ContentslistsavailableatScienceDirect
International Journal of Nursing Studies
journal homepage:www.elsevier.com/ijns
Towards a better understanding of the relationship between feedback and nurses’ work engagement and burnout: A convergent
mixed-methods study on nurses’ attributions about the ‘why’ of feedback
A.P.M. (Suzanne) Giesbers
a,b, Roel L.J. Schouteten
b,∗, Erik Poutsma
b, Beatrice I.J.M. van der Heijden
b,c,g,h,i, Theo van Achterberg
d,e,faJeroen Bosch Hospital, PO Box 90153, NL-5200 ME ’s Hertogenbosch, the Netherlands
bRadboud University, Institute for Management Research, PO Box 9108, NL-6500 HK Nijmegen, the Netherlands
cOpen University of the Netherlands, School of Management, PO Box 2960, NL-6401 DL Heerlen, the Netherlands
dKU Leuven, Department of Public Health and Primary Care, Academic Centre for Nursing and Midwifery, Kapucijnenvoer 35, 30 0 0 Leuven, Belgium
eRadboud University Medical Centre, Radboud Institute for Health Sciences, Scientific Center for Quality of Healthcare, PO Box 9101, NL-6500 HB Nijmegen, the Netherlands
fUppsala University, Department of Public Health and Caring Sciences, Box 256, 751 05 Uppsala, Sweden
gGhent University, Faculty of Economics and Business Administration, 90 0 0 Ghent, Belgium
hHubei University, Hubei Business School, Wuhan, 368 Youyi Ave., Wuchang District, Wuhan 430062, China
iKingston University, Kingston Business School, London KT11LQ, United Kingdom
article info
Article history:
Received 14 March 2020
Received in revised form 21 January 2021 Accepted 22 January 2021
Keywords:
HR attributions Feedback
Feedback environment Quality measurement Nursing
Hospital
abstract
Background:Previousstudiesontheeffectsofprovidingfeedbackaboutqualityimprovementmeasuresto nursesshowmixedresultsandthefactorsexplainingthevarianceineffectsarenotyetwell-understood.
Oneofthefactorsthatcouldexplainthevarianceinoutcomesishownursesperceivethefeedback.Itis notthefeedbackpersethatinfluencesnurses,andconsequentlytheirperformance,butrathertheway thefeedbackisperceived.
Objectives:ThisarticleaimstoenhanceourunderstandingofHumanResourceattributionsandemployee engagementandburnoutinafeedbackenvironment.Anin-depthstudyofnurses’attributionsaboutthe
‘why’offeedbackonqualitymeasurements,anditsrelationtoengagementandburnout,wasperformed.
DesignandMethods:Aconvergentmixed-methods, multiplecasestudydesign wasused.Evidencewas drawnfromfourcomparablesurgicalwardswithinthreeteachinghospitalsintheNetherlandsthatvol- unteeredtoparticipateinthisstudy.Nursesoneachwardwereprovidedwithoralandwrittenfeedback onqualitymeasurementseverytwoweeks,overafourmonthperiod.Afterthisperiod,anonlinesurvey was distributedtoallthenurses(n =184)onthefourparticipatingwards.Datawerecollectedfrom 91nurses.Paralleltothesurvey,individual,semi-structuredface-to-faceinterviewswereconductedwith eightnursesandtheirwardmanagerineachward,resultingininterviewdatafrom32nursesandfour wardmanagers.
Results: Resultsshowthatnurses– bothasagroupandindividually– makevaryingattributionsabout theirmanagers’purposeinprovidingfeedbackonqualitymeasurements.Thefeedbackenvironment is associatedtonurses’attributionsandtheseattributionsarerelatedtonurses’burnout.
Conclusions:Byshowingthatfeedbackonqualitymeasurementscanbeattributeddifferentlybynurses andthatthefeedbackenvironmentplaysaroleinthis,thestudyprovidesaninterestingmechanismfor explaininghowfeedbackisrelatedtoperformance.Implicationsfortheory,practiceandfutureresearch arediscussed.
© 2021TheAuthor(s).PublishedbyElsevierLtd.
ThisisanopenaccessarticleundertheCCBY-NC-NDlicense (http://creativecommons.org/licenses/by-nc-nd/4.0/)
∗ Corresponding author.
E-mail address: r.schouteten@fm.ru.nl (R.L.J. Schouteten).
https://doi.org/10.1016/j.ijnurstu.2021.103889
0020-7489/© 2021 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license ( http://creativecommons.org/licenses/by-nc-nd/4.0/ )
Whatisalreadyknownaboutthetopic?
• Previous studies show variation in the association between feedback provision to nurses and outcomes including nurses’
engagementandburnoutandqualityofcare.
• Factorsexplainingthisvariationarenotyetwell-understood.
Whatthispaperadds
• Foroutcomesoffeedback itis importanttoconsider thepro- cess of how feedback on quality measurements to nursing teams working in a hospital setting is experienced by the nurses.
• Nursesappeartohavedifferentattributionsforthesame(type of) feedback, which resultin different associations withtheir engagementandburnout.
• A supportive feedback environment is positively related to nurses’attributionsaboutthewhyoffeedbackprovision.
1. Introduction 1.1. Background
Providingfeedback to nursing teams is an importantand fre- quently used strategy for improving clinical performance after qualitymeasurementsinhospitalcare(DeVosetal.,2009).Feed- backonperformanceisgenerallyusedtodrawhealthcareworkers attentionto gaps betweendesired andactual practice inpatient care, and can be defined as “delivering information about clinical performance provided to patient populations over a specified period of timeto professionals, practices or institutions, forthe purpose of improving the team’s or clinician’s insight into the quality of care they provide and improving it when possible” (Ivers et al., 2020).
Themechanismsofhowprovidingfeedbackonperformancewould leadtoimprovedperformance aretoooftenignoredinthelitera- tureon healthcare quality improvement (Tutiet al., 2017). How- ever,from behaviour changeliterature we knowthat feedback is a basic change method that relates to several theories on learn- ingandgoalsetting(Koketal.,2016),withthemostlikelymech- anisms being: 1) that feedback on performance triggers positive change through creating awareness of suboptimal performance;
and 2) that positive feedback in case of improved performance over time can be rewarding and thus stimulate further improve- ment.
Studiesontheeffectsoffeedbackonperformancegenerallyin- dicate that thistype of feedback renders small to moderate im- provements, and that effects can be highlyvariable (Ivers et al., 2014, Tuti et al., 2017). However, factors explaining the variance in effects are not yet well-understood (Christina et al., 2016; Giesbers et al., 2016; Sykes et al., 2018). For instance, whereas Mead et al. (1997) gathered structured evidence that feedback is strongly associated to improved clinical practice, research by McCann etal. (2015) highlighted that professional discretion has been increasingly sundered by a narrow focus on “making the numbers” (ibid., p. 787), resulting in dysfunctional outcomes for workforcemorale.
Such variation in findings may result from a lack of strong guiding theoretical frameworks to study the effects of feedback (Christina et al., 2016). In a systematic review of qualitative re- searchonfeedbackinhealthcare,Brownetal.(2019)developeda theoryforexplainingfactorsthatinfluencefeedbacksuccess.From thistheory,itisevident that feedbackiscomplex andthatmany variablesandtheirmutualconnectionsmightplayimportantroles.
In particular, Brown etal. (2019) distinguished three main kinds of variables: feedback variables (content of feedback andway of delivery), recipient variables (healthcare professional characteris- ticsandbehaviouralresponse)andcontextvariables(organization
characteristics, teamcharacteristics,andimplementation process).
WithinBrownetal.’s(2019)theoretical framework,thisstudyfo- cusesonfeedbackonqualitymeasurements,liketheratesoffalls andtheincidenceofpressureulcers(feedbackvariable), howthis feedbackisperceivedbynurses(recipientvariable)andtheroleof thefeedbackenvironment(contextvariable).
In order to provide structure and direction for the study (Christinaetal., 2016), weposit that perceptionsoffeedbackcan beconsideredtoaffectnurses’behaviourandperformance.Inpar- ticular, it is not the feedback per se that influences nurses, and consequentlytheirperformance,butratherthewaythefeedbackis perceived(e.g.,BowenandOstroff,2004;WrightandNishii,2013).
Especially important for nurses’ perceptions of feedback is the idea that nurses themselves haveregarding thewhy ofthe feed- back,i.e., theattributionsnursesmakeabouttheir manager’spur- pose in providing feedback (Nishii et al., 2008). Although previ- ously scholars already underlined the importance of attributions tounderstandtheimpactofsuchpracticesonemployeeoutcomes (e.g., Peccei et al., 2013; Woodrow and Guest, 2014; Wright and Nishii,2013),sofar, littleempiricalresearchhasbeenundertaken on the impact of attributions of managers’ reasons for feedback practicesonemployeeoutcomes.
A second factor explainingnurses’ perceptionsoffeedback on quality measurements is the feedback environment. The feed- back environment, also called feedback culture (London and Smither, 2002), refers to the overall supportiveness for feedback intheworkplace(Steelmanetal.,2004).Previousresearchshowed thatthefeedbackenvironmentinfluenceshowemployeesperceive feedbackinterventions(Dahlingetal., 2012; Wellsetal.,2007). A feedback environment wherein feedback is properly framed may impacthowemployeesperceivethemotivationforprovidingfeed- back(seealsoIlgenandDavis,2000;Wellsetal.,2007).Afocuson feedback environment entailsincluding the relationship between (ward)managersandnursesasanimportantelementofthefeed- backenvironment,becausemanagersareconsideredtoplayasig- nificant role influencing nurses’ experiences and behaviour, and, therefore, on the quality of safety andcare (Adriaenssens et al., 2017).
The aim of thisstudyis to provide a better understanding of nurses’ attributions about the reasons for providing them with feedback andtherole of wardmanagersincreating asupportive feedbackenvironment,inordertoexplainhowprovidingfeedback onqualitymeasurements tonursingteamsinahospitalsettingis relatedtonurses’engagementandburnout.Followingthereason- ingunderlyingtheJobDemands–Resourcesframework(Bakkerand Demerouti,2017;Bakkeretal., 2014; Demeroutietal.,2001),the effectsofthefeedbackinterventionaremeasuredintermsoftwo specific outcomesthat are important in the light of nurses’ per- formance; work engagement and burnout. Research emphasized theimportance ofthepossiblemediatingroleofengagementand burnout in the relationship betweennursing work environments and outcomes (Laschinger and Leiter, 2006; Van Bogaert et al., 2013). Hence, our research question is: What is the impact of nurses’ attributions ofthe manager’s reasons forproviding them withfeedbackontheir engagementandburnout, andwhatisthe roleoffeedbackenvironmentinthisrelationship?
1.2. Nurses’attributionsaboutthe‘why’offeedback
In times of change, employees will engage in explicit ef- forts of sense making (Weick et al., 2005). Since an interven- tion, such as implementing feedback on quality measurements to nursing teams, comprises a changeprocess, we expect nurses to attempt to make sense of why this feedback is provided to them. This process of sense making is not about the truth and getting it right, but about the development of plausible ‘stories’
(Weicketal.,2005).Weexpectthatnursesmayhavedifferent‘sto- ries’orexplanationsregardingthereasonsforprovidingthemwith feedbackonqualitymeasurements,dependingupontheirinterpre- tations of the purpose of the manager who provided the feed- back. Wells et al. (2007) recognize a similar difference between theintendedandperceivedpurposeofperformancemonitoringto nurses, resulting indiffering explanations forthe purpose of the feedback.Tobetterunderstandnurses’differentexplanations,this articlebuilds onattribution theory and, morespecifically, on the modelofHRattributionsdevelopedbyNishiietal.(2008).Wear- guethatthismodelisrelevant,becauseitisapplicabletoallkinds ofinterventionsforwhichemployee perceptions(i.e.,attributions) connotean importantmechanismforexplainingemployeebehav- ior.Specifically,themodelbyNishiietal.(2008)providesauseful lensformappingvariousattributionsemployeescanmakeregard- inginterventionsinaworkcontextandexplainingemployees’re- actionstothoseinterventions(Alfesetal.,2020).
Research on attributions examines the causal explana- tions people make for their own and others’ behaviours (Kelley, 1973). Inspired by the principles of attribution theory, Nishii et al. (2008) introduced their theoretical model of HR attributions. HR attributions are defined as causal explanations that employees make regarding management’spurposes in using particular practices.Building on Koys’ (1991) work, themodel of Nishii et al. (2008) distinguishes between internal and external HR attributions. Internal HR attributions refer to the perception that HR practices are adopted due to factors for which man- agement is responsible, or factors over which management has control. External HR attributions refer to the perception that HR practicesareadoptedbecausemanagementhasto,duetoexternal constraints. Additionally, Nishii et al. (2008) drew a distinction betweeninternalcommitment-focused HRattributionsthatconnote positive consequences for employees and internal control-focused HRattributionsthatconnotenegativeconsequencesforemployees.
The question that follows is: Which different internal commitment-focused, internal control-focused and external attributions donurses makeabouttheir wardmanager’s purpose inprovidingfeedbackonqualitymeasurements?First,nursesmay believe that their manager’s purpose is to support the nursing teaminitsqualityimprovementendeavour,tomonitorthequality ofcare on the ward,and/or to improve quality-relatedoutcomes forpatients. Thisattributionis consistentwiththe broadlybased ideathatfeedbackallowsprofessionalstobecomeawareoftheir– potentiallysuboptimal-performance,whichmayencouragethem to adjust their behaviour (Flottorp et al., 2010). Second, nurses may believe that it is their manager’s purpose to make nurses’
workmoreattractiveandchallenging.Byinformingnursingteams ontheresultsfromqualitymeasurements,thenursesmaybecome more involved in quality improvement possibly resulting into a more professional work environment. The above attributions are all related to internal, commitment-focused factors and we label themas‘Qualityandnurseenhancementattributions’.
Nurses can also attribute feedback provision on quality mea- surements to different internal, control-focused factors. For in- stance,nursesmaybelievethattheirmanager’spurposeistomake thenursesworkharderortogivethemextrawork,herewithpush- ing themtowards quality improvementobjectivesand/or cost re- duction. We label thistype ofattributions as ‘Cost reduction and nurseexploitationattributions’.
Finally, nurses may attribute feedback provision on quality measurements to different external factors (e.g., healthcare in- spectorates, budget, and pay for performance arrangements etc.) because the introduction of feedback on quality measurements within hospitals is often driven by healthcare reform programs, based onNew Public Management ideology – a rangeof emerg- ing social policy ideas that generally sought to combinethe dy-
namism and customer orientation of the private sector withthe service ethic that is traditionally inherent in the public sector (Hood, 1991). First, nurses maybelieve that their manager’s pur- poseinprovidingfeedbackistoadheretosocietalnormsontrans- parency.Second, nursesmaybelieve thattheirmanager’spurpose istobetteradheretothequalitystandardsimposedonthehospi- tal by organizationslikethehealthcare inspectorateorhealth in- surers.Welabelthesekindsofexternalattributionsas‘Compliance attributions’.
1.3. Nurses’attributionsandtheireffectsonnurses’engagementand burnout
In the Job Demands–Resources theoretical framework (BakkerandDemerouti,2017;Bakkeretal.,2014;Demeroutietal., 2001), work engagement and burnout are central variables ex- plainingjob performance.Workengagementandburnoutaretwo individual outcome variables that representpossible positive and negative effects one’s work and work organization can have on employees. Work engagement is characterized by a high level of energy andstrong identification withone’s job (Bakkeret al., 2014). Burnout,onthe other hand,is characterizedby low levels of energy and poor identification with one’s job (Bakker et al., 2014). These individual-level outcomes mayhave important con- sequences for individual employees as well as for organizations, such ashealthoutcomes(e.g.,depression),motivationaloutcomes (e.g., happiness), and job performance (e.g., customer or patient satisfactionororganizationalcitizenshipbehaviour)(Bakkeretal., 2014).
Research on HR attributions has demonstrated that employ- ees may make varying attributions for the same HR practices (Alfes et al., 2020), and that these attributions are differentially associated with employee outcomes, such as commitment,satis- faction,job strainandengagement(Alfesetal.,2020;Koys, 1991; Nishii et al., 2008, Van de Voorde and Beijer, 2015). Both Nishii et al. (2008) and Van de Voorde and Beijer(2015) found empirical support for a positive relationship between internal, commitment-focusedattributionsandemployeeoutcomes,andfor a negative relationship betweeninternal, control-focused attribu- tions andemployee outcomes. Similarly, Alfesetal.(2020) found evidence for a positive relationship between HR well-being at- tributions, being employees’ interpretation that the organization cares aboutthem, andemployee engagement.They also found a negative relationship between HR performance attributions, that is employees’ interpretation that the organization focuses on highlyefficientwork,andemployeeengagement.Koys(1991) and Nishiietal.(2008),intheir researchontheeffectsofexternalHR attributions on commitmentand satisfaction reported no signifi- cantresults.AccordingtoNishiietal.(2008),externalattributions are unrelated to employee commitment andsatisfaction because employees donot attribute meaningfuldispositional explanations (i.e.,explanationsintermsofinternalfactorswhicharespecificto the management) to management’s effort to comply with exter- nal constraints.However,employeesmayfeel pressuredby exter- nalrequirements,withouthavinganyinfluenceonthese,andthis mayleadtoanegativeeffectonemployeeattitudesandoutcomes.
However, the meta-study by Harvey etal. (2014) showsthat ex- ternalattributionsarelessinfluentialforemployees’attitudesand behavioursthaninternalattributions.
Relyingontheabove,weexpecttofind:(1)apositiverelation- ship between‘Qualityandnurseenhancement attributionsofthe whyoffeedback’andnurses’workengagement,and(2)anegative relationship between‘Costreduction andnurseexploitation attri- butions’ andnurses’work engagement.Forburnout, theseexpec- tationsaremirrored,i.e.,anegativerelationshipwith‘Qualityand nurse enhancement attributions’ anda positive relationship with
‘Cost reduction andnurse exploitation attributions’. Ourexpecta- tions regarding (3) the association between ‘Compliance attribu- tions’andnurses’engagementandburnout,isinitiallyindifferent.
1.4. Theinfluenceofthefeedbackenvironmentonnurses’attributions
Several scholars haveunderlined the importance of the orga- nizationalcontext tobetter understanddifferencesinHR attribu- tions(Nishiietal.,2008;VandeVoordeandBeijer,2015).Accord- ingly,researchaboutsensemakinghasindicatedthat‘stories’tend to be seenas plausiblewhen they tapinto an existing organiza- tional context (Weick et al., 2005). In this article,we investigate howthefeedbackenvironment setbytheward manager(thesu- pervisor feedbackenvironment, hereafter referred to as‘feedback environment’) influencesnurses’ attributionsaboutthemanager’s reasonsforprovidingfeedbackonqualitymeasurements.Following Steelmanetal.(2004),thefeedbackenvironmentis characterized by theperceived credibility ofthe supervisorasfeedback source, thequality of thefeedback, the tactfulnesswithwhich thefeed- backisprovided,theextenttowhichfavourableandunfavourable feedback isprovided, the availability of feedback, andthe extent to which feedback-seeking behaviour is promoted. A supportive feedback environment is one in which high-quality feedback is provided by the supervisor ina tactful andconstructive manner.
Dahlinget al.(2012) found empirical supportfor theproposition thatwithinasupportivefeedbackenvironment,employeeswillde- velop,among other things, a positive view offeedback, a lackof apprehension towardfeedback, abelief that feedbackis valuable, anda sense ofaccountability to act on thefeedback that ispro- vided.
We expect to find:(1) a positive relationship betweena sup- portive feedback environment andattributions that connote pos- itive consequencesfor nurses, being‘Quality andnurse enhance- mentattributions’.Inaddition,weassume:(2)anegativerelation- shipbetweenasupportivefeedbackenvironmentandattributions thatconnotenegativeconsequencesfornurses,being‘Cost reduc- tionandnurseexploitationattributions’.Lastly,wehavenoexpec- tations regarding thedirection of the relationship betweenfeed- backenvironmentandcomplianceattributions.
Feedback environment may also be a moderator for the rela- tionshipbetweenattributionsthatnursesmakeandtheirengage- mentandburnout.Apositivefeedbackenvironmentmayenhance thepositiveeffectof‘qualityandnurseenhancementattributions’
on work engagement, while it may decrease the positive effect of‘costreductionandnurseexploitationattributions’onburnout.
Fig.1depictsthe conceptualframework, summarizingtheexpec- tationsinthisstudy.
2. Method
Our study employed a convergent mixed-methods, multiple casestudy design (Creswell, 2015), in which the qualitative data are used for interpreting the quantitative data (according to the conventionofreportingofmixed-methodsstudies,thisstudyisa
‘QUANT-qual’studywherequalitativedataisusedtointerpretthe resultsofthequantitativestudy(Creswelletal.,2011;Fettersand Freshwater,2015).This designprovideduswitha morecomplete understanding than using either a quantitative or a qualitative design (Creswell, 2015; Östlund et al., 2011) and is increasingly recognized for improvingour understanding of the HRM process (WoodrowandGuest,2014).First,thedesignprovideduswiththe opportunitytoestablishwhetherrelationshipsbetweennurses’at- tributions, their engagement and burnout, and the feedback en- vironment were statisticallysignificant, andhelped usto findan explanation of why such relationships occurred. Second, the de- sign revealed the complexityof nurses’ attributions and enabled
Fig. 1. Conceptual framework underlying the study.
a deeper understanding ofthem. Third,the design enabledusto cross-check ourdataaboutnurses’ attributionsaboutthereasons for providing them withfeedback on quality measurements, en- hancing our confidencein the validity and reliabilityof the out- comes.
Our study draws on evidence from four comparable hospital wards. The nurses on each ward were, regularly provided with feedback on quality measurements duringa fourmonths’ period.
Inthefollowingparagraphs,we willaddressthestepstakenwith regard tothewardselection,thefeedbackintervention,thequan- titativeandqualitativedatacollectionandthedataanalyses.
2.1. Wardselection
For reasonsof comparability,we included onlysurgical wards from one type of hospital, i.e., general, teaching hospitalsin the Netherlands. Tobe able toproperly studyour feedbackinterven- tion,weincludedonlywardswherenurseswerenotprovidedwith regularfeedbackonqualitymeasurementsbefore.Basedonconve- nience sampling,wefound fourwards withinthreedifferenthos- pitalsthatvolunteeredtoparticipateinthisstudy.Thesehospitals areall associatedinacooperationnetworktodevelopsimilarini- tiatives forimprovingthe quality ofhealthcare they deliver.The feedback intervention that we studied wasthe result of a coop- erativeinitiativein thisnetwork. Thehospitalsinourstudywere institutions with the number of beds ranging from 643 to 1070 andwiththenumberofstaff (fte)rangingfrom2640to2915.The numberofnursesworkingontheparticipatingwardsrangedfrom 29 to 69.The participating wards housed patientsfrom different surgical, medicalspecialties. The first ward housedpatients from neurosurgery andorthopaedics,thesecond ward housed patients from lung surgery, the third ward housed patients from general surgery, andthefourthward housedpatientsfromurology, plas- ticsurgeryandgynaecology.
2.2. Feedbackintervention
Based onexistingliteraturethat evaluatestheeffectsofdiffer- ent feedback characteristics, the first author developed a frame- work for the designof feedback on each participatingward. The framework implied that duringafourmonths’ period,thenurses on eachward wereatleast onceevery twoweeksprovided with oral andwrittenfeedbackonqualitymeasurementsatteamlevel,
linked to aclearlycommunicated target. The ward managersub- sequentlydetermined whichqualitymeasurementswere selected, which target was set, how the quality measurements were car- ried out, and exactly when and how feedback was provided to thenurses.Examplesoftheselectedqualitymeasurementsarethe percentageofpatients screenedforthe risk orexistence ofpres- sureulcersatadmissionandthepercentageofpatientswithself- reportedpainscoresgreaterthanseven(onascaleofzerototen).
Allquality measurements were establishedinthe wards,buthad notbeenusedforprovidingfeedbacktonursesbefore.Thenurses ontheparticipatingwardswereinformedaboutthefeedbackchar- acteristics (quality measurements, source, format, frequency) by their ward manager, who also explained that the feedback was aimedatchangingtheirworkbehaviour.Additionally,themanager informedthe nurses aboutthescientific studyinto theeffects of providingthiskindoffeedback.Thefeedbackonqualitymeasure- ments,asintendedby thewardmanagersatthebeginningofthe fourmonths’period,wascomparableforthedifferentwards.
To ensure that the feedback on quality measurements as in- tended matched the feedback as implemented (Woodrow and Guest, 2014; WrightandNishii, 2013), the firstauthorconducted several on-site observations during the four months’ period of feedback provision. As intended, in all wards, the feedback was providedbythewardmanageroraseniornurse.Thewrittenfeed- backwasprovidedintheformofaposterintheteamroom(two wards)and/or asan attachment to a weekly orbi-weeklye-mail (threewards).Inallwards,thecontentofthefeedbackcontained thescoresregardingtheincidenceofpressureulcersandtheper- centagesofpatientswhoexperiencedseverepain.Otherscores,in- cludedinthewrittenfeedback,includethepercentageofpatients screened for risk of malnutrition (three wards), frailty in elderly (twowards),delirium(oneward),oracuteillness(oneward).With respecttothefrequencyoforalfeedback,inconsistencieswiththe feedbackasintendedwere foundontwoofthewards.Theinten- tionwastoprovideoralfeedbackonabi-weeklybasis.However,in twowards,oralfeedbacktothenursesonthesewards,intheform of presentationand discussion during teammeetings or debrief- ings,onlyhappenedoccasionally,whereasintheothertwowards thiskindoforalfeedbacktookplaceatleasteverytwoweeks.
2.3. Quantitativedatacollectionandanalysis
Afterthefourmonths’periodduringwhichregularfeedbackon qualitymeasurementswasprovidedtothenurses,anonlinesurvey wasdistributedtoallthenurses(n=184)onthefourparticipat- ing wards. The wardmanagers together withthe first authorin- formedthe nursesaboutthepurposeofthestudyandmotivated themtofilloutthesurvey.Datawerecollectedfrom91nurses,re- sultinginaresponserateof49.46%.Theaverageageinoursample was37.86years(SD= 11.30)and89.25% werefemales. Theaver- age tenure in the organization was12.59 years, and the average tenure asaqualified nurse was14.35 years.These characteristics ofoursamplearecomparabletothecharacteristicsoftheBIGreg- ister inwhichall active qualified nursesin theNetherlands need tobe registered(CIBG,2021).Inthe BIGregister, theaverageage is43years,themale/femaleratiois13/87andtheaveragetenure asaqualified nurseis 14years.Therefore,it isassumedthat the datasetisrepresentativeofthesamplepopulationinthepartici- patingwards.
MeasuresForallmeasures,seven-pointLikertscaleswereused, rangingfromstrongly disagree/never (1)to stronglyagree/always (7).
• Nurses’ attributions. Building on the model of Nishii et al. (2008), we developed a measure on nurses’
attributions abouttheir ward manager’s purpose in providing
feedback on quality measurements. We pilot-tested our mea- surein two rounds. Ina first round,several practitioners and scholars were asked to provide feedback on the content and wordingoftheitems. Inasecondround,dataonthefeedback measure was collectedfrom 55 nurses who did not work on thewards includedforthisarticle.Inthesecond round,some questionsregardingthecomprehensibilityandcompletenessof ourmeasure were added.Thisresulted ina validandreliable measurethatwasusedforthisstudy1.
For this study we validated the developed measure and con- ducted an exploratory factor analysis using varimax rotation for the itemsrelatedto nurses’attributions.Three factorshadEigen- valuesaboveone(withatotalexplainedvariance61percent)and appeared tocorrespondwiththetypologyofthreeattributiondi- mensions.Thereliabilityforalldimensionswasabovetheaccept- able limit of0.60 forexploratoryresearch (Hairetal., 1998); (1)
‘Qualityandnurseenhancementattributions’(
α
=0.72;4items);(2)‘Cost reductionandnurseexploitationattributions’(
α
= 0.72;3 items); and (3) ‘Compliance attributions’ (
α
= 0.69; 4 items).Example itemsforthesedimensions respectivelyare: “I believe I amprovidedwithfeedbackonqualitymeasurements,becausemy wardmanageraimstoimprovethequalityofpatient care”,“Ibe- lieve Iam provided with feedbackon quality measurements, be- causemywardmanagerwantstomakenurses’workmoreattract- ingandchallenging” (Qualityandnurseenhancementattributions),
“Ibelieve Iam providedwithfeedbackonquality measurements, becausemywardmanagerwanttomakethenursesworkharder”,
“Ibelieve Iam providedwithfeedbackonquality measurements, becausemywardmanagerwantstogivenursesextrawork” (Cost reduction andnurse exploitationattributions)and“IbelieveIam providedwithfeedbackonqualitymeasurementsbecausethehos- pital needs to adhere to quality standardsby the healthcare in- spectorate” (Compliance attributions).Qualityandnurseenhance- mentattributions,andcostreductionandnurseexploitationattri- butionsare groupedtogether since thedistinction betweenthese attributions was not supported by empirical datain previous re- search(Giesbersetal.,2014;Nishiietal.,2008).
• Work engagement comprises a positive, fulfilling work-related state of mind that is characterized by vigour, dedication, and absorption(SchaufeliandBakker,2003).Inthisstudy,worken- gagementwasmeasuredwiththeshortversionoftheUtrecht WorkEngagementScale(UWES-9;SchaufeliandBakker,2003).
An item example was: “I am enthusiastic about my work”.
Cronbach’salphafortheUWESdatainourstudywas0.87.
• Burnoutisdescribedasastateofmentalwearinessthatischar- acterizedbycynicism,exhaustion andlowprofessionalefficacy (SchaufeliandBakker,2004). Burnout wasmeasuredwiththe UtrechtBurnoutScale(UBOS);theDutchversionoftheMaslach Burnout Inventory-General Survey. An item example was: “I feelmentallyexhaustedbymywork”.Cronbach’salphaforthe UBOSwas0.84inourstudy.
• Supervisor feedback environment. Steelman etal. (2004) devel- oped a measure for the feedback environment set by the su- pervisor: the Supervisor Feedback Environment Scale (SFES).
We usedtheshortversion oftheSFES by Rosenetal.(2006). ThisshortversionwastranslatedintoDutchusingthevalidated DutchfullversionoftheSFESofAnseelandLievens(2007).The 18-item short version of the SFES characterizes the feedback environment by source credibility, feedback quality, feedback delivery,providingfavourablefeedback,providingunfavourable feedback, source availability and promoting feedback seeking.
An item example was: “I regularly receive positive feedback
1The pilot study was published.
Table 1
Descriptive statistics of main variables.
Mean Std. deviation Minimum Maximum
Gender (ref. = male) 0 .89 0 .31 0 1
Age (years) 38 .00 11 .30 21 64
Tenure as a qualified nurse (years) 14 .46 11 .29 1 42
Tenure in current hospital (years) 12 .72 10 .23 1 41
Hours per week 27 .81 6 .50 10 36
Supervisor Feedback 5 .15 0 .83 1 .18 6 .72
Burnout 2 .61 0 .67 1 .25 4 .88
Work Engagement 5 .53 0 .75 3 .78 7 .00
Compliance attributions 5 .79 0 .79 3 .75 7 .00
Cost reduction and nurse exploitation attributions 3 .11 1 .21 1 .00 6 .00
Quality and nurse enhancement attributions 4 .85 0 .88 1 .25 6 .25
from my ward manager”. Cronbach’s alpha for the SFES was 0.90inourstudy.
2.4. Quantitativeanalyses
To examine the differencesbetween the different wards with regardtonurses’attributionsabouttheir wardmanager’spurpose inprovidingfeedback,nurses’ engagement,burnoutandthefeed- back environment, an Oneway ANOVA test wasconducted onall studyvariables, followed by a Scheffé post-hoc comparison,hav- ingtheadvantageofbeingconservative.TheScheffé post-hoccom- parisonbetweenthemeansofallstudyvariablesonthedifferent wardsshowedthatnoneofthemeansweresignificantlydifferent (p>0.05).Forthisreason,wedidnotcontrolforwardsinfurther analyses.Therelationshipbetweennurses’attributionsandnurses’
engagement and burnout was examined using linear regression analysis.Linear regressionanalysiswasalso used toexamine the relationshipbetweenthefeedback environmentandnurses’attri- butions. Inaddition,we analysed thepossible moderatingrole of feedbackenvironmentintherelationshipbetweennurses’attribu- tionsandengagementandburnout.However,noneoftheinterac- tions were significant and we decided to presentonly the direct effects.
We used R-square and adjusted R-square to determine the amountofvariationexplained.TheFstatisticwasusedtotestthe significanceofthemodel.The5%levelofsignificancewasusedto determinewhetherthenullhypotheseswereacceptedorrejected.
Wecontrolledforgender,age,tenureasaqualifiednurse(mea- suredinyears)andtenureincurrenthospital(measuredinyears), aswellasworkinghoursperweek(measuredinthequestionnaire asaverageworkinghours per week).Table1presentsdescriptive statisticsofthemainvariables.
2.5. Qualitativedatacollectionandanalysis
After the four months’ period during which regular feedback on quality measurements wasprovided to the nurses, individual, semi-structuredface-to-faceinterviewswereconductedbythefirst author witheight nurses and their ward manager in each ward.
Thenurseswereselectedbythewardmanagerfromallthenurses workingonone specificdaythatwasindicated bytheresearcher.
The researcherrequested the wardmanager to take intoaccount thenurses’genderandageatthisselection,inordertosafeguard a representative sampling strategy. This resulted in a total of 32 nursesandfour wardmanagersbeinginterviewed.Out ofthe32 nurses,27werefemalesandfiveweremales,andtheiraverageage was32.93years(SD=11.66).Outofthefourwardmanagers,three werefemalesandonewasmale.Theinterviewswereconductedat theworkplace andcovered threekey areas:howrespondentsex- periencedthe feedback on quality measurements;what they be- lievedtobetheeffectoffeedback;andthecausalexplanationsre-
gardingthewardmanager’spurposeinusingfeedback2.Interviews lasted between10and40 minutes,with20 minutes,on average.
All participants consented to the interviews being recorded, and all fullinterviewsweretranscribedverbatim.Participantdatawas anonymisedusingtwo-digitcodes.Toanalysethedataforthisarti- cle,contentanalysiswasconductedcontainingthreecyclesofcod- ing,usingAtlas.tisoftwarepackage.Phaseonefocusedonidentify- ingattributionsregardingnurses’perceptionsaboutwhyfeedback was being provided to them. Phase two focused on categorizing thefoundattributionsviaadeductiveapproach.Thisimpliedthat theattributions,followingNishiietal.(2008)framework,werecat- egorizedas‘Qualityandnurseenhancementattributions’,‘Costre- ductionandnurseexploitationattributions’or‘Complianceattribu- tions’. Phase three consisted ofidentifying relationships between the different attributions and explanations for the findings from the quantitative data.Additionally,we formulated agrid to com- pare thedatafromthe differentwards andhospitals. Forcalibra- tion purposes, two interviews were coded independently by the first three authors followed by a thorough discussion of its out- comes.
2.6. Ethicalcode
No formal ethicalapprovalwasneededforthisstudy,because it wasnot within thescopeoftheNetherlands’ MedicalResearch InvolvingHumanSubjectsAct(CentralCommitteeonResearchIn- volvingHumanSubjects,2016).Theresearchershaveconsultedthe
“Ethical Principles of Psychologists and Code of Conduct” (APA, 2002) andhavecompliedwiththeethicalguidelinesoftheinsti- tutionswheretheresearchwasconducted.Informedconsentfrom allparticipantshasbeenobtained.
3. Results
3.1. Nurses’attributionsaboutthe‘why’offeedback
Weusedboththesurveyandinterviewdatatoexploretheat- tributionsnursesmakeabouttheirwardmanager’spurposeinpro- vidingfeedbackonquality measurements.First,we examinedthe descriptive statistics and correlationsdisplayed in Table 2.These results revealedthatnurses asagroup make varyingattributions abouttheirwardmanager’spurposeinprovidingfeedbackonqual- ity measurements. The ‘Compliance attributions’, appeared to be most prevalent. Simultaneously, but to a lesser degree, ‘Quality andnurse enhancementattributions’ cameforwardfromthe sur- veydata.Thesurveydatashowedasignificantcorrelationbetween the‘Complianceattributions’and‘Qualityandnurseenhancement attributions’(seeTable2).The‘Costreductionandnurseexploita- tion attributions’ didnot come forwardstrongly fromthe survey
2This study builds mainly on the third key area. The first two were used for another study (published earlier).
Table 2
Pearson’s r correlations based on the survey data ( N = 91).
α Mean SD 1 2 3 4 5
1 Quality and nurse enhancement attribution 0.72 4.85 0.88
2 Cost reduction and nurse exploitation attribution 0.72 3.11 1.21 −0 .03
3 Compliance attribution 0.69 5.79 0.79 0 .24 ∗ 0 .13
4 Feedback environment a 0.90 5.15 0.83 0 .49 ∗∗ −0 .22 ∗ 0.13
5 Work engagement 0.87 5.53 0.75 0 .19 ∗ 0 .01 0.00 0 .15
6 Burnout 0.84 2.61 0.67 −0 .15 0 .18 ∗ 0.25 ∗∗ −0 .24 ∗∗ −0.59 ∗∗
α= Cronbach’s alpha.
∗p < 0.05 ∗∗p < 0.01 (1-tailed).
a higher scores indicate a more supportive feedback environment.
data.Ingeneral,nursesappearednottobelievethattheywerepro- videdwithfeedbackonqualitymeasurementsbecausetheir ward managerwanted toreducecosts and/or tomakethenurses work harder.
Second,weexamined theinterviewdatatoexplorenurses’at- tributions aboutthe‘why’ offeedback. Comparableto thesurvey results,theinterview datarevealedthat nurses makeboth ‘Com- pliance attributions’ and ‘Quality and nurse enhancement attri- butions’. However, in contrast to the survey results,‘Quality and nurseenhancementattributions’cameforwardmoststronglydur- ingthe interviews.Whenlooking morecloselyatnurses’‘Quality and nurse enhancement attributions’, it seems that thesenurses emphasizedqualityenhancement,andnotnurseenhancement.Ac- tually,duringnone ofthe interviews,the nurses attributedfeed- backonqualitymeasurementstotheirmanager’spurposetomake nurses’ work more attractive andchallenging. Onlya few nurses expressedattributionsthatcouldbecategorizedas‘Costreduction andnurseexploitationattributions’.Theinterviewexcerptsbelow (includingareferencetotheparticipant’scode,jobandward)cap- turetheabove-mentionedtypesofdifferentattributions.Theseex- cerpts also illustrate how one nurse can make a diversity of at- tributions covering multiple attributiondimensions. Forexample, participant23describedhowshebelievedthatfeedbackonqual- itymeasurementsisaimedatbothqualityimprovement-a‘Qual- ityandnurseenhancementattribution’-andcostcontrol-a‘Cost reductionandnurseexploitationattribution’.
Quality and nurse enhancement attribution: “I believe the aim was to bring these things [quality measurements] to the team’sattention. Like‘guys, payattentiontothisandthat’.
Toprevent things. To provide better care.” (participant 33, nurse,ward2)
Cost reduction and nurse exploitation attribution: “The aim is mainly to improve the quality of care. […] It [feedback on quality measurements] is also a way to control your costs.Patientswithpressureulcersorbadmalnutritionwill costmuchmorethan apatientwho walksoutthe hospital whistling.” (participant23,nurse,ward1)
Compliance attribution: “These[quality measurements]are im- portant items a hospital is assessed on, so to say. I think that whenthey looked at howwe were performing, it be- cameclearthatthereismuchroomforimprovement.” (par- ticipant02,nurse,ward3)
During the interviewsthe majority of thenurses appeared to simultaneously make ‘Compliance attributions’ and ‘Quality and nurse enhancement attributions’, which explains the significant correlation fromthesurvey data betweenthesedifferentattribu- tions (seeTable 2).The nurses haddifferentexplanations ofhow
‘Complianceattributions’and‘Qualityandnurseenhancementat- tributions’are linked. Forexample,thefollowingnurse explained that she believedthat compliance with external requirements is alsointheinterestofthequalityofpatientcare:
“I believe it is related to each other: it [performing well on quality measurements] is an obligation from the government, butintheendyouwouldn’tdoitifthepatienthasnointerest inthematter.” (participant17,nurse,ward4)
Another nurse described that the motives for providing feed- back onqualitymeasurements are differentforhospital leveland wardlevel:
“The aim is to make usaware of how we are performing on these quality measurements and what can be improved. […]
Thisisimportantforthepatients’welfare,butitisalsoimpor- tantbecausehospital-wideweneedtomeetlegalrequirements.
[…]Thehighermanagement,whoobviouslydonotworkindi- rectpatientcare,[…]theyfocusonwhatthefiguresare.While forus,itismoreimportanthowthepatientisdoing.” (partici- pant08,nurse,ward3)
3.2. Nurses’attributionsandtheirassociationwithnurses’
engagementandburnout
We mainly used the survey data to examine the relationship between nurses’ attributions and their engagement and burnout.
The outcomes of the regression analysis (see Table 3) indi- cated that compliance attributions were associated with burnout (
β
=0.27;p=0.013).Inother words,whennursesbelievedthat they were provided with feedback on quality measurements be- causethe ward managerhad to,dueto externalconstraints (e.g., qualitystandardsimposedonthehospitalbytheinspectorate),this isrelatedtohigherlevelsofburnout.‘Quality and nurse enhancement attributions’ had very lim- ited meaning for burnout (
β
= -0.09; p = 0.46). Regarding the‘Cost reduction and nurse exploitation attributions’, the results showedsomeeffectonburnout,butwithap-valueabovethresh- old (
β
= 0.18; p = 0.11). In general, the attributions have no important association with work engagement; the F-statistic of the model is also not significant (Adjusted R-square = 0.006, F[9,78]=1.06;p>0.05).Theinterviewdatawasusedtofindanexplanationforthepos- itive relationship between ‘Compliance attributions’ andburnout (cynicism and exhaustion). It seems that nurses felt that exter- nal requirementsputaheavy demandon theirjobs.From this,it seems logicalthat whennursesbelievedtheywere providedwith feedbackonqualitymeasurementsduetoexternalconstraints,this led tocynicism andexhaustion.Forinstance, thefollowingnurse described how she felt pressured by governmental requirements, withouthavinganyinfluenceonthem.
“Therequirementsoftheinspectorateareobviouslyincreasing.
It’stoobadthatwehavelittleinfluenceonthat.Theyinsiston making itdemonstrable,hencethe qualitymeasurements. The requirementsareoftentoohigh,inmyopinion.Howeverthatis something fromthe government,you cannot changethat. […]
Table 3
Outcomes of regression analysis based on the survey data ( N = 91).
Burnout Work engagement
B β 95% confidence
interval Lower / Upper
B β 95% confidence
interval Lower / Upper
Gender (ref. = male) 0 .09 0 .04 −0 .42 / 0.59 0 .43 0 .19 −0.17 / 1.02
Age (years) 0 .03 0 .53 −0 .08 / 0.07 −0 .01 −0 .11 −0.05 / 0.04
Tenure as a qualified nurse (years) −0 .05 −0 .88 ∗ −0 .10 / −0.01 0 .00 0 .07 −0.05 / 0.06
Tenure in hospital (years) 0 .03 0 .50 −0 .00 / 0.07 −0 .01 −0 .11 −0.05 / 0.03
Hours per week 0 .01 0 .10 −0 .02 / 0.04 0 .01 0 .07 −0.02 / 0.04
Supervisor Feedback Environment −0 .16 −0 .20 −0 .35 / 0.03 0 .02 0 .018 −021 / 0.24
Quality and nurse enhancement attributions −0 .07 −0 .09 −0 .24 / 0.11 0 .15 0 .19 −0.05 / 0.16 Cost reduction and nurse exploitation attributions 0 .10 0 .18 −0 .02 / 0.21 −0 .00 −0 .00 −0.14 / 0.14
Compliance attributions 0 .22 0 .27 ∗ 0 .05 / 0.40 −0 .05 −0 .05 −0.25 / 0.16
Measures of model fit
R 2 0 .23 0 .11
Adjusted R 2 0 .14 0 .01
F 2 .55 ∗ 1 .06
B = unstandardised beta, β= standardised beta.
∗ p < 0.05.
Table 4
Outcomes of regression analysis based on the survey data ( N = 91).
Quality and nurse enhancement attributions
Cost reduction and nurse exploitation attributions
Compliance attributions
B β 95% confidence
interval Lower / Upper
B β 95% confidence
interval Lower / Upper
B β 95% confidence
interval Lower / Upper Gender (ref. = male) 0 .21 0 .08 −0.43 / 0.85 −0 .44 −0 .12 −1.41 / 0.52 0 .38 0 .15 −0.28 / 1.04 Age (years) −0 .02 −0 .25 −0.07 / 0.03 −0 .01 −0 .09 −0.08 / 0.07 −0 .00 −0 .01 −0.05 / 0.05 Tenure as a qualified nurse (years) 0 .05 0 .59 −0.01 / 0.10 0 .04 0 .37 −0.04 / 0.12 −0 .01 −0 .07 −0.06 / 0.05 Tenure in hospital (years) −0 .03 −0 .28 −0.07 / 0.02 −0 .06 −0 .47 −0.12 / 0.01 0 .01 0 .09 −0.04 / 0.05 Hours per week 0 .01 0 .07 −0.02 / 0.04 −0 .05 −0 .27 ∗ −0.10 / −0.00 −0 .00 −0 .01 −0.03 / 0.03 Supervisor Feedback Environment 0 .53 0 .50 ∗∗∗ 0.33 / 0.74 −0 .29 −0 .20 −0.60 / 0.02 0 .11 0 .11 −0.10 / 0.32 Measures of model fit
R 2 0 .29 0 .13 0 .05
Adjusted R 2 0 .24 0 .07 0 .00
F 5 .46 ∗∗∗ 2 .03 + 0 .64
B = unstandardised beta, β= standardised beta
∗p < 0.05 ∗∗∗p < 0.001.
SometimesIbelievethey[theinspectorate]aregoingtoofarin whattheywantustodo.” (participant06,nurse,ward3) Anothernursereportedonhowgovernmentalrequirementsare inconflictwithherjobsatisfaction:
“I believe it [performing well on quality measurements] is partly obligatory by law. It is obligatory, so we have to pay attentionto it. The hospital wouldbe crazy to say “the min- ister can comeup withanything, butwe are not doing that.”
So,Ibelieveprovidingfeedbackonthesequalitymeasurements comesfromthatdirection. Iguessitwill alsoimprovequality.
However,whenyoulookatmyworksituation,whathastobe done onthejob, itdoesnot improvemy jobsatisfaction.It is inconflictwiththat.” (participant24,nurse,ward1)
3.3. Theassociationbetweenthefeedbackenvironmentandnurses’
attributions
Moreover, we used the survey data to examine the associa- tion between the feedback environment and nurses’ attributions about the manager’s reasons for providing them with feedback.
The outcomes of the regression analysis (see Table 4), indicated thattheexpectedrelationshipsbetweenthefeedbackenvironment andattributionswereconfirmedwithourdata.Asupportivefeed- back environment set by the ward manager was positively re- latedto ‘Quality andnurse enhancement attributions’(
β
= 0.50,p < 0.001) and negatively related to nurses’ ‘Cost reduction and nurseexploitationattributions’(
β
=−0.20,p=0.062),albeitwith a p-valueabove thethreshold.Feedback environmenthadlimited meaningfor‘Complianceattributions’(β
=0.11,p=0.32)andthe overallmodeldidnotexplainmuchofthevariationin‘Compliance attributions’(R-square=0.05,F[6,81]=0.64,p>0.05).Oursurveyresultsshowedarelationshipbetweenthefeedback environment andnurses’attributions.However, thedatafromthe interviews with the ward managers indicated that a third vari- ablemaybe relevantinthisrelationship: thewardmanagers’ac- tual purpose in providing feedback on quality measurements. It could be that nurses’ attributions will more likely match their wardmanager’smotivationswithinasupportivefeedbackenviron- ment.Noneofthewardmanagersappearedtoexplicitlydescribea reduction incostsasoneoftheirpurposes,inprovidingfeedback on quality measurements. Ward managers’ purposes inproviding feedbackwasmainlytoimprovethequalityofnursingcareand/or tomakenurses’workmoreattractive(‘Qualityandnurseenhance- ment’) andasa‘side-effect’adheretoexternal constraints,asthe followingquotedisplays.
“The aim is toimprove thequality of care,especially the im- provements that are obliged. By providing feedback we can achieve rapidresults.I’minfavourofthat.I’minfavourofev- erythingthatleadstoclarificationforthenurses,forourselves and clarifies the possibilities for improvements. […] It [feed- back] showedwe were performing very well. That’s alsonice
tohearfora change.That’snot whyyoudo this, butit’snice toseeweareontherighttrack.Andwhenyouseeyouarenot yetonthe righttrack,to dosomething withthatinformation.
[…]Withthesequalitymeasurementswecansay,asahospital, weare performing well.I’mpartofthishospital.” (participant 10,wardmanager,ward4)
Anotherward manager explained that her purposein provid- ingfeedbackonqualitymeasurementswastoimprovethequality ofcareby making nursesaware oftheir low performanceon the qualitymeasurements.
“It’s myopinion that people remainedstuckin thebeliefthat they were performing very well. At times, I got quite sick of that. Really, I think that’s very extraordinary. [...] Iwanted to make them aware of the fact that they were not performing thatwell.Thatthisisthefuture.Providinggoodcareisnotonly aboutpamperingpatients.Weshouldalsopayattentiontopa- tients inanother way[referring tothe quality measurements]
whichis better forthe quality ofcareand forpatient safety.”
(participant19,wardmanager,ward1)
4. Discussion
The purpose of thisstudy wasto enhance our understanding ofHRattributions,byexploringtheattributionsthatnurses make aboutwhyfeedbackonqualitymeasurementsisprovidedtothem, andwhethertheseattributions arerelatedto thenurses’engage- mentandburnout. Additionally,weexplored therole ofthefeed- backenvironmentsetbythewardmanageronthestrengthofthis relationship. Our study comprised a convergent mixed-methods approach, combining both quantitative and qualitative methods, followingafeedbackinterventioninfourhospitalwards.
Ourfindings indicate that nurses asa group andindividually, makevarying attributionsforthe samefeedbackon qualitymea- surements,andthat theseattributionsappeartobedifferentlyas- sociated with burnout. ‘Quality and nurse enhancement attribu- tions’, i.e., nurses’ perceptionsthat feedback is provided to them inordertoimprovequalityofpatientcareand/ortheirwell-being, are negatively associated withburnout. ‘Complianceattributions’, i.e., nurses’ perceptionsthat thefeedback is provided tothem in order to comply with external regulations, are positively associ- ated with burnout. The latter relationship may be explained by thefactthat nursesexperiencegovernmental requirementsasjob demands. Many nurses appear to simultaneously make ‘Quality andnurseenhancementattributions’and‘Complianceattributions’, forwhichthey havedifferentrationales.Additionally,ourfindings showthat asupportivefeedbackenvironmentispositivelyassoci- atedwith ‘Qualityandnurseenhancement attributions’ andneg- atively with ‘Cost reduction and nurse exploitation attributions’
(nurses’perceptionsthat thefeedback isprovided tothem inor- dertosavecostsandmakethemworkharder).
4.1. Theoreticalimplications
Responding to the call for more scholarly knowledge in this field by Tutiet al. (2017),our findings shed light on the impor- tance of the process of how feedback on quality measurements to nursingteams workingin a hospital settingis experienced by thenurses.Morespecifically,followingBrownetal.’s(2019)frame- workregardingimportantfactors thatinfluence feedbacksuccess, we have studied how feedback on quality measurements (feed- backvariable)isattributedbynurses(recipientvariable)withinits feedbackenvironment(contextvariable),andhowthisfeedbackis associated withnurses’ engagement andburnout. First, our find- ings suggest that it is relevant to consider attribution processes
in order to better understand the effects of feedback interven- tions (Christina et al., 2016). Employees can have different attri- butionsforthe same(typeof) feedback,which mayresultindif- ferentassociationswiththeirengagementandburnout.Ourstudy also confirmsthat the distinctionbetween internal commitment- focused, internal control-focused andexternal attributions is rel- evant and provides a good starting-point for more elaborate re- searchonattributionsaboutfeedback.Incontrasttopastresearch done byKoys(1991)andNishiietal.(2008),ourfindingsindicate thatexternalattributionscanbesignificantlyandpositivelyassoci- atedwithemployee burnout.Moreover,thefeedbackenvironment doesnotmoderatethiseffect,i.e.,thefeedbackenvironmentdoes notweakenthepositiverelationshipbetweenexternalqualitycon- trol and burnout. In our view, this external attribution may be mediated by feelingsof limitedpersonal control andof helpless- ness assuggestedbytheresearch ofSparrandSonnentag (2008). Thisoutcomeindicatesthatpersonalcontrolandlimitedhelpless- ness atwork isan important resource in an advantageous feed- backenvironment.Wesuggestthatfutureresearchonattributions should therefore take theimportant variablesof personal control andhelplessnessintoaccount.
Our study also showsthat an individual employee can make multiple attributions related to its different dimensions for the same (typeof) feedback. Forexample,our findings show that an individualnurse,atthesametime,believedthatshewasprovided withfeedbackonqualitymeasurementsbothbecausethehospital neededto adheretoquality standardsimposedby thehealthcare inspectorate, and because her ward manager wanted to improve thequalityofpatientcare.Althoughthepossibilityofmultipleat- tributions wasleft open inprevious research onattributions (see forinstance, Nishiietal., 2008; Van de VoordeandBeijer,2015), it has not been explicitly addressed in previous scholarly work.
Moreover, the possible effectsof multiple attributionsmay inter- act. Theoutcomesofourstudyconfirmthat abetter understand- ing ofmultiple attributions andtheir associationswithemployee engagementandburnoutprovidesaninterestingavenueforfuture research.
Second,ourfindingsconfirmthatthecontextvariable‘feedback environment’ is relatedto employees’attributions aboutthe rea- sonsforprovidingthemwithfeedback.Morespecifically,ourfind- ingsindicate thattherelationship betweenasupportivefeedback environmentandnurses’attributionsmaybepartiallyexplainedby thewardmanager’sactualpurposeinprovidingfeedbackonqual- ity measurements. An interesting possibility that should be fur- therexamined,isthatnurses’attributionsaremorelikelytomatch their wardmanager’s purposewithin asupportive feedbackenvi- ronment.
Byshowingthattheprocessofimplementingfeedbackonqual- itymeasurementscanbeattributeddifferentlyby(groupsof)indi- vidual nurses, andthat the feedback environment and the man- ager’srole therein,play arole inthis, ourstudyshedsmorelight on the mechanism explaining the effects of feedback on perfor- mance(Iversetal.,2014;Tutietal.,2017).Futureresearchinthis domainshouldfocusonidentifyingadditionalindividualvariables that possibly influence employees’attributions aboutthemotiva- tionforprovidingthemwithfeedback.Moreresearchisneededto betterunderstandtheinfluenceofnurses’feedbackorientation,or nurses’individualpropensity toseekandutilizefeedback.Empiri- calworkbyGabrieletal.(2014)hasshownthatasupportivefeed- back environment is beneficial for employeesthat are favourably orientedtowardsfeedback,yetcanbeharmfulforemployeesthat do not necessarily want to receive or use feedback. Addition- ally, the kind of feedback (delivery) that is used may influence nurses’ perceptions, as literature has suggested that supportive feedback, rather than punitivefeedback, positively influences the effectsoffeedbackinterventions(Christinaetal.,2016).Finally,an
