Effects of a leaflet on breast cancer screening knowledge, explicit attitudes, and implicit associations

(1)

Effects

of

a

lea

ﬂet

on

breast

cancer

screening

knowledge,

explicit

attitudes,

and

implicit

associations

Lindy

M. Kregting

a,

*,

Nicolien

T. van

Ravesteyn

a

,

Wolfert

Spijker

b

,

Tessa

Dierks

a

,

Clare

A. Aitken

a

,

H. Amarens

Geuzinge

a

,

Ida

J. Korfage

a

a_Department_of_Public_Health,_Erasmus_MC,_University_Medical_Center_Rotterdam,_the_Netherlands b

BevolkingsonderzoekZuid-West,Rotterdam,theNetherlands

ARTICLE INFO Articlehistory:

Received20November2019 Receivedinrevisedform25May2020 Accepted29June2020 Keywords: Breastcancer Massscreening Healthknowledge Implicitassociations Screeningattendance ABSTRACT

Objective: To assess the effect of aninformation leaﬂet on knowledge, explicit attitudes, implicit associations,andattendanceforbreastcancerscreening.

Methods:Dutchwomen(aged49–75years)wereapproachedthreemonthsbeforetheirbreastcancer screeninginvitation.Afterprovidinginformedconsent,participantswererandomisedtoreceivingthe informationleaﬂet(interventioncondition)ornot(controlcondition).Screeningknowledge,explicit attitudes,andimplicitassociationswereassessedthroughweb-basedquestionnaires,atbaselineand twoweekslater.Actualscreeningattendancedatawerecollected.

Results:Intotal,988womencompletedbothquestionnaires.Participantsintheleaﬂetconditionscored higheronknowledge(9.9versus9.6,p<0.001,scale0 11),andmoreoftenhadpositiveexplicitattitudes (97%versus95%,p=0.03),thanthoseinthecontrolcondition.Thiscontrastwasbiggeramongﬁrst-time invitees.Implicitassociationswerenotcorrelatedwithexplicitattitudesorattendance.Explicitattitudes weremoderatelycorrelatedwithattendance(r=.30,p<0.001).

Conclusion:Theinformationleaﬂetledtomoreknowledgeandmorepositiveexplicitattitudes.Implicit associationstowardsbreastcancerscreeningwerenotcorrelatedwithattendance.

PracticeImplications:Encouragementtolearnaboutthescreeningprogrammecanincreaselevelsof knowledgeofinviteesandthereforesupporttheirdecision-makingaboutparticipation.Thismightbe especiallyrelevantforﬁrst-timeinvitees.

1.Introduction

Individualbehaviourisshapedbyattitudes[1].Twodifferent types of attitudes can be distinguished: explicit attitudes and implicitassociations[2].Explicitattitudesaredeliberateandare presentattheconsciouslevel[1,2].Peopleareconsciousoftheir explicitattitudesandareabletoself-reportthem.Incontrastto explicit attitudes, implicit associations can inﬂuenceand guide behaviourwithoutpeople’sconsciousawareness,theycanresultin spontaneousorautomaticbehaviour[1,2].Explicitattitudesand implicitassociationscanbecontradictory[2].Ithasbeenshown that implicit associations can affect consumer behaviour and decision-making [3–5], but littleis knowntowhat extentthey affect medical decision-making about, for instance, cancer screening.

Participation in population-based breast cancer screening programmes is voluntary and usually free-of-charge. In the Netherlands, eligible women (ages 50–75) receive a personal invitationeachscreeningroundaccompaniedbyaninformation leaflet about the procedure, and harms and benefits of breast cancerscreening.Theinformationisaimedatenablingwomento makeaninformedchoiceaboutwhetherornottoparticipateinthe screening[6,7].However,itisuncleartowhatextentthecurrent informationleaflet(2018)contributestotheknowledgeofwomen, andwhetheriteffectsexplicitattitudes,implicitassociations,and attendance.

Attendanceratesofbreastcancerscreeningprogrammesinthe Netherlands,England,Finland,andtheUSAslightlydecreasedover thepastyears(e.g.theNetherlands:from82.4%in2007to76.6%in 2018) [8–12]. To betterunderstand this decrease and theway women decidetoparticipatein breast cancerscreeningor not, more insight into knowledge, explicit attitudes, and implicit associationsisuseful.Itiscurrentlyunknownifandtowhatextent attendancetothebreastcancerscreeningprogrammeisassociated withexplicitattitudesorimplicitassociation.

* Correspondingauthorat:Dr.Molewaterplein40,Dep.PublicHealth,3015GD Rotterdam,theNetherlands.

E-mailaddress:l.kregting@erasmusmc.nl(L.M.Kregting). https://doi.org/10.1016/j.pec.2020.06.032

ContentslistsavailableatScienceDirect

Patient

Education

and

Counseling

(2)

The aim of this study was to examine the inﬂuence of an information leaﬂet onthelevel ofdecision-relevant knowledge aboutbreast cancer screening, explicit attitudes and screening attendanceamongwomeninvitedforbreastcancerscreening.This studyalsoaimedtoinvestigatetheassociationbetweenexplicit attitudesaswellasimplicitassociationstowardstheDutchbreast cancerscreeningprogrammeandattendance.

2.Methods

2.1.Population

Women,aged49–75,livingintheSouthWestscreeningregion oftheNetherlands,whowereduetobeinvitedforbreastcancer screeningwereapproachedtoparticipateinthisstudybyajoint letterfromthelocalscreeningorganisation‘Bevolkingsonderzoek Zuid-West’andErasmusMC.ThelettersweresentinNovember andDecember2018andincludedstudyinformation,aninvitation toparticipate,andaninformedconsentform.Womenwhowere registered at the screening organisation as ‘not willing to participateinresearch’werenotapproachedtoparticipateinthis study.Duringﬁveto10yearsfollowingabreastcancerdiagnosis womenare not invited for theregular screening program, and thereforethis groupof women was not includedin our study. Having no email address or internet access was an exclusion criterion.

2.2.Samplesizecalculation

Samplesizecalculationindicatedthat834womenneededto participatetobeabletoshowaneffectinresponsetimewith80% powerandstatisticalsigniﬁcanceof0.05.Basedonthe participa-tionrateinapreviousstudyevaluatingthescreeningprogramme, weexpectedaparticipationrateofabout30%amongscreening attendersandabout10%amongnon-attenders[6].

Potentialparticipantswereselectedbytheregionalscreening organisation(Bevolkingsonderzoek Zuid-West) based onpostal code.Toreacharepresentablepopulationofparticipatingand non-participatingwomen,womenwhohad declinedparticipationin previous screening rounds were oversampled. In total, 5568 women were invited, of which 1211 (22 %) women had not participatedinpreviousscreeningrounds,3817(68%)womenhad participatedinpreviousscreeningrounds,and540(10%)women weretoreceivetheirﬁrstscreeninginvitation[12].

2.3.Design

Womenwhoprovidedconsentandtheire-mailaddresswere randomisedtotheinterventioncondition(leaflet)orthecontrol condition(noleaflet) bycomputer-generated randomnumbers. Subsequently,alinktoaweb-basedquestionnairewassenttothe participants by e-mail. The questionnaire started with a short introduction to the Dutch national breast cancer screening programme and contained questions regarding intention to participate, explicit attitudes, knowledge about the screening programme,reasonstoparticipateornot,and demographics.A primingtask was included toassess implicit associations. Two weeksaftercompletingtheirfirstquestionnaire,participantsin theinterventiongroupwereaskedtoreadanonlineinformation leaflet(seebelow).Participantsinthecontrolconditiondidnot receivethisleaflet.Then,allparticipantswereaskedtocomplete thesecondquestionnaire(seeFig.1).Subsequently,followingthe regularinvitationschedule,allrespondentsreceivedaninvitation toparticipateinthebreastcancerscreeningprogrammeandthe informationleaflet.

Since it is notalways feasible toassess actualparticipation, previous studies concerning informed choice, often assessed intention to participate as a proxy for actual participation. Althoughstronglycorrelated,intentiontoparticipateinscreening isnotnecessarilysimilartoactualscreeningattendanceandcanbe considered to bemore inﬂuenced by explicit attitudes [13,14]. Therefore,thisstudywillstudytheeffectsofbothintentionand actualattendance.Conditional onprovided consent,attendance data for this screening round were provided by the screening organisation.Collectionofattendancedatatookplacetwotothree monthsaftertheplannedscreeningappointment.

2.3.1.Intervention

Womenin theinterventiongroupwereprovidedtheofficial breast cancer screening information leaflet from the Dutch NationalInstituteforPublicHealthandtheEnvironment(RIVM). Theleafletwas developedbasedontheopinionofexpertswho recommended“[using]simpletextswithoutnumericalvaluesto presentinformationondifficulttopicsasfalsepositivesand over-diagnosis”(6).Therefore,theleafletwasdesignedtoincreasethe levelofgistknowledge,i.e.“theabilitytoidentifytheessential points of the information presented”, rather than verbatim knowledge,i.e.“theabilitytocorrectlyreadnumbersfromgraphs” [15].

TheJanuary2018version(appendixFig. 1)containsinformation about the screening invitation, the screening process, possible screening outcomes, and the benefits and harms of screening. Unlikemostofficialinformationleaflets,potentialharmssuchas overdiagnosis,overtreatment,false-negatives,andintervalcancers weredescribedexplicitly[16].

Fig.1.Datacollectiontimeline. xxx–xxx

(3)

2.3.2.Contentofthequestionnaires

The baseline questionnaire included demographic questions about screening history, living situation, educational level, employmentstatus,andhomelanguageoftherespondent.

Gistknowledgeaboutthebreastcancerscreeningprogramme wasdeterminedusing11statements(basedonexpert consulta-tions);responseoptionswere‘true’,‘false’,or‘Idon’tknow’[6].Inthe absence of an agreed external criterion to define ‘sufficient’ knowledge,itwasoperationalisedasaminimumofeightcorrect answers[6,17–19]. Participants’explicit attitudestowardsbreast cancerscreeningweremeasuredthroughanattitudesscalederived fromthemultidimensionalinformedchoicemeasure ofMarteau etal.[20]. It contained six cognitiveitemsregardingthebreast cancer screeningprogramme,suchas‘Ithinkparticipationinthebreast cancerscreeningprogrammewithinthreemonthsformewouldbe useless/useful’.Participantsrespondedon7-point-likertscales.In accordancewithguidelines,missingitemsontheattitudes scale wereimputedbyindividuals’meanscore,ifatleast50%oftheitems hadbeencompleted[21].Theresultsweretransformedtoa0–100 scaleandcategorisedasnegative(<50)orpositive(50)attitudes. Participantswereaskedhowlikelytheyweretoparticipatein thebreastcancerscreeningprogrammeiftheywouldreceivean invitationwithin the comingthree months.The answers were givenona7-point-likertscale.Scores1and2wereclassifiedasa negativeintention,3–5asaneutral intention,and6 and7 asa positiveintention.

FollowingthemodelofMarteauetal.,awomanwasconsidered to have made an informed choice when she had sufﬁcient knowledgeaboutthebreastcancerscreeningprogram,apositive attitudetowardsparticipatinginthisprogram,andparticipatedin theprogramme,orwhenshehadsufﬁcientknowledge,anegative attitudeanddidnotparticipateintheprogramme[20].

Toassessparticipants’implicitassociations,aprimingtaskwas used.Primingtasksarewidelyusedinsocialcognitionresearch, and were originally developed to assess implicit associations towardssocialgroupsoractivities[3–5,22].Duringprimingtasks, peopleareshownprimes(picturesorwords)ofatopicofinterest

followedbytargetwords.Thetargetwordsusedaredistinctively positiveornegative.Theparticipantsareaskedtorespondtothe targetwordandindicateifitwaspositiveornegative.Thetask reliesontheassumptionthattheprimeautomaticallyactivatesan evaluation, and that if primes and target words are strongly associatedintheparticipant’smind,theparticipantwillreactmore quickly[23].Therefore,theresponsetimetothetaskwasassumed tobeshortestwhentheparticipantstronglyassociatedtheprime withthepresentedtarget[22].

Intheprimingtask,ascreening,neutralandnon-wordprime wereused.Theprimewordschosenhadtobeshort,simpleand representative.Forthescreeningprime,_{“Röntgenfoto”}(X-ray)was foundtobetoolonganddifﬁcultforquickreadingandlesstypical forbreastcancerscreening.Therefore,weoptedfor“Borstfoto’s” (breastX-rays/pictures)whichwasamoresimpleandclearreferral to breast cancer screening. Since this prime was crucial, we checkedwithhealthcareproviders,apatientorganisationandthe collaboratinglocalscreeningorganisation(BOZW)whetherthey agreed. Theneutral prime, “Brievenbus” (mailbox), was chosen because of its neutral meaning and because it had the same amountofsyllablesasthescreeningprime.Thenon-wordprime wasarandomorderofconsonantsataboutthesamelengthasthe otherprimes(“Fjnmpklzv”).

Each of the primes was shown on the computer screen, followedaftera100msintervalbyatargetword.Thetargetwords couldbepositive or negative(forexample‘good’or ‘bad’). The respondents were asked to state as quickly and accurately as possiblewhethertheshown targetwas positiveor negativeby pressingaspeciﬁckeyontheirkeyboard(i.e.thekeys“L”and“A”, respectively).Thecompleteprimingtaskconsistedof24 combi-nations ofprimesand targets,inwhich allcombinationsofthe threeprimesandeighttargetswerepresentedonce,inarandom order.Duetomisconceptionsregardingoneofthetargets(double meaning in Dutch) the response times for this target were excludedforanalyses.

Responsetimes consideredtobe toofast(quicker than 300 milliseconds (ms)) or too slow (slower than 3000ms) were

Fig.2. Flowchart.

(4)

excluded [24]. Also, response times were excluded in case of incorrectresponses,e.g.incasethepositivekey“L”waspressed after thenegative word “bad”. Implicit associations were then calculatedperprimebysubtractingtheaverageresponsetimesfor the negative targets from the average response times for the positivetargets.

2.4.Statisticalanalyses

T-tests and chi-square tests were performed to test for differencesbetweenthetwo randomisedgroups in attendance, explicitattitudes,implicitassociations,knowledge,andinformed choice.Subsequently,Pearson’s,Phi,andCramer’sVcorrelations weremeasuredbetweenimplicitassociations,explicitattitudes, intention to participate, attendance, knowledge about breast cancerscreening,levelofeducation,previousinvitationforbreast cancerscreening,previousattendanceinbreastcancerscreening, andpreviousreferralsbasedonbreastcancerscreeningresults.

Subgroupanalyseswereperformedforparticipantswhowere invitedforthenationalbreastcancerscreeningprogrammeforthe firsttime.Thesefirst-timeinviteeswereidentifiedbasedon self-reportingto not have had a previous invitation. Differences in responsetimesbetweenleftandrighthandedrespondentswere alsotested.

Repeated measures ANOVAs were performed to test for differencesinresponsetimebetweentheprimesandtargetsin theprimingtask.Theinteractionterm(“prime*target”)wasalso included. Subsequently, a repeated measures ANCOVA was

performedtakingintoaccountthecovariatesthatweresigniﬁcant inthecorrelationsanalyses.

AllanalyseswereperformedinIBMSPSSStatistics,version24 andstatisticalsigniﬁcancewassetatα=0.05.

3.Results

3.1.Backgroundcharacteristics

In total, 5568 study invitations were sent out and 1372 informed consent forms were received (response rate 25 %) (Fig.2).Ofthese,25werereceivedtoolateand35wereinvalid.The 1312includedparticipantswererandomisedtotheleaﬂet(n=703, 54%)andthecontrolcondition(n=609,46%).

Thirty-fivewomen(2.7%)wereexcludedduetounknownor invalid emailaddresses, and 28 (2.1 %) womenwithdrew from participationafterbeingsentthefirstquestionnaireduetolackof time or technical issues. In total, 1073 participants (83 %) completed the first questionnaire. Afterbeing sent the second questionnaire, another six participants (<1%) withdrew from participation. In total, 988 participants (92 %) completed the secondquestionnaire.

Data-analyses included 988 participants; 531 in the leaﬂet conditionand457inthecontrolcondition.Ofthese,904(92%)also gave consent to collect attendance data from the screening organisation. Baseline characteristics of the two randomised groups weresimilar(Table1).Participants wereonaverage60 yearsofagerangingfrom49to75inbothconditions.

Table1

Participantcharacteristics(n=988).

Leaﬂetcondition(n=531) Controlcondition(n=457) p-value

Age Mean(SD) 60.1(6.7) 59.9(6.9) 0.15 Range 48.5-75.0 49.0-74.9 Missing 0 0 Educationallevel(n,%) High 146(28) 123(27) 0.44 Middle 287(55) 239(53) Low 87(17) 90(20) Missing 11 5

Languagespokenathome(n,%)

Dutch 454(96) 407(98) 0.13

Dutchandother 8(2) 2(1)

Other 10(2) 5(1)

Missing 59 43

Livingsituation(n,%)

Withpartner 441(84) 366(80) 0.21

Notwithpartner 87(17) 89(20)

Missing 3 2 Workingstatus(n,%) Paidwork 293(60) 250(59) 0.85 Nopaidwork 64(13) 53(13) Retired 129(27) 119(28) Missing 45 35

Previouslyinvitedtoparticipateinbreastcancerscreening(n,%)

Yes 464(88) 397(87) 0.77

Donotremember 6(1) 4(1)

No 58(11) 56(12)

Missing 3 0

Previouslyparticipatedinbreastcancerscreening(n,%ofinvited)

Yes 427(93) 374(94) 0.40

No 32(7) 22(6)

Missing 5 1

Previouslyreferredforfurtherdiagnostics(n,%ofparticipated)

Yes 62(15) 46(12) 0.36

No 365(86) 328(88)

Missing 0 0

(5)

3.2.Informedchoice

Atbaseline,theaverageknowledgescorewas9.3(onascaleof 0–11).Atfollow-up,adifferencewasseenbetweentheleafletand control condition (9.9 versus 9.6, respectively, p<0.001). This resulted in 94 % and 91 % of participants having sufficient knowledgeinthetworespectivegroups(p=0.09).Atbaseline,96% hadpositiveexplicitattitudes,atfollow-upthesepercentageswere 97% inthe leafletcondition and95 %in thecontrol condition (p=0.03). No differences in screening attendance were found between the leaflet and control condition (90 % versus 88 %, respectively,p=0.46).

Intotal,718women(80%)madeaninformedchoice.Ofthem, 701madethedecisiontoparticipateinscreeningand17notto participate,seeFig.3A.Abouthalfoftheuninformedchoiceswere due to insufficient knowledge. Differences in informed choice betweenthetwoconditionswerenotsignificant(i.e.intheleaflet condition84%madetheinformedchoicetoparticipateand2%not toparticipatecomparedto78%and3%inthecontrolcondition, p=0.07).

3.2.1.Subgroupanalysesofﬁrst-timeinvitees

Atbaseline,80%offirst-timeinviteeshadsufficient decision-relevantknowledgeversus89%ofwomeninthetotalpopulation (AppendixtableA1).Afterreadingtheleaflet,93%offirst-time inviteesreportedsufficientknowledgeversus77%of first-time inviteesinthecontrolcondition.Theattendanceratewas 83 % amongthefirst-timeinviteesversus89%inthetotalpopulation. Therateofwomenwithpositiveexplicitattitudeswassimilarfor first-timeinviteesandthetotalpopulation(97%versus96%).

Thelowerlevelofknowledgeandthelowerattendanceresulted inalowerproportionofﬁrst-timeinviteeswhomadeaninformed choice(66%).Againabouthalfoftheuninformedchoicesweredue toinsufﬁcientknowledge(Fig.3B).

3.3.Implicitassociations

Atbaseline,505womencompletedtheprimingtask.However, 26 of them withdrew, only partially completed the second questionnaire, or were lost to follow-up. Therefore, baseline primingtaskdataof479(48%of988)womenwereanalysed.At follow-up,primingtaskdataof522(53%)womenwereanalysed (Table 2). Participants pressed the correct key (i.e. the key correspondingtothetarget) 87–89%ofthetime.Nosigniﬁcant differenceinaccurateresponseswasseenbetweentheconditions.

On average,responseswerea littlequicker(i.e.response times were shorter) when positive or neutral targets followed the screening prime versus negativetargets,resulting in a positive meandifferenceinresponsetimes(17.9msand34.2ms, respec-tively) atbaseline. Forthenon-word prime,responseswereon average slower for positive targets than for negative targets, resulting in a negative mean difference in response times (-26.7ms).Thistrendwas alsoseenfor theleaﬂetconditionat follow-up.Inthecontrolcondition,averageresponseswereslower for positive targetsfor allthree primes.However,the standard deviationswerelargeforallmeandifferences.Nodifferenceswere seenbetweenthetwo conditions.Nodifferenceswerefoundin responsetimesbetweenleftandrighthandedparticipants(results notshown).

Nocorrelationswerefoundbetweenimplicitassociationsand explicitattitudesorbetweenimplicitassociationsandintentionto participate (Table 3). Also, no correlation was found between implicitassociationsandattendance(r=.05,p=0.33).

Explicit attitudes werefoundto bestronglycorrelated with intentiontoparticipate,andmoderatelywithattendance(Table3). Intention toparticipate was foundtobe moderately correlated withattendance.Amoderatecorrelationwasalsofoundbetween attendanceandpreviousparticipation.Intentionandattendance werefoundnottobecorrelatedwithimplicitassociations.

RepeatedmeasuresANOVAdidnotshowanysignificantprime effects, target effects or interaction effects for prime*target (Table 4). Thus there were no differences in average response timesbetweenthedifferentprimes,betweenthedifferenttargets, and between certain combinations of primes and targets. No significantdifferencewasfoundfortheinteractionterm prime*-condition meaning that there were no differences in response times for the different primes betweenthe leaflet and control condition.

4.Discussion

4.1.Discussion

Theresultsofourstudyshowthatwomenwhowereprovided withtheinformationleaﬂetreportedbetterknowledge,andmore oftenpositiveexplicitattitudes.Thiscontrastwaslargeramong ﬁrst-timeinvitees.Implicitassociationswerenotassociatedwith explicitattitudestowardsbreastcancerscreening.Explicitattitude was found to be associated with attendance, while implicit associationswerenot.

Fig.3.ClassificationofinformedchoiceaccordingtoMarteauetal.[18].A)totalbaselinepopulationB)subgroupbaselineanalysesoffirst-timeinvitees.Gray:informed choice,Blanc:noinformedchoice*Percentagesareroundedoff,sotheymaynotaddto100%.Toviewthisfigureincolour,pleaseaccesstheonlineversion.

(6)

In an earlier study oninformed choice in theDutch breast cancer screening programme a rate of 88 % informed choices amongfirst-timeinviteeswasreported[6].Thisfindingwasbased onintentiontoparticipateratherthanactualattendance,which mayexplainthedifferencewiththe66%asfoundinthecurrent study.Thepositiveeffectofthelea_fletonknowledgeandinformed choiceconfirmsthefindingsoftwo studiesinAustralia[17,25]. Onestudyfoundthatwomenwhoreceivedadecisionaidleafletfor breastcancerscreeningwithevidence-based informationabout breast cancer mortality reduction, over-detection, and false positivesinscreeninghadmoreknowledgeandmoreoftenmade aninformedchoice thanwomenwho received asimilar leaflet withoutinformation aboutover-detection[25].Theotherstudy foundthat40-year-oldwomenwhoreceivedanonlinedecisionaid regardingbreastcancerscreeningweremoreknowledgeableand lesslikelytobeuncertainabouttheirintentiontoparticipatethan womenwhodidnotreceivetheaid,althoughthisstudyfoundno differenceininformedchoicebetweenthetwogroups[17].Both studies found a reduction in intention to participate in the screening when women received an extensive decision aid, howeverthis wasnot thecase inourstudy[17,25].Our results showedthataskingwomenexplicitlytoreadthelea_fletincreased theirlevelofknowledge,butdidnotdeterthemfromparticipating inscreening.Theuseoftheofficialinformationleafletasdesigned by the Dutch National Institute for Public health and the Environment(RIVM) is a strength of this study. This leaflet is already routinely provided to every woman invited for breast cancerscreeningintheNetherlands.Nofurtherimplementationis neededtoseetheeffects found,whereas,inotherstudies,new leafletsordecision-aidsweredevelopedwithinthestudythatmay notbeimplementedbypolicymakers[25–27].Furtherresearch should be aimed at motivating women to read the leaflet, to increaseitspotentialeffect.

Ameta-analysesof126studiesfoundthatcorrelationsbetween implicitassociationsandexplicitattitudestendedtobesmalland wereevenmorereducedwhentheyconsideredsociallysensitive topics[28].Thisstudydidnotﬁndacorrelationbetweenimplicit associationsandexplicitattitudetowardsbreastcancerscreening. Possibly participants may have felt a pressure to give socially desiredanswers, which madethetopicpartlysociallysensitive therefore the meta-analysis is in line with our ﬁndings. No correlationwasfoundbetweenimplicitassociationsandintention toparticipate or attendancein breast cancerscreening. Thisis similartotheresultsofKorfageetal.,whofoundnocorrelation between implicit associations and intention to participate in cervicalcancerscreening[14].

Thisstudyisuniqueinanalysinginformedchoiceinscreening using actual attendance data. So far, studies used intentionto participateasaproxyforactualattendance.Thecorrelationfound betweenintentiontoparticipateandattendancewasonlyr=.42 (p<0.001). Thismeans that there was anassociation between intention to participate and attendance, but that a number of participantshadanintentionthatwasdeviatingfromtheiractual attendance.Therefore,wethinkitisastrengthofthisstudythat actualattendancedatawasused.Aweaknessofthisstudywasthat onlyabouthalfof theparticipantscompletedtheprimingtask. This reduced thepower of theanalyses and could have ledto selection bias. Comparing demographics, the participants who completedtheprimingtaskweremoreoftenhighereducated,less often retired, and more often ﬁrst-time invitees than the participantswho didnotcomplete thepriming task.No differ-enceswereseeninage,livingconditions,andprevious participa-tion in the screening programme. Reasons why participants completedthequestionnaire,but didnot completethepriming taskwerenotfullyknown,althoughsomeparticipantsreported medicalortechnicaldifﬁculties.Alimitationofthisstudyisthatit

Table2

Implicitassociations,explicitattitudes,knowledge,intentiontoparticipate,attendance,andinformedchoiceofbreastcancerscreeningatbaselineandfollow-up,splitfor leaﬂetandcontrolgroup.

Baseline Follow-up

n=988 Leaﬂetn=531 Controln=457 P-value

Implicitassociations n(%)

Participantscompletingtheprimingtask 479(48) 300(57) 222(49) <0.01

Meandifferenceinresponsetimeinmilliseconds(SD)

Screeningprime&negativetargetminusscreeningprime&positivetarget 17.9(431) 3.6(432) 2.7(436) 0.99 Neutralprime&negativetargetminusneutralprime&positivetarget 34.2(377) 24.9(401) 38.6(367) 0.35 Nonwordprime&negativetargetminusnonwordprime&positivetarget 26.7(415) 13.7(434) 28.8(419) 0.59

Mean%(SD)

Accurateresponsestotargetwords 87(18) 89(17) 87(18) 0.14

Explicitattitudes n(%) Positive 945(96) 516(97) 432(95) 0.03 Negative 41(4) 14(3) 24(5) Missing 2 1 1 Levelsofknowledge(0 11) Mean(range) 9.3(2–11) 9.9(4–11) 9.6(2–11) <0.001 Sufﬁcientknowledge(8)n(%) 869(89) 486(94) 409(91) 0.09 Intentiontoparticipate n(%) Positive 929(94) 507(96) 423(93) 0.15 Neutral 33(3) 15(3) 21(5) Negative 25(3) 9(2) 13(3) Missing 1 0 0 Participation n(%) Participated 803(89) 437(90) 366(88) 0.46

Didnotparticipate 101(11) 51(11) 50(12)

Missing 84 43 41

Informedchoice n(%)

Yes,informedchoicetoparticipateinscreening 701(78) 403(84) 321(78) 0.07

Yes,informedchoicenottoparticipateinscreening 17(2) 8(2) 11(3)

No,notaninformedchoice 186(21) 71(15) 82(20)

Missing 84 49 43

(7)

was not possible to address women who opted out from the breast cancerscreening programme. Alsothe fact that 89 %of the participants in this study attended the screening programme versus 76% in this speciﬁc regionin the previous screeninground,96%reportedapositiveexplicitattitude,96 %

had a positive intention toparticipate,and 89 % had sufﬁcient knowledge at baseline indicates that study participants were probably morepositive about breastcancer screeningthan the averagepopulation[29]. Still,thetworandomised groupswere comparable.

Table3

Correlationsbetweenintention,attendance,implicitassociationsandexplicitattitudesregardingbreastcancerscreeningandeducationallevel,screeninghistory,knowledge aboutthescreeningprogramme.

Baseline,entiregroup(n=988) Implicit

associations

Explicit attitude

Intention Attendance Implicitassociations,i.e.thedifferenceinresponsetimeto[screeningprime&negativetarget]versus

[screeningprime&positivetarget]

– .03 (p=0.83) .05 (p=0.31) .05 (p=0.33) Explicitattitude .03(p=0.83) – .64 (p<0.001) .30 (p<0.001) Intentiontoparticipate .05(p=0.31) .64 (p<0.001) – .42 (p<0.001) Attendance .05(p=0.33) .30 (p<0.001) .42 (p<0.001) – Knowledge .08(p=0.08) .03 (p=0.40) .03 (p=0.35) .03 (p=0.32) Educationallevel .01(p=0.80) .18 (p<0.001) .09 (p<0.01) .01 (P=0.95)

Previouslyinvitedtoparticipateinbreastcancerscreening .05(p=0.30) .00

(p=0.92) .02 (p=0.45)

.07 (p=0.03)

Previousparticipationinbreastcancerscreening .04(p=0.43) .37

(p<0.001) .59 (p<0.001)

.44 (p<0.001)

Previouslyreferredforfurtherdiagnostics .05(p=0.39) .00

(p=0.95) .01 (p=0.83)

.09 (p=0.02) Follow-up,leaﬂetcondition(n=531)

Implicit associations

Explicit attitude

– .01 (p=0.90) .01 (p=0.86) .05 (p=0.45) Explicitattitude .01(p=0.90) – .67 (p<0.001) .30 (p<0.001) Intentiontoparticipate .01(p=0.86) .67 (p<0.001) – .40 (p<0.001) Attendance .05(p=0.45) .30 (p<0.001) .40 (p<0.001) – Knowledge .01(p=0.94) .06 (p=0.19) .07 (p=0.10) .07 (p=0.15) Educationallevel .10(p=0.09) .22 (p<0.001) .10 (p=0.02) .08 (p=0.20)

(p=0.68) .07 (p=0.13)

.02 (p=0.62)

(p<0.001) .52 (p<0.001)

.47 (p<0.001)

(p=0.56) .01 (p=0.82)

.10 (p=0.05) Follow-up,controlcondition(n=457)

Implicit associations

Explicit attitude

– .01 (p=0.92) .09 (p=0.20) .04 (p=0.61) Explicitattitude .01(p=0.92) – .69 (p<0.001) .33 (p<0.01) Intentiontoparticipate .09(p=0.20) .69 (p<0.001) – .47 (p<0.001) Attendance .04(p=0.61) .33 (p<0.001) .47 (p<0.001) – Knowledge .04(p=0.54) .01 (p=0.83) .07 (p=0.12) .04 (p=0.46) Educationallevel .08(p=0.24) .15 (p=0.001) .09 (p=0.05) .07 (p=0.41)

(p=0.29) .07 (p=0.17)

.12 (p=0.02)

(p<0.001) .56 (p<0.001)

.40 (p<0.001)

(p=0.61) .03 (p=0.64) .08 (p=0.16) xxx–xxx

(8)

Itisimportantthatwomenhavesufficientdecision-relevant knowledgeaboutthe advantagesanddisadvantages of partici-patingin breastcancer screeningand are enabled tomakean informedchoice [7]. It could beargued that this is especially importantwhentheymakethisdecisionforthefirsttime,since future attendance had been shown to be strongly related to attendanceatthefirstscreeninground[30].Ourresultsindicate thattheinformation leafletincreasestheknowledge ofwomen aboutthebreastcancerscreeningprogramme.Thiseffectwasthe largest in the subgroup of first-time invitees. Although most participants in this study had been invited for the screening programmemultipletimesbeforeandthereforehadreceivedthis (orasimilar)informationleafletpreviously,thisstudystillfound anincreaseinknowledgeafterreceivingtheleaflet.Possiblynot all women read the leaflet when they receive it with the invitation or they may have forgotten details over time. We expectedthatinthecontextofthestudy,participantsweremore likelytoreadtheleafletmoreintensivelythanwhentheyreceived it with the screening appointment invitation. Therefore, the foundeffectsonknowledgeandexplicitattitudesmaybesmaller inpractice.

4.2.Practiceimplications

Theinformationleaﬂetcanhelpincreaseknowledgeaboutthe screeningprogrammeandtherebyincreasethenumberofwomen making an informed choice. This is especially important for womenwhoareinvitedfortheﬁrsttime,becausetheirlevelof knowledgeislower.

We recommend to raise attention towards and interest in readingtheofficialleaflet.Thisisimportanttokeepwomen up-to-date about changes and insights concerning the screening programme. Next to that, new research can explore how informationcanbestbeprovided.Differentmodesofdelivering information towomen can be studied,such asinfographics or movies, as well as exploring the use of different distribution channelssuchasemail,publishinginlocalnewspapers,viasocial media,orviacommunitygroups.Abarriermightbethatwomen areinvitedforbreastcancerscreeningbienniallyoveraperiodof 24yearsandarethereforepotentiallynotinterestedingathering informationeverytimetheyareinvited.Possibly,more personal-ised information can be offered to first-time invitees and previouslyinvitedparticipants.

4.3.Conclusion

In conclusion, providing an information leaflet to women invitedforbreastcancerscreeningledtoslightlyhigherlevelsof knowledge,andmorewomenwithpositiveexplicitattitudes,in particularamongstwomenwhowereinvitedforthefirsttime.In first-timeinviteesbaselineknowledgewaslessoftensufficient,but theleafletincreasedthis.Intentiontoparticipateandattendance seemtobeassociatedwithexplicit attitude,however, notwith implicitassociations.

Funding

This research was funded by the Centre for Population Screening(CvB)oftheDutchNationalInstituteforPublicHealth andtheEnvironment(RIVM).TheCvBwasnotinanywayinvolved inthestudydesign,inthecollection,analysisandinterpretationof thedata,inthewritingofthereportorinthedecisiontosubmitthe paperforpublication.

CRediTauthorshipcontributionstatement

LindyM.Kregting:Datacuration,Formalanalysis, Investiga-tion,Methodology,Visualization,Writing-originaldraft,Writing -review&editing.NicolienT.vanRavesteyn:Conceptualization, Methodology,Supervision, Writing - review &editing. Wolfert Spijker: Conceptualization, Writing - review & editing. Tessa Dierks: Conceptualization, Writing - original draft, Writing -review &editing.ClareA. Aitken:Conceptualization,Writing -review & editing. H. Amarens Geuzinge: Conceptualization, Writing - review & editing. Ida J. Korfage: Conceptualization, Formalanalysis,Fundingacquisition,Investigation,Methodology, Supervision,Writing-originaldraft,Writing-review&editing. DeclarationofCompetingInterest

None.

Acknowledgements

Wewanttothankallthewomenwhoparticipatedinthisstudy andcompletedthequestionnaires.Wewouldalsoliketothankthe regionalscreeningorganisation‘BevolkingsonderzoekZuid-West’ for their co-operation in contacting potential participants and gatheringtheattendancedata.

AppendixA.Supplementarydata

Supplementarymaterialrelatedtothisarticlecanbefound,in the online version, at doi:https://doi.org/10.1016/j. pec.2020.06.032.

References

[1]M.Perugini,Predictivemodelsofimplicitandexplicitattitudes,Br.J.Soc. Psychol.44(Pt1)(2005)29–45.

[2]T.D.Wilson,S.Lindsey,TYS,Amodelofdualattitudes,Psychol.Rev. 107(2000) 101–126.

[3]A.Dijksterhuis, P.Smith,R. vanBaaren, D. Wigboldus,The unconscious consumer:effectsofenvironmentonconsumerbehavior,J.Consum.Psychol. 15(3)(2005)193–202.

[4]R.Holland, M.deVries,B.Hermsen,A.van Knippenberg,Mood andthe attitude-behaviorlink:thehappyactonimpulse,thesadthinktwice,Soc. Psychol.Personal.Sci.3(3)(2012)356–364.

[5]M. Strick, R. van Baaren, R. Holland, A. van Knippenberg, Humor in advertisementsenhancesproductlikingbymereassociation,J.Exp. Psychol.15(1)(2009)35–45.

[6]H.vanAgt,J.Fracheboud,A.vanderSteen,H.deKoning,Dowomenmakean informedchoiceaboutparticipatinginbreastcancerscreening?Asurvey Table4

Repeatedmeasuresanalyses(ANOVA).

F-test Degreesoffreedom p-value Effectsize Baseline Prime 0.420 2;402 0.66 0.002 Target 0.327 1;403 0.57 0.001 Prime*target 2.078 2;402 0.13 0.010 ANCOVA1 Prime 0.469 2;324 0.63 0.003 Target 1.982 1;325 0.16 0.006 Prime*target 1.571 2;324 0.21 0.010 Follow-up Prime 0.086 2;431 0.92 0.000 Target 0.761 1;432 0.38 0.002 Prime*target 0.216 2;431 0.81 0.001 Prime*condition 2.402 2;431 0.09 0.011 ANCOVA1 Prime 2.255 2;358 0.11 0.012 Target 0.278 1;359 0.60 0.001 Prime*target 1.080 2;358 0.34 0.006 Prime*condition 2.835 2;358 0.06 0.016 1

correctedforthevariableseducationallevelandpreviousparticipationinbreast cancerscreening.

(9)

amongwomeninvitedforaﬁrstmammographyscreeningexamination, PatientEduc.Couns.89(2)(2012)353–359.

[7]J.Hersch, J.Jansen, K.McCaffery,Decision-makingabout mammographic screening:pursuinginformedchoice,Climacteric21(3)(2018)209–213. [8]BreastCancerScreenigEngland2016-17,Screening&ImmunisationsTeam

NHSDigital,2018[Availablefrom:https://ﬁles.digital.nhs.uk/pdf/m/f/ breast_screening_programme__england__2016-17_-_report__v2.pdf] accessedat18-03-2019.

[9]A.Anttila,M.Lehtinen,S.Heinavaara,M.Pankakoski,T.Sarkeala,BreastCancer ScreeningProgrammainFinland-AnnualReview2017,(2017).

[10]EarlyDetection-BreastCancerScreening,NationalCancerInstituteUSA,2015 [Availablefrom:https://progressreport.cancer.gov/detection/breast_cancer] accessedat12-02-2019.

[11]LETB-monitor2008,LETB,2010[Availablefrom:https://www.rivm.nl/sites/ default/ﬁles/2018-11/LETB%20monitor%20verslagjaar%202008.pdf]accessed at15-01-2019.

[12] Monitorbevolkingsonderzoekborstkanker2017/2018,IKNL,2019[Available from:https://www.rivm.nl/sites/default/ﬁles/2019-12/Monitor%

20bevolkingsonderzoek%20borstkanker%202017-2018.pdf]accessedat 20-05-2020.

[13]L.vanDam,I.J.Korfage,E.J.Kuipers,L.Hol,A.H.vanRoon,J.C.Reijerink,etal., Whatinﬂuencesthedecisiontoparticipateincolorectalcancerscreeningwith faecaloccultbloodtestingandsigmoidoscopy?Eur.J.Cancer49(10)(2013) 2321–2330.

[14]IjKorfage,EwdeKwaadsteniet,A.vanVoorst,AmStiggelbout,M.deVries,Ah. Pieterse,Preferencesforcervicalcancerscreening:theroleofimplicit associations,PatientEduc.Couns.101(2)(2018)310–317.

[15]StHawley,B.Zikmund-Fisher,P.Ubel,A.Jancovic,T.Lucas,A.Fagerlin,The impactoftheformatofgraphicalpresentationonhealth-relatedknowledge andtreatmentchoices,PatientEduc.Couns.73(3)(2008)448–455. [16]BevolkingsonderzoekBorstkankerUitnodiging,RIVM,2018[Availablefrom:

https://www.rivm.nl/documenten/folder-bevolkingsonderzoek-borstkanker-nl]accessedat09-09-2019.

[17]E.Mathieu,AlBarratt,K.McGeechan,HmDavey,K.Howard,N.Houssami, Helpingwomenmakechoicesaboutmammographyscreening:anonline randomizedtrialofadecisionaidfor40-year-oldwomen,PatientEduc.Couns. 81(1)(2010)63–72.

[18]S.Michie,E.Dormandy,Tm.Marteau,Themulti-dimensionalmeasureof informedchoice:avalidationstudy,PatientEduc.Couns.48(1)(2002)87–91. [19]A.Ghanouni,C.Renzi,S.F.Meisel,J.Waller,Commonmethodsofmeasuring’ informedchoice’inscreeningparticipation:challengesandfuturedirections, Prev.Med.Rep.4(2016)601–607.

[20]T.Marteau,E.Dormandy,S.Michie,Ameasureofinformedchoice,Health Expect.4(2)(2001)99–108.

[21]J.E.WareJr,K.K.Snow,M.Kosinski,B.Gandek,NewEnglandMedicalC,the HealthI.SF-36HealthSurvey:ManualandInterpretationGuide,TheHealth Institute,NewEnglandMedicalCenter,Boston,Mass,1997.

[22]A.G.Greenwald,D.E.McGhee,J.L.Schwartz,Measuringindividualdifferences inimplicitcognition:theimplicitassociationtest,J.Pers.Soc.Psychol.74(6) (1998)1464–1480.

[23]R.H.Fazio,M.A.Olson,Implicitmeasuresinsocialcognition.research:their meaninganduse,Annu.Rev.Psychol.54(2003)297–327.

[24]R.H.Fazio,D.M.Sanbonmatsu,M.C.Powell,F.R.Kardes,Ontheautomatic activationofattitudes,J.Pers.Soc.Psychol.50(2)(1986)229–238. [25]J.Hersch,A.Barratt,J.Jansen,L.Irwig,K.McGeechan,G.Jacklyn,etal.,Useofa

decisionaidincludinginformationonoverdetectiontosupportinformed choiceaboutbreastcancerscreening:arandomisedcontrolledtrial,Lancet 385(9978)(2015)1642–1652.

[26]M.J.Perez-Lacasta,M.Martinez-Alonso,M.Garcia,M.Sala,L.Perestelo-Perez, C.Vidal,etal.,Effectofinformationaboutthebeneﬁtsandharmsof mammographyonwomen’sdecisionmaking:theInforMarandomised controlledtrial,PLoSOne14(3)(2019)e0214057.

[27]M.Carles,M.Martinez-Alonso,A.Pons,M.J.Perez-Lacasta,L.Perestelo-Perez, M.Sala,etal.,Theeffectofinformationaboutthebeneﬁtsandharmsof mammographyonwomen’sdecision-making:studyprotocolfora randomizedcontrolledtrial,Trials18(1)(2017)426.

[28]W.Hofmann,B.Gawronski,T.Gschwendner,H.Le,M.Schmitt,Ameta-analysis onthecorrelationbetweentheimplicitassociationtestandexplicitself-report measures,Pers.Soc.Psychol.Bull.31(10)(2005)1369–1385.

[29]DeelnamedataBevolkingsonderzoekBorstkankerNederlandregioZuid-West, BevolkingsonderzoekZuid-West,2018.

[30]W.Scaf-Klomp,F.L.P.vanSonderen,R.Stewart,J.A.A.M.vanDijck,W.J.A.van denHeuvel,Complianceafter17yearsofbreastCancerscreening,J.Med. Screen.2(4)(1995)195–199.