatorsforprocess‐orientation:decentraliseddecisionmaking(ρ=‐0.11;p=0.44),agreementsaboutprocessownerships(ρ=‐0.02;p=0.92),hospitalemployeeshavingaprocess‐orientedview(ρ=0.06;p=0.67),availabilityofprotocols forspecificpatientgroups(t= ‐0.22;p=0.83)or forroutingofpatients(t=‐1.82;p=0.08).4.4 DiscussionThepurposeofthisstudywastotestwhethercase‐mixbasedreimbursem*ntstimulates activities to develop care programmes, and eventually a process‐orientedwayofworking.SEManalysisrevealsthattheadoptionofthecase‐mixreimbursem*nt systemwithin the hospitals’ budgetingprocess, via apositiveeffect on the number of process‐oriented performance measures used byclinicians, leads to a higher frequency of activities to establish careprogrammes.Internalprocess‐basedbudgetingwasnotconfirmedtostimulatethese activities directly. It is possible that clinicians need process‐orientedperformance data before they feel compelled to optimise care processes, towork together, and to establish care programmes. This view is supported bypreviousresearchofLehtonen(2007)18 thatshowed that the implementation

thesis_Vos_17x24_v01.pdf

Chapter4 │89

of case‐mix reimbursem*nt in the hospitals’ budgeting process is stronglydependentontheinvolvementofclinicians.Results on indicators for process‐orientation show, however, thatimplementationofcareprogrammesinthesurveyedhospitalshasnotyetbeenaccompanied by process‐orientation. It seems that hospitals do implementprotocols for the diagnosis, treatment, and routing of patients, but thatemployees frequently do not acquire a process‐oriented view. The change offocus from function to process, and all additional arrangements (e.g.decentralisationofresponsibilities),mayrequiremoretime.Itcouldbedifficultfor clinicians, aswell as for nurses, to find the appropriate balance betweentheir traditionally strong orientation to professional values and theircommitmenttodevelopthepowerandprestigeoftheirprofession3;andamorecollaborativeprocess‐orientationthatisneededtoeffectuatecareprogrammes.The lacking process‐orientation can also be related to the low percentage(10%) of hospitals in which clinicians use process‐related performancemeasures to monitor and adjust the organisation of care delivery (i.e. thenumber of patient handovers). The use of these kinds of performanceindicatorsbycliniciansisimportantforthedevelopmentofaprocess‐orientedview, because they provide insight into the process of care (i.e. steps to betakentodiagnoseandtreatpatients)andinformationaboutwaste(i.e.waitingtimes,unnecessaryconsultations).At the time of this research, only one‐third of the hospitals applied internalprocess‐based budgeting to any extent. Because the case‐mix‐basedreimbursem*nt system gradually replaces the fee‐for‐service system in theNetherlands, it can be expected thatmore hospitalswill apply process‐basedbudgetingandtoagreaterextentinthefuture,whichinturnwillcontributetothedevelopment ofmore careprogrammes. This expectation is strengthenedby the results of a recent longitudinal analysis on the development of Dutchhospitalqualitymanagementsystems.Thisanalysisrevealedthatfinancialandnon‐financial policy measures increased the development levels of Dutchhospitalqualitymanagementsystemssince1995,whichincludesanincreaseintheuseofprotocols forspecificpatientordiagnosisgroups, protocols for theroutingofpatientsfromadmissiontodischarge,andmanagementinformationsystems providing periodic overviews of care provision and outcomes27.Assuming that this trend continues,we expect that policymeasures (like theintroduction of case‐mix reimbursem*nt) contribute to the furtherdevelopmentofprocess‐orientedcaredeliveryviatheincreaseduseofprocess‐oriented performance indicators and the development of care programmes.Thismayhelptoshiftthefocusfromfunctiontoprocessinhospitals,becauseit

thesis_Vos_17x24_v01.pdf

90│ Towardsprocess­orientedcaredeliveryinhospitals

stimulateshealthcareprofessionalsindifferentdepartmentstocollaborateandtoachievecommongoals.Although this study gives a valuable insight in the effect of case‐mixreimbursem*nt on the development of a process‐oriented way of working,threekey limitations inthestudyshouldbementioned.First,wemadeuseofself‐reported data gathered using a questionnaire. As a consequence, socialdesirability bias could have influenced our study results. It is a generallimitationofquestionnaires that respondentshave the tendency to reply inamannerthatwillbeviewedfavourablybyothers.Therefore,someoftheresultscouldgiveatoooptimisticpicture.Second,theanswerschiefexecutivesgivetoquestionsabout issues related to characteristics of decentralisedmembersoftheorganisation(suchastheprocess‐orientedviewofemployees)reflecttheirpersonal,potentiallybiased,interpretation.Therefore,futureresearchneedstoextendthenumberofquestionnaireaddresseestodecentralisedorganisationalmembers.Thethirdlimitation,whichisinherentinSEM,isthatagoodfitbyamodelconsistentwithonecausalhypothesisdoesnotruleoutanequallygoodfitbyanothermodelconsistentwithadifferentcausalhypothesis.Careshouldalwaysbe takenwhenmakingclaimsofcausality, especially sincewedidnotcollectdataatmultipletimepoints.4.5 ConclusionsThis study demonstrates that hospital management can stimulate thedevelopment of care programmes by the adoption of the case‐mixreimbursem*ntsystemwithinhospitals’budgetingprocesses.Thismeansthatcase‐mix reimbursem*ntdoespotentially improvecarecoordinationand thatdecouplingbetweenhospitals’budgetingsystemsandinternaloperationsdoesnottakeplace.Futureresearchisrecommendedtoconfirmthis findingandtodeterminewhethertheestablishmentofcareprogrammeswill in timeindeedleadtoamoreprocess‐orientedviewofprofessionals.

thesis_Vos_17x24_v01.pdf

Chapter4 │91

References1. InstituteofMedicine. Crossing thequality chasm: anewhealth system for the

21thcentury.Washington,DC:NationalAcademyPress;2001.2. Lega F, DePietro C: Converging patterns in hospital organization: beyond the

professionalbureaucracy.HealthPolicy2005,74(3):261‐281.3. Vera A, Kuntz L: Process‐based organization design and hospital efficiency.

HealthCareManagementReview2007,32(1):55‐65.4. Ludwig M, Van Merode F, Groot W: Principal agent relationships and the

efficiencyofhospitals.TheEuropeanJournalofHealthEconomics2010,11(3):291‐304.

5. DavisK:Payingforcareepisodesandcarecoordination.NewEnglandJournalofMedicine2007,356(11):1166‐1168.

6. Kaplan RS, Cooper R: Cost and effect: Using integrated cost systems to driveprofitabilityandperformance.Boston:HarvardBusinessSchoolPress;1997.

7. Van Merode GG: A prelude of 2004 Antwerp quality conference: targets andtarget values ‐ integrating qualitymanagement and costing. Accreditation andQualityAssurance2004,9:168‐171.

8. FolmerK,MotE:DiagnosisandtreatmentcombinationsinDutchhospitals.TheHague:CPBNetherlandsBureauforEconomicPolicyAnalysis;2003.

9. Custers T, Arah OA, Klazinga NS: Is there a business case for quality in TheNetherlands?Acriticalanalysisoftherecentreformsofthehealthcaresystem.HealthPolicy2007,82(2):226‐239.

10. Gemmel P, Vandeale D, Tambeur W: Hospital Process Orientation (HPO): thedevelopment of a measurement tool. Total Quality Management & BusinessExcellence2008,19(12):1207‐1217.

11. Gemmel P, Vandaele D, Tambeur W: Hospital process orientation (HPO): thedevelopmentofameasurementtool.Gent:GhentUniversity,Belgium;2007.

12. HammerM,ChampyJ:Reengineeringthecorporation:Amanifestoforbusinessrevolution.London:NicholasBrealey;1993.

13. McCormack KP, Johnson WC: Business process orientation: gaining the e‐businesscompetitiveadvantage.BocaRaton:CRCPressLLC;2001.

14. SchnellerES,SmeltzerLR:Strategicmanagementofthehealthcaresupplychain.SanFrancisco:Jossey‐Bass;2006.

15. BrockKA,ValeSJ,CottonSM:Theeffectoftheintroductionofacase‐mixbasedfunding model of rehabilitation for severe stroke: an Australian experience.ArchivesofPhysicalMedicineandRehabilitation2007,88:827‐832.

16. Theurl E, Winner H: The impact of hospital financing on the length of stay:evidenceofaustria.HealthPolicy2007,82:375‐389.

17. ForsbergE,AxelssonR,ArnetzB:Financialincentivesinhealthcare.Theimpactofperformance‐basedreimbursem*nt.HealthPolicy2001,58:243‐262.

thesis_Vos_17x24_v01.pdf

92│ Towardsprocess­orientedcaredeliveryinhospitals

18. LehtonenT:DRG‐basedprospectivepricingandcase‐mixaccounting‐Exploringmechanisms of successful implementation. Management Accounting Research2007,18(367‐95).

19. CovaleskiMA,DirsmithMW,Michelman JE:An institutional theoryperspectiveon the DRG framework, case‐mix accounting systems and health careorganizations.AccountingOrganizationsandSociety1993:65‐80.

20. Vissers JMH,BeechR:Healthoperationsmanagement:patient flow logistics inhealthcare.London:Routledge;2005.

21. Berg M, Schellekens W, Bergen C: Bridging the quality chasm: integratingprofessionalandorganizationalapproachestoquality.InternationalJournalforQualityinHealthCare2005,17(1):75‐82.

22. BerwickDM,JamesB,CoyeMJ:ConnectionsbetweenqualitymeasurementandImprovement.MedicalCare2003,41(1(suppl)):I30‐38.

23. DeVosM,GraafmansW,KooistraM,MeijboomB,VanDerVoortP,WestertG:Using quality indicators to improve hospital care: a review of the literature.InternationalJournalforQualityinHealthCare2009,21(2):119‐129.

24. Nolan TW, Schall MW, Berwick DM, Roessner J: Reducing delays and waitingtimes throughout the healthcare system. Boston: Institute for HealthcareImprovement;1996.

25. Grol R, Wensing M, Eccles M: Improving patient care: the implementation ofchangeinclinicalpractice.Edinburgh:Elsevier;2005.

26. LikerJK:TheToyotaway:14managementprinciplesfromtheworld'sgreatestmanufacturer.NewYork[etc.]:McGraw‐Hill;2004.

27. DückersMLA,MakaiP,VosL,GroenewegenPP,WagnerC:LongitudinalanalysisonthedevelopmentofhospitalqualitysystemsintheNetherlands.InternationalJournalforQualityinHealthCare2009,21(5):330‐340.

28. PlsekPE:Systematicdesignofhealthcareprocesses.QualityinHealthCare1997,6:40‐48.

29. D'AgostinoRB,BelangerA,D'AgostinoRB, Jr.:A suggestion forusingpowerfulandinformativetestsofnormality.AmericanStatistician1990,44:316‐321.

30. DeCarloLT:On themeaning anduseof kurtosis. PsychologicalMethods1997,2(3):292‐307.

31. Looney SW: How to use tests for univariate normality to assess multivariatenormality.AmericanStatistician1995,49:64‐70.

32. BentlerPM,BonettDG:Significance testsandgoodnessof fit in theanalysisofcovariancestructures.PsychologicalBulletin1980,88:588‐606.

33. Browne MW, Cudeck R: Alternative ways of assessing model fit. In Testingstructural equation models. Edited by Bollen KA, Long JS. Newbury Park, CA:SAGE;1993:136‐162.

34. Byrne BM: Structural equation modeling with AMOS. Mahwah, NJ: LawrenceErlbaum;2001.

thesis_Vos_17x24_v01.pdf

│93

Chapter5

Howtoimplementprocess­orientedcare

acasestudyontheimplementationofprocess­orientedin­hospitalstrokecare

Thischapterwaspublishedas:Leti Vos L, Robert J. van Oostenbrugge, Martien Limburg, Godefridus G. van Merode,Siebren Groothuis. How to implement process‐oriented care: a case study on theimplementation of process‐oriented in‐hospital stroke care. Accreditation and qualityassurance2009:14:5‐13.

thesis_Vos_17x24_v01.pdf

thesis_Vos_17x24_v01.pdf

Chapter5 │95

5.1 IntroductionDutch hospitals are currently in the midst of a transition towards process‐oriented organisation, in order to realise optimal and undisturbed careprocesses. Hospitals in the Netherlands traditionally have a functionalstructure. Similar capacities aregrouped indepartments (units), for example,surgeons in thesurgerydepartment,clinicalchemistry in laboratories,andsoon1. The task differentiation and specialisation of physicians and, to a lesserdegree, also nurses, are the main reasons for the typical functionalorganisationaldesignofhospitals2, 3. Inafunctionallyorganisedhospital,eachdepartmentstrivestooptimiseitsleveloffunctioning,butisunabletointegrateitsservicestobestmeettheneedsofpatients4,i.e.onedepartmentisnotabletotune its processes to those in other departments. In a process‐orientedorganisation,processesaremappedsothattaskresponsibilitiesaredescribedwithafocusonprocesses.Inorganisationsofthistype,functionalbordersarecrossed, and all members of different departments are encouraged tocollaborateandachievecommongoals5.Therearetwowaystoimplementprocess‐orientedorganisationofcare:A. Implementationofcoordinationmeasures(suchascareprogrammes);orB. Implementationofclinicaldirectorates3,5.In the former, horizontal processes are placed on top of the existing verticalstructure,without changing the functional organisation5. The implementationof clinical directorates, on the other hand, requires changes in the hospitalstructure. Clinical directorates are intermediate organisational arrangementsthrough which certain parts of larger hospitals are managed. The hospitalorganisation is then divided into multidisciplinary organisational units,bringing togetherphysicians,nurses,paramedical,administrative,andclericalstaff.Thesemultidisciplinaryorganisationalunitshandleabusinessprocessascomprehensively as possible and have relatively few interdependencieswithotherclinicaldirectorates.Withbothstrategies,itisnecessarytoovercomethefunctionaldivisionoflabour3,6.Until now, the implementation of process‐oriented organisation of care inDutch hospitals has been limited to the implementation of coordinationmeasures.Theimplementationofdirectoratesishamperedby,forexample:‐ Political and ethical obligations, which prevent hospitals from deleting

services, and focussing on strategically important serviceswith the samefreedomasfirmsinotherindustries3;and

thesis_Vos_17x24_v01.pdf

96│ Towardsprocess­orientedcaredeliveryinhospitals

‐ Physicians’strongorientationtoprofessionalvaluesandtheircommitmentto developing the power and prestige of their profession, which oftenconflicts with managerial goals associated with process‐basedorganisation3.

Theintroductionofprocess‐orientationinafunctionalstructureiscomplex.Ina functional structure, the organisation of care delivery traditionally focusesonlyoncertainelementsofcareratherthanthewholecareprocess.Theresultis a very complex system of flows and queues. There are many interactionsbetweenworkplaces, and thereforemany transfers occur. As a consequence,thethroughputtimesofpatientflowsareunpredictable,andthequalityofcareis less than optimal4. In order to implement process‐oriented care in afunctionalstructure,managementandhealthcareprofessionalsmusthandleanadditionalfocus:besidestheirfocusonoptimisingunitperformance,theyalsoneedtofocusonthecontrolofcareprocessesacrosshospitalunits.Tooptimiseunitperformance(utilisationofresourceswithinunits),caredemandandcaredeliveryhavetobeadjusted.Tomaximisethecontrolofcareprocesses,wastein the patient flow over functional borders of separate units must beeliminated. Waste can be considered as all activities, such as waiting andmovement of staff and patients, that do not add value to a care process7. Tomaintain focus on both these aspects, management and health careprofessionals have to find the appropriate trade‐off between optimal unitperformance (utilisation of resources within units), and the level of serviceprovided to stroke patients (short waiting time for diagnostic tests orconsultationwithothermedicalspecialties,andthroughputtimes).In this article, we report on a case study of the implementation of process‐orientedcareusingcoordinationmeasuresforanin‐hospitalstrokeunit.5.1.1 Casestudy:process­orientedin­hospitalstrokecareIn 2004 the University Hospital Maastricht (UHM) decided to implementprocess‐orientation forstrokepatients inorder toadjust theirservices to theneedsof thesepatients and tooptimise theorganisationof in‐hospital strokeunit care. At that time the average length of hospital stay (LOS) for strokepatientswas12days,which,accordingtothemanagementofthedepartment,surpassed theoptimalLOS.According to themanagementof thedepartment,the reason for this was a lack of process‐orientation, resulting in pooradjustmentof caredelivery to theneedsof thepatientsand inefficientuseofhospital resources8.The importanceofwell organised stroke care at a stroke

thesis_Vos_17x24_v01.pdf

Chapter5 │97

unit (specialised hospital ward) has been highlighted by several studies9‐12.Admissiontoastrokeunitleadstoimprovedhealth,functionaloutcomes,andsurvival9,13.Incontrasttoawarenessofthequalitativecomponentsofeffectivestrokeunitcare,lessattentionhasbeenpaidtotheoperationalmanagementofstrokeunitcare.Littleisknownabouteffectivemeasuresthatcouldcontrolthewholecareprocessfromaprocesspointofview,simultaneouslyoptimisingtheefficiencyofthedepartmentsinvolved.Inthecaseof in‐hospitalstrokecare, thehospital isconfrontedwithanacutecaredemand.Over70%ofstrokepatientsareadmittedtohospitalintheacutephase,resulting inuncertaintyaboutthetypeandextentofcareneeded.Thisrequires that the hospital has spare (reactive) capacity. Besides uncertaintyoverarrivaltimes,thetypeofcareneededbypatientsatthestrokeunitisalsocomplex due to the multitude of co‐morbidities, which in turn requires theinvolvement of several medical specialties and paramedical disciplines. Theuncertainty and complexity of care for stroke patients complicate theadjustment between care demand and care delivery. In the case of process‐orientation this adjustment will also be dependent on the collaborationbetween several departments, all ofwhich are involved in delivering care tostroke patients (neurologists, diagnostic capacities, and allied health careprofessionals such as occupational therapists and physiotherapists). Thecentralideaofprocess‐orientationis,afterall,theoptimisationofpatientflowover the functional borders of the separate units involved. In this study, weexplored the possibilities of implementing process‐orientation usingcoordinationmeasures.5.2 MethodsInordertoimplementprocess‐orientationforcareinanin‐hospitalstrokeunit,weusedatime‐seriesdesignconsistingoffoursteps:1. Process‐analysis;2. Identificationofbottlenecks;3. Setting goals for process‐oriented care and selection of coordination

measurestorealiseprocess‐orientedcare;4. Implementationofcoordinationmeasuresandevaluationofeffects.Eachofthesestepsisdiscussedinmoredetailbelow.Step1:Process­analysisIn 2004 we undertook a process‐analysis of the stroke care process. OurretrospectiveanalysiswasconfinedtopatientsadmittedtotheUHMbetween

thesis_Vos_17x24_v01.pdf

98│ Towardsprocess­orientedcaredeliveryinhospitals

June and December 2003. These patientswere selected retrospectively fromthe stroke service database of theUHM.Analysis of the existing careprocesscomprisedseveralsteps:‐ Analysis of the UHM protocol and guidelines of the Dutch Institute for

HealthcareImprovement(CBO)forstroketreatment14;‐ Observation of the care process and structured interviewswith the care

providersinvolved;‐ Quantitative description of the processes based on the records of stroke

patients.Wecollectedthefollowingdemographicdataonallincludedpatients:age,sex,stroketype(basedontheinternationalstatisticalclassificationofdiseases9threvision‐ICD‐9‐cm15: ischemic stroke [ICD‐9‐cm 434.9, 435, 436] andintracerebral haemorrhage [ICD‐9‐cm 431]), length of stay (time periodbetweenadmissionanddischarge, indays),wardofadmission(strokeunitorother),anddischargedestination(home,nursinghome,rehabilitationfacility).The medical files, the hospital patient information system, and paramedicaladministrations were checked for the activities listed to determine whetherthesetookplaceaccordingtoprotocol.Foreachdiagnostictestortherapeuticactivity we determined the day of request and day of performance. Wecalculated the average time between request and performance. Theinterrelationship between the day of multidisciplinarymeeting and LOS wascomputed per discharge destination with Spearman’s correlation coefficient.We visualised the average care process in a Gantt chart‐a graphicalrepresentationofthedurationoftasksagainsttheprogressionofthesetasks16thataidsunderstandingoftheworkflowinthestrokecareprocess.Step2:IdentificationofbottlenecksTheperformanceof the in‐hospital careprocessof strokepatientsbefore theredesignwasdeterminedby comparing treatmentprotocols according to theUHM protocol and Dutch stroke guidelines14 with actual patient treatmentschedules obtained from patient records. We also examined whether allrequired actions were taken, and elucidated time intervals between theseactions.Step3:Goalsforprocess­orientedcareandselectionofinterventionsWe evaluated the identified bottlenecks with all health care professionals,includingalliedhealthworkers,withtheobjectiveofimprovingpatientflowinin‐hospital stroke care in order to deliver optimal care. From an operationsmanagementpointof view, optimal care canbedeliveredonlywith the righttrade‐off between optimal unit performance (utilisation of resources in

thesis_Vos_17x24_v01.pdf

Chapter5 │99

individual units) and the service level provided to stroke patients (shortwaitingandthroughputtimes).Step4:ImplementationofredesignandevaluationofeffectsInJanuary2006theredesignedstrokecareprocesstorealiseprocess‐orientedcarewas implemented.Process‐analysisand selectionof interventions tookalong time becausemany different health care professionalswere involved intheprocessandtheimplementationoftheprocess‐orientationmeantashiftinresponsibilitiesforsomeoftheprofessionalsinvolved.Also,staffeducationwasneeded,whichagaintooktime.Retrospective analysis to determine the performance after redesign wasperformedforpatientsadmittedtothestrokeunitoftheUHMbetweenJanuaryandApril2006.Thelengthofthisintervalwaslimitedto4monthsforpracticalreasons. Demographic data from all included patients were gathered. First,demographicdata (sex,age,anddiagnosis)werecomparedwith independentsamplesbyt‐test.Next,wecomparedtheperformanceofthecareprocessforstrokepatientsbeforeandafterimplementationofprocess‐orientedcare.Dataconcerning the LOS of patients admitted before and after the redesign werecomparedwiththeMann–WhitneyUtest.5.3 ResultsThe following sections present the results of the case study according to thefourstepsofthetime‐seriesdesign.Step1:Process­analysisWeincludedhundredstrokepatients.Table5.1showsthedemographicdataofthe population studied. The average age was 74.0 years [standard deviation(sd) = 9.5 years], 56%weremale, 85%had an ischemic stroke, and 15% anintracerebralhaemorrhage.The lengthof stayvaried from3 to 33days,withtheaverage lengthofstaybeing12.0days(sd=6.3days).Duetoshortageofcapacityonthestrokeunit,thirty‐onepatientswereinitiallyadmittedtoanon‐specialisedwardoftheUHMandsubsequentlytransferredtothestrokeunit.The frequency of application of diagnostic examinations and consultations isshowninTable5.2.Thecareprocessstartsintheemergencyunit.Afterhistorytaking, physical examination, standard blood testing, ECG, and brain CT‐scan,patients are usually admitted to the stroke unit but in some cases first toanother ward (all within the first 24 h). The rehabilitation process, which

thesis_Vos_17x24_v01.pdf

100│ Towardsprocess­orientedcaredeliveryinhospitals

includes mobilisation, physical therapy, occupational therapy, and speechtherapy,startsinthefollowingdays,andadditionalexaminationstakeplaceonaverageonday4,havingbeenrequestedonday2.

Table5.1 Characteristics of patients included before and after implementation of process‐

orientation Before

(2004)

After

(2006)

Numberofpatientsincluded: 100 51- Intracerebralhaemorrhage(%) 15(15) 7(14)- Ischemicstroke(%) 85(85) 44(86)Sex(number):

- Male(%) 56(56) 25(49)- Female(%)

44(44) 26(51)

Averageageinyears(sd) 74.0(9.5) 73.6(12.9)

Lengthofhospitalstay*(days):

- Average(sd) 12.0(6.3) 7.3(5.1)- Range 3‐33 1‐30*TotaldurationofhospitalstayTheGantt chartof theexisting careprocess (Figure5.1) showsat least threedays in hospital in which relatively little activity takes place. These are notclinically necessary ‘watch and wait’ periods. This suggests that the hospitalstaycouldbereducedbyatleastthesethreedaysinthecaseofstablepatients.Required diagnostic tests‐serum, cholesterol, blood testing, and chest X‐ray‐werenotperformedin20%and34%ofcases,respectively.InthecaseofchestX‐ray, neurologists considered the protocol to be outdated14. Theduration ofthehospitalstayofpatientsdischargedtotheirownhomecorrelatedpositivelywith the number of days between day of admission and discussion in amultidisciplinarymeeting (Spearman’s ρ= 0.33, n = 37, p =0.05). Length ofhospitalstayofpatientsdischargedtoothercarefacilitieswasnotaffectedbythenumberofdaysbetweenadmissionanddiscussioninthemultidisciplinarymeeting. Patientswaiting for placement in a rehabilitation facility or nursinghomehadanextendedstayinhospital(3–8days).Thenumberofdayswithoutmedical or diagnostic intervention (with the exception of treatments such asoccupational therapy) is higher (3–9 days) for the latter patients than forpatientswhoweredischargedtohome.

thesis_Vos_17x24_v01.pdf

Table5.2

Frequencyofapplicationofdiagnosticexaminationsandmedicalconsultationsin100consecutivestrokepatientsadm

ittedtothestrokeunit

ofUniversityHospitalM

aastricht(UH

M)

IntracranialHaemorrhage(n=15)

Ischem

icstroke(n=85)

Standardactivity

accordingprotocol

Number

%

Applicationduration

(days)a,(sd)

Possibledayof

applicationb

Dayofapplicationin

careprocessc(sd)

Standardactivity

accordingprotocol

Number

%

Applicationduration

(days)a,(sd)

Possibledayof

applicationb

Dayofapplicationin

careprocessc(sd)

BrainCT‐scan

X15

100

0(0.3)

00(0.0)

X85

100

0(0.0)

00(0.4)

StandardBlood

Testing

X15

100

0(0.0)

00(0.2)

X85

100

0(0.0)

00(0.2)

ECG

X15

100

0(0.0)

00(0.2)

X85

100

0(0.0)

00(0.2)

SerumCholesterol

Testing

‐2

13

1(0.0)

‐3(2.8)

Xd

39

46

1(0.0)

03(3.4)

ChestX‐ray

‐6

40

1(0.1)

‐3(2.0)

X29

34

1(0.1)

03(3.8)

24‐hourECG

‐2

13

4(4.2)

‐3(1.4)

‐11

13

4(2.7)

‐3(1.5)

Speechtherapy

‐7

47

0e

‐2e

‐

50

59

0e

‐2e

Physicaltherapy

X11

73

2e

02e

X72

85

2e

02e

Beforeredesign.t=0

Occupationaltherapy

X12

80

2e

02e

X

70

82

2e

02e

aApplicationduration:timebetweenwritingoutanorderforandtheperformanceofadiagnostictestortherapeuticactivities;bpossibledayof

application:dayincareprocessonwhichordersforstandarddiagnostictestsortherapeuticactivitiescanbewrittenaccordingprotocol,day0istheday

ofadm

ission;c

dayofapplication:dayincareprocessonwhichordersfordiagnostictestsortherapeuticactivitiesarewritten;d

standardforfemale

patientsyoungerthan75yearsandmalepatientsyoungerthan70years;e basedoninterviewresults.

thesis_Vos_17x24_v01.pdf

Chapter5 │ 101

Figure5.1

Timepathofcareprocessof100consecutivestrokepatientsadmittedtothestrokeunitofUniversityHospitalM

aastricht(UH

M).Indicated

aretheaveragetim

e(indays)andrangesbetweenadmissionandperform

anceofeachactivity

thesis_Vos_17x24_v01.pdf

102│ Towardsprocess­orientedcaredeliveryinhospitals

Chapter5 │103

Step2:IdentificationofbottlenecksAnalysis of existing performance helped identify four main bottlenecks foroptimalorganisation.Thefirstbottleneckisalackofbedcapacityonthestrokeunit.One‐thirdofourpatientscouldnotbeadmittedtothestrokeunit.Patientsspendingover fivedays inanon‐specialisedwardresidedsignificantly longerinhospital(Mann–WhitneyU test:Z=‐2.9,p<0.05).Amoreefficientorgani‐sationofhospitalstayinthestrokeunitmayresultintreatingmorepatientsatthe strokeunit, and so reduce theuseofnon‐specialisedwards.Unnecessarydelays in requesting orders for diagnostic procedures, andwaiting times fordiagnostictestsandconsultationbyalliedhealthprofessionalsformthesecondbottleneck. When a patient has been admitted, orders for all standardprocedures should bewritten immediately by the neurologist. Yet, in reality,makingtheserequeststakesanaverageofthreedays.Thelowfrequencyofthemultidisciplinarymeeting(onceaweek)causesanincreaseinthelengthofstayfor patients thatwill be discharged to their homes. Finally, there arewaitingtimesforadmissiontorehabilitationandnursingfacilities.Step3:Goalsforprocess­orientedcareandidentificationofcoordinationmeasuresWeevaluatedthefourmainbottleneckswiththeobjectiveofimprovingpatientflowofin‐hospitalstrokecareinordertodeliveroptimalcarecharacterisedbyan appropriate trade‐off between optimal unit performance (utilisation ofresources in the units) and the level of service provided to stroke patients(short waiting and throughput times). In our case, the bed capacity of thestrokeunitwasa significantbottleneck.On theotherhand,waiting times fordiagnostic procedures and nursing facilities increase the LOS in hospital. InordertodecreasetheLOSandtoeliminatewaitingtimesasmuchaspossible,transfer from the hospital to a nursing facility also had to be subject toimprovement.Thehealthcareprofessionalsinvolvedsetthefollowinggoalsfortheimplementationofprocess‐orientation:1. AnaverageLOSintheUHMoffivedays;and2. Less than 5% of all stroke patients having to be admitted to a non‐

specialisedwardincaseofcapacityshortageonthestrokeunit.The following coordinationmeasureswere selected inorder toachieve thesegoals:‐ Useof anupdatedprotocol forpatients admitted to the strokeunit,with

clear instructions for standard diagnostic procedures and procedures onrequest. The chest X‐ray was dropped as a standard activity, as thisprocedureisnolongerconsideredbestpractice;

thesis_Vos_17x24_v01.pdf

104│ Towardsprocess­orientedcaredeliveryinhospitals

‐ OutsourcingofalliedprofessionalassessmenttotherehabilitationunitofanursingfacilityinordertoeliminatewaitingtimesforthisassessmentandtreatmentintheUHM,andduplicationofcaredeliverybytheUHMandtherehabilitationunitforstrokepatientsatanursingfacilityorrehabilitationcentre. As soon as patients are stabilised, theywill be transferred to thenursing rehabilitation unit of a nursing facility, in order to beginassessmentandrehabilitationasquicklyaspossible. In theUHMpatientsreceive care of allied health care professionals on demand. In this waypatientsdonothavetowait forplacement inanursingfacility, theyhavedirectaccess;

‐ Simplificationofdischargeplanning.Theweeklymultidisciplinarymeetinghas been removed from the agenda in the UHM because of the delayingeffect.Afterredesign,continuousmonitoringofthepatientisperformedbytheneurologist(inconsultationwithallinvolveddisciplines)toassessthepatient’s condition and decide on discharge to a rehabilitation unit of anursing home. Care delivery in the rehabilitation unit is attuned to theneeds of the stroke patient. Patients whose condition allows it, will bedischargeddirectlytotheirhome.

Theredesignof thecareprocess is illustrated inFigure5.2. Performancewasdetermined bymeasuring the outcomes of the care process on the followingformulatedgoals,whichactedasproxiesforoptimalunitperformanceandthelevelofserviceprovidedtostrokepatients:1. LOSintheUHM;and2. Percentageofstrokepatientsinitiallyadmittedtoanon‐specialisedward.These performance indicators were chosen because they oblige theprofessionals involved towork together as a group. If theydo not cooperate,stroke carewill not be optimisedandwaiting timeswill not be reduced, andconsequentlytherewillbenoreductioninLOS.

thesis_Vos_17x24_v01.pdf

Figure5.2

Redesignofthecareprocessofstrokepatientsadm

ittedtotheUH

Mstrokeunit

Occupationaltherapy(treatment)

Admission

Carotid

ultrasound

Discharge

Standard

bloodtesting

ECG

BrainCTscan

24­hourECG

Day0

Day1

Day2

Day3

Day4

Day5

Day6t/m9

Day10

Day11

Hom

e

Param

edical

asessm

entand

treatment

Multidisciplinary

meetingand

dismissal

planning(decision

takenonday5,28

and56ofstay)

Discharge

Nodischarge

Nursinghom

eorhom

eforthe

elderly

Hom

e(withhom

ecare)

Rehabilitationunitofanursinghom

e‘Klevarie’

=activityaccordingtoprotocol

=activityonrequestw

henindicated

UniversityHospitalMaastricht

Echocardio­

graphy

Chest

X­ray

BrainCTscan

Cardiologist

Serum

cholesterol

Speechtherapy(treatment)

Physicaltherapy(treatment)

Rehablilitation

unitfor

assessment/

revalidation

thesis_Vos_17x24_v01.pdf

Chapter5 │105

106│ Towardsprocess­orientedcaredeliveryinhospitals

Step4:ImplementationofredesignandevaluationofeffectsBetweenJanuaryandApril2006,fifty‐onepatientswereincludedinthestudytoassesstheeffectoftheredesign.Theaverageagewas73.6years(sd=12.9years), 49% of these patients were male, 86% of the patients suffered anischemic stroke and 14%had a intracerebral hemorrhage (Table 5.1). Therewerenosignificantdifferences indemographicdataofpatient groups inbothperiods(age,t‐test:t=0.22,df=149,p>0.05;sex,t‐test:t=0.58,df=149,p>0.05;diagnosis,t‐test:t=0.21,df=149,p>0.05).Table 5.3 shows the results of performance indicators of the care processbefore and after implementation of the coordinationmeasures. The length ofstayofpatientsadmittedaftertheredesignvariedfrom1to30days(average7.3days, sd=5.1days).After exclusionof threepatients thatwere explicitlyindicated fornursinghomecare, theaverage lengthofstaywas6.3days.Thelengthofstayofpatientsadmittedtothestrokeunitwassignificantlyshorterafterimplementationofthecoordinationmeasures(Mann–WhitneyUtest:Z=‐5.0,p<0.01).Onlyonepatienthadtobeadmittedtoanotherwardbecauseofalackofcapacityonthestrokeunit.Thiswassignificantlyfewercomparedtothenumberofpatientsadmittedtononspecialisedwardbeforetheredesign(t‐test:t=8.0,df=149,p<0.01).Table5.3 Performance of stroke careprocess in theUHMbefore and after implementation of

process‐orientation

Before

(2004)

After

(2006)

Sig.

Averagelengthofstay(sd)indays 12.0±6.3days 7.3±5.1days p<0.01

Percentageofpatientsadmittedtoanonspecialisedward

31% 2% p<0.01

5.4 DiscussionInthisstudyweexploredthepossibilitiesofimplementingprocess‐orientationforcareinanin‐hospitalstrokeunitusingcoordinationmeasures.Inprocess‐oriented care, the performance of the whole care process is reviewed. Tomaximise performance of a care process from the point of view of process‐orientedcare,waste(definedasunnecessarystepsinthecareprocess)hadtobeeliminated4,7.Besidesoptimisingthelevelofserviceprovided(shortwaitingand throughput times), we also needed to take the unit performance intoaccount, sinceprocess‐oriented carewas tobe implemented ina functionallyorganised hospital. In this case study, implementation of coordination

thesis_Vos_17x24_v01.pdf

Chapter5 │107

measuresledtobetterperformanceofin‐hospitalstrokeunitcareintermsofLOS (throughput time) and thenumberofpatients that couldbe admitted tothe stroke unit. Although the chosen process performance indicators (LOS,number of patients admitted to non‐specialised wards) are perhaps not thebest proxies for process‐orientation, we conclude that the coordinationmeasuresappliedherewereuseful incoordinatingtheactivitiesof thestrokecareunitfromaprocesspointofviewforseveralreasons.Firstly, standardisation of the in‐hospital stroke care process facilitated thedecision‐making process by making clear which medical and diagnosticprocedures were required by each patient and by whom these had to beperformed in the care process. This led to better control of the in‐hospitalstrokecareprocessintime.Asaconsequence,lessflexibilityintermsofsparecapacitywasneeded,thusoptimisingunitperformance.Secondly, the transfer of allied professional assessment and treatment to therehabilitationunitofanursingfacilityreducedwastecausedbyduplicationofwork and unnecessary waiting times for allied professional assessment andtreatmentintheUHM.Besidesthereductionofwaste,itsimplifiedtheplanningprocess andmade clearwhich aspects of carewere the responsibility of thespecialistservices.Inaddition,thefactthatthenumberofpatientsthatcouldbeadmitteddirectlyto the stroke unit increased after implementation reduced the complexity ofplanningofthecareprocessandincreasedthequalityofcare,sinceadmissionto a stroke unit leads to improved health, functional outcomes and survival.9Thereductioninthenumberofpatientsadmittedtonon‐specialisedwardsalsoeliminatedwastefromtheprocessbecauseprofessionalsnolongerhadtowalkto the other side of the hospital to care for their patients. We showed thatapplication of the described coordination measures contributed to both theadjustment of care demand and care delivery, as streamlining of the processreduced uncertainty and complexity. Uncertainty and, as a consequence, theneed for flexibility are reduced by making the process more transparent.Complexity is reduced by standardisation and reducingwaste. Care activitiesarenowdeliveredinamoreintegratedmanner.Furthermore,itwouldappearthat all involved professionals now see their care delivery as part of a careprocess. This means that they have shifted their focus from optimisation ofperformanceoftheirunittoimprovingprocessperformanceoverall.Althoughthisstudywasquitetime‐consuming,allstepstakenwereconsideredessential.VisualisingthecareprocessinGanttchartscreatesasenseofurgency

thesis_Vos_17x24_v01.pdf

108│ Towardsprocess­orientedcaredeliveryinhospitals

among the involved professionals to streamline processes because they canactuallyseewhatishappeningwiththepatientduringtheirhospitalstay.Limitations of our study include the fact that the process‐analysis andevaluation of effects did not take place at the same time of year. Seasonalfactors could have influenced our results. Also, the fact that the time spanbetween process‐analysis and evaluation of effects was quite long fororganisationalreasonscouldmeanthatchangesinhospitalpersonnelorotherchangesinhospitalpracticecouldhavetakenplaceandinfluencedourresults.However,we do not consider that these limitations biased our results aswedescribed the process in both periods in detail and were not aware of anychangesinthehospitalprocessotherthanthoseinitiatedbyourstudy.5.5 ConclusionThe present case study demonstrates that process‐orientation usingcoordinationmeasurescanbeimplementedforin‐hospitalstrokeunitcare,andcan improveperformanceby reducing theuncertainty and complexityof thistype of care. The method used to decide which coordination measures areneeded could be applied to other patient groups. The coordinationmeasuresthat the UHM has taken to optimise in‐hospital stroke care will not beautomatically applicable to other care processes because of the specificcomponents of each care process. However, the general principles of visualrepresentation,bottleneckreduction,andeliminationofwasteanduncertaintywouldbeveryusefulinmanysettings.

thesis_Vos_17x24_v01.pdf

Chapter5 │109

References

1. Van Merode GG: A prelude of 2004 Antwerp quality conference: targets andtarget values ‐ integrating quality management and costing. Accreditation andQualityAssurance2004,9:168‐171.

2. Braithwaite J: Strategic management and organisational structure:Transformationalprocessesatworkinhospitals.AustralianHealthReview1993,16:383‐404.

3. Vera A, Kuntz L: Process‐based organization design and hospital efficiency.HealthCareManagementReview2007,32(1):55‐65.

4. Van Merode F, Molema H, Goldschmidt H: GUM and six sigma approachespositionedasdeterministictoolsinqualitytargetengineering.AccreditationandQualityAssurance2004,10:32‐36.

5. Gemmel P, Vandeale D, Tambeur W: Hospital process orientation (HPO): thedevelopment of a measurement tool. Total Quality Management & BusinessExcellence2008,19(12):1207‐1217.

6. Vanhaverbeke W, Torremans H: Organizational structure in process‐basedorganizations.KnowledgeandProcessManagement1999,6(1):41‐52.

7. Liker JK:TheToyotaway:14managementprinciples fromtheworld'sgreatestmanufacturer.NewYork[etc.]:McGraw‐Hill;2004.

8. LimburgM,VosL,VanOostenbruggeR,VanMerodeGG,GroothuisS:Causesofinefficient stroke unit ‐bed use‐ possibilities for freeing up capacity.CerebrovascularDiseases2005,19(supplement2):114‐115.

9. Stroke Unit Trialists' Collaboration. Organised inpatient (stroke unit) care forstroke.TheCochraneDatabaseofSystematicReviews2001,Issue3.

10. KjellströmT,NorrvingB,A.S:Helsingborgdeclaration2006onEuropeanstrokestrategies.CerebrovascularDiseases2007,23:229‐241.

11. LanghorneP,PollockA:Whatare thecomponentsofeffectivestrokeunitcare?AgeandAgeing2002,31(5):365‐371.

12. LeysD,RingelsteinB,KasteM,HackW:Themaincomponentsofstrokeunitcare:result of a European expert survey. Cerebrovascular Diseases 2007,2007(23):344‐352.

13. BraininM,OlsenTS,ChamorroA,DienerHC,FerroJ,HennericiMG,LanghorneP,Sivenius J: Organization of stroke care: education, referral, emergencymanagement and imaging, stroke units and rehabilitation. CerebrovascularDiseases2004,17(suppl.2):1‐14.

14. LimburgM, TuutMK: CBO‐guideline 'Stroke' (inDutch).NederlandsTijdschriftvoorGeneeskunde2000,144(22):1058‐1062.

thesis_Vos_17x24_v01.pdf

110│ Towardsprocess­orientedcaredeliveryinhospitals

15. Manual of the international statistical classification of diseases, injuries, andcausesofdeath:basedontherecommendationsoftheninthrevisionconference,1975, and adopted by the 29th World Health Assembly. World HealthOrganization,Geneva(1977‐1978).

16. Ozcan YA: Quantitative methods in health care management: techniques andapplications.SanFrancisco:Jossey‐Bass;2005.

thesis_Vos_17x24_v01.pdf

│111

Chapter6

Applyingthequalityimprovementcollaborativemethodtoprocessredesign

amultiplecasestudy

Thischapterwaspublishedas:LetiVos,MichelL.A.Dückers,CordulaWagner,GodefridusG.VanMerode.Applyingthequality improvement collaborativemethod toprocess redesign: amultiple case study.ImplementationScience2010,5(1):19.

thesis_Vos_17x24_v01.pdf

thesis_Vos_17x24_v01.pdf

Chapter6 │113

6.1 BackgroundQuality improvement collaboratives (QICs) are used increasingly in manycountriestoachieve large‐scale improvements inperformanceandtoprovidespecificremediestoovercomethetypicallyslowdiffusionofmedicalandhealthcare innovations1‐3. A QIC is a multifacetedmethod that seeks to implementevidence‐based practice through sharing knowledge with others in a similarsetting over a short period of time4.Within theQICmethod, external changeagents provide collaborative project teams from different health caredepartmentsororganisationswithaclearvisionforidealcareinthetopicareaand a set of specific changes that may improve system performancesignificantly5, 6.Projectteamsalsolearnfromtheexternalchangeagentaboutthemodelforimprovement.Themodelforimprovementincorporatesfourkeyelements6: specific andmeasurable aims; measures of improvement that aretrackedovertime;keychangesthatwillresultinthedesiredimprovement;andseriesofparalleltestingplan‐do‐study‐act(PDSA)cycles.Eachseriesinvolvesatestofonechange idea(Figure6.1,partA)7.Onthebasisoftheresultsof thefirst testofoneseries,aprojectteamcandecidetorefinethechange idea(incase thechange ideaworks in theircontext)or tostartanew testseriesofanew change idea (in case the test did not lead to the desired result). ThesePDSAcyclesshouldbeshortbutsignificant,testingabigchangeideainashorttimeframesothatateamcanidentifywaystoimproveorchangetheidea8.InFigure6.2,anexampleisgiventoillustratethemodelforimprovement.Inadditiontotherelativelyefficientuseofexternalchangeagentsupportandthe exchange of change ideas as well as the model for improvement, thestrengthoftheQICmethodseemstobethatcollaborativeprojectteamsshareexperiences of making changes, which accelerates the rate of improvement(peerstimulus)3.However,despitethewidespreaduseofQICs,arecentreviewontheirimpactindicates that evidence is positive but limited, and the effects cannot bepredicted with certainty5. This apparent inconsistency requires a deeperunderstandingofhowandwhyQICswork.Thereforeitisnecessarytoexplorethe‘blackbox’oftheinterventionandtostudythedeterminantsofsuccessorfailureof theQICmethod5, 9. In thisarticle,wecontribute tothisbyassessingthe applicability of this quality improvement method to process redesign.Processredesignaimstoimprovetheorganisationofcaredeliveryintermsofwaiting times in a patient’s care trajectory. From other studies it is alreadyknown that the QIC method can be successfully applied to improve theorganisationof caredelivery in specificdepartments, such as emergency and

thesis_Vos_17x24_v01.pdf

114│ Towardsprocess­orientedcaredeliveryinhospitals

surgerydepartments8,10.But,toourknowledge,itisunknownwhethertheQICmethoditselfisapplicableforimplementingcomplexprocessredesigns,whichaimtochangepatternsofinteractionbetweendepartmentsinordertoachievespeedyandeffectivecarefromapatient’sperspective11.Therefore,weexploredinthisstudywhethertheQICmethodwasappliedtocomplexprocessredesignprojectsinaprocessredesigncollaborativeintheNetherlands.Figure6.1 TestingandimplementingchangesusingPDSAcycles

A P

S D

A P

S D

A

P

S

D

A

PS

D

DetailDesign

thesis_Vos_17x24_v01.pdf

Chapter6 │115

Figure6.2 Applyingthemodelforimprovement,anexample

6.2 MethodsThecollaborativedescribedinthispaperwaspartoftheDutchnationalqualityimprovementprogramme‘SnellerBeter’(‘BetterFaster’),whichbeganin2004asaninitiativefromtheMinistryofHealthandtheDutchHospitalAssociation.‘Sneller Beter’ aimed to realise substantial and appealing performanceimprovements in threegroupsofeightDutchhospitals in theareasofpatientlogistics and safety. These twenty‐four hospitals were enrolled in theprogrammebyaselectionprocedurethatassessedtheorganisationalsupport,commitment for participation, availability of personnel, time to realiseimprovements,andexperiencewithimprovementprojects.Eachgroupofeighthospitals joined theprogramme for 2 years (2004 to2006, 2005 to 2007, or2006 to 2008) and participated in several QICs on different topics (e.g.,pressureulcers,processredesign)12.The process redesign collaborative evaluated in this study represented thethird group of eight hospitals. The overall aim of this collaborative was toreducethetimebetweenthefirstvisittotheoutpatients’clinicandthestartoftreatmentand/ortoreducethe lengthof in‐hospitalstayby30%forselected

Reducingwaitingtimesforpatientswithinthewaitingroomofmedicalspecialists1. Aim:Tostart90%oftheplannedpatientconsultationswithmedicalspecialistsat

thetimetheyareplannedforduringeachday.2. Measureofimprovement: NumberofconsultationsstartedontimeatdayX ‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐ TotalnumberofconsultationsatdayX3. Changeidea:Medicalspecialistshandintheirpagersduringconsultationhoursto

reducedisruptionsduringconsultationwithpatients.4. Seriesofparalleltestingplan­do­study­act(PDSA)cycles:Medicalspecialists

testduringonedaywhetherthehandinginoftheirpagersreducesdisruptionsoftheirconsultations,andthuspreventsdelaysoftheirconsultationswithpatients.BasedontheresultsofafirstPDSA‐cycle(testingwhether‘handinginpagers’worksornot),aprojectteamcandecidetorefinethechangeidea(if‘handinginpagers’inthiscontextseemstowork)ortoreplacethechangeideabyanotherone(if‘handinginpagers’didnotincreasethenumberofconsultationsthatstartedontimeandifitisunlikelythatiteverwill).

thesis_Vos_17x24_v01.pdf

116│ Towardsprocess­orientedcaredeliveryinhospitals

patientgroups13.Eighteenprojectteamsfromtheeightparticipatinghospitalsjoined this collaborative, which started in October 2006. Seventeen of theseteams agreed to participate in our independent evaluation. The enrolment ofprojectteamswithintheevaluatedQICdifferedperhospital.Projectteamstookpartontheirowninitiativeorwereenrolledbythehospitalboard,butalwaysinagreementwiththeexternalchangeagent.6.2.1 ProcessredesigncollaborativeThe evaluated collaborative used a step‐by‐step guide, which included themodelforimprovement(seeFigure6.3).Thisstep‐by‐stepguidewasprovidedbytheexternalchangeagent.Nexttothis,theexternalchangeagentorganisedfive collaborative meetings to inform teams about the step‐by‐step guide aswellasaboutchangesthathaveworkedatothersites.Thepresentedevidencefor improvement focused mainly on the introduction of a one‐stop‐shop, inwhich various visits per patient (diagnostic examinations, consultations, andpreoperativescreening)areplannedforasingleday,withtheaimofreducingthe throughput time of the diagnostic trajectory. Examples of other processredesign change ideas that were provided are: the standardisation of careprocesses in order to reduce variation, the reduction of the number ofunnecessarystepsincareprocesses(donotprovidecareforwhichthereisnoevidence of efficacy), the reduction of the number of planning moments orhandoversinacareprocesssothatfewerhealthcareworkersareinvolvedinthe process, and that each worker is involved only once per iteration of aprocess. The change agent alsoprovided awebsite enablingproject teams toshareinformation.AlthoughitisrecommendedforQICstotestabigchangeideainoneseriesoftesting cycles8, the external change agent advised splitting up every plannedchange into smaller ones that could be tested instantaneously in a series oftestingcyclesbasedontheexperiencesofothercollaboratives(Figure6.1,partB).Bydoingso,theexternalchangeagenttriedtoensurethatteamsspenttheirinitial resources on testing changes instead of dealing with barriers andresistancetochange.

thesis_Vos_17x24_v01.pdf

Figure6.3

Step‐by‐stepguideusedintheprocessredesigncollaborativeincludingthemodelforimprovem

ent

thesis_Vos_17x24_v01.pdf

Chapter6 │117

118│ Towardsprocess­orientedcaredeliveryinhospitals

6.2.2 DatacollectionTo explore the applicability of the QIC method, we evaluated the processredesigncollaborative inamultiple casestudydesign14usingcomplementaryqualitativedatacollectionmethods.Weanalysedtheprocessredesignteameducationmanualtolearnmoreaboutthe provided change ideas and step‐by‐step guide. Further,we held a surveyamong hospital staffmemberswho took part in the project implementations(projectstaffmembers)(n=17)andamongprojectleaders(n=17)togatherdataonprojectcharacteristicsandaims,compositionoftheprojectteams,andproject plans (including (planned) changes, project progress, and theapplication of the model for improvement). The surveys also includedquestions about team organisation (including a clear task division, selfresponsibility for progress, good compliance to arrangements, goodcommunication and coordination, be in charge of implementation),organisational support (including support of strategic management,organisational willingness to change) and external change agent support(including sufficient support and supply of instruments, transfer of valuableinsights),because it isknown from literature that thesearepreconditions forsuccessfuluseof theQICmethod12, 15, 16. In thesurveyamongproject leaders,we included a validated questionnaire to assess these three preconditions15.Projectstaffmemberswereaskedtoratetheamountoforganisationalsupportandexternalchangeagentsupportonascaleof0to10.Questionnairesweresent to respondents 1 year after the start of the collaborative (September2007),andsixteenprojectstaffmembers(response=94%)andelevenprojectleaders(response=65%)completedandreturnedthem.Wealsointerviewedallprojectstaffmembers(n=17)aftertheyreturnedthequestionnaire betweenOctober andDecember 2007. Interview themeswere:change agent support (provided best practices, change concepts, and qualityimprovementmethods), shared experiences between teams, and applicabilityofthemodelforimprovement.In addition, we observed the guidance and training offered by the externalchange agent during meetings and training sessions of the process redesigncollaborative.Theobservationsprovideduswithcontextfortheanalysisofthequestionnairesandinterviews.Finally, we analysed the results reported on the outcome and intermediatemeasures set by the external change agent, who collected these results in a

thesis_Vos_17x24_v01.pdf

Chapter6 │119

‘Sneller Beter’ database and, at our request, provided us with these data(December2007).Allgatheredinformationwasusedtodescribethecollaborativeprocessandtoassess the applicability of the QIC method to process redesign. Additionalinformation about the preconditionswas gathered to evaluatewhether thesecouldhaveinfluencedtheresults.6.3 Results6.3.1 Characteristics of the process redesign projects within the

collaborativeTable 6.1 gives an overview of the characteristics of the process redesignprojects.Fifteenprojectteamschosetoredesignanelectivecareprocess.Eightofthoseprojectsinvolvedcareforcancerpatients.Twoprojectteamschosetoredesignanacutecareprocess.Allprojectteamsintendedtomakeimprovementsinwaitingtimesanddelays,butindifferentareas(accesstimes,throughputtimesofdiagnostictrajectories,and/or length of stay) and for different types of patient groups. Themedianvalue of the volume of the involved patient groups was 150 patients a year(range 17 to 1,000). The number of medical departments involved in theredesignedcareprocesswasonaveragethreeandvariedperprojectfromonetoeightdepartments.Inseveninstances,notallmedicaldepartmentsinvolvedparticipatedintheprojectteam.

thesis_Vos_17x24_v01.pdf

Table6.1

Characteristicsofenrolledprocessredesignprojects

Processtoberedesigned

1,2

No.

PatientGroup

Volumeofpatient

group1(patients/yr)

Acute(A)

orelective

(E)1,2

Accessto

care

Diagnostic

trajectory

(out­patients

clinic)

In­hospital

stay

Involvedmedicaldepartments

(description*,n**)1,2,3

1.

Acutestom

ach

complaints

200

A‐

‐+

Internalmedicine;Radiology;

Pathology

3(2)

2.

Breastcancer

120

E‐

+‐

Oncology;Surgery;Radiology

3(2)

3.

Breastcancer

250

E+

+‐

Oncology;Surgery;Radiology

3(?)

4.

ChronicO

bstructive

PulmonaryDisease

?E

‐+

+Lungdiseases

1(1)

5.

Coloncancer

110

E+

++

Gastroenterology;Surgery;

Oncology;Anaesthesiology;

Radiology

5(4)

6.

Coloncancer

80

E+

+‐

Gastroenterology;Surgery;

Radiology;Pathology

4(?)

7.

Coloncancer

150

E‐

+‐

Gastroenterology;Surgery;

Radiology;Anaesthesiology;

Oncology

5(2)

8.

Head‐andneckcancer

650

E+

++

Ear,NoseandThroat;Radiology;

Jawsurgery;Radiotherapy;

Oncology;Pathology;

Anaesthesiology;PlasticSurgery

8(5)

thesis_Vos_17x24_v01.pdf

120│ Towardsprocess­orientedcaredeliveryinhospitals

Processtoberedesigned

1,2

No.

PatientGroup

Volumeofpatient

group1(patients/yr)

Acute(A)

orelective

(E)1,2

Accessto

care

Diagnostic

trajectory

(out­patients

clinic)

In­hospital

stay

Involvedmedicaldepartments

(description*,n**)1,2,3

9.

Hem

aturia

130

E+

++

Urology;Radiology

2(2)

10.

Lungcancer

400

E‐

+‐

Lungdiseases;Radiology;

Surgery;Pathology;

Anaesthesiology

5(1)

11.

Oesophagealatresia

(children)

17

A‐

‐+

PaediatricSurgery;Intensivist;

Radiology

3(2)

12.

OpenChestSurgery

1,000

E+

‐+

ThoraxSurgery;

Anaesthesiology

2(2)

13.

SmallOrthopaedic

interventions

250

E+

+‐

Orthopaedics;Radiology

2(2)

14.

SmallOrthopaedic

interventions

>200

E+

+‐

Orthopaedics;Anaesthesiology

2(1)

15.

BenignProstate

Hypertrophy

100

E‐

++

Urology

1(1)

16.

Coloncancer

100

E+

++

Surgery;Gastroenterology;

Radiology;Oncology

4(1)

17.

Varicoseveins

150

E+

+‐

Surgery;Dermatology

2(2)

+Yes,‐N

o;*inbold:medicaldepartmentsthatarerepresentedbyamedicalspecialistintheprojectteam;**numberofmedicaldepartmentsinvolved

(num

berofm

edicaldepartmentsrepresentedinprojectteam

).1 Datasource:interview

samongprojectstaffmem

bers.2Datasource:surveyam

ongproject

staffm

embers.3Datasource:surveyam

ongprojectleaders.

thesis_Vos_17x24_v01.pdf

Chapter6 │121

122│ Towardsprocess­orientedcaredeliveryinhospitals

6.3.2 PresenceofpreconditionsforsuccessfuluseoftheQICmethodTheproject leadersandprojectstaffmembersofsixprojectteamssharedtheopinion that preconditions for successful use of the QIC method ‐ i.e., ‘teamorganisation’, ‘organisational support’, and ‘external change agent support’ ‐were sufficiently present (project no. 1, 4, 6, 10, 16, and 17). The remainingproject teamsshowadiversepictureof thepresenceof thepreconditions. Ingeneral,almostallprojectteamswerepositiveabouttheorganisationoftheirprojectteam.One‐halfoftheprojectteamshadtheopinionthatsupportfromtheirorganisationand/orexternalchangeagentsupportwaslacking.6.3.3 EvaluationofthecollaborativeprocessThissectiondescribes thecollaborativeprocessaccording to thestep‐by‐stepguideprovidedtotheprocessredesigncollaborative(seeFigure6.3).Step1:ProcessmappingofthepresentcareprocessAllprojectsstartedwithaprocessanalysisoftheexistingcareprocess.Sixteenoftheseventeenprojectsperformedabaselinemeasurement.Step2:SettinggoalsfortheredesignThebaselinemeasurementandideasaboutthedesiredcareprocessformedtheinput for the project aims and changes that needed to be implemented.Althoughallprojectteamsformulatedprojectaims,onlyfourteenformulatedatleastonespecificandmeasurableaim(range0to7,average2)(seeTable6.2).Step3:SelectingmeasuresforimprovementAftersettingaims,thenextstepwastoestablishmeasuresthatwouldindicatewhetherachangeledtoanimprovement.Withoneexception,allprojectteamsmade use of one or more of the outcome measures provided for the effectmeasurement.Theprovidedintermediatemeasurewasusedbyelevenprojectteams(Table6.2).Forthreeteams,thismeasure(numberofvisitstooutpatientclinic)wasnotapplicablebecausetheseprojectsinvolvedonlytheredesignofin‐hospitalstay.Fortwoprojectteams,theprovidedintermediatemeasurewasnot applicable because it was not related to the project aims: namely, theprojectdidnotstrivetoreducethenumberofvisits.Eight project teams established additional outcome measures: for example,timebetweenseveraldiagnosticexaminationswithinthediagnostictrajectory.Sixprojectteamsappointedintermediateand/orprocessmeasurestoestablish

thesis_Vos_17x24_v01.pdf

Chapter6 │123

whether a process change was accomplished, for instance: Is the date ofsurgery planned directly after setting the diagnosis, yes or no? Five projectsused no additional intermediate or process measure at all. Reasons for notusing project‐specific measures were that teams thought the providedmeasuresgaveenoughinsighttoknowwhetherachangeisanimprovementorbecause their project aims were not consideredmeasurable (e.g., qualitativeaims such as a standardised discharge planning, or appointing one contactpersonforthepatientduringthewholecareprocess).Step4:SelectingchangeideasThe main change idea, the one‐stop‐shop, presented in the collaborativemeetingswasapplicableforelevenprojectteams(Table6.2).Twoofthemdidnot succeed in combining the visits in one day due to organisationalcharacteristics,thenatureoftheneededdiagnostics,and/ortheburdenofthediagnostics to the patients. Six project teams thought the evidence was notapplicablebecausetheyalreadycombinedallvisitsinthediagnostictrajectoryintoone;theydidnotredesignadiagnostictrajectoryattheoutpatients’clinic;orthelongthroughputtimewasnotaresultofmanyvisitsbutofalongwaitinglist for one specific diagnostic examination. All project teams applied one ormoreof theotherprovidedchangeconcepts toredesign theircareprocesses.Applicationofthesechangeideasrequiredthatprojectteamsfirstinvestigatedthe causes of waiting times and delays in the redesigned process and thentailored the change ideas to their own setting. However, according to theproject staff, tailoringchange ideasprovedmoredifficult in careprocesses inwhich more medical departments were involved, and accordingly moredisagreement existed between the involved medical departments about thechangesthathadtobemade.Steps5and6:TestingandimplementingchangesDuring the interviews, project staff members were asked whether they hadappliedthePDSAcycleforchange.FiveconfirmedthattheirprojectteamusedorwasgoingtousethePDSAcycle.However,thesefiveprojectteamsdidnotsplit up every planned change in smaller changes as the change agentsuggested.Further,staffmembersofthesefiveprojectteamsindicatedthatthePDSAcyclewasnotorwouldnotbeperformedinarapidcyclicalmodebecauseboththepreparationforthetestaswellasthetestofthechangeitselfwastimeconsuming. Because the patient groupswere relatively small, a testing cycletook considerable time evenwhen the number of patients per testing periodwasscaleddown.TheuseofthePDSAcyclewasalsohamperedbythefactthathospitalinformationsystemsprovedunabletogeneratedataontheappointed

thesis_Vos_17x24_v01.pdf

124│ Towardsprocess­orientedcaredeliveryinhospitals

measureswhenmorehospitaldepartmentswere involved.Asaconsequence,projectteamshadtogatherdatabyhand,whichwastimeconsuming.TheteamsthatdidnotuseorwerenotgoingtousePDSAforimplementation(n = 10) chose to change the organisation of the care process radically byimplementing their ‘newlydesignedprocess’ at oncewithout first testing theindividual changes. According to these project teams, testing change ideaswithin a short timeframewasnot applicable to their situation becauseof thenumberofmedicaldepartmentsinvolvedand/orthesmallnumberofpatientsinvolved in their redesign.Another reason fornot testing in rapidcycleswasthefeelingthatatestcouldfailduetonon‐optimalconditionswhensupportingprocesseswerenotoptimised.Forexample,theteamimplementingchangesinthecareforopenchestsurgerypatientsconsidereditimpossibletotestanewoperating room planning process. Changing the planning system for theoperating room would necessitate adjusting all the supporting processes,includingtheworkinghoursof the teamsandhowtheroomswereprepared.Any testing before the altering of supporting processes would be massivelydisruptive.Step7:EffectmeasurementThree project teams performed an effect measurement and reachedcollaborative goals (Table 6.2). The other project teams, including those thatused thePDSA cycle, hadnot yetmeasured any interim results byDecember2007(1yearafterthestartoftheQIC).Thereforeitisunknownwhethertheyreachedthecollaborativegoals.

thesis_Vos_17x24_v01.pdf

Table6.2

Applicationofthemodelforimprovem

entintheenrolledprocessredesignprojects

Measuresofim

provement1

Keychanges1

,2,3

Keyelements

ofthemodel

for

improvement

Providedby

externalchange

agent

Establishedbythe

projectteam

Evidencefor

improvement

(one­stop­shop)in

redesign?

Supplied

change

concepts

used?

Effect

measurement

(collaborative

goalsreached?)?4

No.PatientGroup

Specificand

measurable

aims(n)1

Outcom

eInter­

mediateOutcom

eProcess

and/or

inter­

mediate

Yes/

No

Comments

PDSA

1,3

1.

Acutestom

ach

complaints

+(1)

+n.a.

++

‐n.a.

+‐

‐(?)

2.

Breastcancer

‐(0)

++

‐‐

‐Already

implem

ented

+‐

‐(?)

3.

Breastcancer

+(1)

‐‐

‐‐

‐Already

implem

ented

+.

‐(?)

4.

Chronic

Obstructive

Pulmonary

Disease

+(1)

+‐

‐+

‐One‐stop‐shopis

nosolutionforthe

existing

bottleneck

++

‐(?)

5.

Coloncancer

+(4)

++

+‐

+‐

+‐

‐(?)

6.

Coloncancer

+(1)

++

‐‐

+‐

+‐

‐(?)

7.

Coloncancer

‐(0)

++

+‐

+‐

++/‐*

‐(?)

8.

Head‐and

neckcancer

+(7)

++

++

+‐

+‐

‐(?)

9.

Hem

aturia

+(2)

++

‐‐

+‐

+.

‐(?)

thesis_Vos_17x24_v01.pdf

Chapter6 │125

Measuresofim

provement1

Keychanges1

,2,3

Keyelements

ofthemodel

for

improvement

Providedby

externalchange

agent

Establishedbythe

projectteam

Evidencefor

improvement

(one­stop­shop)in

redesign?

Supplied

change

concepts

used?

Effect

measurement

(collaborative

goalsreached?)?4

No.PatientGroup

Specificand

measurable

aims(n)1

Outcom

eInter­

mediateOutcom

eProcess

and/or

inter­

mediate

Yes/

No

Comments

PDSA

1,3

10.Lungcancer

+(2)

+‐

+‐

+/‐

Three‐stop‐shop

++

‐(?)

11.Oesophageal

atresia(children)

‐(0)

+n.a.

‐‐

‐n.a.

+‐

‐(?)

12.OpenChest

Surgery

+(6)

+n.a.

++

‐n.a.

+‐

‐(?)

13.Sm

all

Orthopaedic

interventions

+(2)

++

++

+‐

++

‐(?)

14.Sm

all

Orthopaedic

interventions

+(3)

++

‐+

+‐

++

‐(?)

15.BenignProstate

Hypertrophy

+(2)

++

‐‐

+‐

+‐

+(+)

16.Coloncancer

+(5)

++

+‐

+/‐

Three‐stop‐shop

+‐

+(+)

17.Varicoseveins

+(5)

++

‐‐

+‐

+‐

+(+)

1 Datasource:surveyam

ongprojectstaffmem

bers.2Datasource:surveyam

ongprojectleaders.3Datasource:interview

sam

ongprojectstaffmem

bers.4

Datasource:SnellerB

eterdatabase.+Yes,‐No,.missingdata,n.a.nonapplicable,becauseprojectonlyinvolvesin‐hospitalcare.*Thisprojectteamused

PDSAfortestingandimplem

entingaselectionofthechanges.

thesis_Vos_17x24_v01.pdf

126│ Towardsprocess­orientedcaredeliveryinhospitals

Chapter6 │127

From this description of the collaborative process we can identify severaldifficulties experienced by the project teams in applying the QIC method toprocess redesign. First, the adoption of change ideas and the accompanyingmeasures provided by the external change agent, appeared not (directly)applicable for these collaborative project teams. Project teams had to tailorchangeideastotheirowncontextorcouldnotusetheprovidedchangeideasatall.Second, theadoptionof themodel for improvementbytheproject teamswashampered. Project teams were not capable of testing change ideas within ashorttimeframeusingPDSAcyclesdueto:theneedfortailoringchangeideasto their own context, and the complexity of aligning several interests ofinvolved medical departments; the small volumes of the involved patientgroups;andhospitalinformationsystemsthatprovedunabletogeneratedataontheappointedmeasures.Third,projectteamsdidnotexperiencepeerstimulus.Allcollaborativeprojectteams intended to make improvements on an administrative subject, but indifferentpartsofcareprocesses(accesstimes,throughputtimesofdiagnostictrajectories, and/or length of stay) for different types of patient groups. As aconsequence,projectteamssawfewsimilaritiesbetweentheirprojects,rarelysharedexperiences,anddemonstratednocompetitivebehaviour.Further,anumberofprojectteamsperceiveda lackoforganisationalsupportandexternalchangeagentsupport.However,theprojectteamsthatsucceededin implementing changes (projects 15, 16, and 17) shared the opinion thatpreconditions for successful use of the QICmethod ‐ i.e., ‘team organisation’,‘organisationalsupport’,and‘externalchangeagentsupport’‐wereingeneralsufficiently present. Only organisational support lacked in one of the threeprojectteams(project15).6.4 DiscussionFrom the results it seems that in the evaluated collaborative theQICmethodwasnotused.Apparently,itdidnotempowerprojectteamstoimplementtheirprocessredesigninashorttimeframe.Asaconsequence,thisstudycouldnotshowwhethertheQICmethodcaneffectivelycontributetoprocessredesign,ifused. The description of the collaborative process provides us with valuableinformationaboutthedifficultiesexperiencedbytheprojectteamsinapplyingthe QIC method to process redesign. In this section, we will discuss

thesis_Vos_17x24_v01.pdf

128│ Towardsprocess­orientedcaredeliveryinhospitals

explanations for thesedifficulties,which concentrateon a lackof fit betweenthe QIC method and process redesign, a non‐optimal application of the QICmethod,andnon‐optimalconditionsforusingtheQICmethod.6.4.1 Non­optimalfitbetweentheQICmethodandprocessredesignFirst,a lotoftheprojectteamsneededcustomisedsolutions fortheirprocessredesign, while the QICmethod aims to spread standardised evidence‐basedpracticesorchangeideastoservemanyteamsatthesametimewithalimitednumberofexternalchangeagents.AccordingtotheQICmethodcollaborativeprojectteamsshouldbenefitoftheexchangeofthestandardisedchangeideasin such a way that they can eliminate much of the investigative work onproblem analysis and change ideas in comparison with traditional qualityproject teams3. For example, in aQIC for pressureulcers, an external changeagent can provide concrete best practices from pressure ulcer guidelines toperfecttheelementsofcare,suchas‘minimiseskinpressurethroughtheuseofa positioning schedule for clients with an identified risk for pressure ulcerdevelopment’. This best practice can then be tested and, if it works, beimplemented directly in every setting. Process redesign, however, calls forcustomised solutions because project teams need to handle context‐specificcausesofwaitingtimesanddelaysincareprocessesdeterminedbytheexistinginteractionpatternsbetweendepartments intheirhospital.Projectteamscantherefore not test the standard change ideas provided by the change agentwithina short time framebuthave to investigate thecausesofwaiting timesanddelaysand to tailor change ideas to theirownsetting.As a consequence,thecollaborativecannoteliminatethe investigativeworkonproblemanalysisandprofitfromstandardchangeideasprovidedbytheexternalchangeagentastheQICmethodprescribes.Second, the model for improvement, and especially the PDSA cycle, seemedinappropriate to test intended changes within a short timeframe. The QICmethod assumes that testing one big change idea lowers the resistance to achange because clinicians aremore likely to be reassured that the change iseffective8, 17. This assumption ignores the fact that testing changes that affectseveraldepartmentsmayleadtomoreconsultationbeforetestingachangeandthustoanincreasedpossibilityofresistancetoachange.Thishappenedinthehospitals involved as result of their functional structure, in which everydepartment has its own responsibilities and tries to optimise its ownfunctioning. These functional boundaries hampered, for example, theadjustment of the department schedules needed to realise a ‘one‐stop‐shop’.

thesis_Vos_17x24_v01.pdf

Chapter6 │129

After all,more relationships are affected, andmore different interests play arole.Asaresult,projectteamscouldonlystarttestingafterabuy‐inorpoliticalsolution. In this study, the complexity of aligning department schedules andinterestsbecamemoreapparentwhenthenumberofdepartmentsinvolvedina care process increased. The project teams might have improved thecollaboration across boundaries if they had included in their team amedicalspecialistfromallmedicaldepartment(s)involved.However,theneedforbuy‐in solutions before testing a change could also be due to the fact that theexternal changeagentadvisedsplittingupeveryplannedchange intosmallerchanges.Althoughsmallerchangescanreducetheriskoffailure,italsolowerstheexpectationsofthebenefitsofachange.Unclearorsmallerbenefitsdonotstimulatemedicaldepartmentstoinvestinmakingchanges.Difficulties in using the PDSA cycle meant that most teams decided toimplement changes without testing them. Subsequently, teams did not getfeedbackontheworktheyweredoinganddidnotexperienceamomentumofchange18. It is known from previous studies that consistent ongoingmeasurementisrequiredtotellwhetherchangesbeingmadeareleadingtoanimprovement, and to provide basis for continued action19, 20. Because of thislackof feedback, teamswerenotstimulated toadaptanotherchange idea forimprovement,whichinturnsloweddowntheimplementationofchanges.AlthoughthedifficultieswiththeuseofthePDSAcycleare(almost)inevitableinprocessredesignprojects infunctionallyorganisedhospitals, theuseofthePDSAcouldbe improvedby takingcareof somepreconditions.First,hospitalinformation systems should be able to generate data on the appointedmeasures. Second, the number of patients involved in the care process thatneed to be redesigned has to be big enough to test a change idea within anumberofdays.6.4.2 Non­optimalapplicationoftheQICmethodNext to the non‐optimal fit between the QIC method and process redesign,difficultiescanalsobeduetotheselectionprocessofthecollaborativeprojectteams. The external change agent includedproject teams in the collaborativethat worked on different parts of care processes (access times, throughputtimes of diagnostic trajectories, and/or length of stay) for different types ofpatient groups, while the QIC method aims to implement evidence‐basedpracticethroughsharingknowledgewithothersinasimilarsetting4.Probably,theexternalchangeagentcouldhaveprovidedpeerstimulusif ithadselected

thesis_Vos_17x24_v01.pdf

130│ Towardsprocess­orientedcaredeliveryinhospitals

project teams that worked on comparable process redesign projects withcomparablegoals.Nevertheless, lackofpeerstimuluscanalso occurbetweencomparable redesign projects because of the existence of context‐specificcausesofdelaysandwaitingtimes.6.4.3Non­optimalconditionsforusingtheQICmethodNext to hospital information systems to generate data on outcome,intermediate and processmeasures, complex process redesign projects needsupport to change interaction patterns between involved departments. Anumberofprojectteamsperceivedalackoforganisationalandexternalchangeagentsupport,despitethefactsthatallprojectteamsreceivedexternalchangeagent support and the participating hospitals were enrolled in the ‘SnellerBeter’ programme by a selection procedure that assessed the organisationalsupport. Unfortunately,we could not identify factors that contributed to thisperceivedlackoforganisationalandexternalchangeagentsupport.6.4.4 LimitationsThis study aimed to assess the applicability of the QIC method for processredesign. Although we think the findings of this study provide usefulinformation for future collaboratives, the results need to be interpretedwithcaution. The findingsof this evaluation couldbe influencednegatively by theselectionprocessofboththecollaborativeprojectteamsandthecareprocessestoberedesigned.Forinstance,notallteamsparticipatedinthecollaborativeona voluntary basis. Unfortunately, we could not determine with certainty towhich project teams this applied and how this influenced the collaborativeprocess.Another limitation is that the gathered data are not complete. However,observations during meetings and training sessions of the process redesigncollaborative showed us that themissing data of project leaders and projectstaff members are not related to poor performing project teams and/ororganisational support. The poor availability of effect measurements oncollaborativegoalscanbecontributedtothefactthatitisnotfeasibleformanyproject teams to redesign, implement, and perform an effect measurementwithinayear,andtothenon‐optimalfitbetweentheprinciplesoftheusedQICmethodandprocessredesign.

thesis_Vos_17x24_v01.pdf

Chapter6 │131

6.5 ConclusionThis study showed that the need for tailoring standard change ideas to thecontext of collaborativeproject teams, and the complexity of aligning severalinterests of involved medical departments, hampered the use of the QICmethodforprocessredesign.WecannotdeterminewhethertheQICmethodisappropriate for process redesign. As result of the selection process forparticipationofprojectteamsbytheexternalchangeagentpeerstimuluswasnon‐optimal.Furtherprojectteamsfeltthatpreconditionsforsuccessfuluseofthe QIC method were lacking. Therefore, additional research into theapplicabilityoftheQICmethodforprocessredesignisneeded.

thesis_Vos_17x24_v01.pdf

132│ Towardsprocess­orientedcaredeliveryinhospitals

References1. Baker GR: Collaborating for improvement: The institute for healthcare

improvement'sbreakthroughseries.NewMedicine1997,1:5‐8.2. Lindenauer PK: Effects of quality improvement collaboratives. BMJ 2008,

336(7659):1448‐1449.3. Øvretveit J, Bate P, Cleary P, Cretin S, Gustafson D, McInnes K, McLeod H,

MolfenterT,PlsekP,RobertGetal:Qualitycollaboratives:lessonsfromresearch.QualityandSafetyinHealthCare2002,11(4):345‐351.

4. NewtonPJ,HalcombEJ,DavidsonPM,DennissAR:Barriersandfacilitatorstotheimplementation of the collaborative method: reflections from a single site.QualityandSafetyinHealthCare2007,16(6):409‐414.

5. SchoutenLM,HulscherME,VanEverdingenJJ,HuijsmanR,GrolRP:Evidencefortheimpactofqualityimprovementcollaboratives:systematicreview.BMJ2008,336(7659):1491‐1494.

6. The breakthrough series: IHI’s collaborativemodel for achieving breakthroughimprovement.IHIinnovationserieswhitepaper.Boston:InstituteforHealthcareImprovement;2003(Availableonwww.IHI.org).

7. Langley GJ, Nolan KM, Nolan TW, Norman CL, Provost LP: The improvementguide: a practical approach to enhancing organizational performance. SanFrancisco:Jossey‐Bass;1996.

8. Nolan TW, Schall MW, Berwick DM, Roessner J: Reducing delays and waitingtimes throughout the healthcare system. Boston: Institute for HealthcareImprovement;1996.

9. MittmanBS:Creatingtheevidencebaseforqualityimprovementcollaboratives.AnnalsofInternalMedicine2004,140:897‐901.

10. KerrD,BevanH,GowlandB,Penny J,BerwickD:Redesigningcancercare.BMJ2002,324(7330):164‐166.

11. Loco*ck L: Health care redesign: meaning, origins and application. Quality andSafetyinHealthCare2003,12(1):53‐57.

12. DückersMLA,SpreeuwenbergP,WagnerC,GroenewegenPP:Exploringtheblackbox of quality improvement collaboratives: modelling relations betweenconditions, applied changes and outcomes. Implementation Science 2009,4(1):74.

13. Rouppe van der Voort M, Stoffer M, Zuiderent‐Jerak T, Janssen S, Berg M:Breakthrough process redesign III: 2006‐2007 Better Faster pillar 3, T2S2 enT3S1 (in Dutch). Utrecht/Rotterdam/Utrecht: Quality Institute for Health CareCBO, Institute of Health Policy and Management of the Erasmus UniversityRotterdam,OrderofMedicalSpecialists;2006.

14. YinRK:Casestudyresearch:designandmethods.3rdedition.ThousandOaks,CA(etc.):Sage;2003.

thesis_Vos_17x24_v01.pdf

Chapter6 │133

15. DückersML,WagnerC,GroenewegenPP:Developingandtestinganinstrumenttomeasure thepresenceof conditions for successful implementationofqualityimprovementcollaboratives.BMCHealthServicesResearch2008,8:172.

16. Dückers MLA, Wagner C, Groenewegen PP: Conditions for a sector wide,knowledge based, improvement programme in the Dutch hospital care (inDutch).ActaHospitalia2005,45(3):37‐54.

17. Resar R: Why we need to learn standardisation. Australian Family Physician2005,34(1‐2):67‐68.

18. Plsek PE: Quality improvement methods in clinical medicine. Pediatrics 1999,103(1SupplE):203‐214.

19. Berwick DM: Developing and testing changes in delivery of care. Annals ofInternalMedicine1998,128(8):651‐656.

20. LeapeLL,RogersG,HannaD,GriswoldP,FedericoF,FennCA,BatesDW,KirleL,Clarridge BR: Developing and implementing new safe practices: voluntaryadoption through statewide collaboratives. Quality and Safety in Health Care2006,15(4):289‐295.

21. Gaucher EJ, Coffey RJ: Breakthrough performance: accelerating thetransformationofhealthcareorganizations.SanFrancisco:Jossey‐Bass;2000.

thesis_Vos_17x24_v01.pdf

thesis_Vos_17x24_v01.pdf

│135

Chapter7

Generaldiscussionandconclusion

thesis_Vos_17x24_v01.pdf

thesis_Vos_17x24_v01.pdf

Chapter7 │137

7.1 IntroductionTraditionally, hospitals have a functional organisation structure. Within thisstructure, organisational departments are built around the skills andknowledgeoftheirprofessionals,likeinternalmedicine,radiology,andsoon1‐3.Thesedepartmentsfrequentlyoperateasseparate‘silos’withtheirowngoals,anddonotexchange information for theplanningandcontrolofpatient flow(functional operating control system)1, 2.Asa consequence, each care activitywithinachainofcareactivitiesforthediagnosisandtreatmentofapatientwillbe planned on the basis of availability of resources, after the previous careactivityiscompleted.Thishasledtoqualityproblemslikewaitingtimes,poorcoordinationofcare,andawasteofresources.National and international attention to these quality problems has madehospital management and medical specialists aware that the way care isdelivered needs to be reorganised. However, little is known about how toimprovetheorganisationofcaredeliveryintermsofqualityandefficiency.Inaddition, it isnotknownhownewwaysoforganisingcaredeliveryshouldbeintroduced into a hospital organisation. Until now, many of the ideas forimprovements of coordination and process control have been derived fromresearch in industry, including thesuccessfulconceptreferred toas ‘businessprocessorientation’.Thisconcept’sbreakthroughoccurredatthebeginningofthe 1990s under the name ‘Business Process Reengineering’4. Successfulexamplesoftheapplicationofbusinessprocessorientationfortheorganisationof care delivery to specific patient groups are known, for instance, from theDutch programme ‘Better Faster’5, from Australia’s ‘Redesigning HospitalCare’6, and from the United States’ ‘Reducing delays and waiting timesthroughout the system’7. It is unknown, however, whether business processorientationorprocess‐orientedcaredeliverycanalsobeappliedsuccessfullyathospitallevel.Therefore, this thesis aimed to test thehypothesis that the implementationofprocess­oriented caredelivery leads tobetteroutcomes in termsofqualityandefficiencyathospitallevel.Inordertotestthecentralhypothesisofthisthesis,theoverallresearchquestionis:‘Is the implementation of a process­oriented logistical concept effective forimprovingqualityandefficiencyofcaredeliveryathospitallevel?’In thestudiesdescribed inthepreviouschapters, the followingsub‐questionswereaddressed:

thesis_Vos_17x24_v01.pdf

138│ Towardsprocess­orientedcaredeliveryinhospitals

1. ‘Whichprocess­orientedlogisticalconceptsdohospitalsimplementtodeliverprocess­oriented care,andhoweffectiveare these in improving thequalityandefficiencyofcaredelivery?’

2. ‘Doesevaluationoftheflexibilityofhospitalbuildinglayoutcontributetothe

implementationofanewlogisticalconcept?’3. ‘Doprocess­basedpaymentscontribute to the implementationofaprocess­

orientedlogisticalconcept?’4. ‘Do improvementand innovationmethodscontributetothe implementation

ofaprocess­orientedlogisticalconcept?’The first sub‐question assessed which process‐oriented logistical conceptshospitals implement todeliverprocess‐orientedcare,andtheeffectivenessoftheselogisticalconceptsinimprovingthequalityandefficiencyofcaredelivery.Sub‐questions 2‐4 examined ways to introduce process‐oriented logisticalconceptseffectivelyintoahospitalorganisation.In this chapter, themain findings on these sub‐questions are presented first.Next, the meaning of these findings with regard to the general researchquestionisdiscussed.Subsequently,thegeneralresearchquestionisanswered.The chapter ends with recommendations for future research and for theimplementationofprocess‐orientedorganisationofcaredelivery.7.2 Mainfindings1. ‘Whichprocess­orientedlogisticalconceptsdohospitalsimplementtodeliver

process­oriented care,andhoweffectiveare these in improving thequalityandefficiencyofcaredelivery?’

The literature review inChapter2 reportedanddiscussed theexperiencesoffive hospitals with respect to implementing a process‐oriented logisticalconcept at organisational level. This literature review showed that hospitalschose between twomain approaches to improve the integration of functionsaroundcareprocessesandtobecomeprocess‐orientedorganisations:A. The implementation of a process‐oriented operational control system

withoutchangingthehospitals’basicstructure;andB. The implementation of a process‐oriented basic structure, in which the

composition of hospital departments is based on the needs of specificpatientgroupsinsteadofonthetypeofmedicalspecialties.

thesis_Vos_17x24_v01.pdf

Chapter7 │139

On thebasisof the results, it seems thatunderpreconditions theadoptionofeither approach to become process‐oriented can possibly lead to improvedquality and efficiency of care delivery. First, the implementation of eachapproachhasto focuson logisticallyhom*ogeneouspatientgroups.Bottom‐upknowledge on the part ofmedical professionals is essential to identify thesepatient groups. The review showed that, ultimately, hospitals with an initialfocus on overarching processing steps with respect to patients’ medicaldiagnosis and treatmentmainly implemented improvements in departmentalprocesses instead of improvements that made the patient flow smoother.Second, the implementation has to be accompanied by a committed processmanagementtochangepreviouspatternsrelatingtothefunctionaldivisionoftasks.The literature review also provided insight into factors that hampered theimplementationofbothapproaches.Aneffectiveimplementationofaprocess‐oriented operational control system (A) requires that department managersandmedical specialistsbothgivepriority to thecoordination ofcaredeliverybetweendepartments above the capacityutilisationof eachdepartment.Thisproveddifficult,sincemedicaldepartmentsinfunctionalorganisationshavetomaximisetheirownoutput.Aneffectiveimplementationofaprocess‐orientedbasic structure based on multidisciplinary patient‐focused departments (B)requiresthathospitalsfocusonstrategicallyimportantservices.However,thecharacteristicsofhospitalcaremakethisdifficultbecausepoliticalandethicalobligations prevent hospitals from refusing to deliver services to groups ofpatientsfromwithintheirservicearea.Unfortunately, we were not able to judge which of the two approaches(implementation of a process‐oriented operational control system orimplementationofaprocess‐orientedbasicstructure)deliversthebestresultsandunderwhatcirc*mstances.Forsuchanassessment,morespecificstudiesareneeded.2. ‘Doesevaluationoftheflexibilityofhospitalbuildinglayoutcontributetothe

implementationofanewlogisticalconcept?’Chapter 3 described an evaluation method for the assessment of hospitalbuilding layout, and its contribution to the implementation of new logisticalconcepts is tested ina case study.Usingcomputer simulation techniques, thedescribed evaluation method assessed hospital building layout from theviewpoint of operations management. The aim of such an evaluation is toensurethatthebuildinglayoutsupportstheefficientandeffectiveoperatingofcurrentandfuturecareprocesses.Inthecasestudy,anewlydesignedhospital

thesis_Vos_17x24_v01.pdf

140│ Towardsprocess­orientedcaredeliveryinhospitals

layout was assessed on its flexibility to meet future requirements of newlogistical concepts and changes in patient mix. The results showed that achange in intensity, direction, or volume of the flow as a result of a newlogistical concept or a changed patientmix led to congestions in the patientflow.Inotherwords,thedesignofthehospitalbuildinglayoutwasnotflexibleenough to deal with the changes. In this case, however, the layout of thebuildingcouldeasilybeadjustedbecauseitwasstillinthedesignphase.Theevaluationofabuilding layout inthedesignphase isofgreat importancefromtheperspectiveofoperationsmanagement.Inthisphase,itisstillpossibletoadjustthelayouttomakeitmoreflexibletomeetfuturerequirementsofnewlogistical conceptsandchanges inpatientmix,and it thereforecontributes toan effective implementation of future logistical concepts. In addition, theevaluationmethodcanalsobeusedtoassesstheimplicationsofnewlogisticalconcepts in existing hospital buildings. Ex‐ante evaluation of new logisticalconceptscanpreventtheoccurrenceofcongestionsintheoverallpatientflow,whichaffectwaitingandthroughputtimes,andthusthequalityandefficiencyofcaredelivery.We conclude that an evaluationof the flexibility of a hospital building layoutusingsimulationtechniquescontributestoaneffectiveimplementationofnewlogisticalconcepts,andpreventstheoccurrenceofconflicts.3. ‘Doprocess­basedpaymentscontribute to the implementationofaprocess­

orientedlogisticalconcept?’Chapter4testedwhethertheuseofprocess‐basedpayments–process‐basedinternal hospital budgeting – contributes to the implementation of process‐oriented logistical concepts, andmore specifically to the implementationof aprocess‐orientedoperationalcontrolsystem.AnalysesofdatainvolvingDutchhospitals demonstrated that process‐based budgeting – the allocation offinancial means on the basis of the total care of a patient during an acuteepisode of illness – stimulatedmedical professionals to use process‐orientedperformancemeasures,which in turn had a positive effect on the number ofactivities to develop process‐oriented operational control systems. Althoughprocess‐based budgeting did not stimulate directly the implementation of aprocess‐orientedlogisticalconcept,itseemsthatitisavaluabletooltobridgeinternal conflicts of interests concerning departments and medicalprofessionals,whichhampercollaborationacrossdepartmentalboundaries. Itstimulated cooperation between departments and medical professionals toorganise care delivery in a process‐oriented way, even though theimplementationwas not yet accompanied by adapted internal organisational

thesis_Vos_17x24_v01.pdf

Chapter7 │141

activities:forinstance,moredecentraliseddecision‐making,agreementsaboutprocess ownerships, and a process‐oriented view on the part of medicalprofessionals.Theresultsof thisstudyshowthatprocess‐basedpaymentscontribute to theimplementation of process‐oriented logistical concepts, and they also help toalign the interests of departments and medical professionals with topmanagement’swishtoimplementaprocess‐orientedconcept.4. ‘Do improvementand innovationmethodscontributetothe implementation

ofaprocess­orientedlogisticalconcept?’Chapters5and6assessedtheeffectsoftwotypesofredesignmethodsontheimplementation of process‐oriented logistical concepts at process level.Chapter 5 dealt with a process innovation method: that is, one intended toradicallyredesignandimproveworkprocesses.Thismethodisusedtochangethe traditionally functional operating system of in‐hospital care for strokepatients towards a process‐orientedoperating system, the aimofwhich is toimprove the quality and efficiency of care delivery. Themethod consisted offoursteps:processanalysis,identificationofbottlenecks,designoftheprocess‐orientedoperatingsystem,andimplementationofthesystem.Resultsindicatethat implementation of the process‐oriented operating system led to asignificantly shorter length of hospital stay. As a result,more stroke patientscouldbeadmittedtothespecialisedward.Chapter6concernedtheapplicabilityoftheQualityImprovementCollaborative(QIC) method for the redesign of work processes. With the help of externalfacilitators andpeer stimulus, thismethod is intended to changea functionaloperating system incrementally to a process‐oriented operating system forspecific patient groups. The method was assessed in a multiple case studydesign. Within the evaluated QIC, external change agents provided eighteencollaborativeproject teams fromvarious hospitalswith a clear vision for theredesign of their operating system, along with a set of specific changes thatcouldsignificantly improvethepatientflow.Resultsof theevaluationshowedthat–forseveralreasons–thisapplicationoftheQICmethoddidnotsucceedin empoweringproject teams to implement a redesign in a short time frame.First, there seems to be a non‐optimal fit between the QIC method and theredesignofworkprocesses.WhiletheQICmethodaimstoserveasmanyteamsas possible with standardised change ideas, the redesign of work processescalls for customised solutions,becauseproject teamsneed tohandle context‐specificcausesofwaitingtimesanddelaysincareprocessesdeterminedbytheexisting interaction patterns between departments. Second, the QIC method

thesis_Vos_17x24_v01.pdf

142│ Towardsprocess­orientedcaredeliveryinhospitals

wasnotoptimallyapplied.Theexternalchangeagentincludedprojectteamsinthe collaborative that worked on different aspects of care processes (accesstimes, throughput times of diagnostic trajectories, and/or length of stay) forvarious types of patient groups and different patient volumes, while the QICmethodaimstoimplementevidence‐basedpracticethroughsharingknowledgewithotherteamsthathavesimilargoals.Third,projectteamswereconfrontedwith non‐optimal conditions for applying the QIC method: for instance, anumberofprojectteamsperceivedalackoforganisationalandexternalchangeagentsupport.Theseresultsshowthattheneedtotailorstandardchangeideasto the context of collaborative project teams, plus the complexity of aligninginterests of involved departments, hamper the use of the QIC method forprocess redesign. To conclude, due to the non‐optimal application of theQICmethodand thenon‐optimal conditions forusing theQICmethod,we cannotdeterminewhethertheQICmethodisappropriatefortheprocessredesignofcareprocessesinhospitals.FromtheresultsofthetwostudiesdescribedinChapters5and6,itseemsthatprocess innovation methods contributed more than the evaluated processimprovement method (QIC) to the implementation of a process‐orientedlogistical concept. This is probably because the implementation requirescustomised solutions that fit context‐specific causes of waiting times anddelays.While the thoroughprocessanalysisaspartof theprocess innovationmethodresultedintheimplementationofcustomisedinterventionstochangethe pattern of interaction between professionals and to redesign the overallwork flow, the process improvement method tried to implement standardchangeideasthatworkedforothers.Moreover,hospitalsthatusedtheprocessimprovement method frequently focused on parts (e.g. the diagnostic ortreatmenttrajectory)ratherthanonwholecareprocesses.Inthedecisionastowhat form work processes should take, the process innovation methoddescribedgaveinvolvedmedicalprofessionalsmoreautonomythandidtheQICmethod.Therefore,theprocessinnovationmethodusedseemsalsomoreabletodealwithpossibleconflictsbetweentheautonomyofmedicalprofessionalsandtherequirementsofaprocess‐orientedlogisticalconcept.7.3 DiscussionofthefindingsThere are successful examples of the application of the business processorientation concept for the organisation of care delivery to specific patientgroups.Thisthesisexaminedwhetherprocess‐orientationcanalsobeapplied

thesis_Vos_17x24_v01.pdf

Chapter7 │143

successfullyathospitallevel,andexploredwaystointroduceprocess‐orientedcaredelivery.Theresearchreconfirmsthattheimplementationofprocess‐orientedlogisticalconceptscanreducethroughputtimesforspecificpatientgroups(Chapters1,5, and 6), which leads to a gain in quality for patients and to efficiency forspecificcareprocesses.Thesepatientgroupswillnotonlyhaveshorterwaitingtimesbetweenexaminationandtreatmentbuttheirconditionswillbetreatedtimely and adequately, andwith fewer resources (e.g. number of bed days)8.Similarbenefitsofprocess‐orientedlogisticalconceptscouldnotbedeterminedat hospital level. The literature review in Chapter 2 was unable to identifyenough high‐quality studies that assessed the implementation of process‐oriented logistical concepts. Therefore,we intended to test the effects of theimplementation of process‐oriented delivery empirically by evaluating theresults of the national quality improvement programme ‘Better Faster’ withregard to its aim to accelerate the transition towards process‐oriented caredelivery in participating hospitals9. Data on outcomes were provided byprogramme officers of ‘Better Faster’. These data proved to be incomplete,however,anditwasthusnotpossibletoassesseffects.Theincompletedataareprobablydueto implementationfailuresandtothe fact thathospitalsneededmore time to implement process‐oriented logistical concepts. The limitedscientific foundation of underlying assumptions of ‘Better Faster’9 may havecontributed to the implementation failures. Nevertheless, on the basis ofavailabledataandadditional research, this thesiswasable to identify certainsuboptimal applications of process‐orientated logistical concepts andintroductorystrategiesthatmayhinderaneffectiveimplementationofprocess‐orientedcaredelivery.7.3.1 Suboptimalapplicationofprocess­orientedlogisticalconceptsNosystemviewOnthebasisofthisresearch, itseemsthathospitalsdonotacquireasystemsviewwhen they try tobecomeprocess‐orientedorganisations.Mosthospitalswork towards thisgoalbymeansof thegradualorpartial implementationofprocess‐oriented control systems or care programmes (Chapter 2). Theseprogrammeshelphospitalstoovercomethefunctionaldivisionoflabourandtoimprove collaborationbetweendepartments.However, this approach ignorestheimportanceofanoptimalperformanceoftheseparatedepartments,whichincorporates a risk for sub‐optimisation of the entire hospital system. Thispotential sub‐optimising effect is, for example, caused by the fact that when

thesis_Vos_17x24_v01.pdf

144│ Towardsprocess­orientedcaredeliveryinhospitals

introducingcareprogrammes,hospitalsoftenreserveslotsfor specificpatientgroupsintheplanningofscarceresources10,likeCT,MRI,andPETscans,andinthe planning of outpatients’ clinics. Although the use of these ‘priority lanes’reducesthroughputandwaitingtimesforspecificpatientgroups,itmayleadtoan inefficient use of resources and more waiting times and delays for otherpatients:first,becauseahighernumberofwaitinglinesincreasesthetimeandenergyneededtocontrolthem;andsecond,becausethewaitinglinesaremoresensitive to fluctuations in the demand for specific places11. However, aninformationexchangebetweenmedicalprofessionals–onethat isenabledbyprocess‐orientedcontrolmechanisms(careprogrammes)inafunctionalbasicstructure or by the implementation of a process‐oriented basic structure(multidisciplinarydepartments)–couldprovidedepartmentswithinformationaboutwhatcapacitiesareneededatwhattime.Theycouldusethisinformationto optimise the organisation of departments. A precondition then is thathospitals be capable of a twofold focus: one on the optimisation of careprocessesandtheotherontheoptimisationofdepartments12.Weconcludethatsolutionschosenbythehospitalsareatrade‐offbetweentheoptimisation of care processes and that of department efficiency. Hospitalstried to optimise care processes by optimising parts of processes and thecoordinationoftransferpointsinthechainofcare.Thecoordinationoftransferpoints frequentlyrequires thereservationof specificplaces.This in turnsub‐optimised theplanningofdepartments.Therefore, inourview it seemsmorefruitful to prevent this trade‐off, which means focusing the entire hospitalstructureonpatientprocessesanddeletingtransferpoints.ToolittleattentiontocontingencyfactorsFurther,ourresearchshowsthathospitalsdonottakecontingencyfactorsintoaccount consistently in applying process‐oriented logistical concepts. It isknown from operationsmanagement literature that contingency factorsmayaffectthesuccessofinterventionsthatattempttoimprovethemanagementofoperations13, 14.According toHall (1999)15 andDonaldson (2001)16, themostimportant contingency factors of a logistical concept are strategy, size, taskuncertainty, and technology. Each hospital has a unique combination ofcontingency factors, and thus of opportunities and limitations. In agreementwith Hall (1999)15 and Donaldson (2001)16, the success of process‐orientedcaredeliveryforasinglepatientgroupisdependentontheorganisationalgoalsand on ways to reach these goals (strategy), the volume of a patient group(size),thecertaintythatstepsthataremostlytakenorthatshouldbetakeninacareprocessforaspecificpatientgroupcanbepredicted(taskcertainty),andthetechnologyofhospitalcare.

thesis_Vos_17x24_v01.pdf

Chapter7 │145

However,hospitalsdonotcarefullyconsiderthe fitbetweenprocess‐orientedlogisticalconceptsandthesecontingencyfactors.Forexample,inChapter6wesaw thathospitals tried to implementprocess‐oriented caredelivery forverysmall patient groups.We know from the literature, however, that due to thehigh level of standardisation and the dedicated investment, the effectiveoperating of a process‐oriented logistical concept calls for a high volume ofpatients who require the same chain of services to answer their health careneeds17.Inaddition, it isdifficultforageneralhospitaltomaintainagroupofmedical professionals specialised in care delivery for only a small group ofpatients. The increasing specialisation within medical disciplines during thelastdecenniaduetotherapid increase inmedicalknowledgeandtherapeuticoptionshasmadethisproblemworse18.Ithasresultedinmedicalprofessionalshaving a limitedwork area,meaning they can only deliver subtaskswithin acare process. Thus, if high volumes are not reached, the organisationalinvestmentwillbehigherthanthereturn.In conclusion, the importance of the fit between logistical concepts andcontingency factors means that process‐oriented logistical concepts are nopanacea for a hospital under all circ*mstances. There are circ*mstances inwhichtheuseofafunctionallogisticalconceptismoreeffectivethanthatofaprocess‐orientedlogisticalconcept.Afterall,disparitiesintheenvironmentarelikelytocallfordifferentresponsesinthedesignofalogisticalconcept19.Asaconsequence,different logisticalconceptswouldremainamonghospitals,andtherewouldnotbea‘one‐size‐fits‐all’solution.7.3.2 Suboptimal strategies to implement process­oriented logistical

conceptsItbecameapparentinseveralchaptersofthisthesisthattheimplementationofprocess‐oriented logistical concepts in hospitals is not accompanied by anappropriate ex‐ante hospital management analysis of the consequences of achange of logistical system and the unintended effects for the differentcomponentsof thehospital system (topmanagement, divisions/departments,medical professionals within a hospital building). This prevented hospitalmanagement from handling the conflicts that emerged between components.As a consequence, the implementation of process‐oriented logistical conceptswasobstructed.

thesis_Vos_17x24_v01.pdf

146│ Towardsprocess­orientedcaredeliveryinhospitals

Hospitalbuildinglayoutandprocess­orientedlogisticalconceptsChapter 3 showed that the implementation of process‐oriented logisticalconceptscanbehamperedbythephysicallayoutofabuilding.Changesintheintensity, direction, or volume of patient flow as a result of a new logisticalconcept led to congestions of the patient flow within the designed hospitalbuilding.Expertshavenotedthatitisextremelyhardtoachieveasituationinwhichthephysicallayoutsupportsthefunctionalityofanewlogisticalconceptinpre‐existingbuildings20.Duringthelifeofahospitalbuilding,changesinthehospitalenvironmentoccurthatinfluenceitsstrategy,whichinturnaffectsthelogistical concept with regard to planning and to controlling patient flow.Therefore, itseems importanttocreate flexibility in thephysical layoutwhenhospital buildings are designed: for example, by the standardisation ofconsultationroomsortheintroductionofsinglepatientrooms21.Afterall, therunningofhospitalbuildingsisextremelyexpensiveduetotheinvestmentsinneeded technology,andbuildingsneed to last formanyyears.The layoutcanonlybechangedifmajorinvestmentsaremade.Thisunderlinestheimportanceofaflexiblebuildinglayouttoenablehospitalstorespondtochange.Chapter3of this thesisdemonstrated that simulation techniques canbeveryhelpful toassessabuildingdesignintermsofitsflexibilitywithregardtohandlingfuturechangesinpatientflow.Goalsofdepartmentsandofprocess­orientedlogisticalconceptsInChapters1and6, it appeared that the implementationofprocess‐orientedlogisticalconceptswashamperedbecausedepartmentalgoals(i.e.maximisingdepartment output) were not adjusted accordingly (i.e. maximisingperformanceofcareprocessesintermsofwaitingtimesanddelays).Afterall,itwas not in the interests of the departments and medical professionals toimprove care coordination andwaiting times and delays, since they are onlyresponsibleforandbenefitfromtheoutputoftheirowndepartments.Chapter4showedthatthe introductionofprocess‐basedbudgeting–theallocationoffinancial means on the basis of the total care of a patient during an acuteepisodeofillness–canstimulatedepartmentsandmedicalprofessionalstouseprocess‐oriented outcome measures, which in turn encourages them toimplementprocess‐oriented logisticalconcepts.Fromthis, itseemsimportantthathospitalmanagementalsoadaptsreward‐andinformationsystemstothenewlogisticalconcept.Thiscanberelatedtothefactthatachangeneedstobeseenbyallinvolvedpartiesasanimprovementintheexistingpractice22,23.Ingeneral,anewlogisticalconceptwillonlybeeffectivelyimplementedifitseemstoofferadvantages.Benefitscanincludebetterhealthcareoutcomes,financialbenefits,workpleasure,oranincreaseinstatusforpeoplewhoadoptthenewworkingmethods.Atanearlystage,whenselectinganddevelopingalogistical

thesis_Vos_17x24_v01.pdf

Chapter7 │147

concept,itisimportanttoconsiderthesefactorsandthepossibleresponsesofthe target group in order to increase the likelihood of a successfulimplementation23. Hence, the implementation of a process‐oriented logisticalconcept has to be accompanied by sufficient incentives for involveddepartmentsandstaff.Routines of medical professionals and process­oriented logisticalconceptsChapter 4 showed that although many Dutch hospitals have implementedprotocolsforthediagnosis,treatment,androutingofpatients,thesearenotyetaccompanied by decentralised decision‐making, agreements about processownerships,andaprocess‐orientedviewonthepartofmedicalprofessionals.Thus, formal structures – protocols – have been created, but internalorganisational activities – routines of medical professionals – have not beenadjusted. It is known from literature that implementation of innovations likeprocess‐orientedcare logisticalconceptsdonotrunsmoothly, nomatterhowgreat the advantages, if the innovation is not consistentwith existing normsandvalues, orwith theexperiencesofprofessionalswhohave toadjust theirworkprocesses22,23.Due to the strong institutionalisation of the health care sector in theNetherlands, it isdifficult toalter the routinesofmedicalprofessionals, sinceresponsibilities and competences of doctors, nurses, and allied health careprofessionalsareboundby law24.Asaconsequence,only limitedpossibilitiesexist to change the responsibilities of involved medical professionals withincareprocesses. Improvingtheorganisationofcaredeliveryrequiresachangeinworkprocesses,modifyingtheinteractionbetweendepartmentsandmedicalprofessionals25,which includes changes in thedivisionof responsibilities anddecision‐makingcapabilities.Thisisnotalwayscompatiblewithexistingrules.Thelawimpedestheintroductionofworkforceflexibility:for instance,nursescannot take over the tasks of medical doctors with the same freedom thatemployees switch tasks in industry. However, workforce flexibility couldreduce the need for transfers and care coordination between medicalprofessionalsinhospitals26.An additional barrier to changing existing routines is that there are noreadymade solutions thatdescribe exactlywhat formprocesses for a specificpatient group in a specific context should take23. Because improvements inworkprocessescannotbederivedfromguidelines,protocols,andprocedures,athoroughproblemanalysis isneeded to tailor changes to the specific context(Chapter 5). Formulating the change itself thus forms part of the

thesis_Vos_17x24_v01.pdf

148│ Towardsprocess­orientedcaredeliveryinhospitals

implementation process, and this enables a process of bottom‐up learninginsteadof top‐down implementation.Medicalprofessionalsneed toengage informulating the change, so that they support the new logistical concept andhelpensurethatitwillbeaccepted.Realisationoftheimplementationprocessmust, however, be a joint effort on the part of medical professionals andmanagement. Medical professionals frequently lack experience in applyingprocess innovation methods, and they already face huge demands on theirtime; hence, justifiably, they may not always be willing to prioritise time‐consumingprocessredesignoverspendingtimewithpatients.7.3.3 ReflectionsWefoundthathospitalscannotordonotfullyimplementtheprocess‐orientedlogisticalconcepts:they lackasystemview,which in turn leadstoatrade‐offbetweenprocess and department optimisation; they do not take contingencyfactorsintoaccount,whichinturnleadstoasuboptimalfitbetweenlogisticalconcept and hospital environment; and they do not accompany theimplementation with an appropriate adaptation of internal organisationalactivities (i.e. reward‐ and information systems, routines of medicalprofessionals), which subsequently leads to conflicts that hamper thesuccessful implementationof logisticalconcepts.Possibleexplanationscanbefound in the institutional hospital environment and in the characteristics ofhospitalcare.TheinstitutionalhospitalenvironmentThe optimal implementation of process‐oriented logistical concepts isobstructed by the need for hospitals to obtain institutional legitimacy, uponwhich the support of normative authorities, the approval of legalbodies, andthesurvivalchancesofhospitalsdepend27.Thislegitimacyiscontingentonhowthehospitalcomplieswiththreeinstitutionalpressures27,28:cultural‐cognitiveisomorphic (‘Everyone else in the hospital sector is doing it’), regulativeisomorphic (‘Government enforces it by rules and regulations, backed bysanctions and enforcement agencies’), and normative isomorphic (‘It hasbecome a prevailing standard in hospital practice, and is thus expected fromhospitals’).Asaconsequence,theimplementationofprocess‐orientedlogisticalconcepts by hospitals is not necessarily motivated by their contribution toeffectiveness,butratherbytheircontributiontolegitimacy29.Thismeansthathospitalssometimesmakesuboptimalchoicesthatdonotautomaticallyleadtomoreefficientandqualitativecaredelivery: insteadof lookingatwhat isbestforthehospitalitself,theyconformtoinstitutionalpressures.Thiscanleadto

thesis_Vos_17x24_v01.pdf

Chapter7 │149

isomorphism of hospitals, although it is known from the contingency theorythatthereisnobestwaytoorganise,thatanywayoforganisingisnotequallyeffective, and that the best way to organise depends on the nature of theenvironment to which the organisation relates30. This means that hospitalsshouldhavedifferentcustomisedlogisticalconceptsdependentonthehospitals’strategy, size, volume, and technology. In addition, there exists a risk ofdecoupling between the formal structures created by hospitals to complysymbolicallywithinstitutionalforceslikethecreationofcareprogrammesandthe internal organisational activities. Thus, hospitals may formally establishcareprogrammeswhiledepartmentsandmedicalprofessionalsdonotchangetheirworkingmethods.The aforementioned situation means that hospitals do frequently notimplement process‐oriented logistical concepts thoroughly, although it doesnot mean that these concepts cannot be fully applied. Due to institutionalpressures to become process‐oriented organisations, hospitals startimprovementprojectswithoutfirstacquiringasystemview.Asaconsequence,hospitals try to perfect individual care processes by optimising parts ofprocesses and by coordinating transfer points. These transfer points causebreakdowns in the coordination of care processes31: namely, coordination oftransfer points offers a solution for one patient group, but may cause thesituation for other patient groups to deteriorate. As such, the transfer pointspresent obstacles to the successful implementation of process‐oriented caredelivery at hospital level. Thus, from a system point of view it is better toeliminatetransferpoints.However,hospitalshavedifficultiesgaininganinsightintotheoperatingoftheirentirehospitalsystem,whichisneededtoeffectivelyreducetransferpoints.Therefore,futureresearchshouldfocusonmethodsthatachieve a clear view of the complexity of a hospital system and of thepossibilitiestoreducetransferpointswithincareprocesses.HospitalcarecharacteristicsItseemsthatthespecificcharacteristicsofhospitalcarepreventhospitalsfromoptimally implementingprocess‐orientedlogisticalconcepts.Theorganisationof service delivery in hospitals is complex. Due to the number of differentillnesses, treatments, and preferences of patients and their medicalprofessionals,hospital care incorporatesmuchmorevariation than industrialprocesses.Inaddition,manypatientshavemorethanoneproblem32,33,whichcalls for different kinds of services, and sometimes at the same time34.Furthermore, there exists considerable medical practice variation or inter‐doctorvariation35.Medicalpracticevariationsaresystematicdifferencesinthestandardised incidence rate of clinical procedures in a specific population36.

thesis_Vos_17x24_v01.pdf

150│ Towardsprocess­orientedcaredeliveryinhospitals

Thesevariationscanalsohaveaneffectontheallocationofresources,andcanresult from different approaches to the same clinical problem or frominappropriateclinicaldecision‐making.Although much of the variation can be eliminated by, for example, anaggregation of the needs of patients and elimination of avoidable practicevariation, the amount of variation still exceeds that of industry. A thoroughelimination of variation requires the standardisation ofmedical practice anddoingawaywithpatientgroups37.However,hospitalsexperiencepoliticalandethicalobligationsthatpreventthemfromdroppingcertainservicesandfromfocusing on strategically important ones2. At the same time, the geographicalserviceareaofhospitals is restrictedbynorms foraccessibilityofcare.Thus,becauseoftheirlimitedservicearea,hospitalsarepreventedfromservinghighvolumes of patients who need the same care; at the same time, to workefficiently,hospitalsneedhighvolumesofpatientswhor*quirethesamechainof services17. In addition, hospitals wishing to standardise medical practiceexperienceresistancefrommedicalprofessionals,andlackeffectivemethodstocounteractit.Recentresearchhasshown,forexample,thattheintroductionofuniformguidelinesdidnotreducepracticevariation38.This variation in hospital care hinders the prediction of the demand forservices, which in turn prevents departments from adjusting their capacityplanningonactualneeds.Asaconsequence,itismoreefficientfordepartmentstocreatewaitinglines,sinceinthatwaytheycanmaximisetheircapacityuseofthe departments. Another problem caused by the variation is that hospitalshave to deal with patients that need standard care and those that needcustomisedcare.Itisalreadyknownfromoperationsmanagementtheorythatthe delivery of standard and customised care is hard to combinewithin oneorganisationwithoneculture,onesetofnorms,staff,incentives,andoperatingsystems,sincetherepetitiveworkofcaredeliverytostandardpatientsandthecreation of one‐of‐a‐kind services are such different operations34. Medicalprofessionals delivering customised care need to be highly skilled, and oftenpossess tacit knowledge that is difficult to transfer. Tacit knowledge is lessimportantwhenprocessstepsareknownaheadoftime.Underthepresentcirc*mstances,fullapplicationofprocess‐orientedlogisticalconcepts does not seem to be feasible. To apply the concepts successfully,hospitalshavetonegotiateethicalandpoliticalobligations,eliminateservices,reduce variation in medical practice, enlarge the geographical services area,andfocuscaredeliveryonalimitednumberofgroupsofstandardpatientswho

thesis_Vos_17x24_v01.pdf

Chapter7 │151

needaspecific treatment,orona limitednumberofmultidisciplinarypatientgroups.If a hospital focuses on the care delivery to patients needing a specificprocedure, it can choose to implement process‐oriented control systems thatestablish the sequence of care activities and the responsibilities ofprofessionalsinvolvedinthediagnosisandtreatment.Anexampleofahospitalthat decided to serve only a limited segment of a total population needing aspecificprocedure istheShouldiceHerniaCentre inCanada. Inahospital likeShouldice, only patients who fit the highly standardised care process areoffered treatment; all others are rejected34. For example, inguinal herniapatientsseekingarepairatShouldicefilloutapreadmissionformthathelpstoidentify those who do not fit in the segment served. Those not suitable forShouldice’s highly standardised process are counselled to seek treatmentelsewhere: for instance, inhospitals that focusoncaredeliverytocustomisedpatients.This selectionmakes itpossible to standardise thecareprocess toahighdegree.If a hospital decides to focus on care delivery to specific groups ofmultidisciplinary patients, the implementation of a process‐oriented basicstructure–ahospitalstructure that is focusedtowardspatientprocesses– isprobably more appropriate. After all, patients frommultidisciplinary groupsmayneed standard aswell as customised care. A health careprocess is thencraftedforeachpatientbylinkingasetofstandardprocessestogetherafteraninitial evaluation34. Each standard process may be provided by the same orseparateprofessionals.Whatmakesthecareuniquelysuitedtoeachpatientisthe combinationof thecomponents; the roleof themedicalprofessionalwhoperforms the initial assessment is then that of a designer. The need fortransfers and care coordination can be reduced if medical professionals areflexible: namely, when they can perform different types of activities26.Moreover, flexibility also increases performance, since it enables hospitals tobalance and distribute the workload effectively because staff can beswitched39,40. Collaboration ofmedical professionals and coordination of caredeliveryis facilitatedbytheeliminationofdepartmentalbordersbetweenthemedicalprofessionals2.Itshouldbenotedthat institutionalcirc*mstances in theNetherlandsarenotall in favour of specialisation. On one hand, regulative pressures enforcehospitals tospecialise inorder toreachhighqualityhealthcaredelivery.TheHealthCareInspectorate(IGZ)41,forexample,setsstandardsfortheminimumnumber of procedures that a hospital has to perform to keep its license for

thesis_Vos_17x24_v01.pdf

152│ Towardsprocess­orientedcaredeliveryinhospitals

certainprocedures.Ontheotherhand,opposingforcesexistintheinstitutionalcontext.Hospitalsneed,forexample,tocomplywiththeDutchCompetitionAct,which isenforcedbyNetherlandsCompetitionAuthority(NMA).AccordingtotheDutchCompetitionAct,hospitalsmaynotparticipateincartels,forexample,by making price agreements, sharing markets, and restricting production42.Fromthis,itseemsthathospitalsareobligedtotreatallpatientsregardlessoftheir care demand, and cannot restrict themselves to the delivery of certainprocedures or to care delivery to specific patient groups. Until this dilemmabetweenpressurestospecialiseandthepressuretocompeteisresolved,Dutchhospitals are de facto limited in their development towards specialisedhospitals, and thus in effective implementation of process‐oriented logisticalconcepts.7.4 ConclusionOn the basis of the findings, we cannot confirm the central hypothesis thatprocess­oriented care delivery leads to better outcomes in terms of the qualityand efficiency of care delivery at hospital level. More research is needed toidentify the circ*mstances under which the implementation of process‐orientedcaredeliveryisorisnoteffective.However,ourresearchdidprovideindications thataneffective implementationofprocess‐orientedcaredeliveryrequires a ‘product’ focus of hospitals: namely, restricting care delivery to alimitednumberofgroupsofstandardpatientswhoneedaspecificprocedureoronalimitednumberofspecificmultidisciplinarypatientgroups,topreventhospitals from needing to trade‐off process and department optimisation. Inthisrespect,hospitalsarefacedwithtwodilemmas:1. Theyareexpectedtoprovidecare foreverypatient,butat thesametime

theymustprovidecareefficiently,whichcallsforspecialisation;2. Caremustbeaccessiblewithinareasonabledistanceandtime,butat the

sametimehospitalsneedasufficientvolumeofthesametypeofpatients,whichcallsforalargerservicearea.

To enable this, it might be necessary that hospitals negotiate with theirinstitutional environment expectations and ethical obligations, redefine theirservice area, and cooperate with other hospitals. As soon as and insofar asinstitutional circ*mstances allow it, hospitals can effectively implementprocess‐oriented caredelivery.A condition for success is thathospitalsmakeanappropriateanalysisoftheconsequencesofachangedlogisticalsystemforthedifferentcomponentsofthehospitalsystem,followedbyactionsthataligninternal organisational activities and interests with the process‐orientedlogisticalconcept.

thesis_Vos_17x24_v01.pdf

Chapter7 │153

7.5 RecommendationsThemainfindingsofthisthesisprovideaninsightintotheeffectivenessoftheimplementation of process‐oriented logistical concepts and supportingmeasuresforthequalityandefficiencyofcaredelivery.Basedonthefindings,several recommendations can be made with regard to future research onlogistical concepts for caredelivery inhospitals aswell aswith regard to theimplementationofprocessorientation.7.5.1FutureresearchThisthesisfocusedmainlyontheevaluationofexistinginitiativestoimplementprocess‐orientedcaredelivery.Asaresultoftheresearch,doubtshavearisenabout the applicability of this type of concept for hospitals that serve aheterogeneouspatientpopulation.Future research should therefore first focus on the preconditions for aneffective application of process‐oriented logistical concepts in health care toensureappropriateapplicationata laterdate.Alternatively, forhospitalsthatserveaheterogeneouspatientpopulation,itwouldbeinterestingtoexploretheapplicability of other logistical concepts such as task or care componentstandardisationforanimprovementintheirqualityandefficiency.It also seems important to developmethods that provide an insight into thecomplexity of a hospital system and into the effects of changes in logisticalconcepts for the efficient and effective operating of care processes. Avisualisationofthecomplexityofsystemsgiveshospitalsbetteropportunitiesto take appropriate actions for process optimisation. Simulation techniquesmay be helpful, because they can estimate effects of changes on interrelatedactivities. This can prevent the occurrence of unexpected and undesirableeffectsduringthereal‐lifeimplementationofchangesinlogisticalconcepts43.Next,researchisneededintopossibilitiestoreducetransferpointsinhospitals,because thesemake it difficult tomaintain the streaming of processes in anefficient and optimal way17. As described, the diversity in the chain of careprevents hospitals from reducing transfer points when a process‐orientedlogisticalconceptisimplemented.Further,researchisneededtoestablishwhetherprocess‐orientedcaredeliverycanbe implementedmoreeffectively if agroupofhospitals cooperates.After

thesis_Vos_17x24_v01.pdf

154│ Towardsprocess­orientedcaredeliveryinhospitals

all,agroupofhospitalshasalargerservicearea,whichincreasesthechanceofstandardisedcaredeliverytoallpatients.Finally, future research has to take into account patient preferences and themedical outcomes of care delivery. It is assumed within this thesis that areductioninwaitingandthroughputtimes,coupledwithcollaborationbetweenmedical professionals, increases the quality of care delivery. However, fastercaredeliverymaynotalwaysbebetter.Insomecases,forexamplewhencancerisdiagnosed,atoo‐shortthroughputtimebetweendiagnosisandtreatmentisnotdesirablebecauseofthepsychologicalimpactonpatients.Moreover,wedidnotinvestigatethemedicaloutcomesofcaredeliveryitself.Althoughweexpectthat themedical outcomes of care delivery will be positively affected as theresultofabetterorganisationofcaredelivery,futureresearchneedstoconfirmthis.7.5.2HospitalpracticeDespite a lack of evidence on the effectiveness of process‐oriented caredelivery, several programmes have been initiated worldwide during the lastdecennium to help hospitals make their organisation of care delivery moreprocess‐oriented.Thisstudyhasattemptedtofillthegapbetweenevidenceandpractice. On the basis of the results, we can offer the followingrecommendations.It seems obvious fromour research that not all hospitals can benefit equallyfrom the implementation of process‐oriented logistical concepts. Theeffectivenessdependsonthefitthatcanbereachedbetweencharacteristicsofthe environment and the concept that is applied. The challenge is to find thebestmatch.Therefore,werecommendthathospitalscarefullyconsiderthe fitbetweenprocess‐oriented logisticalconceptsandtheirenvironmental factors:theseincludethepossibilitytofocusonalimitednumberofgroupsofstandardpatients who need a specific treatment or on a limited number ofmultidisciplinarypatientgroups;thevolumeoftheservedpatientgroups;andthepredictabilityofthecareprocesses.Aneffectiveimplementationofprocess‐orientedlogisticalconceptsseemstorequirethespecialisationinthedeliveryofcaretoalimitednumberofgroupsofstandardpatientswhoneedaspecificprocedure,ortoalimitednumberofspecificmultidisciplinarypatientgroups.From a geographical point of view, specialisation appears to be a feasibleoption. The demand for specialised hospitals seems big enough in theNetherlands, with its approximately 17million inhabitants who live in close

thesis_Vos_17x24_v01.pdf

Chapter7 │155

proximity to one another. The good infrastructure in the Netherlands alsofavoursthespecialisationofhospitals.For an effective implementation of process‐oriented logistical concepts,hospitalsneedasystemviewoftheirownorganisation.Specificconditionsandcontingenciesrequirecustomised implementation insuchawaythatconflictsbetweenhospitalsystemcomponentsareprevented.Toincreasethelikelihoodof successful implementation, we recommend the use of a balanced set ofincentives,suchasreward‐andinformationsystems.Hospitals must determine for themselves the institutional requirements thatmay prevent them from specialising. At the moment, hospitals are beingencouragedtospecialise,butatthesametimeareobligedbytheNetherlandsCompetition Authority to treat all patients regardless of their care demand.Hospitalsneedtoresolvethisdilemma.

thesis_Vos_17x24_v01.pdf

156│ Towardsprocess­orientedcaredeliveryinhospitals

References1. Lega F, DePietro C: Converging patterns in hospital organization: beyond the

professionalbureaucracy.HealthPolicy2005,74(3):261‐281.2. Vera A, Kuntz L: Process‐based organization design and hospital efficiency.

HealthCareManagementReview2007,32(1):55‐65.3. Van Merode GG: A prelude of 2004 Antwerp quality conference: targets and

target values ‐ integrating quality management and costing. Accreditation andQualityAssurance2004,9:168‐171.

4. HammerM,Champy J:Reengineering thecorporation:amanifesto forbusinessrevolution.London:NicholasBrealey;1993.

5. Consortium Better Faster pillar 3. National strategy Better Faster pillar 3 (inDutch). Utrecht/Rotterdam/Utrecht: Quality Institute for Health Care CBO,InstituteofHealthPolicyandManagementoftheErasmusUniversityRotterdam,OrderofMedicalSpecialists;2004.

6. McGrath KM, Bennett DM, Ben Tovim DI, Boyages SC, Lyons NJ, O'Connell TJ:Implementing and sustaining transformational change in health care: lessonslearntabout clinicalprocess redesign.Medical JournalofAustralia2008,188(6Suppl):S32‐35.

7. Nolan TW, Schall MW, Berwick DM, Roessner J: Reducing delays and waitingtimes throughout the healthcare system. Boston: Institute for HealthcareImprovement;1996.

8. Berg M, Schellekens W, Bergen C: Bridging the quality chasm: integratingprofessional andorganizational approaches toquality. International Journal forQualityinHealthCare2005,17(1):75‐82.

9. VosL,DückersM,DeBruijnM,WagnerC,GroenewegenPP,VanMerodeF:Toaprocess‐orientedhospital:desiredeffectsofa logistic improvementprogramme(inDutch).TijdschriftvoorGezondheidswetenschappen2008,86(4):207‐215.

10. Rouppe van der Voort M, Stoffer M, Zuiderent‐Jerak T, Janssen S, Berg M:Breakthrough process redesign III: 2006‐2007 Better Faster pillar 3, T2S2 enT3S1 (in Dutch). Utrecht/Rotterdam/Utrecht: Quality Institute for Health CareCBO, Institute of Health Policy and Management of the Erasmus UniversityRotterdam,OrderofMedicalSpecialists;2006.

11. Vermeulen IB, Bohte SM, Elkhuizen SG, Lameris JS, Bakker PJM, La Poutré JA:Adaptive optimisation of hospital resource planning. In Proceedings 11thConferenceonArtificialIntelligenceinMedicineinEurope,AIME2007.

12. DeVriesG, VanTuijlH (Eds.):Health care under pressure: vital tension in themiddle of organisations, a business approach (in Dutch). Houten: Bohn StafleuvanLoghum;2006.

13. Lawrence P, Lorsch J: Differentiation and integration in complex organisations.AdministrativeScienceQuarterly1967,12:1‐30.

thesis_Vos_17x24_v01.pdf

Chapter7 │157

14. Sousa R, Voss CA: Contingency research in operations management practices.JournalofOperationsManagement2008,26(6):697‐713.

15. HallRH:Organizations:structures,processesandoutcomes.NewJersey:PrenticeHall;1999.

16. Donaldson L: The contingency theory of organizations. Thousand Oaks: Sage;2001.

17. Van Merode F, Molema H, Goldschmidt H: GUM and six sigma approachespositionedasdeterministictoolsinqualitytargetengineering.AccreditationandQualityAssurance2004,10:32‐36.

18. LeeKH:Thehospitalmovement‐acomplexadaptiveresponsetofragmentationofcareinhospitals.AnnalsAcademyofMedicineSingapore2008,37(2).

19. Mintzberg H: The structuring of organizations: a synthesis of research.EnglewoodCliffs,N.J.:Prentice‐Hall;1979.

20. Bevan H, Glenn R, Bate P, Maher L,Wells J: Using a design approach to assistlarge‐scale organizational change: ‘10 high impact changes’ to improve thenational health service in England. The Journal of Applied Behavioral Science2007,43(1):135‐152.

21. Van Merode GG, Verreusel R, Vrieze KOJ, Zeemering S: Constructing newhospitals: the effects of innovative care organization. In Health SciencesSimulationConferenceProceedings:2004;SanDiego;2004:25‐28.

22. Dückers MLA, Wagner C, Groenewegen PP: Conditions for a sector wide,knowledge based, improvement programme in the Dutch hospital care (inDutch).ActaHospitalia2005,45(3):37‐54.

23. Grol R, Wensing M, Eccles M: Improving patient care: the implementation ofchangeinclinicalpractice.Edinburgh:Elsevier;2005.

24. MinistryofHealth,WelfareandSport:Theindividualhealthcareprofessionsact:International Publication Series Health, Welfare and Sport nr. 10. The Hague;2001.

25. BerwickDM,NolanTW: Physicians as leaders in improvinghealth care: a newseriesinannalsofinternalmedicine.AnnalsofInternalMedicine1998,128:289‐292.

26. Desombre T, Kelliher C,Macfarlane F, OzbilginM: Re‐organizingwork roles inhealth care: evidence from the implementation of functional flexibility. BritishJournalofManagement2006,17:139‐151.

27. ScottWR, RuefM,Mendel PJ, Caronna CA: Institutional change and healthcareorganizations:fromprofessionaldominancetomanagedcare.London:UniversityofChicagoPress;2000.

28. Powell WW, DiMaggio PJ: The new institutionalism in organizational analysis.Chicago,IL:TheUniversityofChicagoPress;1991.

thesis_Vos_17x24_v01.pdf

158│ Towardsprocess­orientedcaredeliveryinhospitals

29. Kenis P: Waiting lists in Dutch health care: an analysis from an organizationtheoretical perspective. Journal ofHealth, Organization andManagement 2006,20(4):294‐308.

30. ScottWR:Organisations:rational,naturalandopensystems.NewYersey:UpperSaddleRiver;2003.

31. Tavena S, Grote G, Easty A, Plattner B: Decoding the perioperative processbreakdowns: a theoretical model and implications for system design.InternationalJournalofMedicalInformatics2010,79(1):14‐30.

32. Schellevis FG: Multimorbidity in general practice (in Dutch). Bijblijven 2009,25(3):25‐30.

33. WagnerEM:Meetingtheneedsofchronicallyillpeople.BMJ2001,323:945‐946.34. BohmerRMJ:Medicine's service challenge:blending customand standard care.

HealthCareManagementReview2005,30(4):322‐330.35. WestertGP,GroenewegenPP,BoshuizenHC,SpreeuwenbergPM,SteultjensMP:

Medicalpracticevariationsinhospitalcare;timetrendsofaspatialphenomenon.Health&Place2004,10(3):215‐220.

36. FernandezLA,MartinJM,delCastilloJD,GasparOS,MillanJI,LozanoMJ,KeenoyED: Sources of influence on medical practice. Journal of Epidemiology andCommunityHealth2000,54(8):623‐630.

37. Morton A, Cornwell J:What's the difference between a hospital and a bottlingfactory?BMJ2009,339:428‐430.

38. De Jong JD, Groenewegen PP, Spreeuwenberg P, Schellevis F, Westert GP: Doguidelines create uniformity in medical practice? Social Sciences & Medicine2010,70(2):209‐216.

39. LiLLX,KingBE:Ahealthcarestaffdecisionmodelconsideringtheeffectsofstaffcross‐training.HealthCareManagementScience1999,2(1):53‐61.

40. MolemaJJW:Hospitalsystemdesign:creatingsupplyflexibilitytomatchdemandvariability. Maastricht University, Health Organization, Policy and Economics;2009.

41. TheHealth Care Inspectorate: http://www.igz.nl/english/(accessed at 6March2010)

42. The Netherlands Competition Authority: http://www.nma.nl/engels/home/index.asp(accessedat6March2010)

43. Groothuis S: Simulation in health care: lessons learned. PhDThesis.MaastrichtUniversity;2002.

thesis_Vos_17x24_v01.pdf

│159

Summary

thesis_Vos_17x24_v01.pdf

thesis_Vos_17x24_v01.pdf

Summary │161

‘Towardsprocess­orientedcaredeliveryinhospitals’Patients frequently have to wait for long periods before they are seen bymedicalspecialists;theythenneedtowaitagainfordiagnosticexaminationsortreatment. Problems with regard to quality, such as waiting times and poorcoordination of care often occur because hospital departments operate asseparated‘silos’,eachwithitsowngoals,anddonotexchangeinformationfortheplanningandcontrolofpatientflow.Nationalandinternationalattentiontothesequalityproblemshasmadehospitalmanagementandmedicalspecialistsaware that the way in which care is delivered needs to be reorganised.However, little is known about how to improve the organisation of caredeliveryintermsofqualityandefficiency.Itisalsonotknownhownewwaysofstructuringcaredeliveryneedtobeintroducedintoahospitalorganisation.Until now,many of the ideas for improvements in coordination and processcontrol have been derived from industry, including the successful conceptreferredtoas ‘businessprocessorientation’.Theconcept’sbreakthroughtookplace at the beginning of the 1990s under the name ‘Business ProcessReengineering’. Successful examples of the application of business processorientationfor theorganisationofcaredelivery tospecificpatientgroupsareknown, for instance, from the Dutch programme ‘Better Faster’, fromAustralia’s ‘RedesigningHospital Care’, and from theUnitedStates’ ‘Reducingdelays and waiting times throughout the system’. It is unknown, however,whetherthisprocess‐orientedcaredeliverycanalsobeappliedsuccessfullyathospital level. Therefore, this thesis aims to test the hypothesis that theimplementation of process­oriented care delivery leads to better outcomes intermsofqualityandefficiencyathospitallevel.Chapter1describesthebackgroundofthisthesisandliststhemainresearchquestions. To assess the effectiveness of process‐oriented care delivery athospitallevelintermsofqualityandefficiency,wefirstperformedaliteraturereview;thisreviewispresentedinChapter2.Thereviewdiscusseshospitals’experienceswithimplementingaprocess‐orientedlogisticalconcept. Itshowsthathospitalschoosebetweentwomainapproaches:A. Implementationofaprocess‐orientedoperationalcontrol systemwithout

changingthehospital’sexistingfunctionalbasicstructure;andB. Implementation of a process‐oriented basic structure, in which the

composition of departments is based on the needs of specific patientgroupsinsteadofonthetypeofmedicalspecialties.

Thereviewpointedoutthataneffectiveimplementationofaprocess‐orientedoperational control system (A) requires that both departmentmanagers and

thesis_Vos_17x24_v01.pdf

162│ Towardsprocess­orientedcaredeliveryinhospitals

medical specialists givepriority to the coordinationof caredeliverybetweendepartments above the capacity utilisation of each department. This proveddifficult,however,sincemedicaldepartmentsinfunctionalorganisationshaveto maximise their own output. An effective implementation of a process‐oriented basic structure based on multidisciplinary patient‐focuseddepartments (B) requires that hospitals focus on strategically importantservices. Unfortunately, this literature review could not identify studies ofsufficientquality todeterminewhichof the twoapproaches (A orB)deliversthebestresultsandunderwhatcirc*mstances.Followingonfromtheassessmentoftheeffectivenessofprocess‐orientedcaredeliveryathospitallevel,Chapters3‐6ofthisthesisfocusonwaystointroduceprocess‐orientedlogisticalconceptsintoahospitalorganisation.Chapter3 describes an evaluationmethod to assess the fit betweenhospitalbuildinglayoutandaprocess‐orientedlogisticalconcept.Itscontributiontotheimplementation of new logistical concepts is tested in a case study. Thedescribedmethod aims to ensure that topmanagement’swish to implementprocess‐oriented care delivery is not hampered by a conflict between thebuilding layout and specifics of the new logistical concepts. Using computersimulationtechniques,theevaluationmethodassesseshospitalbuildinglayoutfromtheviewpointofoperationsmanagement.Theaimofsuchanevaluationistoensurethatthebuildinglayoutcurrentlysupportstheefficientandeffectiveoperatingof careprocesses, andwill do so in the future. In the case study, anewly designed hospital layoutwas assessed on its flexibility tomeet futurerequirementsofnewlogisticalconceptsandchangesinpatientmix.Theresultsofthecasestudyshowedthatachangeinintensity,direction,orvolumeoftheflow as a result of a new logistical concept or a changed patient mix led tocongestions in the patient flow. In other words, the design of the hospitalbuildinglayoutwasnotflexibleenoughtodealwithanewlogisticalconceptora changedpatientmix. In this case,however, the layoutof thebuilding couldeasilybeadjustedbecauseitwasstillinthedesignphase.Chapter4 testswhether theuseofprocess‐basedpayments –process‐basedinternal hospital budgeting – contributes to the implementation of process‐oriented logistical concepts. Process‐based budgeting – the allocation offinancialmeansonthebasisofperformanceofcareprocesses–aimstoensurethat thecoordinationofcaredeliverybetweendepartments isnotobstructedby responsibilities atdivisionordepartment level.Analysesofdata involvingDutch hospitals demonstrated that process‐based budgeting stimulatedmedicalprofessionalstouseprocess‐orientedperformancemeasures,whichin

thesis_Vos_17x24_v01.pdf

Summary │163

turn had a positive effect on the number of activities to develop process‐oriented operational control systems. Although process‐based budgeting didnot directly stimulate the implementation of a process‐oriented logisticalconcept, itseemsthat it isavaluable tool tobridgethose internalconflictsofinterests between departments and medical professionals that hampercollaboration across departmental boundaries. It stimulated cooperationbetweendepartmentsandmedicalprofessionalstoorganisecaredeliveryinaprocess‐oriented way, even though the implementation was not yetaccompaniedby adapted internal organisational activities: for instance,moredecentralised decision‐making, agreements about process ownerships, and aprocess‐orientedviewonthepartofmedicalprofessionals.InChapter5andChapter6of this thesis, theeffectsof two typesof redesignmethods on the implementation of process‐oriented logistical concepts atprocess levelareassessed.Theseredesignmethodsaimtochangepatternsofinteractions between professionals in order to develop and implement newroutinesthatfitprocess‐orientedlogisticalconcepts.Chapter 5 applies a process innovation method – intended to radicallyredesignand improveworkprocesses– tochangethe traditionally functionaloperating system of in‐hospital care for stroke patients towards a process‐orientedoperatingsystemwiththeaimtoimprovethequalityandefficiencyofcaredelivery.Thismethodincludesfoursteps:processanalysis, identificationof bottlenecks, design of the process‐oriented operating system, andimplementation of the system. Results indicate that implementation led to asignificantly shorter length of hospital stay. As a result,more stroke patientscouldbeadmittedtothespecialisedward.Chapter6 uses theQuality Improvement Collaborative (QIC)method for theredesign of work processes to change a functional operating systemincrementallytowardsaprocess‐orientedoperatingsystemforspecificpatientgroups with the help of external facilitators and peer stimulus. Thismethodwasassessedinamultiplecasestudydesign.WithintheevaluatedQIC,externalchange agents provided eighteen collaborative project teams from varioushospitalswitha clearvision for the redesignof theiroperating system,alongwith a set of specific changes that could significantly improve patient flow.Results of the multiple case study showed that the perceived need to tailorstandard change ideas to local contexts and the complexity of aligning theinterestsofinvolveddepartmentshampereduseoftheQICmethodforprocessredesign by the project teams. As result of the selection process for theparticipationofprojectteamsbytheexternalchangeagent,peerstimuluswas

thesis_Vos_17x24_v01.pdf

164│ Towardsprocess­orientedcaredeliveryinhospitals

non‐optimal.Theexternalchangeagentincludedprojectteamsintheevaluatedcollaborative that worked on different aspects of care processes for varioustypes of patient groups and different patient volumes, while the QICmethodology aimed to implement evidence‐based practice through sharingknowledgewith other teams having similar goals. Further, project teams feltthat preconditions for the successful use of the QIC method were lacking.Therefore, additional research into the applicability of the QIC method forprocessredesignisneeded.Chapter7summarisesthemainfindingsofthestudiespresentedinthisthesis,and takes the findings into consideration. The research reconfirms thatimplementationofprocess‐orientedlogisticalconceptscanreducethroughputtimes for specific patient groups, which leads to a gain in quality care forpatientsandtoefficiencyforspecificcareprocesses.Notonlywillthesepatientgroups have shorter waiting times between examination and treatment buttheirconditionswillbe treatedtimelyandadequately,andwill require fewerresources (e.g. number of bed days). Similar benefits of process‐orientedlogistical concepts could not be determined at hospital level. The literaturereview in Chapter 2 was unable to identify enough high‐quality studies thatassessed the implementation of process‐oriented logistical concepts, and anempiricalassessmentfailedduetothelimitedavailabilityofdata.However,theresearchshowedthathospitalscannotordonotfully implementtheprocess‐oriented logisticalconcepts: they lackasystemview,which in turn leads toatrade‐offbetweenprocessanddepartmentoptimisation;theydonottakeintoaccount the fit between contingency factors (strategy, volume, task certainty,and technology) and process‐oriented logistical concepts when deciding toimplement a process‐oriented care delivery (process‐oriented logisticalconceptsarenopanaceaforhospitalsunderallcirc*mstances);andtheydonotaccompany the implementation with an appropriate adaptation of internalorganisational activities (i.e. reward‐ and information systems, routines ofmedicalprofessionals),whichinturnleadstoconflictsthathampersuccessfulimplementationoflogisticalconcepts.Possibleexplanationscanbefoundintheinstitutional hospital environment and in the characteristics of hospital care.The institutional environment forces hospitals to comply with cultural‐cognitive (‘Everyone else in the hospital sector is doing it’), regulative(‘Governmentenforceshospitalstodoitbyregulationsandlegalisation,backedby sanctions and enforcement agencies’), and normative (‘It has become aprevailing standard in hospital practice and thus expected from hospitals’)pressures to obtain institutional legitimacy. As a consequence, theimplementation of process‐oriented logistical concepts by hospitals is notnecessarilymotivatedbytheircontributiontoeffectiveness,butratherbytheir

thesis_Vos_17x24_v01.pdf

Summary │165

contribution to legitimacy. This means that hospitals sometimes makesuboptimal choices that do not automatically lead to more efficient andqualitative care delivery: instead of looking at what is best for the hospitalitself,theyconformtoinstitutionalpressures.Thiscanlead toisomorphismofhospitalsandtodecouplingbetweentheimplementationoflogisticalconceptsandadaptationofinternalorganisationalactivities.Isomorphismisnotdesired,sinceitisknownfromthecontingencytheorythathospitalsneedtocustomisetheir logistical concept to their environment. Decoupling leads toineffectiveness of the introduced logistical concept. In addition to theinstitutional environment, current hospital care characteristics also limit theimplementation of process‐oriented logistical concepts. Ethical and politicalobligationsandresistancetostandardisingmedicalpracticepreventhospitalsfromfocusingonalimitednumberofgroupsofstandardpatientswhor*quireaspecific procedure, or on a limited number of groups of specific multi‐disciplinarypatients.On thebasisof the findings,wecannotconfirm thecentralhypothesisof thisthesisthatprocess­orientedcaredeliveryleadstobetteroutcomesintermsofthequalityandefficiencyofcaredeliveryathospitallevel.Moreresearchisneededto identify the circ*mstances under which the implementation of process‐orientedcaredeliveryisorisnoteffective.However,ourresearchdidprovideindications thataneffective implementationofprocess‐orientedcaredeliveryrequires a ‘product’ focus of hospitals: namely, restricting care delivery to alimitednumberofgroupsofstandardpatientswhoneedaspecificprocedureor to a limited number of groups of specific multidisciplinary patients, toprevent hospitals from needing to trade‐off process and departmentoptimisation.Inthisrespect,hospitalsarefacedwithtwodilemmas:1. Theyareexpectedtoprovidecare foreverypatient,butat thesametime

theymustprovidecareefficiently,whichcallsforspecialisation;2. Caremustbeaccessiblewithinareasonabledistanceandtime,butat the

sametimehospitalsneedasufficientvolumeofthesametypeofpatients,whichcallsforalargerservicearea.

To enable this, it might be necessary that hospitals negotiate with theirinstitutional environment expectations and ethical obligations, redefine theirservice area, and cooperate with other hospitals. As soon as and insofar asinstitutional circ*mstances allow it, hospitals can effectively implementprocess‐oriented caredelivery.A condition for success is thathospitalsmakeanappropriateanalysisoftheconsequencesofachangedlogisticalsystemforthedifferentcomponentsofthehospitalsystem,followedbyactionsthataligninternal organisational activities and interests with the process‐orientedlogisticalconcept.

thesis_Vos_17x24_v01.pdf

thesis_Vos_17x24_v01.pdf

│167

Samenvatting

thesis_Vos_17x24_v01.pdf

thesis_Vos_17x24_v01.pdf

Samenvatting │169

‘Naareenprocesgerichteorganisatievanziekenhuiszorg’Patiëntenmoetenregelmatiglangwachtenopeenafspraakbijdespecialist,opdiagnostisch onderzoek en op behandeling. Deze wachttijden en de slechteafstemming tussenbetrokkenzorgverleners ineenzorgprocesvandepatiënttreden vaak op omdat Nederlandse ziekenhuizen van oudsher functioneelgeorganiseerdzijn.Ineenfunctionelestructuurstaatnietdezorgvraagvaneenpatiënt maar het zorgaanbod, georganiseerd in specialistische afdelingen,centraal. Informatie‐uitwisseling over de planning en beheersing vanpatiëntenstromen tussen verschillende specialistische afdelingen is in eendergelijkefunctioneleziekenhuisstructuurbeperkt.Nationaleeninternationaleaandacht voor kortere wachttijden en zorgcoördinatie heeft specialisten enziekenhuismanagement er bewust van gemaakt dat de zorgverlening andersmoetwordeningericht.Erisechterweinigbekendovermogelijkhedenomdekwaliteit en doelmatigheid van de organisatie van zorg te verbeteren. Ookbestaaterslechtsbeperktekennisoverdewijzewaaropnieuwemanierenvanzorgorganisatie het beste kunnen worden ingevoerd in de dagelijksezorgpraktijk.Tot nu toe komen veel van de ideeën ter verbetering van de organisatie vanzorg uit de industrie, zoals het ‘business process orientation’ concept. Dedoorbraakvandit conceptvondplaats inhetbeginvande jaren90onderdenaam ‘Business process reengineering’. Succesvolle voorbeelden van detoepassing van dit concept op de inrichting van zorg voor specifiekepatiëntengroepenzijnbekenduithetNederlandseverbeterprogramma‘SnellerBeter’,uithetAustralische‘RedesigningHospitalCare’,enuithetAmerikaanse‘Reducing delays andwaiting times throughout the system’. Maar het is nietbekendofditconcept,hetprocesgerichtorganiserenvanzorg,metsucceskanworden toegepast op het niveau van een heel ziekenhuis. Daarom beoogt ditproefschrift de volgende centrale hypothese te toetsen: ‘de invoering van eenprocesgerichte organisatie van zorg leidt op ziekenhuisniveau tot betereresultatenophetgebiedvankwaliteitendoelmatigheid’.Hoofdstuk 1 beschrijft de achtergrond van het proefschrift en deonderzoeksvragen. Om inzicht te krijgen in de effectiviteit van eenprocesgerichteorganisatievanzorgopziekenhuisniveauiseenliteratuurstudieuitgevoerd. De resultaten van deze literatuurstudie zijn beschreven inhoofdstuk 2. De literatuurstudie laat zien dat ziekenhuizen op tweeverschillende manieren toewerken naar een procesgerichte organisatie vanzorg:

thesis_Vos_17x24_v01.pdf

170│ Towardsprocess­orientedcaredeliveryinhospitals

A. Door de invoering van procesgerichte besturingssystemen, zoalszorgpaden,binnendebestaandefunctioneleziekenhuisstructuur;en

B. Doorde invoeringvaneenprocesgerichteziekenhuisstructuur,waarindeafdelingen zijn samengesteld op basis van de behoeften van specifiekepatiëntengroepeninplaatsvanopmedischespecialismen.

Uit de literatuurstudie bleek dat een effectieve invoering van procesgerichtebesturingssystemen (A) vereist dat managers en medisch specialisten meerprioriteitgevenaaneengoedecoördinatievanzorgaanpatiënteninplaatsvanaaneenoptimaalcapaciteitsgebruikvanhunafdeling.Moeilijkheidhierbijisdatspecialistische afdelingen in functionele ziekenhuisorganisaties afgerekendwordenophunproductie.Vooreeneffectieveinvoeringvaneenprocesgerichteziekenhuisstructuur (B) bleek een focus op strategisch belangrijke patiënten‐groepenessentieel.Deliteratuurstudiekonnietgenoegvalideenbetrouwbarestudiesvindenomvasttestellenwelkevandetweebenaderingen(AofB)debesteresultatenonderwelkeomstandighedenlevert.In aansluiting op deze evaluatie van de effectiviteit van een procesgerichteorganisatie van zorg op ziekenhuisniveau, onderzoeken de hoofdstukken 3‐6vanditproefschriftde toegevoegdewaardevanondersteunendemaatregelenbij de introductie van een procesgerichte organisatie van zorg in eenziekenhuisorganisatie.Hoofdstuk3beschrijftentesteenevaluatiemethodedienagaatofdewenstothet invoerenvaneenprocesgerichteorganisatievanzorgwordt ondersteunddoorde lay‐out vanhet gebouw.Bovendien toetst demethodede bruikbaar‐heidvanhetgebouwindetoekomst.Hetnutvandezeevaluatiemethodebijdeinvoering van nieuwe logistieke concepten is getest in een casestudie. Decasestudie evalueerde de flexibiliteit van een nieuw ziekenhuisontwerp tenaanzien van nieuwe eisen van toekomstige logistieke concepten en vanveranderingen in de samenstelling van de patiëntenpopulatie. De casestudietoondeaandateenveranderingindeintensiteit,derichtingofhetvolumevande patiëntenstroom als gevolg van een nieuw logistiek concept of van eenverandering van de patiëntenpopulatie tot opstoppingen in het ziekenhuisleidde.Hetontwerpvanhetziekenhuiswasdusniet flexibelgenoeg.Opbasisvandeuitkomstenvandeevaluatiekonhetziekenhuisontwerpnogmakkelijkwordenaangepastenhetoptredenvanopstoppingenwordenvoorkomen.Hoofdstuk4gaatnaofdetoekenningvanfinanciëlemiddelenperzorgproces–procesgerichtebudgettering–binnenhetziekenhuisbijdraagtaandeinvoeringvan procesgerichte logistieke concepten.Met procesgerichte budgetteringwilmen bereiken dat de coördinatie van zorg niet wordt gehinderd door

thesis_Vos_17x24_v01.pdf

Samenvatting │171

verantwoordelijkheid van divisies en specialistische afdelingen om doorproductie hun budget veilig te stellen. De analyse van Nederlandseziekenhuisdata toonde aan dat procesgerichte budgettering medischeprofessionals stimuleerde om procesgerichte prestatie‐indicatoren tegebruiken.Hetgebruikvanprocesgerichtepresentatie‐indicatorenbevorderdevervolgens weer activiteiten gericht op het realiseren van procesgerichtebesturingssystemen.Ondanksdat procesgerichte budgetteringdusniet directbijdraagtaandeinvoeringvanprocesgerichtebesturingssystemen,lijktheteenwaardevol instrument om interne belangenconflicten tussen specialistischeafdelingenteoverbruggen.Hetstimuleertdesamenwerkingtussenafdelingendienodigisomdezorgprocesgerichtteorganiseren.Hoofdstuk5en6beschrijvendetoegevoegdewaardevantweemethodenvoorherontwerp van zorgprocessen van specifieke patiëntengroepen. Dezemethoden beogen interactiepatronen tussen medische professionals teveranderen zodat nieuwe routines ontstaan die passen bij procesgerichtelogistiekeconcepten.Hoofdstuk 5 toetst het effect van een procesinnovatie methode, diewerkprocessen opnieuw ontwerpt, op de kwaliteit en doelmatigheid van deziekenhuiszorg voor CVA‐patiënten. Deze methode bevat vier stappen:procesanalyse, identificatie van knelpunten, ontwerp van een procesgerichtbesturingssysteem en de implementatie hiervan. De toepassing van dezemethode leidde tot een significant korter ziekenhuisverblijf. Gevolg hiervanwas datmeer patiënten toegang konden krijgen tot de gespecialiseerde CVAafdeling.Hoofdstuk 6 evalueert de stapsgewijze verandering van een functioneelbesturingssysteem naar een procesgericht besturingssysteem voor specifiekepatiëntengroepenmetbehulpvandeDoorbraakmethode.Dit iseenmethodewaarin teams met professionals van verschillende organisaties eengezamenlijke trainingvolgenomspecifiekeprobleemgebiedenaan tepakken.Doorbraakprojecten maken gebruik van een kort cyclische verbetermethodegebaseerd op het formuleren van concrete doelen (plan), het toepassen vanspecifieke interventies (do) en het aanhoudendmeten en bestuderen van destandvanzakenaandehandvanprestatie‐indicatoren(study)omtebepalenwaar bijsturing wenselijk is (act). In het geëvalueerde doorbraakprojectadviseerden externe veranderagenten achttien deelnemende teams over hetherontwerp van zorgprocessen voor verschillende patiëntengroepen envoorzagen zij de teams van gerichte veranderideeën. Deze methode werdgeëvalueerdineenmultiplecasestudie.Demultiplecasestudielietziendathet

thesis_Vos_17x24_v01.pdf

172│ Towardsprocess­orientedcaredeliveryinhospitals

gebruikvandeDoorbraakmethodewerdbelemmerddoordenoodzaakvandeteamsomdeaangereikteveranderideeënaantepassenaanhuneigensituatieen door de belangenafweging van betrokken specialistische afdelingen.Bovendien bleef de uitwisseling van ervaringen tussen de teams beperktdoordatdeteamswerktenaanverschillendezorgprocessenvoorverschillendepatiëntengroepen van verschillende volumes, terwijl de Doorbraakmethodeuitgaat van kennisuitwisseling tussen teams met vergelijkbare doelen.Daarnaast ontbraken volgens projectteams voorwaarden voor een succesvolgebruik van de Doorbraakmethode. Hierdoor kunnen op basis van ditdoorbraakproject niet met zekerheid uitspraken worden gedaan over dewaarde ervan.Daarom is additioneel onderzoeknodig omde toepasbaarheidvandeDoorbraakmethodevoorprocesherontwerptebevestigen.Hoofdstuk 7 geeft een overzicht van de bevindingen van de verschillendestudiesuit dit proefschrift en bediscussieertdeze in het licht vande centralehypothese. Het onderzoek bevestigt opnieuw dat de invoering van eenprocesgerichteorganisatiekanleidentotkwaliteitswinstvoorpatiëntenentotdoelmatigheidswinst voor specifieke zorgprocessen. Als gevolg van deprocesgerichte organisatie zijnwachttijden tussendiagnostisch onderzoek enbehandelingkorterenwordenpatiëntentijdigenadequaatbehandeldbijeenkleinere inzet van middelen, bijvoorbeeld door een reductie in het aantalligdagen. Op ziekenhuisniveau konden vergelijkbare voordelen niet wordenvastgesteld. De literatuurstudie kon niet genoeg studies van goede kwaliteithierovervinden.Daarnaastfaaldeeenempirischonderzoeknaardeinvoeringvaneenprocesgerichteorganisatievanzorgdoorgebrekkigedata.Welkwamnaarvorendatziekenhuizenprocesgerichte logistiekeconceptennietvolledigwillenofkunneninvoeren.Allereerstmissendeziekenhuizeneensysteemvisie,wat leidt tot een compromis tussen de optimalisatie van processen enafdelingen.Ook besteden de ziekenhuizen geen of teweinig aandacht aan derelatie tussen contingentiefactoren (zoals strategie, volume, taakzekerheid entechnologie) enprocesgerichte logistieke conceptenwanneer zij overgaan totde invoering van een procesgerichte organisatie van zorg (procesgerichtelogistieke concepten kunnen immers alleen werken onder bepaaldevoorwaarden). Tenslotte combineren ziekenhuizen de invoering vanprocesgerichte logistieke concepten niet met een aanpassing van de interneorganisatie (beloningsstructuren, informatiesystemen en medische routines),watleidttotconflictendiedesuccesvolleinvoeringvandelogistiekeconceptenbelemmeren.Mogelijkeverklaringenhiervoorkunnengevondenwordenindeinstitutionele omgeving van het ziekenhuis en de kenmerken vanziekenhuiszorg.Deinstitutioneleomgevingdwingtziekenhuizentevoldoenaancultureel‐cognitieve (‘iedereen doet het’), regulatieve (‘wet‐ en regelgeving

thesis_Vos_17x24_v01.pdf

Samenvatting │173

vereist het’) en normatieve (‘het is standaard ziekenhuispraktijk’) druk voorhet verkrijgen van institutionele legitimiteit. De invoering van procesgerichtelogistieke conceptenwordt daardoor niet noodzakelijkerwijs ingegeven doorde bijdrage aan de effectiviteit maar eerder door de bijdrage aan deinstitutionele legitimiteit. Dit betekent dat ziekenhuizen soms suboptimalekeuzenmakendienietautomatischleidentotgroteredoelmatigheidenbeterekwaliteit van de organisatie van zorg: ziekenhuizen conformeren zich aaninstitutionele druk in plaats van dat zij kiezen wat het beste is voor hetziekenhuis zelf. Dit kan leiden tot isomorfisme van ziekenhuizen en totontkoppeling tussende invoeringvan logistiekeconceptenenaanpassingvande interne ziekenhuisorganisatie. Isomorfisme is ongewenst omdathet uit decontingentietheoriebekendisdatziekenhuizenhunlogistiekeconceptmoetenaanpassen aan hun omgeving. Ontkoppeling leidt tot ineffectiviteit van hetgeïntroduceerdelogistiekeconcept.Deinvoeringvanprocesgerichtelogistiekeconceptenwordtnietalleenbeperktdoordeinstitutioneleomgeving,maarookdoor de kenmerken van de huidige ziekenhuiszorg. Ethische en politiekeverplichtingen en weerstand tegen het standaardiseren van de medischepraktijk verhinderen dat ziekenhuizen zich richten op een beperkt aantalgroepenstandaardpatiëntendieeenspecifiekeprocedurenodighebbenofopeenbeperktaantalgroepenvanspecifiekemultidisciplinairepatiënten.Op grond van de bevindingen kunnen we de centrale hypothese van ditproefschrift ‘de invoeringvaneenprocesgerichteorganisatievanzorg leidt totbetere resultaten op het gebied van kwaliteit en doelmatigheid opziekenhuisniveau’ niet bevestigen. Er is meer onderzoek nodig om deomstandigheden,waaronder de invoering van een procesgerichte organisatievanzorgalofnieteffectiefis,teidentificeren.Inonsonderzoekzijnindicatiesgevondendateeneffectieve invoeringvaneenprocesgerichteorganisatievanzorg vraagt om een ‘product’ focus van ziekenhuizen: dat wil zeggen, hetbeperken van de zorgverlening tot een beperkt aantal groepen standaardpatiëntendieeenspecifiekeprocedurenodighebbenofopeenbeperktaantalgroepen van specifieke multidisciplinaire patiënten, om te voorkomen datziekenhuizen een compromis moeten sluiten tussen de optimalisatie vanprocessen en de optimalisatie van afdelingen. In dit opzicht wordenziekenhuizengeconfronteerdmettweedilemma’s:1. De maatschappij verwacht dat ziekenhuizen zorg verlenen aan iedere

patiëntmaar tegelijkertijdmoetdezezorgefficiëntverleendworden,watvraagtomspecialisatie;

2. Zorg moet beschikbaar zijn binnen redelijke afstand en tijd maartegelijkertijd hebben ziekenhuizen een voldoende volume van hetzelfdesoortpatiëntennodig,watvraagtomeengroterverzorgingsgebied.

thesis_Vos_17x24_v01.pdf

174│ Towardsprocess­orientedcaredeliveryinhospitals

Om dit mogelijk te maken kan het nodig zijn dat ziekenhuizen met huninstitutionele omgeving onderhandelen over verwachtingen en ethischeverplichtingen,hunverzorgingsgebiedopnieuwbepalen,ensamenwerkenmetandereziekenhuizen.Wanneerenvoorzoverinstitutioneleomstandighedendittoestaankunnenziekenhuizeneffectiefeenprocesgerichteorganisatievanzorginvoeren.Voorwaardevoorsuccesisdatziekenhuizeneengedegenanalysevandegevolgenvaneenveranderd logistieksystemmakenvoordeverschillendecomponenten van het ziekenhuissysteem, gevolgd door acties die de interneorganisatieactiviteiten en belangen in overeenstemming brengen met hetprocesgerichtelogistiekeconcept.

thesis_Vos_17x24_v01.pdf

│175

Dankwoord

thesis_Vos_17x24_v01.pdf

thesis_Vos_17x24_v01.pdf

Dankwoord │177

Ditproefschrift istotstandgekomenmethulpensteunvanheelveelmensenuitmijnprofessioneleomgeving, familieenvriendenkring;eenaantalvanhenwilikhierinhetbijzondernoemen.Frits:Jevertrouweninmijtijdenshetschrijvenhebikenormgewaardeerd.DefysiekeafstandtussenMaastrichtenDenHaagheefteengoedesamenwerkingniet in de weg gestaan; jij stond te allen tijde voor mijn vragen klaar enbeantwoordde die altijd super snel. Je vele suggesties en kritische vragenhebbenmijsteeds,totaanheteindetoe,aanhetdenkengezet.Ikhebveelvanjegeleerd.Cordula enPeter: Ikhebhet zeer gewaardeerddat ikvanhetNIVELdekansheb gekregen aanmijn proefschrift te werken en ook af te ronden. Hartelijkdankvoorjullieadviezenenderuimtedieikkreegommijneigenideeënvoorditproefschriftuittewerken.Beoordelingscommissie:GraagwilikopdezeplekookeenwoordvandankaanuitsprekenaanProf.dr.C.Spreeuwenberg,Prof.dr.H.J.J.M.Berden,Prof.dr. J.van Engelshoven, Prof. dr. J.A.M. Maarse en Prof. dr. J. de Vries voor debeoordelingvanmijnproefschrift.Paranimfen:Michel en Olga, onze gezellige tijd bij het NIVEL zal ik niet snelvergeten.Michel,mijn‘SnellerBeter’collega,hetwaseenfeestommetjesamentewerken. Olga, ik heb goede herinneringen aan onze gesprekken tijdens delaatstefasevanonzepromotietrajecten.Fantastischdatjulliemijnparanimfenwillenzijn.Coauteurs: Jullie hulp bij de analyses en kritische revisies vanmijn artikelenwaren zeer welkom. Siebren, jouw interesse in mijn promotie en verdereloopbaanhebikalsbijzonderervaren.‘Vormgevers’ vanmijn proefschrift: Christel, dankzij jouw ruime ervaring enoog voor detail ziet de opmaak van mijn proefschrift er prachtig uit. Marc,bedanktvoorhetmooieontwerpvandekaftvanmijnproefschrift.NIVEL en LUMC collega’s: Jullie morele steun, de gezelligheid tussen onzewerkzaamhedendoor,enjulliebelangstellingvoormijnpromotietrajecthebikheelprettiggevonden.Vriend(inn)en, Elfjes, Hans en Ria, en familie: Hartelijk dank voor jullievriendschap,interesseinmijnwerkenvooralookdeafleidingvanmijnwerk.

thesis_Vos_17x24_v01.pdf

178│ Towardsprocess­orientedcaredeliveryinhospitals

Juul: Mam, jouw en papa’s onvoorwaardelijke steun, aanmoediging enbetrokkenheid, van begin tot het eind, hebben ongelooflijk veel voor mijbetekend. Helaas kunnen we niet meer samen met papa van dit momentgenieten.Rob:Meerdandriejaarlanghebjijmijgesteund,ookalbetekendedatvoorjoudatikindeweekendenheelvaakachtermijnlaptopzat.Dankvoorjeeindelozegeduldengoedezorgen.

thesis_Vos_17x24_v01.pdf

│179

CurriculumVitae

thesis_Vos_17x24_v01.pdf

thesis_Vos_17x24_v01.pdf

CurriculumVitae │181

LetiVoswasbornon22March1980inLeidschendam,theNetherlands.AfterfinishingGrammar School in 1998, she studiedHealth Sciences atMaastrichtUniversity in the Netherlands. In August 2004 she graduated withspecialisations in Health Policy and Administration and Movement Sciences.Aftergraduating,LetiworkedasaresearcherintheFacultyofHealth,Medicine,andLifeSciencesatMaastrichtUniversityMedicalCentre+,andatNIVEL, theNetherlandsInstituteforHealthServicesResearch.Herresearchconcentratedonoperationsmanagement inhospitals.Letiwas involved inseveralprojects,including researchon the improvementof careprocesses for strokepatients;the relation between hospital building layout and logistical concepts forcontrollingpatientflow;andtheevaluationofanationalqualityimprovementprogrammefor improvingpatientsafetyand logistics inhospitals. Inadditiontoherresearchactivities,duringtheperiodfromOctober2005untilJune2007,Leti worked as programme assistant for the Health Care Efficiency ResearchProgramme at the Netherlands Organisation for Health Research andDevelopment(ZonMw).InJanuary2010,LetiwasappointedImplementationFellowintheDepartmentofMedicalDecisionMakingattheUniversityMedicalCentreinLeiden.

thesis_Vos_17x24_v01.pdf

thesis_Vos_17x24_v01.pdf

│183

Listofpublications

thesis_Vos_17x24_v01.pdf

thesis_Vos_17x24_v01.pdf

Listofpublications │185

ArticlesandabstractsLeti Vos, Judith D. de Jong, Peter Spreeuwenberg, Emile C. Curfs, Peter P.Groenewegen.Patients’useofandpreferencesforprimaryandsecondarycare:a multinomial multilevel regression model (revised version submitted forpublication).Leti Vos, Sarah E. Chalmers, Michel L. A. Dückers, Peter P. Groenewegen,Cordula Wagner, Godefridus G. van Merode. Towards an organisation‐wideprocess‐oriented organisation of care: a literature review (revised versionsubmittedforpublication).Michel L.A. Dückers, Cordula Wagner, Leti Vos and Peter P. Groenewegen.Organisationaldevelopment,sustainability,anddiffusionofinnovationswithinhospitals participating in a multilevel quality collaborative (revised versionsubmittedforpublication).LetiVos,MichelL.A.Dückers,CordulaWagner,GodefridusG.vanMerode.Doescase‐mixbasedreimbursem*ntstimulatethedevelopmentofprocess‐orientedcaredelivery?HealthPolicy(acceptedforpublication).Arno van Raak, Siebren Groothuis, Robert van der Aa,Martien Limburg, LetiVos.Shiftingstrokecarefromthehospitaltothenursinghome:explainingtheoutcomesof aDutch case. JournalofEvaluation inClinicalPractice (acceptedforpublication).Leti Vos, Michel L.A. Dückers, Cordula Wagner, Godefridus G. van Merode.Applyingthequalityimprovementcollaborativemethodtoprocessredesign:amultiplecasestudy.ImplementationScience2010,5(1):19.Cordula Wagner, Michel L.A. Dückers, Peter Makai P, Leti Vos, Peter P.Groenewegen. National policy changes and evolving hospital qualitymanagementsystems. ISQua's(The InternationalSociety forQuality inHealthCare)26thInternationalConference2009;Dublin,Ireland,October11‐14.Leti Vos, Michel Dückers, Cordula Wagner. Blijft het vliegwiel draaien?Resultatenvaneenfollow‐upmeting.BestPracticesZorg2009:2:35‐40.

thesis_Vos_17x24_v01.pdf

186│ Towardsprocess­orientedcaredeliveryinhospitals

Michel L.A.Dückers, PeterMakai, Leti Vos, PeterP. Groenewegen, C.Wagner.Longitudinal analysis on the development of hospital quality managementsystems in the Netherlands. International Journal for Quality in Health Care2009:21(5):330‐340.Leti Vos, Robert J. van Oostenbrugge, Martien Limburg, Godefridus G. vanMerode, Siebren Groothuis. How to implement process‐oriented care: a casestudy on the implementation of process‐oriented in‐hospital stroke care.Accreditationandqualityassurance2009:14:5‐13.Leti Vos, Michel L.A. Dückers, CordulaWagner, Godefridus G. vanMerode. Abreakthrough collaborative for process redesign: does it work? ISQua's (TheInternationalSocietyforQualityinHealthCare)25thInternationalConference2008;Copenhagen,Denmark,October20‐22.Michel L.A. Dückers, Cordula Wagner, Leti Vos, Peter P. Groenewegen.Organization development, sustainability and dissemination of innovationswithinhospitals.ISQua's(TheInternationalSocietyforQualityinHealthCare)25thInternationalConference2008;Copenhagen,Denmark,October20‐22.LetiVos,MichelDückers,CordulaWagner.SnellerBeter,watkunnenweervanleren?KwaliteitinZorg2008:6:10‐14.Leti Vos, Michel L.A. Dückers, Monique de Bruijn, Cordula Wagner, Peter P.Groenewegen,GodefridusG.vanMerode.Naareenprocesgestuurdziekenhuis:beoogde effecten van een logistiek verbeterprogramma. Tijdschrift voorGezondheidswetenschappen2008(86):4:207‐215.Leti Vos, Siebren Groothuis, Godefridus G. van Merode. Evaluating hospitalredesign from an operations management perspective. Health CareManagementScience2007:10:357‐64.Leti Vos, Jessika van Kammen, Karen van Liere, Carolien Bouma, JettyHoeksema.Knowledge synthesis foruseofPET research: variety of interests,one view? Health Technology Assessment International (HTAi) 2007;Barcelona,Spain,17–20June2007.JessikavanKammen,LetiVosandJettyHoeksema.Knowledgesynthesis:frominforming to involving. Health Technology Assessment International (HTAi)2007;Barcelona,Spain,17–20June2007.

thesis_Vos_17x24_v01.pdf

Listofpublications │187

SiebrenGroothuis, LetiVos,GodefridusG. vanMerode.Anewhospital layoutanalysed using simulation. The Society for Modeling and SimulationInternational SCS, 2007 Western MultiConference WMC; San Francisco,California,USA,22‐26January2007.Martien Limburg, Leti Vos, Robert J. van Oostenbrugge, Godefridus G. vanMerode, Siebren Groothuis. Causes of inefficient stroke‐unit‐bed use ‐possibilitiesforfreeingupcapacity.CerebrovascularDiseases2005;19(Suppl.2)114‐115.FactsheetsandreportsLetiVos, JudithD.de Jong.Percentageoverstappersvanzorgverzekeraar3%:Ouderenwisselennauwelijksvanzorgverzekeraar.Utrecht:NIVEL,2009.Leti Vos, Michel Dückers, Cordula Wagner. Evaluatie Sneller Beter pijler 3:geleerdelessenvanziekenhuizen.Utrecht,NIVEL,2008(Rapport,ISBN978‐90‐6905‐920‐4).Leti Vos, Michel Dückers, Cordula Wagner. Evaluatie Sneller Beter pijler 3:resultaten van een verbeterprogramma voor ziekenhuizen. Utrecht, NIVEL,2008(Rapport,ISBN978‐90‐6905‐916‐7).

thesis_Vos_17x24_v01.pdf

thesis_Vos_17x24_v01.pdf

│189

Notes

thesis_Vos_17x24_v01.pdf

190│ Towardsprocess­orientedcaredeliveryinhospitals

thesis_Vos_17x24_v01.pdf

Notes │191

thesis_Vos_17x24_v01.pdf

thesis_Vos_17x24_v01.pdf