How incident reporting systems can stimulate social and participative learning

ContentslistsavailableatScienceDirect

Health

Policy

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

How

incident

reporting

systems

can

stimulate

social

and

participative

learning:

A

mixed-methods

study

David

de

Kam

_,

_Josje

_Kok

_,

_Kor

_Grit

_,

_Ian

_Leistikow

_,

_Maurice

_Vlemminx

_,

_Roland

_Bal

a_{ErasmusSchoolofHealthPolicy&Management,ErasmusUniversityRotterdam,P.OBox1738,3000DR,Rotterdam,theNetherlands} b_{MedicalSpecialistCare,HealthandYouthCareInspectorate,Stadsplateau1,3521AZ,Utrecht,theNetherlands}

a

r

t

i

c

l

e

i

n

f

o

Articlehistory:

Received17October2019

Receivedinrevisedform11May2020 Accepted17May2020

Keywords: Incidentreporting Regulation

Organizationaltheory Socialandparticipativelearning Patientsafety

a

b

s

t

r

a

c

t

Incidentreportingsystems(IRSs)havebeenwidelyadoptedinhealthcare,callingfortheinvestigation ofseriousincidentstounderstandwhatcausespatientharm.Inthisarticle,westudyhowtheDutchIRS contributedtosocialandparticipativelearningfromincidents.Weintegratequantitativeandqualitative datainamixed-methodsdesign.Between1July2013and31March2019,Dutchhospitalsreportedand investigated4667incidents.Healthcareinspectorsscoredallinvestigationstoassesshospitals’learning processfollowingincidents.Weanalysedifandonwhataspectshospitalsimprovedovertime. Addition-ally,wedrawfromsemi-structuredinterviewswithincidentinvestigators,qualitymanagers,healthcare inspectorsandhealthcareprofessionals.Healthcareinspectorsscoreincidentinvestigationreportsbetter overtime,suggestingthathospitalsconductbetterinvestigationsorhavebecomeadeptatwritingreports inlinewithinspectors’expectations.OurqualitativedatasuggeststheIRScontributedtopracticesthat supportsocialandparticipativelearning—theprofessionalisationofincidentinvestigationteams,the increasedinvolvementofpatientsandfamiliesininvestigations—andpracticesthatdonot—notlinking learningfromtheinvestigationteamstothatofprofessionals,notconsistentlymonitoringthe recom-mendationsthatinvestigationsidentify.TheIRSbothhitsandmissesthemark.WelearnedthatIRSs needtoberesponsivetothe(developing)capabilitiesofhealthcareproviderstoinvestigateandlearn fromincidents,iftheIRSistostimulatesocialandparticipativelearningfromincidents.

©2020TheAuthor(s).PublishedbyElsevierB.V.ThisisanopenaccessarticleundertheCCBYlicense (http://creativecommons.org/licenses/by/4.0/).

1. Introduction

The idea that incident reporting holds an important key to improvingsafetyofhealthcareiswell-established.[1,2]Adapted fromhigh-riskindustries,thepremiseofincidentreportingisthat by reporting and investigating incidents,we might understand what causesorcontributes topatient harm, sothat preventive strategiescanbedevisedandhealthcaremadesafer[3,4].Inmany countries,incidentreportingsystems(IRSs)havebeensetupwith theaimtolearnfromincidents[5,6].Researchhasshown, how-ever,thatIRSsstruggletofosterlearning[5,7–9].Inthesestudies, learning fromincidentsis understood asbeing ableto prevent futureincidents,sothatlearningisbelievedtohaveoccurredwhen fewerincidentsarereported.WhentheeffectivityofIRSsis eval-uatedintermsofthenumberofincidentsreported,IRSsfrustrate ordisappoint.[10,11]IRSsfailtodemonstrateprogress,

suggest-∗ Correspondingauthor.

E-mailaddress:dekam@eshpm.eur.nl(D.deKam).

ing thatlearning hasnot occurred[12,13]. Weargue that such evaluationsareproblematicastheyworkwithimpoverished con-ceptualisationsofwhatlearningis—generallyconfusinglearning withperformance [14]—, neglecthow definitions of what con-stitutesincidentsshift[15,16]and areinattentivetohowmore reportedincidentsmightbereflectiveofasafetyminded organ-isationalcultureratherthanpoorperformance[17,18].

IntheNetherlands,theDutchHealthandYouthCare Inspec-torate(further:Inspectorate),thenationalregulatortaskedwith monitoringqualityandsafetyofcare,hasdesignedandmaintains a nationalIRSforhospitals.TheDutchIRSfocusesonhospitals’ learningprocessesfollowingsentinelevents(further:SEs)andwas designed withtheideathatit should‘leadtosocial and partic-ipative learningatthelocallevel’ (seeboxA inappendix Afor thetypeofincidentsreportedintheNetherlandsandtheroleof theInspectorate).[16]Ratherthanassessingwhathospitalslearn fromSEs,theInspectoratemonitorshowhospitalslearnfromSEs, inquiringifhospitalslearntolearnfromSEs.[16]Specifically,the Inspectoratemonitorshospitals’abilitytoinvestigateincidentsand identifyfittingcorrectiveactions.Inordertomonitor‘thequalityof

https://doi.org/10.1016/j.healthpol.2020.05.018

Box1:Scoringinstrumenttoassessthequalityofthe SEanalysisreport.

Item Judgementofinspectors

Process

1 Isthemethodforanalysisspecified?(e.g., rootcauseanalysis(RCA))

Yes No ? Notapplicable 2 Istheinvestigatingcommittee

multidisciplinary?

Yes No ? 3 Aremembersoftheinvestigating

committeeindependent?

Yes No ? 4 Didallpersonneldirectlyinvolved

contribute?

Yes No ? 5 Didotherstaffwithknowledgeaboutthe

careprocesscontribute?

Yes No ? Notapplicable 6 Wasinputsoughtfromthe

patient/relatives?

Yes No ? Notapplicable

Reconstruction

7 Doesthedescriptionoftheeventgivea completepictureoftherelevant‘scenes’?

Yes No ?

Analysis

8 Havetheinvestigatorssearchedrelevant scientificliterature?

Yes No ? Notapplicable 9 Doesthereportstatewhetherapplicable

guidelines/protocolswerefollowed?

Yes No ? Notapplicable 10 Wasexternalexpertiseconsulted? Yes No ?

11 Doesthereportstatewhetherthemedical indicationfortheprovidedcarewas correct?

Yes No ? Notapplicable

12 Hasthequestion‘why’beenasked extensivelyenoughtoanalysethe underlyingcauseandeffect?

Yes No ?

Conclusions

13 Doesthereportidentifyrootcauses? Yes No ? Notapplicable 14 Dotherootcausesfitthereconstruction

andanalysis?

Yes No ? Notapplicable 15 Arecontributingfactorsconsideredand/or

identified?

Yes No ? Notapplicable

Recommendations

16 Doesthereportdocument recommendations?

Yes No ? Notapplicable 17 Dothesecorrectiveactionsaddressthe

identifiedrootcauses?

Yes No ? Notapplicable 18 Arethesecorrectiveactionsformulated

SMART?(Specific,Measurable,Attainable, RealisticandTime-Sensitive)

Yes No ? Notapplicable

Aftercare

19 Istheaftercareforthepatient/relatives described?

Yes No ? Notapplicable 20 Istheaftercarefortheprofessionals

involveddescribed?

Yes No ? Notapplicable 21 Hasthereportbeensharedwiththe

patient/relatives?

Yes No ? Notapplicable

Reactionofthehospitalboard

23 Doestheboardofdirectorsprovidetheir perspectiveontheanalysis,conclusions andrecommendationsinthereport?

Yes No ? Notapplicable

24 Doestheboardofdirectionsengagewith theanalysisandconclusionsofthereport?

Yes No ? Notapplicable 25 Isitstatedhowtheboardofdirectors

ensurestheimplementationofthe recommendationsofthereport?

Yes No ? Notapplicable

thelearningprocess’ofhospitals[16],theInspectoratedevelopeda scoringinstrumentthatsetsforthkeyconditionstoproperly inves-tigateandlearnfromSEs(Box1).Inlinewiththisinstrument,the Inspectoratepublishedaguideline,informinghospitalsonwhatthe Inspectorateexpectsfromaninvestigation.[19]SinceJuly2013, everySEreportedandinvestigatedbyhospitalsisscoredbythe Inspectorate[16].

Inthisarticle,westudytheeffectsoftheDutchIRSonthelocal learningprocessofhospitals.InlinewiththeaimsoftheIRS,we approachlearningfromincidentsasasocialandparticipative prac-tice,drawingonworkofMacrae[7]andRamanujanandGoodman

[14].Learningfromincidents,forMacrae,‘involvespeopleactively reflectingonandreorganisingsharedknowledge,technologiesand practices.It istheseprocessesofactionand reorganisationthat constitutelearningandmustbesupportedthroughinvestigation andimprovement.’[7]ForRamanujanandGoodman,‘learning rep-resentsasharedunderstandingamonggroupmembersofanew courseofactiontominimizeorpreventtherecurrenceofnegative events.(...)Iflearningdoestakeplacefromtheeventanalysis, thisnewrepertoirewouldbeshared,stored,andenactedatthe appropriatetime.’[14]Ourstudyisguidedbythequestion:How doestheDutchIRSstimulatesocialandparticipativelearningfrom incidents?

2. Methods

To answer our research question, we adopted a sequential mixed-methodsstudydesign.Drawingonquantitativeand qualita-tivedata,weaimtogenerateamorecomprehensiveunderstanding oftheeffectsoftheDutchIRS.[20,21]Wepresentandintegrate quantitativedataonscoredSEinvestigationreportsand qualita-tivedataonhowSEinvestigatorsperceivetheeffectsoftheIRSon theirinvestigationpracticesandlearningprocesses.

2.1. Datacollection

2.1.1. DatabaseofSEinvestigationreports

As researchers, we were granted access to an Excel-export that listed4667scoredSEreports, fromall 96 hospitalsinthe Netherlands,between1July2013and31March2019.Wereceived ananonymisedversionandcouldnotlinkhospitalstoindividual SEreports.Thedatabaseshowshowinspectorsscoredeachofthe 25itemsforeachSEinvestigationreport.Ifanitemisadequately addressed,itreceivesa‘yes’andisscoredas‘1.Ifareportdoes notadequatelyaddressanitem,itreceivesa‘no’andisscoredas ‘0.Whenitisuncleartoinspectorswhethersomethingwasorwas notdone,inspectorsscorea‘?’andisscoredas‘0.Ifanitemis deemedinapplicable,itisremovedfromthesetofquestionsthat cometomakeupthetotalscorethereportreceives.Basedonthe itemsscored,eachreportreceivesanoverallscore,expressedasa percentagefrom0%to100%.Multipleinspectorsscoreindividual reportswhicharediscussedinweeklymultidisciplinarymeetings, asaresultofwhichscoresmaybeamended.[35]Givenour inter-estinhowanIRSmightstimulatesocialandparticipativelearning, thedatabasewithscoredSEinvestigationreportspotentially pro-videsanindicationifandonwhatitemshospitalsimprovedtheir capabilitytoinvestigateSEs.Wedrawonqualitativeresearchto understandwhathappensbehindthenumbers.

2.1.2. QualitativeresearchontheeffectsoftheDutchIRS

Since2015,allauthorsexceptMVhavebeeninvolvedinvarious researchprojectsthatstudiedtheeffectsoftheDutchIRS.[33–36] Alloftheseprojectsincludedqualitative,ethnographicresearch. Inall,weconducted73 semi-structuredinterviewsand36hof ethnographicobservations.Inthisarticle,wepresentdatacollected withintwoprojectsspecifically(Table1).Inthefirstproject,the objectivewastoexplorehowhospitalsorganisetheirSE investi-gationpractices,howmanagersandSEinvestigatorsperceivethe effectsofinvestigatingSEsontheirlearningprocessesandwhat challengestheyencounter.Inthesecondproject,followingthefirst andotherresearchprojectsintotheDutchIRS,theobjectivewasto reviewandsynthesisefindingsfromstudiesconductedinthe col-laborativeontheeffectsofIRSonlearningand,withstakeholders, thinkabouthowtheDutchIRScouldbedevelopedfurther.

In both projects, sampling was purposive and while depth was strived for in the first project—aiming to reach data saturation—breadth was strived for in the second

Table1

Researchprojectscharacteristics.

Researchproject Authorsinvolved

infieldwork

Datacollected Project1

Apr2015–Sept 2016

JK 15semi-structuredinterviewsin13Dutchhospitalswithrespondentsinvolvedinorresponsiblefor conductinginvestigationsintoSEs:healthcareprofessionals,incidentinvestigators,qualitymanagersand chairsofinvestigationcommittees.Interviewslastedbetween51–91minutes(total18respondents). Respondentswereapproachedviaemailandinformedabouttheobjectiveoftheresearchinthisemail.Inthe email,thevoluntarynatureofparticipationwasstressed,aswasthefactthatdatawouldbefullyanonymised. Allapproachedrespondentsagreedtoparticipate.Duringinterviewsinternalincidentinvestigationprotocols andrelateddocumentation(meetingminutes,agenda’s,reportformatsetc.)werereviewedandwhen possible/appropriatehardcopieswerecollectedforfurtheranalysis.

Wehavediscussedmethodsusedtoconductthisstudymorein-depthelsewhere.[33] Project2

Jan2017–May 2018

DdKand KG

8semi-structuredinterviewswith(former)healthcareinspectorsinvolvedindesigningand/ormonitoringthe IRS.RespondentsincludedinspectorsinvolvedinscoringSEinvestigationreportsofhospitals,aswellas inspectorsregulatingotherhealthcaresectors(e.g.mentalhealthcare).Interviewslasted57–103minutes (total10respondents).Respondentswereapproachedviaemailandinformedabouttheobjectiveofthe researchinthisemail.Intheemail,thevoluntarynatureofparticipationwasstressed,aswasthefactthat datawouldbefullyanonymised.Allapproachedrespondentsagreedtoparticipate.

Focusgroupswith1)healthcareinspectors(3h),2)healthcaremanagersandprofessionals(3h),3)theDutch MinistryofHealth(1.5h)and4)citizens(5h).Fieldnotesweremadeduringthefocusgroups.

PolicydocumentsoftheInspectorateontheDutchIRSwereanalysedinordertounderstandthehistorical developmentoftheIRS.

Wehavediscussedmethodsusedtoconductthisstudymorein-depthelsewhere.[36]

project—soliciting insights from inspectors supervising a vari-ety of care sectors and otherstakeholders. All semi-structured interviews were structured using interview guides. Interview guideslistedthemesofinterestandwereamendedinlightof find-ingsfromprecedinginterviews.Interviewsweredigitallyrecorded followingrespondents’consentandtranscribedverbatim.

2.2. Dataanalysis

2.2.1. DatabaseofSEinvestigationreports

Descriptivestatisticswereappliedanalysingthe4667scored SE reports. To study changes over time, we obtained how SE reportsscoredoneachofthe25itemsscoredbytheInspectorate perquarter, asthepercentage ofreports adequatelyaddressing eachitem.We alsodeterminedtheaveragefinalscoreawarded to SEreports over time. Following two meetings with inspec-torsand a statisticianoftheInspectorate,who wereintimately familiarwiththedataandwithhowthescoringinstrumentwas developed and used over time, we revisitedthe dataand con-structedgroupsofhospitals.Toconstructthegroups,theinitial year(01−07-2013/01−07-2014)wasusedtocalculatethe aver-agescoreoftheSEreportsbyeachofthe96hospitals.Hospitals thatreportedlessthanthreeSEsduringtheinitialyear,werenot assignedtogroups(n=16hospitals).The80remaininghospitals wereassignedtooneoffourquartiles,basedonaveragescores (Table2).Wemergedthetwogroupsinbetweenthe‘low’(n= 20)and‘high’(n=20)scoringhospitals,referringtothatgroup asthe‘middle’(n=40).Ourreasonsfordoingsoareinformedby theInspectorate’sideasabouthowhospitalsshouldlearnfromSEs. [16,35,36]Forone,theInspectorate‘tailorsitsregulatorypractices tothelearningcapabilitiesandthedevelopmentalstagesof health-careproviders.’[18]Second,conductinggoodSEinvestigationsis thoughttobeaskillthathospitalsdevelopovertime.[16,35,36] So,whilehospitalperformance—intermsofSEscores—mightbe benchmarkedagainstotherhospitalsthatareinsimilar develop-mentalstages,theInspectorateisparticularlyinterestedifhospitals improveovertime.[16,35,36]Toplotthedevelopmentof aver-ageSEscoresforallhospitalsovertimemasksdifferencesbetween hospitals.Therefore,weconstructed4groupsof20hospitalsthat remainstable overtime—the two groups betweenthelow and highscoringhospitalgroupswemergedintoonemiddlegroup. We can expect that group construction based on received SE scoresduringthefirstyearservesasanapproximationof

hospi-tal’slearningcapabilitiesandthedevelopmentalstagestheyare in.

2.2.2. Semi-structuredinterviews

Thetranscribedinterviewswereanalysedwiththeaimto iden-tify themes, performingthematic analysis. [22] The concept of learningassocialandparticipativepracticefunctionedasa sensitiz-ingconceptthatguidedbutdidnotrestrictouranalysis.DdKandJK individuallyanalysedtwointerviewseach,identifyingthemes. Fol-lowingthat,DdKandJKreviewedthecodedmaterialanddeveloped acodingschemethatwasreachedthroughiterativediscussionsand multiplemeetingsbetweenboth authors.DdKandJKcodedthe remaininginterviewswiththecodingschemeinMicrosoftWord, attimesrefiningoraddingcodestothecodingscheme.The cod-ingschemeandthethemesidentifiedwerediscussedamongall authors.Consensuswasreachedoverthecourseoftwomeetings withallauthors.

3. Results

We identified five core themes that we formulate as prac-ticestheIRScancontributeto.RespondentslinkedtheIRSto:1) changedstaffattitudesandincreasedreporting,2)improvedSE investigations,3)participativelearning,4)locallearning,and5) recommendationsthatimprovequalityandsafetyofcare.These themesorderourresultsandwepresentquantitativeand qualita-tivedatapertheme.

3.1. Changedstaffattitudesandincreasedreporting

SeveralhospitalrespondentsreportthattheIRScontributedto changedattitudestowardspatientsafety,helpingtogenerate,as theycallit,‘safetythinking’.

YoulearnsomuchbyinvestigatingSEs;you’lllookatyourown workdifferently.(...)Itisreallybeneficialandthosereports areonething,butwhat Iaminterestedin issafetythinking thatneedstopermeatetheorganisation.Forthattohappen, ithelpstoinvestigateSEs,becauseyou’llforceyourselftodig deep.(Investigationcommitteechair,10−08-2015)

SEinvestigationsareenvisionedasatoolthatcanhelpfoster safetythinking,thatgoesbeyondlearningtopreventincidentsand

Table2

Informationonhospitalgroups,reportedandscoredSEs(01-07-2013/01-07-2014).

Groups Cut-offpointsofthe

groups(averageSE reportscores)

ReportedSEs AverageofSE

reportscore StdDevofSE reportscore Low(n=20) 24.0,64.9 188 572 185 Middle(n=40) 64.9,76.5 355 715 155 High(n=20) 76.5,89.8 188 808 107

refers,rather,toawayinwhichprofessionalsapproachtheirwork,

cognizantofriskstheirworkholds.

Also, respondentscredit theIRS withstressing theneed for

reportingSEs.

R1:WhenIcomparewherewewerefive,sixyearsagowith

today,we’vereallydeveloped.AlsojustintermsoftheSEswe

report.WeneverhadSEs...

R2:(laughs)

R1:Youhadnothingtoworryaboutwhenyouvisitedour

hos-pital;thingsdidnotgowrong...Nowwereport12SEseach

year.(Investigationcommitteechairandincidentinvestigator,

20−9-2016)

Manyhospitalrespondentsstatethattheyreportand

investi-gatemoreSEsnowthaninthepast.Thisissupportedbydataofthe

Inspectoratethatshowshow,since2009,reportedSEshavesteadily

increased(FigureAinappendixA).Thequotealsoshowsthatwhat

(thenumberofreported)SEstellushaschanged.‘Before,’an

inspec-tortoldus‘noSEsmeantyouwerethebestorganisation.Now,when

anorganisationreportsnoSEs,something’snotright’(Inspector,

30−05-2017).Thoughtofasreflectiveofanorganisationalsafety

culture,theamountofreportedSEsbecomesaqualitymetricinits

ownright,butonethatsayslittleabouthoworganisationsareable

tolearnfromthem.[7,23]

3.2. ImprovedSEinvestigations

A key aim of theDutch IRS wasto have hospitals improve theircapabilitytoinvestigateSEsasanimportantsteptowards learningfromSEs.[16]ForhowSEreportsarescoredby inspec-torssince2013,seeFig.1inthistextandfigureB–Ginappendix A.

Wemight concludethat thehighscoringgroupofhospitals already didfairlywell,having many ofthe conditionsfor con-ductingSEanalysisinplaceandthat,particularly,thelowscoring groupofhospitalsdeveloped.FromQ42015onwards,sometwo yearsafterSEreportswerescoredinaccordancetothenew scor-inginstrument,thedevelopmentoftheaverageSEscoresoflow andhighscoringhospitalsintertwine.TheIRSoffersthe opportu-nitytozoominfurther,onspecificitemsscored.Thisispotentially insightfulgiventhatnotallitemsareequallyeasytoperformwell on.Doingwellonsomeitems(e.g.‘Dothecorrectiveactionsaddress theidentifiedrootcauses?’)requiresmoreexpertiseandworkfrom investigationcommitteesthanothers(e.g.‘Isthemethodfor analy-sisspecified?’).Moreover,whileforthefinalscoreofareporteach itemisgrantedequalweight,inspectorsdeemsomeitemsmore importantthanothers.[34]Weselectedthreespecificitemsscored bytheIRSthat,accordingtoinspectors,adequatelyreflectthe capa-bilitytoconductSEinvestigations(seefigureC–EinappendixA.) [34]Astotheweightattributedtotheseitemsbyinspectors,one inspectornotes:

Whathappened[leadinguptoandduringtheSE]hastobeclear (...)soIcantelliftherootcausesareproperlyidentified.This iswhereitstarts;itdeterminesthenextstepsandwhetheror notthesestepsmakesense.(Inspector,1-11-2016)

Fig.1.Presentedherearetheaveragescoreandstandarddeviationofthoseaverage scoresofthelow,middleandhighscoringgroupofhospitalsbetween1July2013 to31March2019(n=4406).Thereisnobigdifferenceintheextenttowhichthe high,middleandlowscoringgroupsaccountforthenumberofreportedSEs;low scoringhospitalsreported1118SEsovertheperiod,themiddlescoringgroupsof hospitals2227(themiddlegroupconsistsof40hospitals,ratherthanthe20in thelowandhighscoringgroups)andhighscoringhospitals1061.Thehighscoring groupofhospitalsonaveragereceived79.8%scoreattheintroductionoftheIRS andreceivea90.0%scoreinQ12019.Thelowscoringgroupofhospitalsonaverage received58.6%scoreattheintroductionoftheIRSandreceivean88.8%scoreinQ1 2019.Themiddlescoringgroupofhospitalsonaveragereceived67.3%scoreatthe introductionoftheIRSandreceivean87.4%scoreinQ12019.Standarddeviation valuesdecreaseovertime.Inthelowscoringhospitalgroups,theaverageSDacross reportsinthefirstyear(Q32013toQ32014)was18.6.Inthefinalyear(Q22018 toQ22019)theaverageSDacrossreportswas7.4.Inthemiddlescoringhospital groups,theaverageSDacrossreportsinthefirstyear(Q32013toQ32014)was 15.1.Inthefinalyear(Q22018toQ22019)theaverageSDacrossreportswas7.2. Inthehighscoringhospitalgroups,theaverageSDacrossreportsinthefirstyear (Q32013toQ32014)was10.2.Inthefinalyear(Q22018toQ22019)theaverage SDacrossreportswas6.2.

Theitemsthatinspectorsemphasisearesequentialinthesense thatoneitembuildsuponthenext.Thequalityofaninvestigation, multipleinspectorsreport,startswithadequatelyaddressingthe ‘why’question(figureC)—sothattherootcausesmightbe iden-tified(figureD)andcorrectiveactionsdevisedthataddressthose rootcauses(figureE).

Whilethedataclearlyshowsprogressofhospitalscoresover time,wecannotdeterminebasedonthisdatawhetherhospitals havebecomebetteratinvestigatingSEsorifhospitalshavebecome moreadeptatwritingSEreportsinlinewiththescoringinstrument oftheInspectorate.Fromourinterviews,weknowrespondentsare wellawareofwhatneedstobeintheSEreport.Also,thescore awardedtoSEreportsisinterpretedbyhospitalrespondentsasa ‘grade’andtheinvestigationbecomesapracticerespondentswant toscorewellon.

IftheInspectoratewantsustonotedownhowmanyhourswe havespentdoing something,or whatevercriteria theyhave thoughtof,wellthenwe addittoourchecklistof thingsto addinthereport.Wewanttoscore100%.(Committeechair, 20-09-2016)

Hospitals have invested in the professionalisation of inves-tigationteams—emphasised and arguedfor in multiplestudies [8,24]—bytrainingtheminmethodsonhowtoconductSE inves-tigationsandbykeepingteamsconsistent,allowinginvestigators todevelopexpertise.But,dedicatedteamsarealsoneededdueto theincreasednumbersofSEsthat arereportedandneed tobe investigated.

Theseinvestigationstakesomuchtime.Medicalspecialistsdo themontheside,whileadedicated[investigation]team devel-opsexperience[withSEinvestigations]sothatthequalityof investigationsis consistent.And yeah, it takesan incredible amountof time... andyou wanttheinvestigations tobeof goodquality.(...)ThesereportsgototheInspectorate.(Medical doctor,18-08-2016)

Ashospitalsincreasinglysetupdedicatedteamsinresponseto increasingnumbersofSEsthatneedtobeinvestigated,coupled tothedesireto‘score’well,conductingSEinvestigationsbecomes aparticularorganisationalactivityandresponsibility,targetedat creatingreportsthatfittherequirementsoftheInspectorate.Input fromconcernedprofessionals,especiallyintherecommendation phase,isoftennottakenseriously.

I:Whatifprofessionalsdon’tagreewiththerootcausesyou’ve identifiedandtherecommendationsyoupropose...Doesthat happen?

R:Yeah,sure,thathappens(laughs).Um,so,withthe investiga-torswe’lllookattheresponse[oftheprofessionals].Whatdowe think?Aretheycorrect?Andarewegoingtochangethat?Ifwe believethatitdoesnotfittheinvestigationweconducted,we donotchangeitinthereport.(Committeechair,28-06-2016) Anotherhospitalrespondent toldusthat whenprofessionals disagreewiththerecommendationsoftheinvestigationteam,the teamiswillingtoconsidertheprofessionals’perspectivewhenit identifies‘errors’inthereport,butthatwhen‘[professionals]think ourrecommendationsareradicalorsomethingelse,well...,it’s ourrecommendation’(Medicaldoctor,18-08-2016).Investigators developrecommendationsinlightofhowtheInspectoratescores them—asfittingtheanalysis—ratherthaniftheycontributetothe qualityandsafetyofcarepractices.

3.3. Participativelearning

Theimportanceofinvolvingpatientsandfamiliesinincident investigationsisincreasinglyrecognisedandisspurredbytheidea thathealthcarecanlearnfromthepatients’andfamilies’ perspec-tives[25–27].IntheDutchIRS,hospitalsareexpectedtoinvolve patientsandfamiliesinSEinvestigationsandassuch,it encour-agedhospitalstowidenthecircleofpeopleabletoparticipatein andcontributetoSEinvestigations.

Yeah,[involvingpatientsandfamiliesinSEinvestigations]it’s somethingwe’vewantedforsometime,thinking‘weneedto dothis,thisisimportant’.Buttoactuallystartdoingit,isquitea bigstep.(...)Soontheonehand,weweremotivatedtoinvolve patientsandfamilies,havingheardhowimportantitisandon theotherhand,thepressurefromtheInspectoratetostartdoing this...,ithelped.(Medicaldoctor,28-06-2016)

Thequantitativedatasuggestthat,in2013,involvingpatients andfamiliesinSEinvestigationswasnocustomarypractice (fig-ureFinappendixA).Similarly,theIRSassessedandcontributed tothedegreetowhichSEinvestigationsreportsaresharedwith patientsand families afterwards (figure G in appendix A). The IRScontributedtothenormalisationofapractice—theincreased involvementofpatientsandfamilies—thatiswidelyarguedfor.

ButinvolvingpatientsandfamiliesinSEinvestigationsisnot thesameaslearningfromthem.TheIRSoperationalisestheneed ‘toengagethepatientorapatientrepresentativeinSEanalysis’ [16]byinquiringif‘inputwassoughtfrompatient/relatives?’The IRSdoesnotspecifywhatconstitutessuch‘input’ortheextentto whichhospitalsneedtoinvolvepatientsandfamilies.Hospitals,in responsetotheIRS’sencouragementtoinvolvepatientsand fami-lies,havedevelopeddifferentwaysoforganisingsaidinvolvement. Typically,however—andwereportonpracticesofpatientand fam-ilyinvolvementinSEinvestigationsmoreextensivelyinourother work[33,37]—incidentinvestigatorspredeterminethescopeand thequestionstheinvestigationneedstoprovideanswersto.

[IncaseofanSE]we[theinvestigativeteam]lookat:whatis thefocusoftheinvestigationandbasedonthat,whatdowe wanttoknow?Wedrafttheresearchquestions.Andthenwe decide,givenallthat,whowewanttospeakto.Weschedule appointmentswiththosepeopleandthen,basically,wehave alltheinformationweneed.(Committeechair,28-06-2016) Patientand family inputand theperceived value thereofis restrictedtotheabilityofpatientsandfamilies tocontributeto theanalysisofSEassetforthbytheIRS.Sometimes,patientsand familiesare‘eyewitnesses’whoprovide‘newfacts’(Incident inves-tigator,20-09-2016),butthisisnotalwaysthecase.

Sometimes,Ireallywonder‘whatcouldthefamilypossiblyadd tothis[analysis]?’Andthen,westillhavetoinvolvethem,for theInspectorate,really.(Incidentinvestigator,12-07-2016) Look,iffamiliesarereallydistanced...orhavenothingtodo [withtheSE],Idon’tthinkyoushouldinvolvethemjustbecause theprotocolsaysyoushould.Ittakesalotoftime;involvement hastobeofvalue.But,ifafamilymemberwasphysicallypresent [atthetimeoftheSE]orreallyplayed apartin theprocess thatledtotheSE,wellyeah,thenitmakessensetoinvolve them.(Medicaldoctorandinvestigationcommitteechair, 16-08-2016)

Moreover, although hospitals are committed to involving patientsand familiesinSEinvestigations,when theperspective of patientsand families doesnotalign withthatof profession-als,investigatorstendtogranttheprofessionalperspectivemore weight.HospitalsalsohavedifferentwaysofsharingSE investi-gationreports;while someshare reportsin full,othersprovide summariestopatientsandfamiliesorarrangeaface-to-face meet-ingwhereintheinvestigation’sfindingsarepresentedtopatients andfamilies.Whilesomehospitalsexplorepossibilitiesformore comprehensive patient and family involvement—e.g. by asking patientsandfamilieswhatkindofquestionstheywouldliketosee theinvestigationaddress—thisinvolvementin SEinvestigations generallyhappensonthehospital’sterms.Clearlythen,theIRS—in inquiringifhospitalssolicitinputfrompatientsandfamilies—does notattendtoordiscernbetweenthedifferentwaysinwhich hos-pitalslooktoinvolvepatientsandfamiliesinSEinvestigations.

3.4. Locallearning

Whileinvestigating SEsisexpectedtogenerate learning,the needtoinvestigateSEsisnotpromptedbythepotentiallearning opportunitiesanSEholdsbutbecauseitissevereintermsofpatient outcome(seeBox1).This,respondentspointout,meansthat organ-isationalresourcesandtimearecommittedtoinvestigatingSEs atthecostofattendingtolesssevereincidentsthatmighthold valuablelearningopportunities.

Ijustcamebackfromaholidayandwantedtogetbacktomy planonhowtotakethese[SEinvestigations]toahigherlevel andthenIsawthreemoreSEsinmyinbox.(...)It’s

frustrat-ing;wewanttodoittherightway...It’slike...running;you cantrainforenduranceorforspeed.Whenyoudobothatthe sametime,you’llgetinjured.Sowealwayshavetoinvestigate moreand,atthesametime,theinvestigationshavetobebetter, becauseeverytimewereceivefeedback[fromtheInspectorate] ‘you’renotdoingthiswellenough’.Andit’smakingmeanxious. Wegettheidea[oftheInspectorate],butwestrugglekeeping up.(Committeechair,10-08-2015)

TheincessantstreamofreportedSEsthatneedtobe investi-gatedbyhospitalscomesatthecostofreflectingonwhatsingular SEstella hospitalaboutitsqualityand safetyof careand how findingsfromparticularinvestigationsmightgenerateaggregated learningatadeeperlevel.Inspectorsreportsimilarexperiences. AshospitalscontinuetoinvestigateandreportonSEs,inspectors havetokeep scoringthem.‘Whatdo alltheseSEstellus?How mightotherorganisationslearnfromthis?(...)Wewanttogetto thosequestions,butwedon’thavethetime.Wearesocaughtup ingettingtheseSEswrappedup...it’soverwhelming’(Inspector, 25-09-2017).

3.5. Recommendationsthatimprovequalityandsafetyofcare

OneoftheaimsoftheDutchIRSwastohavehospitalslearn todevisecorrectiveactionsthatfittheircontext.WhilefigureE seemstosuggesthospitalsareincreasinglycapableofdoingso, recommendationsarescoredinlightofwhetherornottheyfitthe analysis,ratherthaniftheycontributetosafecarepractices.Also, hospitalrespondentsacknowledgethatitisachallengetokeep trackofalltherecommendationsSEinvestigationsidentify.

SometimesIfindoutaparticularrecommendationhasjust van-ished.Thenthereisanewmanagerandnobodyisabletorecall thatrecommendation.(Incidentinvestigator,12-07-2016) Um,wehavealltheserecommendationsinanExcel-sheetand wetrytofollowupontheseeverythreemonths,asking peo-plehowthey’refaring.Attimes,ourannualmeetingwiththe Inspectorateservesasatriggertothink‘oh,right,westillhave todothis’.(Incidentinvestigator,18-05-2016)

Our interviewssuggest thathospitals struggle tokeep track of and evaluate theeffects of the identified recommendations. Respondents suggest that while organisational investment into investigatingSEisconsiderable,followinguponrecommendations aftertheinvestigationdoesnotreceivethesame(structured) atten-tion.

4. Discussion

Indrawingonandintegratingquantitativeandqualitativedata ontheDutchIRS,ourstudysuggeststhattheIRScontributedto arange ofpractices inhospitals.Interms ofitscontributionto socialandparticipativelearningfromSEs,theIRSbothhits and missesthemark.GoingbacktoRamanujanandGoodman’s def-initionofsocialandparticipativelearning,‘learningrepresentsa sharedunderstandingamonggroupmembersofanewcourseof actiontominimizeorpreventtherecurrenceofnegativeevents.’ [14]Ourstudyfindsthatwhilehospitalsinvestinthetrainingof incidentinvestigatorsandwhilehospitalSEinvestigationreports arescoredhigherby inspectorsover time,thelearning process oftheinvestigationteamsisnotorpoorly connectedtothatof theinvolved healthcare professionals. While patients and fam-ilymembersareincreasinglyinvolved,theirinputisnotalways valued byinvestigators. Theinput and perceivedvalue of both patientsandprofessionalsislinkedtotheextenttowhichithelps investigatorsconducttheinvestigationasoutlinedbytheIRS.The ‘sharedunderstandingofanewcourseofaction’thatRamanujan

andGoodmanspeakof,ismostlysharedamongincident inves-tigators,who—onaccountoftheirexpertiseandtheneedforan independentinvestigation—claimownership overthe investiga-tion which can hamper the participation of others and shared learning. Paradoxically, in theattempt toencourage and mea-suresocialandparticipativelearning,theIRSengenderedpractices oflearning thatrestrict who cantrulyparticipate.Investigators can act as gatekeepers of the investigative process; investiga-tionsareorganisationallycordonedoffandparticipationisvalued inlightofthestandardtheInspectorate holdsinvestigationsto. Momentsofreflectionandopportunitiesforaggregatedlearning, meanwhile,arescarcegiventheconsistentpressuretoreportand investigate(forhospitals)aswellasscore(fortheInspectorate) moreSEs.Thisisatrendwecanexpecttocontinueasreporting behaviourhasbecomeaquality metricin itsown right,that is saidtobeindicativeofahospitals’safety-mindednessand trans-parency[7]. While corrective actionsare adequately identified, theyare not consistently monitored or evaluated by hospitals. Also,correctiveactionsareassessedintermsofcoherencewith theSEanalysisrather than ifor how theyare ofvalue forthe practiceofhealthcare professionals. ‘Iflearningdoestakeplace fromtheeventanalysis,’RamanujanandGoodmanfurtherwrite, ‘thisnewrepertoirewouldbeshared,stored,andenactedatthe appropriatetime.’[14]ThedatacollectedthroughtheIRSsheds no light on if and how hospitalsshare, store or appropriately enact this newrepertoire that theinvestigation ideally results in.

Giventhatweknowthatorganisationsinvestinpracticesthat areexternallymonitored,[28,29]itishardlysurprisingthat hospi-talsconsistentlydeliverhigherscoringSEreports.Still,ourfindings resist theinterpretation that theDutch IRS is a tick box exer-cisehospitalshavebecomeincreasinglyadeptat.Askinghospitals whethertheyaskedthepatientandfamilyforinputgenerated dis-cussionsaboutthevalueofpatientand familyinvolvementand hospitalsorganiseforandvaluesuchinvolvementdifferently.[33] Herewewanttopointoutthattheinvolvementofbothpatients andprofessionalsinSEinvestigationsisinstrumentaltothe objec-tive of learning from an SE and that the emotional impact of SEs,onboth patients, families and professionals, isnot accom-modatedforintheseinvestigations.[30,33,37] AsNicolinietal. already pointedout, failingtoengagewithand make roomfor theemotional impact of an SEin favour of thequest for facts and evidencecanactually hamperlearning [30]. Elsewhere,we explorehow ‘beingemotional’renderspatientsand profession-alspronetobeingdisqualifiedascontributingvaluableinputin anSEinvestigation.[37]Now,theIRSdoesinquireintoaftercare practicesofhospitalsfollowinganSE,forbothpatientsand pro-fessionals,thatmightmakeroomforsaidimpact—eveniftheIRS doesnotfollowuponhowthoseaftercarepracticesareorganised andvaluedbythosewhomakeuseofthem.The professionalisa-tionofSEinvestigatorsandthereportstheydeliverisavaluable achievement, even if that also allows a hospital toscore well. Our respondents notethat knowledge aboutpatient safety has increasedasaresultofinvestigations.Butalthoughitis acknowl-edgedthatinvestigatingincidents‘isjustonestepinthepathto improvement’[16],theIRSriskssinglingouttheinvestigationas themostimportantone.ScoringSEreportsasreflectiveof hospi-tals’learningprocessperpetuates,oratleastdoeslittletodispel themistakennotionthat investigatingincidentsis thesameas learningfromincidents[7,14,31].Withtheaimtoencourageand contributetosocialandparticipativelearningfromincidents,the DutchIRSmonitorsadynamicpractice,ratherthananoutcome. However,weconcludethattheIRSdoesnotadequatelyreflectthe dynamicpracticeitmonitors.Nowthattheconditionsfor hospi-talstoproperlyinvestigatetheirSEsseeminplace,theIRSshould redirect its focus to encourage reflection,monitor how shared

understandingdevelopsafteranSEandstressthelinkagebetween investigatingandlearning.WeproposetwowaysinwhichanIRS mightfurtherencouragesharedandparticipativelearningfrom SEs.

First,there is a needtorethink theemphasis on investigat-ingsingularSEs.Investigationsarepronetobecomestand-alone activities,disconnectedfromwiderorganisationalsafetypractices andlearningopportunities.[8,9,32]IntheNetherlands,asinother countries,‘theperimeter[s]ofpatientsafety’[15]keepexpanding asmoreeventsqualifyasSEs[16].Asbothhospitalrespondentsand inspectorsstrugglewiththeamountofSEsthathavetobe investi-gatedandassessed,acontinuedfocusonsingularSEsmightbecome untenable.Especiallyforhospitalsthatconsistentlydemonstrate theabilitytoadequatelyinvestigatesingularSEs,theIRSwoulddo welltoaccommodateanaggregatedlevelofanalysis,encouraging hospitalstoreflectonandlearnfromSEsinconnectiontotheir widersafetypoliciesandpractices[8,9,32].Second,thereisaneed tomovebeyondtheinvestigationpracticesandmonitorhow hos-pitalsuseSEstoimprovedailycarepractices.FollowingRamanujan andGoodman,theIRScanmonitorhowhospitalsworktolinkthe analysisofanSEwithlearningbyposingquestionsthataddress howlearningisshared,storedandenacted[14].Forexample:How didpatientsandfamiliescontributetoyourunderstandingofthe SE?Howdoyoulinkthelearningprocessoftheinvestigationteam totheprofessionals workingwiththeirsolutions? Howdo you institutionaliseandnormalisethesolutionsidentifiedsothatthey areusedinpractice?[14]Suchopenquestionsencourage hospi-talstoreflectonhowinvestigationpractices(ofsingularSEswhen thisiswarrantedoratanaggregatedlevel)aremeaningfultotheir safetypracticesandenablehospitalstodemonstrateownershipof thesepractices.

Ourstudyhassomelimitations.TheDutchIRS’sfocusonsocial and participativelearning of hospitalsfollowing SEs is unique anddevelopedin responsetoproblemsidentified inotherIRSs, sothat ourfindingsarespecifictotheDutchIRS.Still,howthe DutchIRS,asamonitoringinstrument,encouragesandgenerates particularorganisationalpracticesandinvestmentscanbe valu-ableforthedesignandcontinueddevelopmentofIRSsthathave adifferentfocus.Ourfindingscouldhavebeenstrengthenedby theperspectivesof SEinvolved healthcareprofessionalsas well aspatients.InourfocusonhowtheIRSencouragespracticesof socialand participativelearning,weforegrounded theaccounts ofincidentinvestigatorsandcommitteechairs;theprofessional groupsthat,inhospitals,organisetheinvestigativepracticesthat aimto supportsuch learning. By conceptualising learning as a socialandparticipativepractice,wewereabletodemonstratehow IRSscanencouragehospitalstodevelopvaluablepractices. Draw-ingfrombothquantitativeandqualitativedata,wewereableto generateaninsightfulunderstandingoftheeffectsoftheDutch IRS.

5. Conclusion

IRSscan encouragehospitals todevelop and investin prac-tices that contribute to social and participative learning from incidents. IRSs need to be dynamic to accommodate for the improvedlearningcapabilitiesofhealthcareprovidersand encour-agecontinuedimprovement.Whenproviderssucceedinmeeting the demands an IRS set, these demands should be adjusted towardsanextlevel.Continuouslyraisingthebaroraddingnew elementspreventsaplateaueffectthatwoulddiminishthe effec-tiveness of measures over time and stagnate further learning. Assessing and stimulating hospitals’ learning process with the aidof IRSsis a promising strategy, but itssuccessdepends on

theconsistent evaluation of itseffects and its further develop-ment.

Contributors

Theauthorscontributedtothemanuscriptasfollows.DdK,JK, KG,ILandRBcontributedtotheconceptionanddesignofthestudy. DdK,JK,KGandMVwereinvolvedindatacollection.DdK,JK,KG, ILandRBperformeddataanalysisandinterpretation.DdKandJK draftedtheinitialmanuscript.Themanuscriptwascriticallyrevised byDdK,JK,KG,IL,MVandRBandapprovedbyallauthorsbefore beingsubmitted.

Funding

Theresearchprojectsreportedonherewereconductedwithin and supportedby the DutchAcademic Collaborative Centre of Supervision,aresearchcollaborativethatpairsuptheInspectorate withfourresearchinstituteswiththeaimtostudyandenhance theeffectivityoftheInspectorate’sregulatorypractices.Thefirst project wasnot funded. The second project was supportedby ZonMw,theDutchorganisationforHealthResearchand Develop-ment,projectnumber516004604.

Patientconsentforpublication Notrequired.

Ethicsapproval

Theresearchpresenteddidnotrequireethicalapproval.Inthe Netherlands,researchthatdoesnotinvolvesubjectingparticipants tomedicalinterventionordoesnotdictateparticularcodesof con-ductforparticipantsrequiresnoapproval(ccmo.nl).Respondents consentedtobeinterviewed.Contributionswereanonymisedand respondentsweregiventheopportunitytoreview theirquoted material.

Declarationofcompetinginterest

ILandMVworkasinspectorsfortheDutchHealthandYouth CareInspectorate.BothILandMVhavebeenandareinvolvedin theprocessofdesigningtheIRSandMVispartofateamof inspec-torsscoringhospitalSEreports.ILandMVdidnotparticipatein qualitativedatacollection.

Acknowledgements

Ourworkhasbenefitedfromthehelpofothers.BasKluitenberg hasbeenagreathelpinunderstandingandanalysingthe quantita-tivedataontheSEreports.Anearlierdraftofthisarticlewasread byJan-WillemWeenink,whosethoughtfulandgenerousfeedback enabledustoimproveuponourearlierwork.

AppendixA. Supplementarydata

Supplementarydataassociatedwiththisarticlecanbefound,in theonlineversion,athttps://doi.org/10.1016/j.healthpol.2020.05. 018.

References

[1]KohnLT,CorriganJM,DonaldsonMS.Toerrishuman:buildingasaferhealth system.Washington,D.C:NationalAcademiesPress;1999,http://dx.doi.org/ 10.17226/9728.

[2]VincentC.Reportingandlearningsystems.In:VincentC,editor.Patientsafety. Chichester:WileyBlackwell;2010.p.75–95.

[3]BarachP.HowtheNHScanimprovesafetyandlearning.BMJ2000;320:1683–4, http://dx.doi.org/10.1136/bmj.320.7251.1683.

[4]HudsonP.Applyingthelessonsofhighriskindustriestohealthcare.Qualityand SafetyinHealthCare2003;12:7i–12,http://dx.doi.org/10.1136/qhc.12.suppl 1.i7.

[5]MitchellI,SchusterA,SmithK,etal.Patientsafetyincidentreporting:a qual-itativestudyofthoughtsandperceptionsofexperts15yearsafter‘ToErris Human’.BMJQuality&Safety2016;25:92–9, http://dx.doi.org/10.1136/bmjqs-2015-004405.

[6]HowellA-M, BurnsEm, HullL, etal. Internationalrecommendations for nationalpatientsafetyincidentreportingsystems:anexpertDelphi consensus-buildingprocess.BMJQuality&Safety2017;26:150–63,http://dx.doi.org/10. 1136/bmjqs-2015-004456.

[7]Macrae C. The problem with incident reporting. BMJ Quality & Safety 2016;25:71–5,http://dx.doi.org/10.1136/bmjqs-2015-004732.

[8]PeerallyMF,CarrS,WaringJ,etal.Theproblemwithrootcauseanalysis.BMJ Quality&Safety2016,http://dx.doi.org/10.1136/bmjqs-2016-005511, bmjqs-2016-005511.

[9]Stavropoulou C,DohertyC, ToseyP.Howeffective areincident-reporting systems for improving patient safety? A systematic literature review: incident-reportingsystemsforimprovingpatients’safety.MilbankQuarterly 2015;93:826–66,http://dx.doi.org/10.1111/1468-0009.12166.

[10]ShojaniaKg,Marang-vandeMheenPj.Temporaltrendsinpatientsafetyinthe Netherlands:reductionsinpreventableadverseeventsortheendofadverse eventsasausefulmetric?BMJQuality&Safety2015;24:541–4,http://dx.doi. org/10.1136/bmjqs-2015-004461.

[11]ShojaniaKG.Thefrustratingcaseofincident-reportingsystems.Qualityand Safety inHealthCare 2008;17:400–2,http://dx.doi.org/10.1136/qshc.2008. 029496.

[12]ShojaniaKG,ThomasEJ.Trendsinadverseeventsovertime:whyarewenot improving?BMJQuality&Safety2013;22:273–7,http://dx.doi.org/10.1136/ bmjqs-2013-001935.

[13]BainesRJ,LangelaanM,deBruijneMC,etal.Changesinadverseeventratesin hospitalsovertime:alongitudinalretrospectivepatientrecordreviewstudy. BMJQuality&Safety2013;22:290–8, http://dx.doi.org/10.1136/bmjqs-2012-001126.

[14]RamanujamR,GoodmanPS.Thechallengeofcollectivelearningfromevent analysis.SafetyScience2011;49:83–9,http://dx.doi.org/10.1016/j.ssci.2010. 03.019.

[15]VincentC,AmalbertiR.Safetyinhealthcareisamovingtarget.BMJQuality& Safety2015;24:539–40,http://dx.doi.org/10.1136/bmjqs-2015-004403. [16]LeistikowI,MulderS,VesseurJ,etal.Learningfromincidentsinhealthcare:the

journey,notthearrival,matters.BMJQuality&Safety2017;26:252–6,http:// dx.doi.org/10.1136/bmjqs-2015-004853.

[17]WaringJJ.Beyondblame:culturalbarrierstomedicalincidentreporting.Social Science&Medicine2005;60:1927–35,http://dx.doi.org/10.1016/j.socscimed. 2004.08.055.

[18]DutchHealthandYouthCareInspectorate.Meerjarenbeleidsplan2016-2019: gezondvertrouwen.Utrecht:DutchHealthandYouthCareInspectorate, Min-istryofHealth,WelfareandSport;2016.

[19]DutchHealthandYouthCareInspectorate.Richtlijncalamiteitenrapportage. Utrecht:DutchHealthandYouthCareInspectorate,MinistryofHealth,Welfare andSport;2016https://www.igj.nl/onderwerpen/calamiteiten/documenten/ richtlijnen/2016/01/01/richtlijn-calamiteitenrapportage.

[20]GreeneJC,BenjaminL,GoodyearL.Themeritsofmixingmethodsinevaluation. Evaluation2001;7:25–44,http://dx.doi.org/10.1177/13563890122209504. [21]JohnsonRB,OnwuegbuzieAJ,TurnerLA.Towardadefinitionofmixedmethods

research.JournalofMixedMethodsResearch2007;1:112–33,http://dx.doi. org/10.1177/1558689806298224.

[22]GreenJ,ThorogoodN.Qualitativemethodsforhealthresearch,4thedition ThousandOaks,CA:SAGEPublications;2018.

[23]VincentCA.Patientsafety:whataboutthepatient?QualityandSafetyinHealth Care2002;11:76–80,http://dx.doi.org/10.1136/qhc.11.1.76.

[24]NHSEngland,NHSImprovement.TheNHSpatientsafetystrategy.In:Safer culture, safer systems, safer patients; 2019 https://improvement.nhs.uk/ documents/5472/190708PatientSafetyStrategyforwebsitev2.pdf. [25]Iedema R, Allen S. Anatomy of an incident disclosure: the importance

of dialogue.TheJoint CommissionJournalonQuality andPatient Safety 2012;38:435–42,http://dx.doi.org/10.1016/S1553-7250(12)38057-4. [26]FitzsimonsB,CornwellJ.Whatcanwelearnfrompatients’perspectivesonthe

qualityandsafetyofhospitalcare?BMJQuality&Safety2018;27:671–2,http:// dx.doi.org/10.1136/bmjqs-2018-008106.

[27]O’HaraJK,ReynoldsC,MooreS,etal.Whatcanpatientstellusaboutthe qual-ityandsafetyofhospitalcare?FindingsfromaUKmulticentresurveystudy. BMJQuality&Safety2018;27:673–82, http://dx.doi.org/10.1136/bmjqs-2017-006974.

[28]Dahler-LarsenP.Constitutiveeffectsofperformanceindicators:gettingbeyond unintended consequences. Public Management Review 2014;16:969–86, http://dx.doi.org/10.1080/14719037.2013.770058.

[29]Wallenburg I,Quartz J,Bal R.Making hospitals governable: performativ-ity andinstitutional workin rankingpractices. Administration &Society 2019;51:637–63,http://dx.doi.org/10.1177/0095399716680054.

[30]NicoliniD,WaringJ,MengisJ.Thechallengesofundertakingrootcauseanalysis inhealthcare:aqualitativestudy.JournalofHealthServicesResearch&Policy 2011;16:34–41,http://dx.doi.org/10.1258/jhsrp.2010.010092.

[31]AndersonJE,KodateN,WaltersR,etal.Canincidentreportingimprovesafety? Healthcarepractitioners’viewsoftheeffectivenessofincidentreporting. Inter-nationalJournalforQualityinHealthCare2013;25:141–50,http://dx.doi.org/ 10.1093/intqhc/mzs081.

[32]HibbertPD,ThomasMJW,DeakinA,etal.Arerootcauseanalyses recommenda-tionseffectiveandsustainable?Anobservationalstudy.InternationalJournal for Qualityin HealthCare 2018, http://dx.doi.org/10.1093/intqhc/mzx181. PublishedOnlineFirst:16January.

[33]KokJ,LeistikowI,BalR.Patientandfamilyengagementinincident investiga-tions:exploringhospitalmanagerandincidentinvestigators’experiencesand challenges.JournalofHealthServicesResearch&Policy2018;23(4):252–61, http://dx.doi.org/10.1177/1355819618788586.

[34]deKamD,GritK,BalR.Sharedlearningfromincidents:Aqualitativestudyinto theperceivedvalueofanexternalchaironincidentinvestigationcommittees. SafetyScience2019;120:57–66,http://dx.doi.org/10.1016/j.ssci.2019.06.031. [35]Kok J, Leistikow I, Bal R. Pedagogyof regulation:Strategies and

instru-mentstosuperviselearningfromadverseevents.Regulation&Governance 2019;13:470–8,http://dx.doi.org/10.1111/rego.12242.

[36]GritK,deKamD,BouwmanR,HarmsenM,FrieleR,BalR.Kennissynthese Calamiteitentoezicht.Rotterdam:ErasmusUniversiteitRotterdam;2018. [37]KokJ,deKamD,LeistikowI,GritK,BalR.Epistemicinjusticeinincident

investigations:aqualitativestudy.HealthCareAnalysis2020.Submittedfor publication.