Perceived Barriers and Parental Adherence to Recommendations following Child Psychological Assessment

Material Information

Perceived Barriers and Parental Adherence to Recommendations following Child Psychological Assessment
Davis, Eileen Matias
Place of Publication:
[Gainesville, Fla.]
University of Florida
Publication Date:
Physical Description:
1 online resource (127 p.)

Thesis/Dissertation Information

Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Clinical and Health Psychology
Committee Chair:
Committee Co-Chair:
Committee Members:
Graduation Date:


Subjects / Keywords:
Caregivers ( jstor )
Child psychology ( jstor )
Clinical psychology ( jstor )
Parenting ( jstor )
Parents ( jstor )
Professional schools ( jstor )
Psychological assessment ( jstor )
Psychology ( jstor )
Recommendations ( jstor )
Symptomatology ( jstor )
Clinical and Health Psychology -- Dissertations, Academic -- UF
adherence -- assessment -- child -- evaluation -- recommendations
bibliography ( marcgt )
theses ( marcgt )
government publication (state, provincial, terriorial, dependent) ( marcgt )
born-digital ( sobekcm )
Electronic Thesis or Dissertation
Psychology thesis, Ph.D.


Every year approximately 1.5 million youths participate in a psychological evaluation (MacNaughton & Rodrigue, 2001). As a result of these evaluations, psychologists provide families with a broad range of recommendations. Parents and caregivers are then responsible for accessing services, utilizing resources, and navigating systems within which these recommendations are to be implemented. Few researchers have explored factors that contribute to adherence to recommendations following child psychological assessment. What data is available suggests that adherence varies substantially depending on the type of recommendation prescribed, with recommendations for psychological services receiving the lowest levels of compliance across studies. Perceived barriers has also emerged as a consistent predictor of non-adherence across recommendation types (Dreyer, O'Laughlin, Moore, & Milam, 2010; Human & Teglasi, 1993; MacNaughton & Rodrigue, 2001). This study explored predictors of adherence as well as outcomes that may be impacted by increased adherence. Consistent with previous findings, adherence ratings varied by recommendation type, and lowest rates were reported for psychological recommendations. Perceived barriers was the primary predictor of adherence, with more perceived barriers predicting lower adherence. An interaction effect was also identified between recommendation type and adherence, wherein the strength of the relationship between barriers and adherence varied by recommendation type. Child symptom severity and impairment, parenting stress, and feedback modality did not predict adherence rates. Furthermore, higher adherence rates did not predict greater improvement in symptom severity or impairment. These results highlight the importance of implementing procedures to minimize the impact of barriers on parental adherence, such as incorporating a barriers assessment into child psychological assessments in order to tailor recommendations more carefully to the needs and resources of individual families. The lack of a relationship between adherence and changes in symptom severity or impairment should also be explored. Measures that are sufficiently specific to capture meaningful changes in child functioning but broad enough to be utilized with diverse clinical populations are needed. ( en )
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis (Ph.D.)--University of Florida, 2014.
Co-adviser: BOGGS,STEPHEN R.
Electronic Access:
Statement of Responsibility:
by Eileen Matias Davis.

Record Information

Source Institution:
Rights Management:
Applicable rights reserved.
Embargo Date:
LD1780 2014 ( lcc )


This item has the following downloads:

Full Text


ParentalAdherencetoClinicalRecommendations inanADHDEvaluationClinicmA.SamanthaDreyerParkAvenueHealthCareManagementmLizO’Laughlin,JannaMoore,andZacharyMilamIndianaStateUniversityThepresentstudyexaminedperceivedbarrierstoparentaladherence withchildpsychologicalassessmentrecommendations.Eightycaregiversofchildrenreferredtoanattention-deÞcit/hyperactivitydisorder (ADHD)evaluationcliniccompletedatelephoneinterview4to6weeks afterreceivingevaluationfeedback.Caregiversreportedadherenceto 81.5%ofrecommendationsandwereequallylikelytoengageinselfhelprecommendations(i.e.,informationonADHD)andthosefor professional-nonpsychologicalservices(i.e.,medicationconsultation). Caregiverswereleastlikelytofollowthroughonrecommendationsfor psychologicalservices(e.g.,parentalbehaviortraining,individual therapyforchildrenwithsymptomsofanxiety/depression).Higher ratherthanlowerlevelsofparentingstresswereassociatedwith greateradherence.Themostcommonlyreportedbarrierswerelackof timeandperceptionsthatthechild'steacherwasuncooperativewith implementingschool-basedrecommendations. & 2010WileyPeriodicals,Inc.JClinPsychol66:1101Ð1120,2010. Keywords:psychologicalassessment;parentaladherence;ADHD; helpseekingbehavior Formanychildclinicians,assessmentisseenasanecessaryprerequisitefor developing,implementing,andevaluatingthesuccessofanintervention(Mash& Hunsley,2005).Althoughthepurposeandextentofchildassessmentwillvary dependingonthesettingandreasonforreferral,acommongoalofassessmentisto facilitatepositiveoutcomesforthechildandhisorherfamily.Thesuccessofthis goalisdependentontheextenttowhichthetreatmentrecommendationsare followed.Althoughresearchhasexaminedparentalcompliancewithchildtherapy Correspondenceconcerningthisarticleshouldbeaddressedto:LizOLaughlin,PsychologyDepartment, IndianaStateUniversity,TerreHaute,IN47809; JOURNALOFCLINICALPSYCHOLOGY,Vol.66(10),1101--1120(2010) & 2010WileyPeriodicals,Inc. PublishedonlineinWileyOnlineLibrary(


(e.g.,Kazdin&Wassell,1999),thereisverylittleresearchexaminingparental adherencetopsychologicaltreatmentrecommendationsafterchildassessment. Thereisalsoalackofresearchexaminingwhethercompliancewithtreatment recommendationsleadstoimprovedchildfunctioning(Geffken,Keeley,Kellison, Storch,&Rodrigue,2006).Toourknowledge,thereisnocurrentresearchthat speci“callyconsiderscompliancetoassessmentrecommendationsforchildren evaluatedforattention-de“cit/hyperactivitydisorder(ADHD). Aconservativeestimateofprevalencesuggeststhat3%to5%ofchildrenmeet diagnosticcriteriaforADHD(AmericanPsychiatricAssociation,2000).ADHDis oneofthemostcommonmentalhealthconcernsamongchildren(Pliszka,2007). ThereiscurrentlynomedicaltestorgoldstandardtodiagnosisADHD.Inaddition, therearemanyADHDlookalikesorsymptomsassociatedwithADHDthatmay stemfromotherfactors(Adler,Barkley,Wiles,&Ginsberg,2006).Practice guidelinesforADHDconsistentlycallforthoroughevaluation,generallyinvolving informationfrommultipleinformants,tocorrectlydiagnoseandtreatADHD.For example,theAmericanAcademyofPediatricsguidelinesforassessmentofADHD (2000)callforconductingacomprehensiveinterviewwiththeparenttodetermine ageofonsetandruleoutotherexplanationsforthedisorder,andobtainingrating scaledatafromparentsandteacherstoprovideevidenceofsymptomsand impairmentintwoormoresettings.Suchevaluationsareoftenconductedby psychologists,whoprovidefeedbackontheevaluationandrecommendationsto parentsorcaretakers,whothenmakethedecisionaboutthenextstepinaddressing thechildsneeds. Evidence-basedtreatmentsforADHDincludemedication(i.e.,stimulantsand atomoxotine),parentalbehaviortraining,behavioralclassroommanagement,and intensivesummerprogramsfocusedonimprovedpeerinteractions(Pelham& Fabiano,2008).Behavioralinterventionstypicallyinvolvecontingencymanagement proceduresimplementedbyparentsorcaregivers,teachers,andparaprofessional staffinthecaseofsummertreatmentprograms.Dopfneretal.(2004)consideredthe sequenceofmedicationandpsychosocialinterventionsandfoundthatthemajority ofchildreninitiallytreatedwithmedicationonlylaterrequiredadditionalbehavioral interventions.Incontrast,approximatelytwothirdsofchildreninitiallytreatedwith behavioralinterventionsdidnotrequireadditionalpsychopharmacologicaltreatment.Hoza,Kaiser,andHurt(2007)notethatneithermedicationnorbehavioral interventionsareeffectiveforallchildrenandthatbothtypesofinterventionshave limitations.Parentsoftenhaveconcernsaboutgivingmedicationtotheirchildren, andbehavioraltreatmentscanbedif“cultandtimeconsumingtoimplement.In addition,theorderofimplementationwillaffecttheincrementalbene“tofmultiple interventions,suchaswhenmedicationandbehavioralinterventionsarecombined (Hozaetal.). GiventhemanypossiblenegativeoutcomesassociatedwithADHDandthe empiricalsupportfortheef“cacyoftheADHD-focusedtreatmentsdiscussedabove, itisclearthatcompliancewithassessmentrecommendationsafteradiagnosisof ADHDisveryimportant.Althoughmanyofthesupportedinterventionsare administeredbyteachers(e.g.,classroommodi“cations)orphysicians(i.e., medication),parentsorcaregiversmostoftenareresponsiblefordecidingupona courseoftreatmentandseekingoutoradvocatingforchildservicesintheschool andcommunity.Gainingabetterunderstandingoftheparentalorcaregiverlevelsof complianceaswellasobstaclestocompliancearestepstowardabetterintegrationof assessmentandtreatmentpracticeforchildrenwithADHD.1102JournalofClinicalPsychology,October2010JournalofClinicalPsychology DOI:10.1002/jclp


Help-SeekingProcessforParents Researchershavenotedthatfewerthan50%ofchildrenwithmentalhealth problemsreceiveneededservices(e.g.,Kataoka,Zhang,&Wells,2002).Giventhat childrenaregenerallynotself-referred,thiswouldsuggestthatgreaterunderstanding ofparentalhelp-seekingbehaviorisneededtoclosethisgapbetweenneedand intervention.Numerousmodelshavebeenproposedtoexplainvariationsinhelpseekingbehavioramongadultsseekingmedicalservices(i.e.,Anderson,1995)and mentalhealthservicesmorespeci“cally(i.e.,Srebnik,Cauce,&Baydar,1996). Asmallernumberofstudieshaveconsideredhealth-seekingbehavioramongparents seekingservicesfortheirchildren(e.g.,Eiraldi,Mazzuca,Clarke,&Power,2006; Janicke&Finney,2003;Power,Eiraldi,Clarke,&Mazzuca,2005;Srebniketal.). Severalofthesestudieshavebuiltuponpreviousmodelsandincludefour interrelatedstagesofhelpseeking:problemrecognition,decisiontoseekhelp, serviceselection,andserviceutilizationpatterns.Followingisabriefreviewof variablesproposedtoberelatedtoparentalhelp-seekingbehaviorineachofthefour stages. Problemrecognition. Parentrecognitionofproblembehaviorinachildinvolves notonlythelevelofchildimpairmentbutalsothecaregivertolerancefordif“cult behavior(Eiraldietal.,2006;Poweretal.,2005).Eiraldietal.notethatculturehasa signi“cantin”uenceatallstagesofhelpseekingandthatparentsfromdifferent culturesmayhavedifferentthresholdsinperceivingbehaviorastypicalversus abnormal.IntheADHDHelp-SeekingModel(Eiraldietal.),variablesproposedto in”uenceproblemrecognitionincludeobjectiveassessmentofneed(i.e., multimethod,multi-informantassessment),parentalperceivedneedorburden,and parentalcharacteristicssuchasparentalpathology,levelofeducation,andparentchildrelationship.Teachercharacteristicsarealsoproposedtoin”uenceproblem recognitioninthecaseofADHDparticularly,asteachersmaybetheonesto“rst alertparentstoaproblemandguidehelp-seekingbehavior. Decisiontoseekhelp. Onceaparentbecomesawareofaproblem,understanding oftheproblemordisorderandknowledgeofhowtomanagetheproblemareneeded forparentstomakeadecisionaboutseekinghelp(Poweretal.,2005).Compliance withclinicalrecommendationsgenerallyfollowstheproblemrecognitionstage,as childassessmentismostoftentriggeredbyaparents(orateachers)reportofchild behavioraldif“culties.Poweretal.notethathealthlocusofcontrol(i.e.,beliefthat followingthroughonaprocesswillresultinimprovedchildhealth),parentalselfef“cacy,andthelevelofacculturationforimmigrantfamilieswillin”uencethe decisiontoseekmentalhealthservicesforchildren.JanickeandFinney(2003)found thattheinteractionofparentalself-ef“cacyandparentalstressbestpredicted utilizationofprimarycareservicesforchildren.Parentshighinself-ef“cacyandhigh instressweremorelikelytoseekservicesasopposedtoparentswithlesscon“dence intheirparentingabilities.Eiraldietal.(2006)proposethatdemographic characteristics(e.g.,childageandgender)andpsychologicalfactors(e.g.,parental attitudesandexpectationsregardingservices,knowledgeofADHD)willin”uence thedecisiontoseekhelp.InthecaseofADHD,boysaregenerallyidenti“edand referredforservicesmorecommonlythangirls,andimpulsiveandhyperactive behaviorsarenotedatanearlieragethaninattentivebehaviors.Parentsthatlack knowledgeofADHDorhaveconcernsaboutstigmatizationarelesslikelytoseek servicesfortheirchild(Eiraldietal.).Similarly,Corkum,Rimer,andSchachar1103ParentalAdherencetoClinicalRecommendationsinanADHDEvaluationClinicJournalofClinicalPsychology DOI:10.1002/jclp


(1999)foundgreaterknowledgeofADHDtobeassociatedwithgreateracceptance ofparentalbehaviortraining. Serviceselection. Oncethedecisionismadetoseekassistance,parentsmust decideonthetypeofsetting,provider(s),andintervention.Thisstepinthehelpseekingprocesshasagreatdealofrelevanceinconsideringadherenceas recommendationswilloftenincludeinterventionsassociatedwithdifferent providers(e.g.,medicaldoctor,childclinician,teacher)anddifferentsettings(e.g., school,medicalclinic,outpatientmentalhealthclinic).Onepurposeofthepresent studywastoconsiderwhichtypesofrecommendationsweremostlikelytobe followed,thusalsoprovidinginformationonparentchoicesrelatedtoservice selection.Poweretal.(2005)notethattrustintheprovider,socialsupport,andhelpseekingpatterns(i.e.,preferenceforformalvs.informalsettings)in”uenceservice selection.Alowerlevelofbothinterpersonaltrust(i.e.,trustinprovider)and institutionaltrust(i.e.,trustinhealthcaresystem)willcontributetolesslikelihood ofseekingservices.AccordingtoSrebniketal.(1996),socialsupportnetworkscan eitherfacilitateorimpedeserviceutilization.Ethnicminorityparents,inparticular, maybelesslikelytoseekoutformalservicesfrommentalhealthprofessionals withouttheinterventionofatrustedliaisonorgatekeepersuchasateacher,spiritual leader,orphysician(Cauceetal.,2002).TheEiraldietal.(2006)ADHDHelpSeekingModelalsoincludesconsiderationofeconomicfactors(e.g.,health insuranceorMedicaidcoverage)andsocietalfactors(racialbias,discrimination) inparentserviceselectiondecisions. Patternsofserviceuse. GiventhatADHDisachroniccondition,theADHD Help-SeekingModelincludesafourthstagefocusedonuseandadherencetoservices overtime.Eiraldietal.(2006)considerdifferenttypesofservicesettings(e.g., primarycare,mentalhealth,school)forADHDinthisfourthstage.Theynotethat thelevelofdemandassociatedwithaparticularinterventionandserviceintegrity andqualityofcarewillin”uenceparentdecisionsregardingpatternsofserviceuse overtime.Inthecaseofparentcompliancewithassessmentrecommendations,it maybethecasethatparentspickandchoosewhichinterventionstopursuesooner versuslaterbasedonfactorssuchastimedemandsandeaseofaccesstoaparticular service.Forexample,childrenareoftenreferredforassessmentbytheirprimary physician,thusparentsmaybeespeciallylikelytocomplywitharecommendationto consultwithaphysiciantodiscussresultsoftheassessmentwithafamiliarhealth careprovider.Theymaybelesslikelytocomplywitharecommendationtoseek psychologicalservicesbecauseofthegreattimedemand,aswellasperhapsless familiarityorcomfortwithpsychologicalservicesasopposedtomedicalservices. Theparenthelp-seekingmodelsproposedbyEiraldietal.(2006)andPoweretal. (2005)presentausefulfoundationfromwhichtoexamineparentcompliancewith assessmentrecommendations.Inparticular,thesemodelsproposefactorsthatmay bothhelpandhinderhelp-seekingbehavior.Eiraldietal.notethatgiventhelackof researchonparenthelpseekingrelatedtochildrenwithADHD,therelative importanceofthedifferentvariablesinpredictingmovementthroughthestagesis unknown.Theyalsonotethatcultureisapervasivein”uenceacrossallfourstages andthatculturalin”uencevariesbybothgroupandlocation.Recommendation adherencecouldbeconsideredpartoftheserviceselectionstageofhelpseekingas parentshavealreadyrecognizedaproblem(Stage1)thathaspromptedthemtoseek assessmentservices(Stage2).1104JournalofClinicalPsychology,October2010JournalofClinicalPsychology DOI:10.1002/jclp


StagesofChangeModel Anothermodelofbehavioralchangethatmayhaverelevanceinconsidering adherencetopsychologicalevaluationrecommendationisProchaskasstagesof changemodel(Prochaska,DiClemente,&Norcross,1992).Thismodeldescribes“ve stagesofbehavior:precontemplation(unawareofproblemorunwillingtochange), contemplation(recognizeproblemandconsideringchange),preparation(getting readytochangebehavior),action(engagedinbehaviorchange),andmaintenance (sustainingbehaviorchange).Individualsatanearlierstage(i.e.,precontemplation) aremorelikelytoberesistanttobehavioralchange.Compliancewithchild evaluationrecommendationsinvolvesacost-bene“tanalysisassociatedwiththe contemplationstage,inwhichanindividualconsiderswhethertoinitiatefurther treatmentorintervention.Compliancewithclinicalrecommendationscouldalsobe associatedwiththepreparationstage,inwhichtheindividualhasdecidedtotake actionbuthasnotyetbeguntodoso(Prochaskaetal.).Parentsthatseekchild assessmentbecauseofpressurefromothers(e.g.,teacherordoctor)maybeatan earlierstageofchangeandmaybemorelikelytoperceivebarrierstoadherence. BarrierstoTreatmentModels Kazdin,Holland,andCrowleys(1997)barriers-to-treatmentmodelconsiders factorsassociatedwithprematureterminationfromchildandparenttherapyfor externalizingdisorders.ThismodelwaslateradaptedbyMacNaughtonand Rodrigue(2001)toexamineparentcompliancewithchildpsychologicalevaluation recommendations.InKazdinetal.soriginalmodel,typesofbarriersincluded negativeperceptionsofthetreatment,practicalimpedimentstoparticipation,and poorclient-therapistrelationship.Parentperceptionofbarrierswaspredictiveof continuedparticipationintreatmentoverandabovevariablesknowntobe associatedwithprematuretermination(e.g.,socioeconomicdisadvantage,parent historyofantisocialbehavior,single-parentfamilies).Thus,perceptionoffewer barriersseemedtoserveaprotectivefunction. Inmodifyingthebarriers-to-treatmentmodel,MacNaughtonandRodrigue (2001)proposedfourtypesofbarrierstopsychologicalassessmentrecommendation adherence:problemswithaccesstoservices(e.g.,transportationproblems,lackof localprovider);negativeattitudesandbeliefs(e.g.,familymembersunwillingnessto comply,beliefthatrecommendationswillnothelp);schedulingproblems(e.g., parentsunableto“ndtimetofollowthrough);and“nancialproblems(e.g.,lackof insurance,nodiscretionaryincome).Theyalsosuggestedthattypesofrecommendationsstemmingfromchild-focusedassessmentcouldbeclassi“edintofour categories:psychologicalservices(i.e.,anytypeofpsychotherapyorfurther psychologicalassessment);school-basedrecommendations(e.g.,consultingwith theteacher,tutoring);professional-nonpsychologicalrecommendations(i.e.,referral toapediatrician);andactiveself-helprecommendations(i.e.,obtaininginformation onchildsdiagnosis). MacNaughtonandRodrigue(2001)examinedrecommendationadherenceof93 parentsandguardians,whosechildrenwereevaluatedatanoutpatientgeneral mentalhealthclinicforavarietyofreferralproblems.Theyconsideredseveral additionalpredictorsofparentaladherencetorecommendations,includingparents occupation,education,age,race,familyincome,familystructure,abilitytorecall recommendationsatfollow-up,satisfactionwithevaluation,andparentlocus ofcontrol.Childgender,age,historyofpsychologicaltreatmentorassessment,1105ParentalAdherencetoClinicalRecommendationsinanADHDEvaluationClinicJournalofClinicalPsychology DOI:10.1002/jclp


andproblemseveritywerealsoconsidered.Participantswerecontacted4weeksafter receivingevaluationfeedbackandaskedhowmanyandwhichrecommendations theyhadfollowed,howmanyrecommendationstheyrecalled,andiftheyhad perceivedanybarrierstoadherencetotherecommendations.Onaverage,parents reportedadherenceto67%ofallrecommendations,withgreatestadherencetothose forprofessionalnonpsychologicalconsultation,suchasmedicationconsultation (81%adherencerate),andlowestadherencetothoseforpsychologicalservices(47% compliancerate).MacNaughtonandRodriguefoundthatthebestpredictorof compliance,regardlessoftypeofrecommendationmade,wasthenumberofbarriers (regardlessoftype)thatparentsperceived.Noneoftheothervariablesthatthey considered(e.g.,parentcharacteristics,severityofchildproblems,satisfactionwith evaluationservices,recallofrecommendations,orparentallocusofcontrol)were foundtobesigni“cantpredictorsofcompliance. Parentandchildcharacteristicsassociatedwithcompliance. Although MacNaughtonandRodrigue(2001)didnot“ndthatparentorchild characteristicspredictedlevelofcompliance,otherstudieshavefoundthelevelof parentingstress,familysocioeconomicstatus(SES),andlevelofchilddif“cultyto in”uencecompliancewithtreatment.Severalstudieshavefoundthatparents terminatingtreatmentprematurelytendtoreporthigherlevelsofparentingstress (Kazdin&Mazurick,1994;Kazdin&Wassell,1999).However,JanickeandFinney (2003)foundhigherparentingstresstobeassociatedwithseekingchildservices. LowerSEShasalsogenerallybeenassociatedwithlowerlevelsoftreatment complianceoroutcome(Armbruster&Fallon,1994;Kazdin&Mazurick;Kazdin& Wassell),althoughonestudy(King,Hovey,Brand,Wilson,&Ghaziuddin,1997) foundlowerSEStobeassociatedwithhigherlevelsofcompliancewith recommendationsformedicationandindividualtherapyafterdischarge.In addition,MacNaughtonandRodriguefoundnosigni“cantrelationshipbetween SESandrecommendationadherence.Researchexaminingtheimpactofseverityof childdysfunctionontreatmentcompliancehasalsoresultedinmixed“ndings. Althoughseveralstudieshavefoundgreaterlevelofchildimpairmenttobe associatedwithlowerlevelsoftreatmentadherence(i.e.,Kazdin&Mazurick), MacNaughtonandRodrigue,foundthattheseverityofproblembehaviorsdidnot playasigni“cantroleinadherencetoassessmentrecommendations.Onepossibility isthatSESandchilddysfunctionmayhaveagreaterimpactontreatment complianceascomparedwithassessmentrecommendationadherence,perhaps relatedtothegreaterdemandsinvolvedintreatmentadherence(e.g.,severalmonths ofweeklysessions).However,giventhelimitedresearchonrecommendation adherence,furtherinvestigationoftheassociationamongchild,parent,andfamily factorsoncomplianceisneeded. AprimarygoalofthepresentstudywastoexaminewhetherMacNaughtonand Rodrigues(2001)versionofthebarriers-to-treatmentmodelisusefulinunderstandingbarrierstypicallyexperiencedbychildrenandparentsreferredforan ADHDevaluation.Thepresentstudyalsoconsideredparentalstressandthelevelof reportedchildbehavioralproblemsasfactorsassociatedwithparentalcompliance. Consistentwiththe“ndingsofMacNaughtonandRodrigue,itwasalsopredicted thatparentswouldreportadherencetoapproximately70%oftherecommendations andthattherecommendationwithwhichparentswouldmostfrequentlyreport compliancewouldbethatofconsultingwithaphysicianformedications.Toassess whethercompliancewithassessmentrecommendationsresultedinimprovedchild1106JournalofClinicalPsychology,October2010JournalofClinicalPsychology DOI:10.1002/jclp


functioning,wealsoconsideredwhetheraparentalreportofgreaterlevelsof adherencewouldbeassociatedwithgreaterimprovementinchildbehavior.Finally, consistentwiththebarrierstotreatmentmodel,wepredictedthatthenumberof perceivedbarrierswouldpredictparentcompliancewithrecommendationsafter takingintoaccountchildandfamilydemographicvariables.Thelevelofchild behavioraldif“culty,inparticular,waspredictedtobeassociatedwithgreaterlevels ofparentcompliance,asseeninpreviousresearchontreatmentcompliance(e.g., Kazdin&Mazurick,1994).Basedonthe“ndingsofJanickeandFinney,wealso predictedthathigherlevelsofparentingstresswouldcontributetothepredictionof greaterparentcompliancewithtreatmentrecommendations. Method Participants Ninety-“veparentsandcaregiverswhosechildrenwereevaluatedforADHD consentedtoparticipateinthestudy.Childrenwerereferredprimarilybygeneral physicianstobeevaluatedatanADHDevaluationclinic,asubspecialityclinic, whichispartofauniversity-basedtrainingcliniclocatedinamoderatelysized communityintheMidwest.Fifteenparticipantswerelosttofollowup,most commonlybecausetheydidnotrespondtomessagesortheirphonewas disconnected.Parentsandcaregiverslosttofollowupweresigni“cantlyyounger (mean[ M ] 5 31.06,standarddeviation[ SD ] 5 6.33years)thanthosewhocompleted thestudy, M 5 34.88years, SD 5 7.26, F (1,88) 5 3.99, p o . 05, Z25 .05 . Theaverage educationlevelofthemothersofchildrenwhosecaregiverswerelosttofollowup ( M 5 12.43years, SD 5 1.79)alsodifferedsigni“cantlyfromthatofthosewho completedthestudy, M 5 13.75years, SD 5 2.18, F (1,77) 5 4.49, p o .05, Z25 .06. Datawereanalyzedfrom80parentsandcaregiverswhoparticipatedinthefollowupphoneinterviewoveraperiodof18monthsandrangedinagefrom23to59years old( M 5 34.94, SD 5 7.16).Participantcaregiverswerelargelyfemale(88%), Caucasian(95%),andwerebiologicalmothersorfathers.Asmallpercentageof participants(7.5%)werelegalguardiansthatwerenotbiologicalparents.Children oftheparticipants(i.e.,childreferredforevaluation)were60boysand20girls,ages 5to13years( M 5 7.9, SD 5 1.6),whowereprimarilydiagnosedwithADHD (68.8%),OppositionalDe“antDisorder(21.3%),orLearningDisabilities(17.5%). Severalchildrenhadmorethanonediagnosis,andasmallpercentageofchildren (5%)weregivennodiagnosis.Participantreportoffamilyincomerangedfromless than$10,000tomorethan$100,000,however,themostfrequentlyreportedincome rangeswere$10,000to20,000and$20,000to30,000(i.e.,44%ofparticipantsin thesetwogroups).Regardlessofdiagnosis,allassessmentreportsincludedatleast threerecommendationsforaddressinglearningorbehavioralproblemsthat promptedthereferral. Measures ADHDevaluation. ThestandardbatteryusedattheADHDevaluationclinic includescognitiveandacademicachievementtesting,acomputer-basedtestof attention,parentandteacherformsofbehavioralratingscales,parentingstress measure,parentinterview,andchilddevelopmentalquestionnaire.Informationfrom thebehavioralratingscale,aparentingstressmeasure,anddemographic1107ParentalAdherencetoClinicalRecommendationsinanADHDEvaluationClinicJournalofClinicalPsychology DOI:10.1002/jclp


informationwereusedforthisstudy.Allotherinformationwasusedindetermining childdiagnosis,butnotspeci“callyusedforthepresentstudy. Conners-MarchDevelopmentalQuestionnaire(CMDQ). TheCMDQ(Conners &March,1999)isaself-reportmeasuredesignedtoprovideacomprehensive descriptionofachildsdemographicinformation,includingrace,socioeconomic status,familyconstellation,educationalhistory,familyandchildmedicaland psychiatrichistories,andchilddevelopmentalhistory.Informationaboutparent education,race,andhouseholdincomeweregatheredfromthismeasure. BehaviorAssessmentSystemforChildren-ParentResponseScales(BASC-PRS). TheBASC-PRS(Reynolds&Kamphaus,1998)isa130-item,parent-reportmeasure ofchildbehavioralproblems.TheBASC-PRSincludescompositeandindividual scalesofavarietyofbehaviors.ThisstudyutilizedthecompositeExternalizing ProblemsandInternalizingProblemsscales.TheExternalizingProblemsscale providesanassessmentofaggression,hyperactivity,andconductproblems,while theInternalizingProblemsscalemeasuresanxiety,depression,andsomatization. ReynoldsandKamphausreportedthatthecompositescoresshowhighinternal consistency( coef“cientalphas .85to.90)andtest-retestreliabilities( r 5 .88),andthe BASC-PRSdisplaysconcurrentvaliditywithothermeasuresofchildbehavior. Inthepresentsample,caregiverratingsoftheseverityofchildbehaviorwerequite varied,re”ectingthediversityofthesample(i.e.,parentsofchildrendiagnosedwith ADHDoroppositionalde“antdisorder(ODD),aswellasparentsofchildren diagnosedwithneitherdisorder).BASC-PRSinternalizingbehavioralcomposite scores(Tscore)rangedfrom33to95( M 5 52.63, SD 5 13.77),whileexternalizing behavioralcompositescoresrangedfrom38to97( M 5 60.01, SD 5 12.34). ParentingStressIndex(PSI)-ShortForm. ThePSI-ShortForm(Abidin,1995)is basedontheassumptionthatparentingstressisdeterminedbychildandparent characteristicsandinteractionsbetweenparentandchild.Thismeasureincludesfour clinicalsubscales:TotalStress,ParentalDistress(PD),Parent-ChildDysfunctional Interaction(P-CDI),andDif“cultChild(DC).TheTotalStressscaleprovides informationontheoverallparentingstresslevel.Abidinreportsgoodinternal consistencyfortheTotalStressscore( coef“cientalpha 5 .91),aswellasgoodtestrestreliability( r 5 .84).Internalconsistencyforthecurrentsamplewas.96forthe TotalStressscale,.95forParentDistress,.93forParent-Childdysfunctional interaction,and.83fortheDif“cultChildsubscale.ThePSI-SFhasbeenusedin priorstudiesoftreatmentcompliancebyparentsofchildrendiagnosedwithADHD (e.g.,Wellsetal.,2000). AdherenceTelephoneInterviewForm(ATIF;MacNaughton&Rodrigue,2001). TheATIFwasdevelopedbyMacNaughtonandRodrigueforuseintheirstudyof parentalcompliancetorecommendationsgivenintheirchildspsychological evaluation.TheATIFbeginswithanexplanationofwhattheinterviewwillentail. Next,theinterviewerreadseachrecommendationfromthepsychologicalevaluation totheparentorcaregiverandasksifsheorhehascompletedthatrecommendation. Parentsorcaregiversarethenaskedtoindicateifanyofthefollowingbarriersmade itdif“cultforthemtocompletetherecommendation:didntthinkitwouldhelp,no longeraproblem,resourcesnotavailableinmycommunity,transportation, insurance,time,andforgottodoit.Finally,theyareaskedtoprovideanyother reasonsforwhichtheyhaddif“cultyfollowingthroughontherecommendation.1108JournalofClinicalPsychology,October2010JournalofClinicalPsychology DOI:10.1002/jclp


MacNaughtonandRodriguefoundgoodinter-rateragreementonwhetherparents compliedwithrecommendations. Forthepresentstudy,theATIFwasmodi“edslightly.Insteadofthequestion, Didyoucompletethisrecommendation?themodi“edATIFstates:Onascaleofone to“ve,with1beingnotatall,3beingsomewhat,and5beingcompletely, pleasetellmehowmuchyouthinkyoufollowedthisrecommendation.Severalnew questionswerealsoaddedtotheATIF.Parentswereaskedtorate,ona5-point Likertscale,howimportanttheythoughteachrecommendationwas,rangingfrom1 ( notimportant ),to3( somewhatimportant ),to5( extremelyimportant ).Parentswere alsoaskedtoindicateiftheirchildsbehaviorhadchangedsincethefeedbacksession andtoratethelevelofimprovement,rangingfrom1( alittlebetter ),to3( better ),to 5( muchbetter ),orworseningofthebehavior,rangingfrom1( alittleworse ),to3 ( worse ),to5( muchworse ).Inaddition,parentswereaskedtoindicateonthesame 5-pointscalesiftheirchildsin-schoolbehaviorhadchangedbasedonteacher feedback.Finally,parentswereaskedanopen-endedquestionaboutwhatcould havehelpedthemtofollowtherecommendationsbetter. Participantsreceivedthreetoeightrecommendationsperevaluation( M 5 5.15, SD 5 1.1).Ratesofcomplianceweremeasuredintwoways:inadichotomous mannerconsistentwiththeworkofMacNaughtonandRodrigue(2001)andusing aLikertratingthatre”ectedcaregiversreportsofvaryinglevelsofcompliance. Thedichotomouscompliance.Thedichotomouscompliancewascalculatedasthe numberofrecommendationsfollowed(de“nedasinitiatingorcompletingthe recommendation)dividedbynumberofrecommendationsgiven.Aparentwho receivedan80%dichotomouscomplianceratemayhaveamuchlower(i.e.,40%) Likertadherenceratethatre”ectslowerlevelsofcompliance. Procedure AftertheADHDevaluation,twopsychologistsandupperlevelgraduatestudents madediagnosesandrecommendationsforeachchildthroughdiscussionofalltest data.ConsistentwithDiagnosticandStatisticalManualofMentalDisorders, 4thed.(APA,2000),criteria,thelevelofchildfunctionalimpairmentwasassessed duringtheclinicalinterviewandconsideredindiagnosis.Theprimaryevaluator, supervisedbyalicensedpsychologistifagraduatestudent,wrotetheassessment report,includingthediagnosesandrecommendationsandprovidedfeedbacktothe parents.Parentsweregivencopiesofthereport,whichincludedspeci“c recommendationsduringthefeedbacksession.Feedbacksessionslastedapproximately1hour,duringwhichtheevaluatorexplainedthetest“ndingsanddiagnoses, detailsofeachrecommendation,andansweredcaregiverquestions.Allcaregivers weregivendiagnosis-relatedinformationhandouts,informationonparental behaviorinterventions,andthedailyschoolbehavioralreportcardduringfeedback, andtheywereencouragedtocalliftheyhadanyfurtherquestionsregardingthe evaluationorrecommendations. Participantswererecruitedforthestudyaftertheirfeedbacksessionandsigneda writtenconsent,whichindicatedthattheirparticipationwasvoluntaryandthat refusaltoparticipatewouldhavenoimpactonservicesprovided.Participantsalso signedaHIPAAconsentform,allowinginformationfromthechildsADHD evaluation(e.g.,BASCparentform,PSI,informationfromdevelopmental questionnaire)tobeusedinthepresentstudy.Becauseparentandcaregiver compliancewithevaluationrecommendationswasthefocusofthisstudy,child1109ParentalAdherencetoClinicalRecommendationsinanADHDEvaluationClinicJournalofClinicalPsychology DOI:10.1002/jclp


consentwasnotsolicited.Theuniversityinstitutionalreviewboardapprovedall proceduresandmeasuresbeforethestartofthestudy.Participantswereaskedto providecontactinformationfortwoindividuals,whocouldprovideparentcontact informationifthefamilyrelocatedorchangedtheirphonenumber. ParentscompletedallquestionnairesaspartoftheADHDassessmentbattery. Approximately4to6weeksaftertheirfeedbacksession,anupper-levelundergraduatewhowastrainedtoadministerthemodi“edATIFtelephonedtheparents andcaregivers.Theresearchassistantalsoclassi“edrecommendationsintothefour typesidenti“edbyMacNaughtonandRodrigue(2001).Commonrecommendations forallchildrenevaluatedincludedparenteducationonthechildsdiagnosisand parentalbehaviorinterventions,aswellasuseofaschoolbehavioralreportcard. ForchildrengivenadiagnosisofADHD,recommendationsgenerallyincluded consultationwithaphysicianregardingapossiblemedicationtrial,classroom modi“cations,andreferraltoachildclinicianforchildorfamilyinterventions.For the“rst30participants,boththeresearchassistantandtheprimaryinvestigator independentlyclassi“edtherecommendations,yieldinginter-rateragreementof 100%. Results ParentReportofCompliance Consistentwiththe“ndingsofMacNaughtonandRodrigue(2001),theaveragerate ofadherencewasapproximately70%whenconsideringtheextenttowhichparents followingthroughonrecommendations(i.e.,Likertrating).ThemeanLikert complianceratewas67.5%( SD 5 23.1%),whereasusingthedichotomousscalethe meancomplianceratewas81.5%( SD 5 22.4%) . AsseeninTable1,thehighestrate ofcompliancewasforself-helprecommendations.Aone-wayanalysisofvariance revealedasigni“cantdifferenceinratesofcompliancebetweenthetypesof recommendationsforLikertratings, F (3,410) 5 8.24, p o .001, Z25 .06.Resultsofa chi-squareanalysisalsoshowdifferencesincompliancebyrecommendationtypefor dichotomousratings, w25 14.94, degreeoffreedom [ df ] 5 3, p 5 .002, Z 5 .19.Asseen inTable1,thelowestrateofcompliance(72%)wasforpsychologicalservicesas comparedwiththehighestpercentageofcomplianceforactiveself-help(91%). Table1MeanRecommendationComplianceRates%Adherencerate( SD ) TypeofrecommendationDichotomousLikert Activeself-helpa90(30)79(32) Professional-nonpsychologicalb88(33)78(36) School-basedc78(42)61(41) Psychologicalservicesd72(45)58(43) SD 5 standarddeviation.aParenttoinitiateorengageinsomeformofactiveself-helpstrategy.bConsultwithaprofessionalotherthanamentalhealthprofessional.cInvolvingtheschool,tutoring,orschoolacademic-relatedprograms.dAnytypeofpsychotherapyoranotherpsychologicalevaluation.1110JournalofClinicalPsychology,October2010JournalofClinicalPsychology DOI:10.1002/jclp


AsseeninTable2,post-hocanalysesforbothLikertanddichotomousratings showthatparentsfollowedthroughonrecommendationsregardingactiveself-help atahigherratethanschool-basedinterventionsorpsychologicalservices.Post-hoc analysesofLikertratingsshowthatparentsweremorelikelytocomplywith professional/nonpsychological(i.e.,consultwithaphysicianregardingmedication) recommendationsascomparedwithschool-basedorpsychologicalservicerecommendations.However,therewasnosigni“cantdifferenceinratesofcompliance (Likertanddichotomousratings)betweenprofessional/nonpsychologicalandselfhelprecommendations. BarrierstoCompliance Thevastmajorityofrespondents(92.5%)reportedencounteringatleastonebarrier tocompliance( M 5 2.6, SD 5 1.7).AsseeninTable3,themostcommonlyreported barriertocompliancewaslackoftimetocarryouttherecommendation.Theleast reportedbarrierwaslackofinsurancecoverage.Numberofperceivedbarriers reportedwassigni“cantlynegativelycorrelatedwithbothratingsofcompliance (dichotomous r 5 .45, p o .001;Likert r 5 .62, p o .001). ImportanceofRecommendationsandRatingsofChildImprovement Caregiversrated74.9%ofrecommendationsasextremelyimportant( M 5 4.58, SD 5 .85).Therewasnosigni“cantdifferenceinparentratingsofimportanceforthe variouscategoriesofrecommendations.Therewasasigni“cantassociationbetween caregiverratingoftheimportanceoftherecommendationandcompliance(Likert Table2DunnetsCPost-HocPairwiseComparisonsofMeanComplianceRates1234 RecommendationtypeDLDLDLDL 1.School-based…… .12 .18 .11 .17.06.03 2.Activeself-help.12.18……. .02 .01…….16.204.Psychologicalservices .06 .03 .18 .21 .16.20…… Note. D 5 differenceinmeandichotomouscompliancerates;L 5 differenceinmeanLikertcompliance rates.p o .05. Table3CommonBarrierstoComplianceBarrierSubjectsreportingbarrier(%) Time 38.8 Lackofteachercooperation37.5 Resourcesnotavailableinmycommunity28.8 Wantedtotrybehavioralinterventionsbeforemedications23.8 Waitingforappointment18.8 Didnotthinkitwouldhelp13.8 Insurance 8.81111ParentalAdherencetoClinicalRecommendationsinanADHDEvaluationClinicJournalofClinicalPsychology DOI:10.1002/jclp


r 5 .36, p o .001).Likertcompliancerate,butnotdichotomouscompliancerate ( p o .09),wassigni“cantlyassociatedwiththelevelofimprovementreportedby caregivers( r 5 .34, p 5 .007). Themajorityofcaregivers(76.3%)reportedthattheybelievedthattheirchilds behaviorhadimprovedsincereceivingtherecommendations.Approximatelyhalf thecaregivers(49%)reportedthattheyhadreceivedfeedbackfromtheirchilds teacherabouttheirchildsbehaviorsincetheevaluation,and90%oftheseindicated thatchildbehaviorhadimproved. RelationshipBetweenComplianceRatesandChild/FamilyVariables Morethan50%ofparticipantscamefromhouseholdswithincomesof$30,000or less.However,neitherincomenoranyotherdemographicvariablewasfoundtobe signi“cantlyrelatedtocompliancerates.Neitherexternalizingnorinternalizing behavioralcompositescoresweresigni“cantlyrelatedtoratesofcompliance, however,therewasasigni“cantpositivecorrelationbetweenparentingstressand adherencerates(dichotomous r 5 .23, p o .05;Likert r 5 .22, p o .05). Previousresearchhasindicatedthatparentsofchildrenwithexternalizing disordersaremorelikelytoseektherapyservicesascomparedwithparentsof childrenwithotherbehavioralproblems(e.g.,Cohen,Kasen,Brook,&Struening, 1991).Inaddition,parentsofchildrenwithgreaterlevelsofimpairment(i.e.,those meetingcriteriaforADHDorODD)maybemorelikelytofollowthroughwith evaluationrecommendations(Kataokaetal.,2002).Thus,ratesofcomplianceand factorsassociatedwithcompliancewereexaminedforthesubsetofparentsand caretakershavingachilddiagnosedwithADHDorODD( n 5 59).Resultswere similartothosefoundforthetotalsample,exceptnoneofthePSI-SFscaleswere signi“cantlyrelatedtocomplianceinthissubsample.Caregiversreportedamean dichotomousadherencerateof81.48%( SD 5 22.36%)andameanLikert compliancerateof67.50%( SD 5 23.05%).Resultsrevealedthatthetotalbarriers wassigni“cantlyassociatedwithcomplianceratesatmagnitudessimilartothetotal sample(Likert r 5 .56, p o .001). PredictorsofParentCompliance Astepwisehierarchicalregressionwasconducted,enteringdemographicvariables (parentage,education,familyincome)andparentratingsoftheseverityoftheir childrensbehavioralproblemsinthe“rststepandPSI-SFTotalStressscoreand numberofbarriersinthesecondstep,aspredictorsofcompliance(Likert compliance).CorrelationsbetweenthepredictorvariablesarepresentedinTable4. AsseeninTable5,the“rstmodel,whichincludeddemographicinformationonly, wasnotsigni“cant.Thesecondmodel,inwhichparentingstressandnumberof barrierswereincludedaspredictors,wassigni“cantandaccountedfor44%ofthe overallvariance.Totalbarriers,butnotparentingstress,wasasigni“cantindividual predictorofcompliance. Todetermineiftherewasapossiblecurvilinearrelationshipbetweenparenting stressandcompliance(i.e.,ifparentslowandhighonstressweremorelikelyto complywithrecommendations),weconductedacurve“tregressionanalysesusing SPSS(Version16.0),consideringalinearmodel“rst,followedbyamodelincluding bothlinearandquadraticterms.Resultsrevealedasigni“cantlinearrelationship betweenparentingstressandcompliance;Model1: R25 .050, F (1,76) 5 4.02, p o 05.Model2,addingthequadraticterm,wasnolongersigni“cant, R25 .057,1112JournalofClinicalPsychology,October2010JournalofClinicalPsychology DOI:10.1002/jclp


F (2,75) 5 2.26, p 5 .11,andtherewasalmostnoincreaseinvariance,thusindicating nosupportforacurvilinearrelationshipbetweenparentingstressandcompliance. Discussion Thefocusofthecurrentstudywasonrecommendationcompliance,whichmaybe conceptualizedasastageofchangethatprecedesmoreactivebehavioralchangeand focusesonprocessesinvolvedinpreparingtochangebehavior(i.e.,makingan appointmentfortreatment).Fromahelp-seekingmodelperspective,recommendationcompliancemaybeseenaspartofthedecisiontoseekhelpstage(e.g.,decision toengageinfurtherservices)ortheserviceselectionstage(e.g.,decisiontoseek formalorinformalsupports).Byconsideringfactorsassociatedwithcompliance withevaluationrecommendationsandcomparingthesewithfactorspreviously foundtobeassociatedwithtreatmentcompliance,thisstudyexploredwhether differentfactorsorobstaclesareassociatedwiththesedifferentstagesofchange. Caregiversinthecurrentsamplereportedcompliancewithapproximately70%of recommendationswhenconsideringextentofcomplianceorLikertratings.The dichotomouscompliancerateofover80%wassomewhathigherthanexpected, Table5StepwiseHierarchicalRegressionPredictingParentComplianceWithRecommendationsVariableBSEBXtR2Changein R2Step1 Parentage. Parenteducation .02.07 .26 .27 Familyincome. Childexternalizingtotal. .04.04 Step2 Parentage. Parenteducation .01.05 .01 .03 Familyincome. Childexternalizingtotal. PSItotalstress. Totalbarriers .33.05 .64 Note .SE 5 standarderror,PSI 5 ParentingStressIndex.p o .001. Table4CorrelationsAmongParent/FamilyCharacteristics,ParentingStress,andTotalBarriers123456 1.ParentageÂ….15.30 .18 .17 .17 2.IncomeÂ….36.09 .01 .10 3.MothereducationÂ… .29 .18 .12 4.BASC-ExtÂ….53.12 5.PSI-TotalÂ….00 6.Totalbarriers Â… Note. BASC-Ext 5 BehaviorAssessmentSystemforChildren,ExternalizingProblems;PSI-Total 5 ParentingStressIndexTotal.p o .05;p o .01.1113ParentalAdherencetoClinicalRecommendationsinanADHDEvaluationClinicJournalofClinicalPsychology DOI:10.1002/jclp


giventhatMacNaughtonandRodrigue(2001)foundcomplianceratesoflessthan 70%whenusingadichotomousmeasure.Thehighrateofcompliancewith recommendationsforself-helpactivitiesmayhavebeenanartifactoftheADHD clinicfeedbackprocedure.Themostcommonself-helprecommendationintheclinic wasthatparentsandcaregiverseducatethemselvesaboutADHDorODD.This recommendationwas,atleastpartly,accomplishedduringthefeedbacksession,as parentsorcaregiverswereprovidedwithseveralhandoutsonADHD/ODDandthe contentofthesehandoutswasdiscussedinthecontextofsuggestingspeci“cparental behavioralinterventions(e.g.,consistentrules,useofrewardsandconsequences).It maybethatcaregiversinterpretedtheirparticipationinfeedbackascomplyingwith thisrecommendationand,thus,in”atedthecompliancerateforthistypeof recommendation,aswellastheoverallcompliancerate.However,examiningrates ofcompliancebytypeofrecommendationandexcludingself-helprecommendations, therateofLikertanddichotomouscompliancedecreasedonlyslightly(5%for Likertand3%fordichotomous),thussuggestingcompliancewithself-help recommendationsdidnotgreatlyin”atetheoveralladherence.Itisalsopossible thattheparentandcaregiverreportsofcompliancemayhavebeenaffectedbysocial desirabilityfactors.Theymayhavein”atedorexaggeratedlevelsofcompliancein responsetotheresearchassistantsinquiriestopresentthemselvesinafavorable light. CaregiverswhoreportedgreatercomplianceratesontheLikertscalealsoreported greaterlevelsofimprovementintheirchildrensbehavior,andthemajorityof caregiversreportedsuchimprovement.Althoughtheseresultssupportthepremisethat compliancewithtreatmentrecommendationsisassociatedwithimprovedchildwellbeing,thereareotherfactorstoconsider .Caregiverswhocompliedwithrecommendationstolearnmoreabouttheirchildrensdif“cultiesmayhavebecomelesscriticalof theirchildrensbehavior(i.e.,interpretedpoorlisteningskillsbecauseofADHDrelatedde“citsratherthanmisbehavior)and,thus,reportedimprovementbasedon adjustedexpectationsratherthanactu alchildbehavioralchange.Itmaybethat caregiverswhohadmademoreefforttoco mplywithrecommendationsspentmore timeinteractingwiththeirchildrenand,therefore,hadmoreopportunitytoobserve positivechanges.Theyalsomayhavewantedtofeelthattheireffortswerehavingan impactandsoweremoreattunedtorealchangeand/ortheyinterpretedbehavioras improved.Caregiverscompliancewithrecommendationstoutilizebehavioral modi“cationtechniques(e.g.,positivereinforcement)orstarttheirchildrenon medicationmayalsohaveresultedinactualbehavioralchangeinthechild. Otherresearchhassuggestedthatparentsandcaregiversaremorelikelytocomply withrecommendationstoseeaphysicianformedicationthanothertypesof recommendations(Kingetal.,1997;MacNaughton&Rodrigue,2001).This“nding waspartiallysupportedinthepresentstudy,asparentsweremorelikelytocomply withrecommendationstoconsultwithanonpsychologicalprofessional(i.e., physician)thanfollowthroughwithschool-basedorpsychologicalservices recommendationsbasedonLikertratings.Consideringdichotomousratings,there wasnodifferenceincomplianceratesfortheprofessional/nonpsychologicaland othercategoriesofinterventions.Inaddition,bothtypesofratingsshowedthat parentswerejustaslikelytocomplywithself-helprecommendationsasmedication consultationreferrals.ItislikelythattheLikertratingsareamoreaccurate re”ectionofcomplianceandallowgreaterdetectionofrelationshipsbetween variables.Previousresearch(Flamer,1983)hasalsofoundthatLikertscaleswith moreresponseoptionsbetterrepresenttheunderlyingfactororconstructas1114JournalofClinicalPsychology,October2010JournalofClinicalPsychology DOI:10.1002/jclp


comparedwithmeasureswithfewerresponseoptions.Assuch,resultsmaybe interpretedtomeanthatparticipantswereequallylikelytofollowrecommendations forself-helporconsultingwithanonpsychologicalprofessionandmorelikelyto adheretothesetypesofrecommendationsthanschool-basedorpsychological servicesrecommendations. Parentsandcaregiversratedalmostalltherecommendationsasatleast somewhatimportantandalmostthreequartersoftherecommendationsas extremelyimportant.Justascompliancerateswerelikelyaffectedbysocial desirabilityfactors,caregiversmayhavefeltsomepressuretoraterecommendations asbeingatleastsomewhatimportant,becausetheywerespeakingwitharesearch assistantwhowasaf“liatedwiththeclinicthathadprovidedtherecommendations. Itisalsopossiblethatthesampleofcaregiverswhoparticipatedinthisstudywere biasedtoseetherecommendationsproducedfromtheevaluationsasimportant. Theyvoluntarilybroughttheirchildreninfortheevaluations,anditisunlikelythat theywouldhavedonesohadtheynotexpectedtoreceivesomebene“tfromthe evaluation.Notypeofrecommendationwasratedasmoreimportantthananyother type.Recommendationimportancewasalsopositivelyassociatedwithboth complianceratings,indicatingthatcaregiverswerelikelymoremotivatedtocomply withthoserecommendationsthattheybelievedweremostimportant. Contrarytoexpectations,theseverityofchildbehavioraldisorderwasnot associatedwithcompliancerates,evenamongthesub-sampleofparticipantswhose childrenhadbeendiagnosedwithexternalizingdisorders.Thelackofanassociation maybebecauseofthegreaterdemandsassociatedwithcomplyingwithtreatment (e.g.,attendingseveralweeksoftherapysessions)asopposedtoadheringto recommendationsafterapsychologicalevaluation(e.g.,schedulinganappointment withaphysician).Anotherpossibilityisthatfactorsthatin”uencecaregiver compliancedifferbasedonthelevelofmotivationandpreparednessforchange(e.g., StageofChangemodel;Prochaskaetal.,1992).Caregiversinthisstudyhadalready recognizedapotentialproblemwiththeirchild,madethedecisiontoseekassessment services,andfollowedthroughinparticipationintheevaluation.Thus,itseems likelythatthemajorityofparticipantswereinlaterstages(i.e.,preparation,action) asopposedtoearlier(precontemplation)stagesofchange.AsproposedbyPower etal.(2005),theseverityofchildbehavioralproblemsandthelevelofimpairment mayhavemorein”uenceatearlierstagesinthehelp-seekingprocess(i.e., recognizingproblembehavior). Bothparentalstressandthenumberofbarriersweresigni“cantlyassociatedwith compliancerates,withthesetwofactorsaccountingfor40%ofthevariance.As predicted,parentalstresswaspositivelyassociatedwithadherencetoevaluation recommendations.This“ndingisconsistentwithresearchshowingthatindividuals experiencinghigherlevelsofdistressaremorelikelytoseekouttreatment(i.e., Cramer,1999;Janicke&Finney,2003),andfurtherunderscorestheimportanceof specifyingwhethercomplianceresearchisfocusedonpreparationoractionprocesses ofchange.Itmaybethatstresshelpstomotivatecaregiverstofollowrecommendationsfortheirchildrenscare,butoncetheyhavebeguntreatment,thesourcesof thosehigherstresslevelsimpedefurthercompliance(e.g.,lackoffundsimpedes transportationtoweeklyappointments).ConsistentwithMacNaughtonand Rodrigue(2001),caregiversreportinggreaternumbersofperceivedbarriersreported lowerlevelsofcompliance.Thenumberofreportedbarrierswasalsothesinglemost powerfulpredictorofcompliance.Geffkenetal.(2006)suggestthatparent perceptionofbarrierswillnegativelyin”uenceadherenceatallstagesofhelpseeking.1115ParentalAdherencetoClinicalRecommendationsinanADHDEvaluationClinicJournalofClinicalPsychology DOI:10.1002/jclp


Almostallparticipantsreportedexperiencingatleastonebarriertocompliance. Themostcommonlyreportedbarriertocompliancewasalackoftimetocarry outtherecommendations,whichcaregiverselaboratedasmeaningthattheyeither didnothavetheopportunitytocarryoutrecommendationsorwerewaiting foranappointmentwithanotherserviceprovider.Itmaybethatcontacting parentsandcaregivers4to6weeksafterfollow-upwastoosoon,resultingin anoverreportingofthisbarrierandperhapsminimizingthereportofother barriersthatmaynotyethavearisen.Anotherpossibilityisthatparentsreported timeasacommonbarrier,becausetheywerestillcontemplatingtheirnextstepin helpseeking(e.g.,decisiontoseekfurtherassistanceandwhattypeofservice provider).Futureresearchoncompliancewithchildevaluationrecommendations maywishtoexplorefactorsspeci“callyassociatedwiththeserviceselectionstageof helpseeking(e.g.,trustinevaluator,socialsupportandpreferenceforformal/ informalservices). Lackofcooperationfromchildrensteachersincarryingoutschool-based recommendationswasthesecondmostcommonlycitedbarrier.Onepossible explanationforparentaltendencytoattributeblametoteachersmayinvolvetheselfservingbias.Respondingtoquestionsaboutcompliancemayhaveactivatedaselfservingbias(e.g.,Campbell&Sedikides,1999),suchthatparentstendedtoattribute alackofadherencetoothers(i.e.,childsteacher).Itisalsoquitepossiblethat caregiverreportsoflackofteachercooperationwerevalid.Teachersmaynothave hadsuf“cienttimetoimplementaschoolbehavioralreportcardorothersuggested interventions.Otherbarrierstoschool-basedinterventionsmayincludelackof communicationbetweentheparentandteacher,teacher(orparent)perceptionthat theschool-basedrecommendationswerenothelpful,orlackofresourcesto implementthesuggestedintervention(e.g.,teacherfeltsheorhedoesnothavetime tomonitorspeci“cchildbehaviorandcompletethedailyschoolbehavioralreport card).Furtherresearchisneededonbarrierstoschool-basedinterventionsfor childrenwithADHD. LimitationsofStudy CompliancewasassessedusingonlytheATIF,whichisaself-reportmeasure. Geffkenetal.(2006)advocatefordevelopmentofasystematicmeasureofadherence thatcouldthenbeusedtodetermineifadherenceisrelatedtooutcome.Studyresults mayhavebeendifferentifindependentlyveri“ablemeasuresofcompliance,suchas checkingwithphysicianstoseeifappointmentshadbeenscheduledandattended, hadalsobeenconducted.Asmentioned,therelativelyshortamountoftimebetween receivingrecommendationsandthefollow-uptelephonecallresultedinmostparents reportingthattheysimplydidnothaveenoughtimetocomply,andlikelyobscured theimpactofotherbarriertypes. ThemajorityofparticipantswereCaucasianandseekingservicesspeci“callyfor suspectedADHD,thusresultsmaynotgeneralizetoamoreethnicallyand diagnosticallydiversesample.Bussing,Gary,Mills,andGarvan(2007)found differencesbetweenAfricanAmericanandCaucasianparentsinperceptionsof treatmenteffectivenessforADHD,suggestingadherenceratesandperceived barrierstotreatmentmaydifferinamoreculturallydiversegroupofparentsand children.Eiraldietal.(2006)alsosuggestedthatcultureislikelytohaveapervasive in”uenceatallstagesofhelpseeking.Researchhasalsoshownthatparentsofgirls withADHDaremorelikelytoutilizeservicesiftheirdaughtersdisplaysymptomsof1116JournalofClinicalPsychology,October2010JournalofClinicalPsychology DOI:10.1002/jclp


depression,whileboysparentsutilizeservicesiftheirsonsarehavingdif“cultywith schoolworkorhavehighernumbersofADHDsymptoms(Graetz,Sawyer, Baghurst,&Hirte,2006).Assuch,havingalargernumberofgirlsinthesample mayhaveshowncomplianceratedifferencesassociatedwithchildgender.In addition,almostalloftheparticipantswerefemalecaretakers.Thereisverylittle researchthatexaminescaregivergenderandparentinginterventions(e.g.,Arnold, OLeary,&Edwards,1997)andnonethatconsiderscaregivergenderand compliancewithtreatmentand/orassessmentrecommendations.Itispossiblethat theinclusionoffathersorothermalecaretakersmayhavecontributedtodifferent “ndings.Finally,caregiversthatdidnotparticipateinthefollow-upinterviewtended tobeyoungerandlesseducated,thusresultsmaynotfullyre”ectobstaclesandlevels ofcomplianceforthissubgroup. FutureDirectionsandSuggestionsforFutureResearch Futurestudiesmayincludefurtherinquiryregardingtimeasabarriertobetter understandthisobstacle(i.e.,parenttimevs.timeneededtoaccessservices).Itis alsopossiblethatothertypesofbarrierswerenotreportedsimplybecauseparents hadnotyethadtheopportunitytoencounterthem.Increasingthetimebetween feedbackandfollow-upmayincreasethenumberandtypesofbarriersreportedand allowforfurtheranalysisintotherelativeimpactsofdifferentbarriertypes.Thehigh numberofcaregiversreportingdif“cultyimplementingschool-basedinterventions becauseoflackofteachercooperationleadstoquestionsaboutwhyteachersrefused tocooperate,andwhethertheexpectationthatparentsandcaregiversattemptto implementschool-basedrecommendationswithoutacliniciansassistanceis appropriate.Teachersmayalsohavedif“cultyimplementinginterventions,suchas adailybehavioralreportcard,withouttheassistanceofabehavioralconsultant. Futureresearchshouldconsiderassessingteacherperceptionsofbarriersto treatmentrecommendationstobetterunderstandobstaclesintheschoolsetting. Futureresearchmayincludemoreopen-endedquestionsabouthowparentsand caregiversfollowedthroughonrecommendations.Finally,itisrecommendedthat futureresearchshouldutilizeaLikertscalemeasureofadherence,ratherthana dichotomousmeasure,whichwouldallowtheexplorationoffactorsassociatedwith lowerlevelsofcompliance(i.e.,somewhatcomplied),whichcouldbetargetsof interventiontopromotegreaterlevelsofcompliance. ImplicationsandApplications Resultsofthecurrentstudysuggestseveralimplicationsforpractice.TheADHD evaluationclinic,wherethestudywasconducted,isasubspecialtyclinicthat primarilyprovidesassessmentservicesforchildren.Resultssuggestthatclinician assistancemayoftenbenecessaryinfollowingthroughonassessmentrecommendations.Assuch,itappearsthatcaregiversandchildrenwouldbene“tifassessments wereundertakenasthe“rststepinatreatmentprocess,withthesameclinician performingtheassessment,developingrecommendations,andprovidingtreatment. Insuchidealsituations,theclinicianwouldhavegreateropportunitytoaddress barrierstocomplianceandhelpparentsandcaregiverstoimplementschool-based recommendations.Consideringthehelp-seekingmodel,trustinproviderisakey factorpredictingserviceselection,particularlyforethnicminorityparents.Ifthe parenthasasupportiverelationshipwithachildsteacher,principal,orphysician,1117ParentalAdherencetoClinicalRecommendationsinanADHDEvaluationClinicJournalofClinicalPsychology DOI:10.1002/jclp


thenthisindividualmightserveasaliaisontoprovidesupportfortheparentin followingthroughontreatmentrecommendations. Ataminimum,thecurrentresultswouldsuggestthatschool-basedinterventions areunlikelytobesuccessfullyimplementedwithoutassistanceprovidedtothe parent,aswellastheteacher,tofacilitatecommunication,developrealisticgoals, andassistwithongoingtroubleshootingandinterventionmodi“cation.Inasurvey of142elementaryschoolteachers,JonesandChronis-Tuscano(2008)foundthatthe majorityofteachersreportedreceivinglittletrainingrelatedtoADHD.Regular educationteachersinthisstudyalsoreportedlittleincreaseinuseofbehavioral interventionsaftera2.5-hourin-servicetraining,suggestingthatabriefoverview maynotbesuf“cienttoresultinbehavioralchange.Inregardstothepresentstudy, theexpectationthatteacherswouldbeableandwillingtoimplementaschool behavioralreportcardwithonlybriefwritteninstructionsprovidedbyparentsina handoutmayhavebeenunrealistic. Motivationalinterviewingtechniquesduringassessmentandfeedbackmaybe effectiveinincreasingthelikelihoodofcompliancewithrecommendations(e.g., Miller&Rollnick,2002).Consistentwithmotivationalinterviewingtechniques, cliniciansareencouragedtohelpcaregiversunderstandhowfollowingrecommendationswouldhelpthemtoachievetheirspeci“cgoalsand,thereby,increase motivationtowardscompliance.Cliniciansarealsoencouragedtoassesscaregivers thoughtsandfeelingsabouttheirchildrensdiagnosesandaddressanysubsequent resistancetorecommendations.Intheeventthatacaretakerappearsoverwhelmed orminimallymotivatedduringthefeedback(e.g.,asksnoquestions),cliniciansmay stronglyadvisethatparentsseekoutthesupportofachildcliniciantoassistthemin followingthroughontreatmentrecommendations.Obtaininginformedconsentto provideresultsoftestingtothereferralsource(e.g.,childsphysician),aswellasany mentalhealthproviderscurrentlyworkingwiththechild(e.g.,therapistorcase managers),canalsohelptofacilitateconsistencyincare. Parentcompliancewithnonpsychologicalprofessionalconsultationinthepresent studygenerallyreferredtoarecommendationthatparentsconsultwithaphysician regardingpossiblemedicationtotreatsymptomsassociatedwithADHDorODD.If the“rstrecommendationthatparentsadheretoistherecommendationfor medication,andthemedicationleadstoimprovement,thenitispossiblethat parentsmaybelesslikelytofollowthroughwithinterventionsrequiringgreaterlevels ofparenttimeandenergy,suchasbehavioralmodi“cation.However,consideringthe resultsofDopfneretal.(2004),anumberofchildrenwithADHDtreatedwith medicationalonearelikelytorequireadditionalinterventionstoaddressareasof impairment.Thus,itmaybehelpfultodiscusspotentialshort-termandlong-term goalsandoutcomesrelatedtochildfunctioningduringthefeedbacksession. Acentralargumentofthisstudyhasbeenthatpsychologicalassessmentloses muchofitsvalueifparentsfailtoadheretorecommendations.Thepresentresults aregenerallyconsistentwiththatassertion,asgreatercompliancewasassociated withgreaterreportedimprovementinchildrensbehavior.Nonetheless,itmaybe arguedthattheevaluationitselfalsohadvalueasanintervention,asparents reportedgreatestcompliancewithself-helprecommendationsprovidedinthe feedbacksession,supportingFinnandTonsagers(1997)assertionthatevaluations serveatreatmentfunction.Thisstudy,therefore,supportsthecontinueduseof evaluationbypracticingpsychologistsandsuggeststhatevaluatorsshouldbeaware ofthetherapeuticimpactstheycanhaveandconducttheirevaluationsinwaysto capitalizeonthis.1118JournalofClinicalPsychology,October2010JournalofClinicalPsychology DOI:10.1002/jclp


ReferencesAbidin,R.R.(1995).Parentingstressindex:Testmanual.Charlottesville,VA:Pediatric PsychologyPress. Adler,L.A.,Barkley,R.A.,Wilens,T.E.,&Gin sberg,D.L.(2006).Differentialdiagnosisof attention-de“cit/hyperactivitydisorderandc omorbidconditions.PrimaryPsychiatry,13,1…14. AmericanAcademyofPediatrics.(2000).Committeeonqualityimprovementandsubcommitteeonattention-de“cit/hyperactivitydisorder.Diagnosisandevaluationofthe childwithattention-de“cit/hyperactivitydisorder.Pediatrics,105,1158…1170. AmericanPsychiatricAssociation.(2000).Diagnosticandstatisticalmanualofmental disorders(4thed.,textrevision).Washington,DC:APA. Anderson,R.M.(1995).Revisitingthebehavioralmodelandaccesstomedicalcare:Doesit matter?JournalofHealthandSocialBehavior,36,1…10. Armbruster,P.,&Fallon,T.(1994).Clinical,sociodemographic,andsystemsriskfactorsfor attritioninachildrensmentalhealthclinic.AmericanJournalofOrthopsychiatry,64, 577…585. Arnold,E.H.,OLeary,S.G.,&Edwards,G.H.(1997).Fatherinvolvementandself-reported parentingofchildrenwithattentionde“cit-hyperactivitydisorder.JournalofConsulting andClinicalPsychology,65,337…342. Bussing,R.,Gary,A.F.,Mills,T.L.,&Garvan,C.W.(2007).Culturevariationsinparental healthbeliefs,knowledge,andinformationsourcesrelatedtoattention-de“cithyperactivity disorder.JournalofFamilyIssues,28,291…318. Campbell,K.W.,&Sedikides,C.(1999).Self-threatmagni“estheself-servingbias:Ametaanalysisintegration.ReviewofGeneralPsychology,3,23…43. Cauce,A.M.,Domenech-Rodrigue,M.,Paradise,M.,Cochran,B.N.,Shea,J.M., Srebnik,D.,etal.(2002).Contextualin”uencesinmentalhealthhelpseeking:Afocus onethnicminorityyouth.JournalofConsultingandClinicalPsychology,70(1),44…55. Cohen,P.,Kasen,S.,Brook,J.S.,&Struening,E.L.(1991).Diagnosticpredictorsof treatmentpatternsinacohortofadolescents.JournaloftheAmericanAcademyofChild andAdolescentPsychiatry,30,989…993. Conners,C.K.,&March,J.S.(1999).Conners-Marchdevelopmentalquestionnaire.North Tonawanda,NY:Multi-HealthSystems. Corkum,P.,Rimer,P.,&Schachar,R.(1999).Parentalknowledgeofattention-de“cit hyperactivitydisorderandopinionsoftreatmentoptions.Impactonenrollmentand adherencetoa12-monthtreatmenttrial.CanadianJournalofPsychiatry,44,1043…1048. Cramer,K.M.(1999).Psychologicalantecedentstohelp-seekingbehavior:Areanalysisusing pathmodelingstructures.JournalofCounselingPsychology,46(3),381…387. Dopfner,M.,Breuer,D.,Schurmann,S.,Metternich,T.W.,Rademacher,C.,&Lehmkuhl,G. (2004).Effectivenessofanadaptivemultimodaltreatmentinchildrenwithattention-de“cit hyperactivitydisorder-globaloutcome.EuropeanChildandAdolescentPsychiatry, (Suppl.1),117…129. Eiraldi,R.B.,Mazzuca,L.B.,Clarke,A.T.,&Power,T.J.(2006).Serviceutilizationamong ethnicminoritychildrenwithADHD:Amodelofhelp-seekingbehavior.Administration andPolicyinMentalHealth,33(5),607…622. Finn,S.E.,&Tonsager,M.E.(1997).Information-gatheringandtherapeuticmodelsof assessment:Complementaryparadigms.PsychologicalAssessment,9,374…385. Flamer,S.(1983).AssessmentoftheMultitrait-MultimethodMatrixValidityofLikertScales viacon“rmatoryfactoranalysis.MultivariateBehavioralResearch,18,275…308. Geffken,G.R.,Keeley,M.L.,Kellison,I.,Storch,E.A.,&Rodrigue,J.R.(2006).Parental adherencetochildpsychologistrecommendationfrompsychologytesting.Professional Psychology:ResearchandPractice,37,499…505. 1119ParentalAdherencetoClinicalRecommendationsinanADHDEvaluationClinicJournalofClinicalPsychology DOI:10.1002/jclp


Graetz,B.W.,Sawyer,M.G.,Baghurst,P.,&Hirte,C.(2006).Gendercomparisonsofservice useamongyouthwithattention-de“cit/hyperactivitydisorder.JournalofEmotionaland BehavioralDisorders,14,2…11. Hoza,B.,Kaiser,N.M.,&Hurt,E.(2007).Mutlimodaltreatmentsforchildhoodattentionde“cit/hyperactivitydisorder:interpretingoutcomesinthecontextofstudydesigns. Clinical,Child,&FamilyPsychologyReview,10,318…334. Janicke,D.M.,&Finney,J.W.(2003).Childrensprimaryhealthcareservices:Socialcognitivefactorsrelatedtoutilization.JournalofPediatricPsychology,28(8),547…558. Jones,H.A.,&Chronis-Tuscano,A.(2008).Ef“cayofteacherin-servicetrainingfor attention-de“cit/hyperactivitydisorder.PsychologyintheSchools,45(10),918…929. Kataoka,S.H.,Zhang,L.,&Wells,K.G.(2002).Unmetneedformentalhealthcareamong U.S.children:Variationbyethnicityandinsurancestatus.AmericanJournalofPsychiatry, 159,1548…1555. Kazdin,A.E.,Holland,L.,&Crowley,M.(1997).Familyexperienceofbarrierstotreatment andprematureterminationfromchildtherapy.JournalofConsultingandClinical Psychology,65,453…463. Kazdin,A.E.,&Mazurick,J.L.(1994).Droppingoutofchildpsychotherapy:Distinguishing earlyandlatedropoutsoverthecourseoftreatment.JournalofConsultingandClinical Psychology,62,1069…1074. Kazdin,A.E.,&Wassell,G.(1999).Barrierstotreatmentparticipationandtherapeutic changeamongchildrenreferredforconductdisorder.JournalofClinicalChild Psychology,28,160…172. King,C.A.,Hovey,J.,Brand,E.,Wilson,R.,&Ghaziuddin,N.(1997).Suicidaladolescents afterhospitalization:Parentandfamilyimpactsontreatmentfollow-through.Journalof theAmericanAcademyofChildandAdolescentPsychiatry,36,85…93. MacNaughton,K.L.,&Rodrigue,J.R.(2001).Predictingadherencetorecommendationsby parentsofclinic-referredchildren.JournalofConsultingandClinicalPsychology,69, 262…270. Mash,E.J.,&Hunsley,J.(2005).Evidence-basedassessmentofchildandadolescentdisorders: IssuesandChallenges.JournalofClinicalChildandAdolescentPsychology,34,362…379. Miller,W.R.,&Rollnick,S.(2002).Motivationalinterviewing:Preparingpeopleforchange (2nded.).NewYork:Guilford. Pelham,W.E.,&Fabiano,G.A.(2008).Evidence-BasedPsychosocialTreatmentsfor Attention-De“cit/HyperactivityDisorder.JournalofClinicalChild&Adolescent Psychology,37,184…214. Pliszka,S.(2007).Practiceparametersfortheassessmentandtreatmentofchildrenand adolescentswithattention-de“cit/hyperactivitydisorder.JournaloftheAmerican AcademyofChildandAdolescentPsychiatry,46,894…921. Power,T.J.,Eiraldi,R.B.,Clarke,A.T.,&Mazzuca,L.B.(2005).Improvingmentalhealth serviceutilizationforchildrenandadolescents.SchoolPsychologyQuarterly,20(2), 187…205. Prochaska,J.O.,DiClemente,C.C.,&Norcross,J.C.(1992).Insearchofhowpeoplechange: Applicationstoaddictivebehaviors.AmericanPsychologist,47,1102…1114. Reynolds,C.R.,&Kamphaus,R.W.(1998).BASC:Behaviorassessmentsystemforchildren manual.CirclePines,MN:AmericanGuidanceService. Srebnik,D.,Cauce,A.M.,&Baydar,N.(1996).Help-seekingpathwaysforchildrenand adolescents.JournalofEmotionalandBehavioralDisorders,4,210…220. Wells,K.C.,Epstein,J.N.,Hinshaw,S.P.,Conners,C.K.,Klaric,J.,Abikoff,H.B.,etal. (2000).Parentingandfamilystresstreatmentoutcomesinattentionde“cithyperactivity disorder(ADHD):AnempiricalanalysisintheMTAstudy.JournalofAbnormalChild Psychology,28,543…553. 1120JournalofClinicalPsychology,October2010JournalofClinicalPsychology DOI:10.1002/jclp




2014 Eileen Matias Davis


To my husband, son, parents, family, friends, and colleagues without whom this journey would have been lonely and incomplete


ACKNOWLEDGMENTS I thank all the people who have supported me throughout my graduate school journey . I thank my husband, Rusty, whose flexibility, strength, and confidence in me made challenging circumstances manageable. I thank my parents, Maria Elena and Enrique who have always provided me with their unconditional love and support. I thank my godmother, Mabel, and uncle, Guille, who introduced me to places and possibili ties I had never considered for myself , giving me the courage to pursue great opportunities. I thank my sister, Elaine, and her husband, Jose, for their generosity and ongoing support. I thank my son, Max, my niece, Hailey, my godson, Daniel, and my cousin Arianne for filling my life with joy. I thank my great friend, colleague, and former roommate, Jenny, who was an amazing companion throughout graduate school and who continues to be an important part of my life. I thank my labmate, Melissa, whose knowledge and sense of humor were frequently lifesaving . I thank all the people who made the execution of this dissertation possible: Laura, Jessica, Evelyn, and Caroline for running my s tudy while I was away; Liz, Nicole , and the other graduate students and interns for helping with participant recruitment; Bridg et for being my statistics guru; and the wonderful clinical supervisors who were helpful in so many ways during recruitment. I thank my committee members for providing guidance and feedback as I pursue completion of this final graduation requirement. Finally , I thank my mentor, Brenda, who has truly been the ideal mentor. Her supportiveness, laidback nature, and hard work have helped carry me through every milestone with my sanity intact. She has also been a great teacher and model of clinical skill and professionalism. 4


TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................. 4 LIST OF TABLES ............................................................................................................ 8 LIST OF FIGURES ........................................................................................................ 10 ABSTRACT ................................................................................................................... 11 CHAPTER 1 INTRODUCTION .................................................................................................... 13 Unders tanding Parental Helpseeking and Compliance .......................................... 15 Barriers to Treatment ........................................................................................ 16 Stages of Change ............................................................................................. 17 Health Belief Model .......................................................................................... 18 Stages of Help Seeking .................................................................................... 20 Research on Parental Adherence to Clinician Recommendations .......................... 25 Research on Barriers to Adherence ........................................................................ 29 Feedback Following Psychological Assessment ..................................................... 30 Outcomes of Adherence ......................................................................................... 33 Aims and Hypotheses ............................................................................................. 34 Aim 1 ................................................................................................................ 35 Aim 2 ................................................................................................................ 37 Aim 3 ................................................................................................................ 37 Aim 4 ................................................................................................................ 38 Aim 5 and Hypothesis ....................................................................................... 39 Aim 6 and Hypothesis ....................................................................................... 40 5


2 METHODS .............................................................................................................. 41 Design ..................................................................................................................... 41 Participants ............................................................................................................. 41 Inclusion Criteria ............................................................................................... 43 Exclusion Criteria ............................................................................................. 44 Measures ................................................................................................................ 44 Adherence and Perceived Barriers ................................................................... 44 Child Symptom Severity and Functional Impairment ........................................ 46 Parenting Stress ............................................................................................... 48 Procedures ............................................................................................................. 49 3 RESULTS ............................................................................................................... 54 Demographics ......................................................................................................... 54 Preliminary Anal yses .............................................................................................. 55 Primary Analyses .................................................................................................... 59 Aim 1 Analyses ................................................................................................. 59 Aim 2 Analyses ................................................................................................. 68 Aim 3 Analyses ................................................................................................. 70 Aim 4 Analyses ................................................................................................. 71 Aim 5 Analyses ................................................................................................. 72 Aim 6 Analyses ................................................................................................. 73 4 DISCUSSION ....................................................................................................... 104 Summary of Findings ............................................................................................ 104 Challenges and Limitations ................................................................................... 110 Clinical Implications and Future Directions ........................................................... 114 6


APPENDIX: DEMOGRAPHIC QUESTIONNAIRE ...................................................... 120 LIST OF REFERENCES ............................................................................................. 122 BIOGRAPHICAL SKETCH .......................................................................................... 127 7


LIST OF TABLES Table page 3 1 Demographic characteristics — parent and child ................................................. 76 3 2 M ean child and caregiver age for whole sample and for groups A and B ........... 77 3 3 Mean child and caregiver age for participants with and without follow up data .. 78 3 4 Means and standard deviations for BASC2PRS, PIS, and PSI ......................... 79 3 5 Descriptive data for total barriers at 2 months and 4 months for participants with follow up data, total and by recommendation type. ..................................... 80 3 6 Descriptive data for adherence ratings at 2 months and 4 months, total and by recommendation type. ................................................................................... 81 3 7 Number of recommendations for participants with follow up data presented for total sample and by recommendation type. ................................................... 82 3 8 Frequencies and examples for each barrier type. ............................................... 83 3 9 Means, standard deviations, and mean differences for total barriers. ................ 84 3 10 Recommendation t ype as predictor of dichotomous barriers variable. ............... 85 3 11 Mean barriers as predictor of 4 month mean adherence. ................................... 86 3 12 Total barriers as predictor of dichotomous adherence variable at 4month follow up. ............................................................................................................ 87 3 13 Means, standard deviations, and mean differences for adherence ratings. ........ 88 3 14 Recommendation type as predictor of dichotomous adherence at 4 month follow up. ............................................................................................................ 89 3 15 4 mo nth mean adherence and standard deviations per recommendation type at each level of total barriers ............................................................................... 90 3 16 BASC2 PRS composite scores as nonsignificant predictors of 4month adherence. .......................................................................................................... 91 3 17 PIS total score as nonsignificant predictor of 4month adherenc e. .................... 92 3 18 BASC2 PRS composite scores, total barriers, and interaction terms as predictors of 4 month adherence. ....................................................................... 93 8


3 19 PIS total score, total barriers, and interaction terms as predictors of 4month adherence. .......................................................................................................... 94 3 20 PSI Total Stress as nonsignificant predictor of 4month adherence. ................. 95 3 21 PSI Total Stress, total barriers, and interaction terms as predictors of 4month adherence. ............................................................................................... 96 3 22 Mean adherence at first follow up, percent recall, and mean understanding presented separately by feedback modality. ....................................................... 97 3 23 Descriptive data for mean adherence by group and test of significance for group difference. ................................................................................................. 98 3 24 4 month adherence as non significant predictor of BASC2PRS Externalizing Symptoms Index difference score. ..................................................................... 99 3 25 4 month adherence as non significant predictor of BASC2PRS Internalizing Symptoms Index difference score. ................................................................... 100 3 26 4 month adherence as non significant predictor of BASC2PRS Behavioral Symptoms Index difference score. ................................................................... 101 3 27 4 month adherence as non significant predictor of BASC2PRS Adaptive Skills Index difference score. ............................................................................ 102 3 28 4 month adherence as non significant predictor of Perceived Impairment Scale difference score. ..................................................................................... 103 9


LIST OF FIGURES Figure page 2 1 Participation and procedural flow diagram .......................................................... 53 10


Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy PERCEIVED BARRIERS AND PARENTAL ADHERENCE TO RECOMMENDATIONS FOLLOWING CHILD PSYCHOLOGICAL ASSESSMENT By Eileen Matias Davis August 2014 Chair: Brenda Wiens Major: Psychology Every year approximately 1.5 million youths participate in a psychological evaluation (MacNaughton & Rodrigue, 2001). As a result of these evaluations, psychologists provide families with a broad range of recommendations . P arents and caregivers are then responsible for accessing services, utilizing resources, and navigating systems within which these recommendations are to be implemented. F ew researchers have explored factors that contribute to adherence to recommendations following child psychological assessment. What data is available suggests that adherence varies substantially depending on the type of recommendation prescribed, with recommendations for psychological services receiving the lowest levels of compliance across studies. Perceived barriers has also emerged as a consistent predictor of non adherence across rec ommendation types (Dreyer, O’Laughlin, Moore, & Milam, 2010; Human & Teglasi, 1993; MacNaughton & Rodrigue, 2001). This study explored predictors of adherence as well as outcomes that may be impacted by increased adherence. Consistent with previous findi ngs, adherence ratings 11


varied by recommendation type , and lowest rates were reported for psychological recommendations. Perceived barriers was the primary predictor of adherence, with more perceived barriers predicting lower adherence. An interaction effect was also identified between recommendation type and adherence, wherein the strength of the relationship between barriers and adherence varied by recommendation type. Child symptom severity and impairment, parenting stress, and feedback modality did no t predict adherence rates. Furthermore, higher adherence rates did not predict greater improvement in symptom severity or impairment. These results highlight the importance of implementing procedures to minimize the impact of barriers on parental adherence, such as incorporating a barriers assessment into child psychological assessments in order to tailor recommendations more carefully to the needs and resources of individual families. The lack of a relationship between adherence and changes in symptom severity or impairment should also be explored. M easures that are sufficiently specific to capture meaningful changes in child functioning but broad enough to be utilized wi th diverse clinical populations are needed. 12


CHAPTER 1 INTRODUCTION Every year approximately 1.5 million youths in the United States undergo psychological testing (MacNaughton & Rodrigue, 2001) . A comprehensive psychological assessment , which may also be referred to as neuropsychological or psychoeducational testing, depending on its nature and function, is a service that consists of clinicianadministered performance measures (i.e. measures of intellectual, cognitive, and academic functioning) , diagnostic interviews , questionnaires completed by the child and his or her caregiv ers and teachers , computer administered assessment tools , and behavioral observations (Geffken, Keeley, Kellison, Storch, & Rodrigue, 2006) , as well as review of medical and academic records , when available . Some of the functions of psychological assessme nt include the characterization of symptoms for the purposes of differential diagnosis, the description of levels of functioning across a variety of important life domains, and the identification of other patient characteristics that may be relevant for tr eatment and prognosis (Kibiszyn et al., 2000). Based on the results of these assessment s, psychologists provide recommendations for interventions that commonly fall within five primary domains, including recommendations for psychological services, school based recommendations, physician consultation (i.e. medication consultation), referral for other professional nonpsychological services (i.e. , occupational therapy, physical therapy, speech/language therapy ), and active self help recommendations ( e.g., implementation of behavioral strategies in the home, bibliotherapy resources ; Dreyer, Laugh lin, Moore, & Milam, 2010). These recommendations are intended to help improve the child’s 13


functioning, alleviate di stress for the chil d and family, and improve the child’s overall well being (Geffken, Keeley, Kellison, Storch, & Rodrigue, 2006). The degree to which these goals for improved well being are met is likely dependent on whether or not families follow through with the recommendations they receive following assessment . Unfortunately, a lthough the focus on parental attendance and adherence within the context of child therapy has increased over the last fifteen years ( e.g., Kazdin, Holland, & Crowley, 1997; Kazdin & Mazurick , 1994; Kazdin & Wassell, 1999 ) , there remains a significant dearth of research examining the degree to which caregivers follow through with recommendations provided to them following child assessment (Dreyer, Laughlin, Moore, & Milam, 2010). As Geffken and colleagues note, there is also a substantial lack of published research examining whether adherence to recommendations following child assessment corresponds with desired improvements in child functioning (2006) . Considerable resources have been alloc ated over the years to the development of interventions for the treatment of psychosocial problems in youth that have strong support as evidencebased interventions (e.g., Barrett, Farrell, Pina, Peris, & Piacentini, 2008; Chambless & Ollendick, 2001; Davi d Ferndon & Kaslow, 2008; Eyberg, Nelson, & Boggs, 2008; Rogers & Vismara, 2008; Silverman, Ortiz, Viswesvaran, Burns, Kolko, Putnam, & AmayaJackson, 2008; Silverman, Pina, & Viswesvaran, 2008; Waldron & Turner, 2008) . However, these interventions can on ly begin to effect change in children’s symptoms and functioning if they are available to and accessed by families once the need for services has been identified. Researchers have focused their attention on factors that serve as barriers to continued part icipation and adherence to therapeutic techniques with in the context of psychotherapy ( Kazdin, Holland, & Crowley, 14


1997), but very little is yet known about the role of barriers in determining families’ adherence to recommendations following child assessment . The degree to which families understand the results of psychological assessments, their understanding of the recommendations being prescribed and the reasons for which they are prescribed, and factors associated with service accessibility and feasibility of service utilization are undoubtedly important variables to explore in an effort to increase adherence to recommendations. For researchers and clinicians involved in child assessment, it is important to gain a better understand ing of the factors that contribute to whether or not their efforts translate into desired improvements in functioning and overall well being for the child and family . Understanding parents’ help seeking behaviors and the factors that p redict adherence to recommendations , including gaining a better understanding of how different kinds of barriers affect compliance, can he lp researchers and clinicians develop interventions to improve compliance and increase accessibility and feasibility of service utilization for families . Additionally, such interventions may ultimately lead to an increase in the degree to which children are benefiting from available evidencebased interventions . Understanding P arental H elp seeking and C ompliance Dreyer and colleagues propose two behavioral change models that may be relevant for understanding adherence to recommendations following psychological assessment s (2010) – the barriers to treatment model and the stages of change model. In addition to thes e models, the Health Belief Model is a model that is commonly used within health research to explain adherence to recommendations for interventions to 15


treat medical conditions. Finally, the stages of helpseeking model outlined by Power and colleagues (2005) provides a good summary of the help seeking behaviors of families at different stages of the assessment/treatment process and describes factors that contribute to families’ movement between stages. These models are summarized below. Barriers to Trea tment The barriers to treatment model developed by Kazdin and colleagues (Kazdin, Holland, & Crowley, 1997) was designed to explain factors contributing to premature treatment dropout within the context of child psychotherapy. This model includes four pri mary domains within which barriers can be classified, including 1) the experience of stressors and obstacles, 2) a poor relationship with the therapist, 3) perceptions that treatment is irrelevant, and 4) the perception that treatment is too demanding. Th e barriers to treatment model was adapted by MacNaughton and Rodrigue (2001) to help describe the factors that commonly interfere with families’ ability and/or willingness to comply with recommendations they receive following child assessment. These authors identified 1) access problems, 2) negative attitudes and beliefs, 3) scheduling conflicts, and 4) financial limitations as relevant barriers to adherence with in the context of psychological assessment. Stressors and logistical obstacles may interfere in similar ways as they do within the context of child therapy by prev enting families from being able to access the services that were recommended to them during the assessment . Transportation difficulties, lack of available local providers, language barriers, and childcare needs are some common stressors and obstacles that may interfere with a family’s ability to implement the prescribed interventions offered by the child’s 16


psychologist. These access issues may be particularly relevant for families of lower socioeconomic status as well as for individuals living in more rur al areas of the country in which fewer services are available. Negative attitudes and beliefs regarding the utility and desirability of recommendations may also prevent families from pursuing the recommended interventions. These attitudes and beliefs may include the perception that the psychologist lacks credibility, overall distrust in mental health care providers (which may be particularly relevant for certain cultural groups), as well as the belief that the prescribed treatment is not relevant to the c hild’s needs. Competing demands may cause scheduling difficulties for families that do intend to follow through with proposed recommendations but have other responsibilities that interfere with their ability to do so. Finally, financial problems such as a lack of insurance coverage or the inability to meet deductible and copayment requirements may further interfere with families’ adherence to recommendations . Stages of Change Prochaska, DiClemente, and Norcross’s stages of change model describes five st ages of behavior change and how individuals’ responses to treatment recommendations may differ based on their degree of readiness for behavioral change (1992). The five stages are precontemplation (unaware of the need for change; unwillingness to pursue c hange), contemplation (recognition of a problem; consideration of possible change), preparation (intention to engage in behavioral change in the near future; may involve minor behavior change), action (engaged in active steps to change behavior, environment, or experiences to resolve problem), and maintenance (following behavior change, involves relapse prevention and consolidation of changes). Dreyer 17


and colleagues suggest that parental adherence to recommendations following child assessment is associated with the “contemplation” stage for some families who are just learning about the child’s problems and considering whether or not the potential burden associated with initiating treatment is outweighed by the potential benefits (2010). For other families, their stage of readiness for behavior change more closely resembles the “preparedness” phase, as these families have already decided to pursue treatment prior to initiating an assessment . These families enter the child assessment process with hopes of better understanding their child’s problems and obtaining information on how to best focus their efforts in addressing the areas of greatest identified need. Families who have sought psychological testing due to their own perceptions of need for services m ay be in later stages of behavior change while those who are seeking child assessment at the request of others involved in the child’s care may be less aware of the need for services and consequently be in earlier stages of readiness for behavior change. Given the likely association between reason for seeking psychological assessment and stage of change, it is important to assess parents’ motivations for seeking child assessment and us e this information to inform strategies to increase adherence to recommendations. Health Belief Model The Health Belief Model , commonly used within the context of research on medical compliance, describes the relationship between compliance to treatment recommendations and four parental perceptions: perceived severity, perc eived susceptibility, perceived benefits, and perceived barriers (Rosenstock, 1974). As Human and Teglasi (1993) suggest, the four parent perceptions as defined by the 18


Health Belief Model correspond with parental perceptions within the context of psycholo gical assessment. Parental perceptions regarding the severity of their child’s academic, emotional, and interpersonal difficulties , as well as the potential short and long term consequences of these difficulties , correspond with the perceived severity and perceived susceptibility variables of the Health Belief Model. The perceived benefits variable of the model parallels parental belief in the efficacy of the intervention, agreement with the results of the assessment , and understanding of both the results and the suggested interventions within the context of child assessment. Finally, perceived barriers within the Health Belief Model similarly reflects parental beliefs and attitudes regarding the potential burden of seeking interventions, the possible adverse effects to the child, and the parents’ ability to appropriately implement these recommendations given their resources. Although this model was developed to explain parental compliance to recommendations for interventions to trea t medical diagnoses, it is likely that similar factors are at work in determining the degree to which families will comply with recommendations provided by the child’s psychologist. In working with families within the child assessment context, professionals should take care to assess parental beliefs regarding the nature and severity of their children’s presenting problems, their beliefs about the potential benefits of seeking treatment, and psychological and financial factors that ma y interfere with parents’ adherence to recommendations following child assessment. 19


Stages of Help Seeking Combining common features shared by the barriers to treatment, stages of change, and Health Belief models and their various adaptations, Power and colleagues (2005) provide a summary of the three primary stages of families’ help seeking behaviors for the treatment of children’s health and mental health needs. This final summary model provides a good framework within which to explore a variety of f actors contributing to families’ helpseeking behaviors prior to and after seeking a comprehensive psychological assessment . The three stages of help seeking they proposed based on general consensus among researchers include 1) problem recognition, 2) dec ision to seek help, and 3) service selection and utilization. Each stage is described below with emphasis on how it relates to parental adherence to recommendations. Problem r ecognition. In the initial problem recognition stage of helpseeking, factors t hat influence help seeking include parental perceptions regarding the severity of the child’s problems and the impact of these problems on major areas of functioning (including academic performance, peer relationships, and relationships with adults), the d egree of caregiver burden associated with the child’s symptoms, and problem threshold (i.e. , parental tolerance of a given symptom before it is perceived as a problem; Power, Eiraldi, Clarke, & Mazzuca, 2005). Although one may be inclined to assume that parents who seek psychological assessment services recognize that their child is experiencing difficulties, in reality many parents do not seek psychological assessment until it is proposed by someone working closely with the child in another environment such as the child’s school (Jackson, 1991). In fact, parents may be very 20


unaware that the child is having significant difficulties at the time that they come in for assessment. To illustrate this, Teagle (2002) explored parent problem recognition and found that only 39% of parents of children with at least one diagnosis recognized significant difficulties in the child and only 31.7% expressed awareness of the negative impact of the child’s behavioral and emotional problems on others. Variability among par ents in level of problem awareness may be explained by several factors, including parental education level, socioeconomic status, and cultural background; referral concerns; the degree to which the problems are observable to the parent (i.e. , externalizing versus internalizing in nature) ; parental beliefs about the nat ure and cause of child problems (i.e. , the belief that the child’s problems are psychiatric in nature versus the belief that the problems are fully within the child’s control, Mo ses, 2011); the existence of comorbid symptoms; and the extent of associated impairment across areas of functioning. Research also suggests that certain kinds of problems may be more easily recognized by parents than others. For instance, externalizing behaviors such as oppositionality and hyperactivity are more readily identified and perceived as problematic by parents than are less overt symptoms such as depression and anxiety (Teagle, 2002). It would be reasonable to expect that parents who enter the process of child assessment with greater perceived symptom severity and greater concerns regarding degree of impairment may be more likely to comply with recommendations made by the psychologist who conducted the assessment . There are mixed findings, however, regarding the relationship between perceived symptom severity and adherence to treatment recommendations. While some research findings have supported a positive relationship between perceived 2 1


severity of child psychopathology and treatment compliance, (Griest, Forehand, Wells, & McMahon, 1980; Gustafson, McNamara, & Jensen, 1994; Hricik & Keane, 1988; Meichenbaum & Turk, 1987; Teagle, 2002), others have failed to find this relationship (BrookmanFrazee, et al., 2008; Kendall & Sugarman, 1997; MacNaughton & Rodrigue, 2001). Nock and Ferriter (2005) suggested two alternative ways in which parent perceived child symptom severity may influence treatment seeking behaviors. One possibility they proposed is that greater perceived symptom severity leads to greater perceived need for intervention, thus increasing the likelihood of adherence to recommendations for treatment initiation. Conversely, greater perceived symptom severity may also be associated with greater perceived barriers to treatment (i.e., mor e parenting stress, reduced social support, etc.), thus negatively impacting families’ ability and willingness to initiate services. Given the lack of consistency in findings across studies, there is much yet to be understood about the relationship between the different factors associated with problem recognition and parental adherence to recommendations following child assessment. Decision to seek help. The second phase of helpseeking as outlined by Power and colleagues (2005) involves parental familiarity with the problem and its treatment, parental health locus of control (belief that following through with recommendations can improve the child’s well being), parental self efficacy (belief that following through with recommendations will successfully improve the child’s health and well being), and acculturation (the degree to which minority parents have adopted the values of the host culture). The degree to which parents understand the nature of the identified problems 22


and their treatments can vary gr eatly across families. Given this, the assessing child psychologist is in a unique position of influence for directing the course that families take to address their concerns and for providing them with the psychoeducation needed to advocate successfully for their children as they pursue services. The strategies used by the psychologist to help families understand the results of the assessment and the associated recommendations may have a significant impact on whether or not parents follow through with th ese recommendations. Additionally, certain parental beliefs about the nature of the child’s problems and their ability to address them effectively may interfere with compliance. This is another area in which psychologists can intervene with families by i ncreasing their understanding of the problem and its treatment and by empowering them to seek out the services necessary to address their child’s problems. For minority families, it may be especially important to assess for understanding of results and as sociated recommendations and to ensure that the recommendations adequately address problem areas that are most salient to the family and its unique values. Service selection and patterns of use. The final stage of the helpseeking model includes the sel ection of services and patterns of utilization of these services. It is in this phase that parents’ trust in the psychologist and willingness to follow through with recommendations is most important (Power et al, 2005). Psychologists often provide famili es with specific recommendations for how to address the problems identified through a comprehensive assessment . Parents must then decide whether it is worthwhile to follow through with these recommendations as outlined by the psychologist or whether they should instead address their child’s difficulties in some 23


other manner or not address them at all. This decision is likely influenced by several factors, including the degree to which parents agree with the results of the assessment , how much they underst and what the recommended intervention entails, and the degree to which they perceive the recommendation as an appropriate method for achieving goals that are meaningful to the parents. Parental belief in the credibility of a particular treatment and expec tancies for change as a result of the proposed interventions has been shown to significantly predict treatment participation within the context of child psychotherapy (Nock, Ferriter, Holmberg, 2006). Although parental attitudes about treatment recommendations have not been systematically evaluated, it is likely that parental attitudes and beliefs about the recommended interventions following child assessment also predict subsequent adherence to these recommendations. Another factor that has been evaluated as a potential predictor of adherence to recommendations is parental satisfaction with assessment services. Surprisingly, however, these factors do not appear to have a significant relationship with compliance (Human & Teglasi, 1993; MacNaughton & Rodrigue, 2001). It appears, therefore, that the nature of the recommendations given to families and the degree to which they understand and accept these recommendations may be more predictive of subsequent compliance than is overall satisfaction with the assessment process. Th e above framework provides a good basis for organizing research on parental adherence to recommendations that focuses on factors that are relevant at each phase of helpseeking. Factors pertaining to problem recognition, the decis ion to seek help, and service selection and patterns of use will be evaluated in the current study in order 24


to gain a better understanding of factors that contribute to families’ adherence to recommendations following assessment. Research on Parental Adhe rence to Clinician Recommendations Child adherence to clinician recommendations is largely dependent on parental involvement and compliance, as parents are often charged with carrying out specific recommendations in the home or with ensuring that the child or adolescent is complying with clinician recommendations (Nock & Ferriter, 2005). Given this, it is important to understand how parent variables affect overall adherence to recommendations . This is true within the pediatric health field as well as within the field of youth mental health. Researchers have focused largely on parental factors as predictors of adherence to child treatment . Within the literature on pediatric compliance, parental factors such as parental disease knowledge, c aregiver distress, parenting stress , and parent self efficacy have all been found to be associated with compliance and utilization of services in the treatment of different child medical conditions (e.g. diabetes, asthma, HIV ; Chisholm, Atkinson, Donaldson, Noyes, Payne, & Kelnar, 2006; Janicke, D.M., & Finney, J.W., 2003; Marhefka, Tepper, Brown, & Farley, 2006). Similar parent factors have also been associated with treatment attendance and participation with in child psychotherapy , including parental moti vation for treatment , socioeconomic status, parent psychopathology, parent stress, level of agreement with therapist’s explanation of child’s problems, and perceived barriers (Nock & Photos, 2006 ; Patterson & Chamberlain, 1994) . Greater continuity of care and shorter wait list periods (Dierker, Nargiso, Wiseman, & Hoff, 2001; Sherman, Barnum, BuhmanWiggs, & Nyberg, 2009; Sirles, 1990) , as well as greater perceived treatment credibility and expectations for 25


treatment (Nock, Ferriter, & Holmberg, 2006), few er therapeutic relationship problems (Garcia & Weisz, 2002), and stronger parent youth treatment goal agreement (BrookmanFrazee, Gabayan, Haine, & Garland, 2008) are also consistently associated with greater compliance and attendance in psychotherapy . Less is known about the factors that predict initiation of services in response to recommendations following child assessment. Given that parents are usually the ones who are capable of and expected to implement recommendations put forth by the assessing ps ychologist , parental factors have been the focus when evaluating adherence to recommendations following child assessment . MacNaughton and Rodrigue (2001) did not find a significant association between parent satisfaction with the child’s assessment , paren tal locus of control, or parent recall of recommendations and the measured outcome of parental adherence to recommendations. The one parental factor that did most strongly predict adherence was the number of perceived barriers , as more perceived barriers was predictive of poorer adherence. In a sample of youths from the Great Smoky Mountains Study, problem perception, but not the degree to which parents perceived their child’s problems to have an impact on others, significantly predicted initiation of specialty services (GSMS; Teagle, 2002). More recently, Dreyer and colleagues (2010) found a significant association between parenting stress and adherence to recommendations, with greater parenting stress predicting greater adherence. These findings sugges t that the degree of caregiver burden and parent perceived barriers to health utilization are important variables to examine when predicting compliance to recommendations following child psychological assessment . 26


Despite the limited knowledge that has been obtained regarding specific predictors of adherence to assessment recommendations , there is growing evidence to suggest that parental adherence to recommendations does vary across types of recommendations . Overall, adherence to recommendations following psychological assessment appear s to be in the seventy percent range when averaged across recommendation categories (Dreyer, Laughlin, Moore, & Milam, 2010; Geffken, Keeley, Kellison, Storch, & Rodrigue, 2006; Human & Teglasi, 1993; Moore & Symons, 2009) , but notable differences can be seen with regards to the kinds of recommendations to which families adhere. In a sample of 4through 12year olds referred to a psychology outpatient service for internalizing, external izing, developmental, and psychotic disorders, parents were significantly less likely to follow through with recommendations for psychological services than with school based recommendations or recommendations for consultation with nonpsychological profess ionals such as medical doctors ( MacNaughton & Rodrigue, 2001 ) . Follow through with recommendations for self help activities was the lowest and significantly lower than for professiona l nonpsychological consultation. In a separate sample of children with Attention Deficit/Hyperactivity Disorder , the lowest rate of parental adherence was for recommendations for psychological services while the highest rates of adherence were found for active self help (Dreyer, Laughlin, Moore, & Milam, 2010) . Similarly, a dherence to recommendations for behavioral treatment in a sample of children with Autism Spectrum Disorders was appr oximately seventy five percent and was significantly lower than adherence to recommendat ions for medication treatment (Moore & Symons, 2009) . 27


Despite limited research in this area, the consistent finding of lower rates of adherence to recommendations for psychological services is alarming. It does not appear that p ossible reasons for these lower rates of adherence to psychological services have been explored systematically through research. MacNaughton and Rodrigue (2001) suggest some possible explanations, including less parental familiarity and comfort working within the mental health system as compared to the primary care system , challen ges associated with access to mental health services (e.g., insurance authorization and compensation) , and psychological factors (e.g. , parental attitudes and beliefs about the mental health system , low motivation for engaging in the challenging work of tr eating psychological problems, greater parental emphasis on finding causes for problems rather than on treatment, etc.). They also note that adherence to recommendations for psychological services may be more demanding, overall, for families, thus leading to lower rates of adherence when compared to adherence to recommendations for consultation with a physician or school based recommendations. This hypothesis is in line with the barriers to treatment model (Kazdin, Holland, & Crowley, 1997), which suggest s that lower rates of adherence are associated with more perceived barriers to compliance. In light of findings that many children with mental health needs are not receiving appropriate mental health services (Richardson, 2001) , it is important to gain a better understanding of families’ failure to seek psychological services once the need for these services has been identified. Explaining families’ apparent preference for nonpsychological services can help clinicians to develop strategies for increasing families’ willingness and ability to adhere to recommendations for psychological services . 28


Research on B arriers to A dherence The barriers to treatment model has been used extensively to help explain participation and adherence t o treatment recommendations within the context of child psychotherapy. As a measure of barriers to treatment, t he Barriers to Treatment Participation Scale was developed to assess the relationship between perceived treatment barriers and subsequent participation in treatment. Key findings demonstrate a significant relationship between more perceived barriers and important outcomes such as premature termination (Kazdin, Holland, & Crowley, 1997) , treatment acceptability (Kazdin, 2000), and therapeuti c change (Kazdin & Wassell, 2000). Interestingly, perceived barriers is predictive of treatment participation above and beyond the effect of other important child and parent characteristics that have previously been established as predictors of premature dropout and limited therapeutic change (Kazdin, Holland, Crowley, & Breton, 1997). Extending this research into the domain of child assessment, MacNaughton and Rodrigue (2001) examined the relationship between parent perceived barriers and subsequent adherence to recommendations following assessment. One key finding of this study was that the total number of perceived barriers to completing recommendations was the only significant predictor of overall adherence, while specific type of barrier, demographic factors, child symptom severity, satisfaction with services, parent locus of control, and parental ability to recall the recommendations provided during feedback did not predict adherence. In light of this finding, clinicians may find it helpful to assess for the total number of perceived barriers and attempt to address some of these barriers in order to increase the likelihood that families will comply with 29


recommendations. It remains unclear, however, whether barriers interact with other factors su ch as perceived symptom severity and impairment or degree of parenting stress, to predict adherence; thus, the present study will attempt to further elucidate the se relationship s. Feedback F ollowing P sychological A ssessment Following a comprehensive psychological assessment , the child psychologist may generate some form of written report that includes testing results (with or without graphs) as well as diagnostic impressions based on testing data, behavioral observations, and information gathered through the clinical interview and through review of any additional medical and educational records. The written report may also include a list of recommendations for how to address the areas of difficulty identified throughout the assessment . Written reports can be useful for families by providing them with a record of results that can be referenced in the future (Ownby & Wallbrown, 1986) and can be used to recall the results of the assessment and to share these results with o ther professionals who are involved in the child’s care in order to obtain and tailor services to the child’s specific needs (Tallent, 1993). In addition to written feedback , a n oral feedback session may also be provided to families. During an oral feedback session, which can take place in person or on the phone, the results of the assessment , their implic ations, and the recommended interventions are explained in detail. An oral feedback session gives families the opportunity to ask questions and seek cl arification if they have difficulty understanding the results and their implications . During an oral feedback, the assessor also has an opportunity to empathize with parents’ reactions to what can be difficult findings regarding diagnoses and their conseq uences for the child’s 30


current and future functioning (Tharinger, Hersh, Christopher, Finn, Wilkinson, & Tran, 2008) . Whether written or oral, formal feedback from the psychologist following a psychological assessment is necessary in order for families to make use of the asse ssment results and to initiate implementation of recommendations. There is great variability, however, in the nature and thoroughness of the feedback that is provided to families . In some cases, feedback may be of little utility for families if the language used is too technical for them to understand or if the results are simply relayed in a way that is not meaningfu l and does not adequately address the families’ concerns ( Pope, 1992) . Psychologists providing psychological assessme nt services for children have the ethical and professional responsibility of providing families with comprehensive feedback to guide them in making educated decisions about how to address the child’s difficulties and to better understand the rationale behi nd certain treatment approaches that are recommended by the psychologist (Gass & Brown, 1992). Given the important role of a comprehensive feedback in ensuring that the results of the assessment are understood by and useful to families, it is not surprising that most psychologists report engaging in some form of feedback with families at least some of the time (Curry & Hanson, 2010; Tharinger, Hersh, Christopher, Finn, Wilkinson, & Tran, 2008 ). However, t here is significant variability among psychologists with regards to the frequency with which they provide feedback and the kinds of feedback they offer families . A national survey of psychologists suggests that oral feedback is provided following every assessment by only 35% of respondents, and “s ometimes” or “usually” by 57% of responding psychologists (Curry & Hanson, 2010) . With regards to a written summary, 25% of respondents indicated that this form of feedback is provided after 31


every assessment and another 46% reported that they provide a wr itten summary of the results and its implications “sometimes” or “usually . ” The remaining respondents indicated that they rarely or never give verbal or written feedback following assessment . There is a surprising lack of research on the relationship between feedback , or modality of feedback, and important outcome variables such as satisfaction with the assessment process, understanding of the results and recommendations, and compliance with recommendations. It seems reasonable to expect that families who receive comprehensive feedback are better equipped to address the problems that were identified during the assessment than families who do not receive any feedback , as they ma y have greater opportunities for having their questions addressed and seeking further clarification. Even among those families who do receive feedback, t he degree to which parents understand the assessment results and their accompanying recommendations may vary based on the type of feedback that is given . Some psychologists provide written feedback only while others provide oral feedback in addition to a written report (Pinto, 2003) . Although the feedback process has been construed by some clinicians and researchers as a “dynamic, interactive process” that involves both the assessor as well as the family (Pope, 1992) , the degree to which the feedback actually possesses these characteristics can depend on several f actors, including whether the feedback was provided orally or just in writing. It appears logical that a written summary of findings does not provide the same opportunities for a dynamic exchange between the assessor and the child’s parents that an oral feedback session, whether in person or ove r the phone, provides. In an era of managed health care , increasing financial pres sures, and 32


increasing demands on psychologists’ time, psychologists may be less inclined to offer families oral feedback as part of the assessment fee. I t is important , therefore, to understand whether the time taken to provide families with oral feedback following assessment has significant positive incremental benefits for parents’ ability to understand testing results and for their adherence to recommendations , above and beyond the benefits of a written summary of the results . The present study will attempt to examine differences in adherence among families who receive written and oral feedback following child psychological assessment versus those that receive only wri tten feedback. Outcomes of Adherence Recommendations provided to families following child psychological assessment s are intended to help remediate child symptoms and improve child and family functioning, but there is a remarkable lack of research examining the degree to which adherence to recommendations following child assessment is associated with these desired outcomes (Geffken, G.R., Keeley, M.L., Kellison, I., Storch, E.A., & Rodrigue, J.R.). S ince the development of criteria for well established treatments (Chambless, Sanderson, Shoham, et al, 1996), numerous articles have been published that review the evidence base for treatment of a wide range of childhood problems, including mood disorders, anxiety d isorders and obsessive compulsive disorder , disruptive behavior disorders, trauma, substance abuse, and autism ( Barrett, Farrell, Pina, Peris, & Piacentini, 2008; David Ferndon & Kaslow, 2008; Eyberg, Nelson, & Boggs, 2008; Rogers & Vismara, 2008; Silverma n, Ortiz, Viswesvaran, Burns, Kolko, Putnam, & AmayaJackson, 2008; Silverman, Pina, & Viswesvaran, 2008; Waldron & 33


Turner, 2008) . These articles suggest that significant advances have been made in identifying treatments that produce significant improvements in symptoms and functioning for children and adolescents . Given the extensive body of knowledge about treatment efficacy, i t is reasonable to expect that parent initiat ion of services recommended by the assessing psychologist would lead to significant improvements for the youth if the treatments are evidence based. When examining the outcomes associated with adherence to recommendations following child assessment, if an association between compliance and improvements in child symptoms and functioning was found it would suggest that efforts should be made to improve families’ willingness and ability to follow through with these recommendations. Conversely, findings that greater adherence to recommendations does not necessarily lead to signi fi cant improvements for the child would suggest a need for psychologists to better tailor their recommendations to the specific needs of the child and family and to ensure that the recommendations given are based on empirical evidence in order to better address the problems identified throughout the assessment . To date, the relationship between adherence to recommendations and improvements in child functioning has not been examined. Thus, the present study will aim to provide some preliminary data that can inform future research in this area . Aims and Hypotheses A g rowing body of research exists exploring factors that predict attendance and adherence with treatment recommendations within the context of child psychotherapy . However, f ew researchers have explored factors that contribute to adherence to 34


recommendations following child psychological assessment . Given the extensive resources that are devoted to the development of assessment measures and evidence based treatments for academic, emotional, behavioral, and psychological problems in youth (Geffken, G.R., Keeley, M.L., Kellison, I., Storch, E.A., & Rodrigue, J.R.), it is important to understand the factors that determine whether or not families make use of the recommendations that result from child psychological assessments and to determine the degree to which adherence to these recommendations produces desired positive results for the child and family. Aim 1 Researchers who have examined adherence within the child psychological assessment con text have consistently found the lowest rates of adherence to recommendations for psychological services, with barriers to compliance emerging as the most common predictor of nonadherence across recommendation types (Dreyer, O’Laughlin, Moore, & Milam, 2010; Human & Teglasi, 1993; MacNaughton & Rodrigue, 2001; and Moore & Symons, 2009). Possible reasons for these lower rates, however, have not been systematically investigated. It has been suggested that families experience more barriers to adherence to re commendations for psychological services as compared to recommendations for consultation with nonpsychological professionals or school based recommendations . T hus , more barriers may account for the lower rates of adherence to recommendations for psycholog ical services (MacNaughton & Rodrigue, 2001) . As its first aim , the present study set out to fill this gap in the literature by exploring the role of barriers in the relationship between recommendation type and adherence. 35


Hypothesis 1a. It was hypothesized that perceived barriers would vary significantly across recommendation types. More barriers were anticipated for psychological services recommendations than for s chool based recommendations, recommendations for physician consultation , and re commendations for other nonpsychological co nsultation. Hypothesis 1b. Total number of perceived barriers was expected to be significantly related to adherence regardless of recommendation type , with more perceived barriers predicting poorer adherence to recommendations. Hypothesis 1c . Based on previous findings of poorer adherence to recommendations f or psychological services, it was hypothesiz ed that recommendation type would be significant ly related to adherence. Higher levels of adherence were anticipated for s chool based recommendations, recommendations for physician consultation, and recommendations for other nonpsychological consultation than for psychological services. Hypothesis 1d . The relationship between recommendation type and adherence was expected to be moderated by perceived barriers . Specifically , it was hypothesized that more perceived barriers w ould have a greater negative impact on adherence to psychological recommendations than on adherence to other types of recommendations . This hypothesis was based on MacNaughton and Rodrigue’s (2001) suggestion that the acquisition of psychological services may be associated with more logistical, psychological, and financial barriers than are other recommendation types and that these barriers may account for the poorer rates of adherence to recommendations for psychological services. 36


Aim 2 The second aim of the present study was to explore the relationship between two parent perceptions, namely parent ratings of symptom severity and ratings of impairment , and adherence to recommendations. Mixed findings in the literature with regards to the relationship between these variables (BrookmanFrazee, et al., 2008; Griest, Forehand, Wells, & McMahon, 1980; Gustafson, McNamara, & Jensen, 1994; Hricik & Keane, 1988; Kendall & Sugarman, 1997; MacNaughton & Rodrigue, 2001; Meichenbaum & Turk, 1987; Teagle, 2002) s uggest that there may be additional factors that serve as moderators for the strength and direction of these relationships. Hypothesis 2a . It was hypothesized that both perceived severity of symptoms , as well as perceived impairment , would be related to adherence , with greater perceived symptoms and impairment predicting greater adherence to recommendati ons. Hypothesis 2b. It was further hypothesized that the relationship between parent ratings of symptom severity and adherence, as well as the relationship between perceived impairment and adherence, would be moderated by perceived barriers to compliance. In the presence of m ore barriers to implementation , a weaker relationship between perceived symptom severity or perceived impairment and adherence to recommendations was anticipated. In the absence of multiple barriers, however, a stronger relationship betw een perceptions of severity/ impairment and adherence was expected to emerge. Aim 3 As the thir d aim, this study explored the relationship between parenting stress and adherence to recommendations. Dreyer and colleagues (2010) found a positive 37


relationship between parenting stress and adherenc e to recommendations, with greater levels of parenting stress predicting greater adherence. The sample used in Dreyer et al study , however, consisted only of children with a diagnosis of Attention Deficit/Hyperactivity Di sorder. The present study examined the relationship between parenting stress and adherence to recommendations in a more diverse clinical sample. Hypothesis 3a . Based on findings from Dreyer et al, i t is hypothesized that more parenting stress would be associated with greater adherence to recommendations . Hypothesis 3b. Although increased parenting stress may serve as an impetus for families to follow through with recommendations following a child’s psychological assessment, it was hypothesized that the relationship between par enting stress and adherence would be weaker for famili es that experience more perceived barriers to adherence than for those who report fewer barriers , indicative of a moderating effect of perceived barriers . This hypothesis was based on the assumption that some families who experience significant parenting stress as a result of the child’s difficulties may be unwilling or una ble to obtain these services due to the presence of multiple barriers . Aim 4 The information obtained as a result of a child psychological assessment can be very valuable for families seeking guidance on how to address the ir child’s difficulties across a variety of domains. A family’s ability to adhere to these recommendations, however, may be limited by the degree to which parents recall and understand these recommendations after they are provided. Although psychologists generally agree about the importance of providing families with feedback after an assessment, the 38


format in which they provide this feedback can vary across clinicians. The relationship between ty pe of feedback and adherence had not been formally investigated prior to the current study . The fourth aim of this study , therefore, was to examine the relationship between type of feedback provided and both recall of , and adherence to, recommendations . Hypothesis 4a . It was hypothesized that parents who received oral feedback would be better able to recall , and would express greater understanding of , the recommendations provided than families who did not receive this oral feedback. Hypothesis 4b. Families who receive oral feedback have the opportunity to ask questions and seek clarification if they do not adequately understand the results of the assessment . They may also be better able to understand the recommendations provided and thus more lik ely to follow through with these recommendations. Based on this assumption, i t was hypothesized that parents who receive oral feedback in addition to a written report would be more adherent to rec ommendations than parents who did not receive this oral feedback. Aim 5 and Hypothesis Researchers within the child psychotherapy literature have investigated the impact of contact with families on attendance at initial therapy appointments. Findings showed that any contact (e.g., telephone reminder, letter) was associated with lower noshow rates at the initial therapy appointment (Kourany et al., 1990; MacLean et al., 1989). The impact of follow up contact on adherence to recommendations following child psychological assessment, however, had not been previously examined. The 39


present study aimed to explore the impact of a brief telephone contact on subsequent adherence to recommendations. It was hypothesized that families who were contacted via telephone 2 months a fter the child’s assessment would h ave higher rates of adherence to recommendations at the 4 month follow up than families who were not contacted at 2 months. Aim 6 and Hypothesis The final aim was also exploratory in nature and secondary to the aims listed previously. Here, the goal was to obtain some preliminary data regarding whether greater adher ence to recommendations predicted improv ements in parent rated child symptom severity as well as improvements in parent ratings of functioning. As psychologists strive to provide parents with recommendations that are specifica lly tailored to the needs of each particular child, and given the number of evidencebased interventions that have been developed for the treatment of different kinds of childhood problems, it is logical to anticipate that follow through with recommendations for treatment following assessment can result in significant improvements in child symptom manifestati on and functioning. It was hypothesized, therefore, that greater parent al adhere nce to recommendations would be ass ociated with greater reported improvements in child symptoms and functioning. 40


CHAPTER 2 METHODS Design The current study was designed to investigate predictors of adherence to recommendations following child psychological assessment, including perceiv ed barriers, parenting stress, perceived symptom severity, and feedback format. The implications of adherence for child outcomes were also explored. Parents of children and adolescents completed measures assessing parenting stress and child symptom sever ity as well as various demographic factors at the time of the initial assessment and were contacted for follow up interviews at two months and four months post delivery of the psychological report in order to assess for adherence to recommendations, recall and understanding of recommendations, perceived barriers to compliance, and changes in child symptom severity and functioning. Participants Participants were the parents of 90 children and adolescents between the ages of 3 and 17 years referred to either the Psychology Clinic or the Behavioral Health Unit in the Department of Psychiatry at the University of Florida for psychological assessment. This age range includes the majority of children who are seen in the two participating clinics and who remain under the custodial care of their parents. It is assumed that parents of all minors remain at least partly responsible for the implementation of recommendations that result from the child’s psychological assessment. The decision was made to extend the a ge range beyond that used by MacNaughton and Rodrigue (2001) in order to increase the sample size and allow for comparisons between older and younger youths with the understanding that differences 41


may exist in terms of the number and types of barriers faced, as well as the levels of adherence exhibited. Given the greater autonomy exhibited by adolescents as compared to younger children, the youth’s refusal to participate in services may be a significant barrier to adherence faced by families. Treatment refusal by the child has indeed been reported by parents to be a significant barrier to service utilization (Koroloff, Elliott, Koren, & Friesen, 1994). The Psychology Clinic at the University of Florida sees approximately 9 to 11 referrals per week divided among several practitioners. The Behavioral Health Unit, a section of the Department of Psychiatry at the University of Florida, sees 2 patients per week under one psychologist as the licensed supervisor. A large percentage of referrals for these clinics come from a variety of hospital and community resources, including physicians and other healthcare providers, while others are self referrals. Based on the results of power analyses, it was determined that approximately 45 participants (30 in the 2 month follow up group and 15 in the 4monthonly follow up group) would be needed to conduct the desired analyses1. To account for approximately 50% attrition as was experienced in the MacNaughton and Rodrigue (2001) study, 90 (60 and 30, for the 2 and 4month follow up groups respectively) participants were recruited across both clinics. As was anticipated, recruitment efforts 1 Note: Power analyses indicated a need for a total of 76 recommendations to capture medium effect sizes for the primary analyses. It was estimated that each participant would receive an average of 3 recommendations; thus it was determined that approximately 30 participants would be needed to obtain this many recommendations. An additional 15 participants were sought for conducting the betweengroups analyses assessing the relationship between phone contact and adherence. 42


in both clinics were sufficient to meet study goals over the course of approximately four months. Of the 90 participants who signed i nformed consent forms enrolling them in the study, three did not complete or submit initial paperwork and were subsequently withdrawn. One participant refused to participate when contacted via telephone for the follow up portion of the study. Nineteen participants were withdrawn after efforts to contact them for follow up via telephone were unsuccessful following at least four attempts. Another 14 participants were withdrawn because information regarding receipt of the evaluation report was not obtained, as this information was necessary in order to determine the appropriate timing of follow up contact. A total of 31 (52%) Group A participants completed the Time 2 (2 months) follow up. Of these 31 Group A participants, 8 were subsequently unable to be reached via telephone for completion of the 4month follow up interview or the follow up questionnaires. Across the entire study, f orty five participants (50%) completed the 4month follow up telephone interview. Of these, 23 were Group A participants who also completed the 2month follow up interview, while the remaining 22 were Group B participants who were only contacted for the 4month follow up interview. A total of 28 participants (31%) completed all follow up portions of the study, including all required telephone interviews and the follow up BASC2 PRS which was mailed to their homes. Figure 1 presents a summary of participation and drop out rates at each phase of the study. Inclusion Criteria Inclusion criteria for the current study were as follows: Signed informed consent by parent or legal guardian of target child 43


Child or adolescent had to be between 3 and 17 years of age Child must have resided primarily with consenting parent or guardian for at least 6 months prior to consenting date; if parent had joint custody, child must have resided with consenting parent at least 50 percent of the time over the course of 6 months Child was expected to reside with consenting parent or guardian for at least 6 months following consenting date, with at l east 50 percent residency with custodial parent during this time Home or cellular telephone number was provided Parent primary language had to be English Exclusion Criteria Exclusion criteria for the current study were as follows: Parent self reported his tory of memory disorder Parent inability to read or understand study questionnaires Measures Demographic Information A questionnaire (see Appendix ) was completed by parents in which questions were asked pertaining to demographic characteristics of the par ent (age, gender, race/ethnicity, education, marital status, occupation), the family (family income, family structure), and the child (age, gender, race/ethnicity, grade in school). Parents were also asked to provide information regarding the child’s hist ory of psychological assessment, diagnoses, and treatment, including current and past use of psychotropic medications. Adherence and Perceived Barriers The Adherence Telephone Interview Form (ATIF; MacNaughton & Rodrigue, 2001) was used to assess parental recall of and adherence to recommendations 44


provided by the assessing psychologist following a psychological assessment. Perceived barriers to adherence were also obtained using this measure. Good interrater reliability was reported by MacNaughton and Rodrigue for ratings of recall ( interclass correlation coefficient = .77) and adherence ( interclass correlation coefficient = .74), as well as for coding of recommendation type (interclass correlation coefficient = .74). Recommendation types were coded using similar criteria to those employed by MacNaughton and Rodrigu e with one exception. For the present study, the MacNaughton and Rodrigue “professional nonpsychological” category w as divided into two separate categories, 1) “physician consultation,” and 2) “other nonpsychological consultation,” to distinguish between recommendations to seek medical consults with a physician (such as for a medication trial) and recommendations for consultation with other nonpsychological professionals, such as speech and language pathologists or occupational therapists. For the present study, the ATIF was adapted based on changes made by Dreyer and colleagues (2010) in their investigation of adherence to recommendations in an ADHD population. This adaptation utilizes a 5 point Likert scale to measure perceived adherence. Perceived usefulness of each recommendati on was also measured using a 5 point Likert scale in contrast to the dichotomous variable (useful/not useful) used by MacNaughton and Rodrigue (2001), allowing for greater variability in ratings of perceived usefulness. Additionally, parents were asked to rate how well they understood each recommendation using a 5point Likert scale. In line with the Dreyer et al adaptation of the ATIF, a final question was added to the end of the interview asking parents to list factors that they believed would have made adherence to 45


recommendations easier. Below is a detailed description of the adapted ATIF as it was administered for the present study. After explaining what the parent was to expect from the interview, the interviewer asked parents to recall recommendations provided to them during the child’s assessment and rate the usefulness of each recommendation on a 5 point Likert scale. Parents wer e then asked to rate how well they understood the recommendation and their level of adherence to each recommendation. In addition, parents were asked to list any factors they believed may have interfered with their ability to adhere to the recommendation. Following the free recall portion of the interview, the unrecalled recommendations were then listed by the interviewer and parents were asked to rate each of these recommendations on a 5point Likert scale for perceived usefulness, the degree to which th e recommendation was understood, and level of adherence. Parents were also asked to describe any factors that interfered with their ability to adhere to the unrecalled recommendations. Barriers were summed for each recalled and unrecalled recommendation to produce a total barriers variable corresponding with each recommendation. Mean barriers was calculated by dividing the total number of barriers reported by the number of recommendations prescribed. Finally, mean adherence was calculated by adding adherence ratings across recommendations and dividing this sum by the total number of recommendations given. Child Symptom Severity and Functional Impairment At the time of the initial assessment and at the 4month follow up, parents completed measures of c hild symptom severity and functional impairment. First, parents completed the Behavior Assessment System for Children, Second Edition, 46


Parent Rating Scales (BASC2PRS; Reynolds & Kamphaus, 2004) which measures a wide range of behavioral problems, emotional difficulties, and adaptive skills. Three forms of the BASC2PRS were used corresponding with each child’s age (preschool form for ages 3 to 5, child form for ages 6 to 11, and adolescent form for ages 12 to 17). Unlike the Child Behavior Checklist (CBC L) used by MacNaughton and Rodgrigue (2001), the BASC2PRS does not provide a total score but instead includes four composite scales: Externalizing Problems, Internalizing Problems, Behavioral Symptoms Index, and Adaptive Skills. The Externalizing Problems composite scale includes an assessment of aggression, hyperactivity, and conduct problems. The Internalizing Problems scale assesses symptoms of anxiety, depression, and somatization. The Behavioral Symptoms Index reflects the overall level of problem behaviors and includes the following six subscales: Hyperactivity, Aggression, Depression, Attention Problems, Atypicality, and Withdrawal. Finally, the Adaptive Skills composite scale measures adaptability, activities of daily living, functional commun ication, social skills, and leadership abilities. The BASC2PRS has demonstrated good concurrent validity with other measures of child behavior. .High internal consistencies (coefficient alphas = .89 to .95) and test retest reliabilities ( r = .78 to .92) were also reported by Reynolds and Kamphaus for the four PRS composite scales In the current study, the internal consistency of the BASC2PRS composite scales was assessed for the child and adolescent forms at Time 1 only, as there were too few preschool forms completed at Time 1 ( n = 6) and insufficient BASC2PRS forms returned at Time 3 for this analysis (preschool, n = 2; child, n = 14; adolescent, n = 13). Internal consistencies for the present sample were similar to those reported in the 47


BASC2 PRS m anual, with coefficient alphas ranging from .84 to .94 for the four BASC2PRS composite scales. In addition to the BASC2PRS, parents also completed the Perceived Impairment Scale to assess the child’s level of impairment across four major areas of f unctioning (academic, behavior, emotional, and social), as well as parental concerns regarding the possible short and long term consequences of the child’s problems. This scale was developed by Human and Teglasi (1993) to assess the Perceived Severity di mension of the Health Belief Model. It yields a total impairment score obtained by summing ratings across fifteen items. The scale previously demonstrated good internal consistency (coefficient alpha = .84; Human & Teglasi, 1993), although reliability es timates for the full 15 point scale were lower for the present study (Time 1 coefficient alpha = .65, Time 3 coefficient alpha = .63). Parenting Stress Parental stress associated with parenting the target child was assessed using the Parenting Stress Index, Short Form (PSI SF; Abidin, 1995). The PSI SF was developed on the assumption that child and parent characteristics, as well as parent child interaction patterns, determine parenting stress levels. This measure is composed of four clinical subsca les: Total Stress, Parental Distress (PD), Parent Child Dysfunctional Interaction (P CDI), and Difficult Child (DC). For the present study, the Total Stress subscale, an index of overall parenting stress, was used for all analyses. The PSI SF has been used previously in studies examining predictors of adherence to recommendations and treatment compliance (e.g. Dreyer et al., 2010; Wells et al., 2000). Scale developers reported good internal consistency (coefficient alpha = .91) 48


and test retest reliabili ty ( r = .84) for the Total Stress subscale (Abidin, 1995). Good internal consistency was also observed for the Total S tress subscale in the present study (coefficient alpha = .95). Procedures Initial recruitment of participants took place on the day o f each child’s scheduled assessment. Clinicians, who were not part of the research team, made initial contact with families regarding the study. They provided a brief description of the study and asked if the caregivers wished to be approached by study s taff with more details. Families who expressed interest were then approached in the clinic waiting area by the principal investigator or a research assistant and they were given a detailed description of the purposes and procedures of the study, as well a s information about their rights with regard to confidentiality and voluntary withdrawal from the study at any point. All caregivers who met study criteria and who agreed to participate then signed an informed consent form approved by the university’s int ernal review board and were enrolled in the study. Although the parents or guardians of children undergoing psychological evaluations were the primary participants in this study, information was obtained for the children of these caregivers who were under going a psychological evaluation. Consequently, the children were also provided with information about the study and given the opportunity to give informed assent or to refuse participation. Some children were not asked to provide informed assent, includ ing children younger than eight years of age and those who were deemed by their caregiver or by the treating clinician to be unable to comprehend a basic description of the study activities or their rights regarding participation and confidentiality. All youths who were approached 49


agreed to have their caregiver participate in this research. Although records were not kept with regard to rates of initial participation, it is notable that the majority of families approached about this research agreed to part icipate at the point of recruitment. After informed written consent was received, caregivers then completed questionnaires pertaining to child symptom severity/impairment and parenting stress, as well as a demographic questionnaire developed for t he present study (see Appendix ). At the time of recruitment (Time 1), participants were randomized to one of two groups1 (See Figure 21 for procedure flow diagram). The first group (Group A) was contacted for a telephone interview at two time points following delivery of the child’s assessment report (2 months and 4 months) in order to examine adherence to recommendations and barriers to adherence. During completion of the 2month follow up interview (Time 2), in addition to gathering information about adherence and barriers, the interviewer attempted to answer questions and provided clarification regarding the recommendations when appropriate. When the answers to participants’ questions were not known to the interviewer, participants were encouraged to contact their clinicians for additional support. When questions or concerns regarding the results of the evaluation (i.e ., diagnoses, testing results) emerged, participants were directed immediately to their clinicians. The second group (Group B) was not contacted for a 2 month follow up 1 Note: As only 15 participants were s ought for the group that did not receive a 2month follow -up (Group B), only 30 participants (one third of the total participants recruited) were randomized to this group to account for attrition. Randomization was conducted by randomizing participants in groups of three. 50


interview (Time 2), while all families (Groups A and B) were contacted at 4 months aft er delivery of the report (Time 3). At the 4 month follow up (Time 3), adherence and barriers were measured for all participants via a telephone interview, just as was done for Group A at the 2month follow up (Time 2). At Time 3, all parents (Groups A and B) also completed a measure of child symptom severity and a measure of impairment to examine change from the time of the initial assessment, as these measures were also completed at the time of recruitment (Time 1). The child impairment measure (Perceiv ed Impairment Scale) was administered over the phone during the Time 3 telephone contact while the measure of child symptom severity (BASC2PRS) was mailed to families along with a self addressed, stamped envelope that participants used to return completed questionnaires to the investigator. When forms were not returned, research assistants attempted to contact families at least once by telephone to confirm receipt of the questionnaire and to provide reminders, resending forms when necessary. Twenty nine (48%) of the BASC2 PRS forms mailed to families were returned by the completion of the study. Parents who participated in the 2month follow up were given a $5 gift card as compensation at that time. Upon return of the BASC2PRS by mail following the 4month follow up interview, a second $5 gift card was mailed to families for completion of the study. Parents who only participated in the 4month follow up according to the study design were given the full $10 upon return of the BASC2PRS by mail. T he approximate timing of follow ups was determined by tracking each child’s assessment process. Investigators were notified when the final assessment report was completed and delivered to the family, either in person during an in person feedback 51


session o r in the mail. This date was then entered in a database and used to calculate the timing for each follow up interview. 52


Figure 2 1. Participation and procedural flow diagram. BASC2 PRS = Behavior Assessment System for Children, Second Edition—Parent Rating Scale; PIS = Perceived Impairment Scale; PSI = Parenting Stress Index; ATIF R = Adherence Telephone Interview Form —Revised. Group A ( n = 60) Completed all initial questionnaires ( n = 58) Did not complete all initial questionnaires ( n = 2) Time 1: Enrollment ( N = 90) (Demographics, PIS, BASC2 PRS, PSI) Time 2: 2 month follow up (ATIF R) Group A: Contact attempted for Time 2 ( n = 58) Completed 2month ATIF R ( n = 31 ) Refused participation ( n = 1 ) Unable to contact for follow up ( n = 26) Time 3 : 4 month follow up (ATIF R, PIS, BASC2 PRS) Group B ( n = 30) Completed all initial questionnaires ( n = 29) Did not complete all initial questionnaires ( n = 1) Group A: Contact attempted for Time 3 ( n = 31 ) Completed 4month ATIF and PIS ( n = 23) Unable to contact for follow up ( n = 8 ) Group B: Contact attempted for Time 3 ( n = 29 ) Completed 4month ATIF and PIS ( n = 22) Unable to contact for follow up ( n = 7 ) Group A: BASC2 PRS forms mailed ( n = 23 ) Returned 4 month BASC2PRS ( n = 18) Did not return 4 month BASC2PRS ( n = 5 ) Group B: BASC2 PRS forms mailed ( n = 22 ) Returned 4 month BASC2PRS ( n = 10) Did not return 4 month BASC2PRS ( n = 1 2 ) 53


CHAPTER 3 RESULTS Demographics Each participant completed a demographic s questionnaire consisting of questions about the caregiver, the family, and the child being evaluated. Tables 3 1 and 32 present a summary of these data. Demographic information was available for 87 participants, as 3 participants did not complete a dem ographic s questionnaire despite giving their consent for participation. The mean age of caregive rs for the total sample was 40.18 years ( SD = 9.59 ), with no significant difference in mean caregiver age across groups A and B ( p = .42). The majority of caregivers reported their race/ethnicity as Caucasian ( n= 72, 80%), while the remaining participants were African American ( n = 11, 12.2%), Hispanic ( n = 2, 2.2%), Asian American ( n = 2, 2.2%), or “Other” ( n = 2, 2.2%). The reported family yearly income ranged from $6,000 to $10,000 per year (4.4%) on the low end to greater than $61,000 (34.4%) on the high end. There were no significant group differences in family income ( F = .404, p = .53) or parent race/ethnicity ( F = .146, p = . 70). The majority (92%) of participating caregivers were biological or adoptive mothers ( n = 70) or fathers ( n = 12), while the remaining 18 percent identified themselves as a grandmother ( n = 4), guardian ( n = 2), or step parent ( n = 1). The mean child age for the total sample was 10.02 years , with no significant differences between Groups A and B ( n = 89, t (87) = .65 , p = .52 ) . Similar to the caregiver race/ethnicity distribution, the majority of children were Caucasian ( n = 66, 74.2%), followed by African American ( n = 11, 12.4%), Hispanic ( n = 7, 7.9%), Asian American ( n = 2, 2.2%), and “Other” ( n = 3, 3.4%). As seen in Table 3 1, all grade 54


levels were represented in our sample. Sixty one percent ( n = 55) of the children ranged from prekindergarten level through 5th grade, while the remaining youths were in middle school and high school grade levels ( n = 31, 34.2% ). Demographic characteristics were examined separately for participants for whom follow up data was availa ble ( n = 53) versus those who did not complete any follow up portion of the study ( n = 3 6 , Table 3 3 ). Children for whom follow up data was obtained were approximately one year younger ( M = 9.53 years, SD = 3.63) than were the children for whom follow up data was not obtained ( M = 10.75 years, SD = 3.43 ), although this difference did not reach significance ( t = 1.59, p = .76 ). Caregivers who participated in at least one follow up portion of the study were also younger ( M = 38.60 years, SD = 7.95) than were the caregivers who did not complete the 2month or 4month follow ups ( M = 42.48 years, SD = 11.31), and this age difference between groups was statistically significant ( t = 2.08, p < .05). Chi square tests indicated that the likelihood of belonging to the group for which follow up data was obtained did not vary significantly by child sex ( 2 = 1.11, p = .29), caregiver sex ( 2 = .70, p = .40), child race/ethnicity, ( 2 = 2.38, p = .67), caregiver race/ethnicity ( 2 = 2.98, p = .56), o r family income ( 2 = 4.89, p = .56). Preliminary Analyses Multiple demographic factors were entered into two separate regression analyses predicting mean adherence at 2 months and 4 months in order to determine whether any of these factors needed to be controlled for in the study’s primary analyses. With the exception of child age at time of assessment, there were no significant relationships between any of the demographic variables and adherence at 2 months or 4 months (all 55


p ’s > .05), thus the decisi on was made to exclude these demographic factors from primary study analyses. A single significant relationship emerged between child age at time of assessment and 4month mean adherence ( t = 2.33, p = .03), thus child age was controlled for in subsequent analyses for which adherence at 4 months was the primary outcome. Although child age did not significantly predict mean adherence at 2 months, the p value neared significance ( t = 1.86, p > .07), thus child age was also controlled for when predicting 2 month adherence. In both groups, older child age predicted lower rates of parental adherence to recommendations. Child age was not significantly associated with total barriers at 2 months or 4 months ( p ’s > .05). A separate regression analysis was conducted to identify any association between the total number of recommendations prescribed and mean adherence. Contrary to expectations, total recommendations did not predict mean adherence at 2 months ( t = .26, p = .80) or 4 months ( t = .78, p = .80), ind icating that receiving more recommendations was not associated with lower (or higher) rates of overall adherence. On average, parents and guardians recalled 37% of the recommendations prescribed. On a 5point Likert type scale, a mean rating of 4 was giv en for the usefulness of recommendations, while a mean rating of 4.86 was obtained for level of understanding across recommendations. Whether or not the recommendation was recalled, perceived usefulness ratings, and ratings of understanding were entered i n a bootstrapped regression analysis as predictors of 4month adherence, controlling for child age and mean adherence. Adherence ratings at 4 months were slightly higher for recalled recommendations ( M = 3.88) as compared to unrecalled recommendations ( M = 3.79), but this difference was not statistically significant ( t = 1.13, p = .26). This finding of non56


significant differences in adherence rates for recalled and unrecalled recommendations was also true when each recommendation type was examined separately. In contrast, ratings of usefulness ( t = 6.01, p <.001) and understanding ( t = 3.67, p <.001) significantly predicted 4 month adherence across recommendations. Interestingly, these relationships were not consistent across recommendation types. Perc eived usefulness was significantly positively associated with adherence at 4 months for school based recommendations ( t = 3.24, p < .01), recommendations for physician consultation/services ( t = 2.57, p < .05), recommendations for other nonpsychological p rofessional consultation/services ( t = 2.53, p < .05), and active self help ( t = 4.15, p < .001), but not for psychological recommendations ( t = .43, p = .67). Parent ratings of their level of understanding of recommendations w ere positively related to 4 month adherence only for recommendations pertaining to physician consultation/services ( t = 2.28, p < .05) and active self help ( t = 2.13, p < .05), but not for psychological recommendations ( t = 1.28, p = .21), school based recommendations ( t = 1.19, p = .24), or recommendations for other nonpsychological professional consultation/services ( t = 1.49, p = .14). Descriptive data was also obtained for questionnaires completed by caregivers at Time 1 and Time 3. Means and standard deviations for the BASC2 PR S (Time 1 and Time 3), PIS (Time 1 and Time 3), and PSI (Time 1) are presented in Table 3 4 . Notably, mean scores for all BASC2 composite scales are below the clinical cut off score of 70, suggesting this sample does not exhibit particularly severe diffic ulties as measured by the BASC2. C orrelational analyses did not show a significant relationship between the number of recommendations prescribed and severity of child problems or 57


impairment as measured by ratings on any of the BASC2PRS composite scales or the PIS Total Score at Time 1 or Time 3 (all p ’s > .05). D istributional properties of all independent and dependent variables were also examined to determine whether any transformations were needed prior to conducting primary analyses. The normality of each variable was determined through visual inspection of distributional graphs (i.e., histograms and stem andleaf plots), by examining the magnitude of the skewness and kurtosis, and by evaluating the significance level of the Kolmogorov Smirnov and Shap ir o Wilk tests of normality. When significant violations of the normality assumption were observed, efforts were made to normalize the data utilizing transformation techniques, including use of the log base 10, square root, and reciprocal transformations . If a transformation was successful in reducing the non normality of data, the new transformed variable was used to replace the original variable in all analyses. However, when transformations were not successful, such as was the case with the adherence ratings and total barriers variables utilized in Aim 1, bootstrapping was utilized to increase the validity of more traditional parametric techniques (i.e., ANOVA, linear regression) using the original, untransformed variable. In addition to conducting A NOVA and regression analyses with bootstrapping, corroborating support for findings from bootstrapped linear regressions and ANOVAs was obtained by dichotomizing dependent variables and subsequently examining these new variables using logistic regression, an alternative regression approach that does not assume the normal distribution of dependent variables. 58


Primary Analyses Aim 1 Analyses Four month interview data was used to complete all Aim 1 analyses, as there was more complete data available for this time point (45 participants) than for the 2month follow up (31 participants), giving the data more power to detect significant relationships. However, both 2month and 4month descriptive information is presented for review in Tables 3 5 and 3 6 . Each r ecommendation was classified into one of five categories: 1) psychological services, 2) school based recommendations, 3) physician consultation/services, 4) other nonpsychological professional consultation/services, and 5) active self help . A total of 4 86 recommendations were prescribed to the participants of this study for whom the evaluation report was received by study staff ( n = 80) . No information regarding number and type of recommendations is available for the 10 participants for whom the report evaluation was not obtained from the participants’ clinician. A cross participants who participated in at least one follow up portion of the study ( n = 53) , 308 total recommendations were prescribed. S ixteen percent of thes e 308 recommendations ( n = 48 ) wer e for psychological services, 26% ( n = 80 ) were for school based recommendations, 13% ( n = 39) were for phy sician consultation/services, 19% ( n = 59 ) were for other nonpsychological professional consultation, and 27% ( n = 82) involved active self help activities (Table 3 7 ). Mean adherence across all recommendations was 3.97 ( SD = 1.40) at the 2month follow up and 3.85 ( SD = 1.51) at the 4 month follow up, and these ratings varied by recommendation type: psychological services, M = 3.31, SD = 1.65; school based recommendations, M = 4.13, SD = 1.41; physician consultation/services, M = 4.47, SD = 1.06; other non59


psychological professional consultation, M = 3.45, SD = 1.78; and active self help, M = 3.87, SD = 1.37. Barriers reported by caregivers were categorized into one of six categories: 1) access problems ( n = 49), 2) financial problems ( n = 25), 3) competing time or scheduling demands ( n = 31), 4) caregiver characteristics ( n = 19), 5) child characteristics ( n = 26), or 6) knowledge pro blems ( n = 4). See Table 38 for criteria used to determine appropriate category classifications. A total of 154 barriers were reported across recommendations. As is noted in Table 3 8 , the most common barriers reported were those pertaining to access p roblems (32%) and competing time or scheduling demands (20%). A greater number of barriers per recommendation was reported for recommendations for psychological services at the 2 month ( n = 27, M = .93 ) and 4 month follow ups ( n = 38 , M = 0 . 90 ; Table 3 5 ) . Hypothesis 1a analysis. For analyses 1a, 1c, and 1d, a separate line of data was created for each recommendation prescribed, allowing comparisons to be made across individual recommendations rather than across participants. Using recommendation type as the predictor variable, a bootstrapped univariate analysis of variance (ANOVA) was conducted to test the hypothesis that total perceived barriers would differ significantly across recommendation types. Means and standard deviations for the total number of barriers reported for each recommendation type are presented in Table 3 5 . A distributional analysis of the total barriers variable revealed significant deviations from normality, as the variable was highly positively skewed due to frequent ratings of zero barriers. Transformation approaches were unsuccessful at improving the distribution of this variable. Due to the violation of normality assumptions, a traditional 60


ANOVA, which relies on the assumption of a normally distributed outcome variable, was not appropriate. Instead, in order to retain the variance afforded by the 5point adherence scale a bootstrapped ANOVA was employed rather than relying on other nonparametric tests that would have required significantly altering the adherence variable. The bootstrapped ANOVA was then followed up with a logistic regression analysis using a dichotomous adherence variable designed to demonstrate the ability for recommendation type to predict the presence or absence of reported barriers. Results of both analyses are reported below. ANOVA with bootstrapping results revealed significant differences in total barriers across recommendation types after controlling for child age ( F = 2.97, p = .02). Post hoc analyses with Bonferroni correction were then conducted to determine whether total barriers reported for psychological recommendations differed significantly from the total barriers reported for other recommendation types, testing the hypothesis that psychological recommendations are associated with more perceived barriers than other types of recommendations. In order to conduct these analyses, child age was first removed from the model where it was initially entered as a covariate, as post hoc analyses cannot be obtained in an A NOVA when covariates are included. Results of the post hoc analyses revealed that recommendations for psychological services were associated with more perceived barriers than school based recommendations ( p = .03), but not when compared to other types of recommendations prescribed (Tabl e 3 9 ) . While there were no a priori hypotheses regarding differences in perceived barriers across other types of recommendations, it is notable that perceived barriers did not 61


differ significantly for any of the pair wise comparisons that did not include recommendations for psychological services. For the follow up logistic regression analysis, the total barriers variable was recoded into a dichotomous variable representing the presence (coded as 1) or absence (coded as 0) of reported barriers, and this variable was then entered as the dependent variable in the analysis. Table 3 5 presents the frequency with which at least one barrier was reported for each recommendation type. Seventy one percent ( N = 30) of recommendations for psychological services were associated with at least one barrier at the 4 month follow up, while this percentage ranged from 42% to 49% for the other four recommendation types . The categorical recommendation type variable was then entered in the sec ond block of a logistic regression following child age in the first block. The addition of recommendation type to the model’s second block resulted in a decrease in the 2 log likelihood of 11.61, representing a significant change from the previous model including child age only [ 2 ( 2 , N = 258) = 25.26, p < .001; Table 3 10]. Overall, the recommendation type variable significantly predicted reported barriers (Wald test = 10.65, p < .05). To determine which recommendation types differed significantly from one another, individual beta weights for each recommendation type were examined. Each recommendation type was compared to recommendations for psychological services which was set as the reference category for the analysis. Results indicate that recommendations for psychological services were more likely to be associated with at least one barrier than school based interventions ( B = 1.39, p < .01), physician consultation/services ( B = 1.14, p <.05), and active self help ( B = 1.09, p < 62


.05), but not other nonpsychological professional consultation/services ( B = .86, p = .07). Hypothesis 1b analysis. The hypothesis that a negative relationship would exist between perceived barriers and adherence was tested using a hierarchical regression analys is. For this analysis an average score was calculated for perceived barriers by adding the total number of reported barriers across recommendations and dividing this sum by the number of recommendations provided. The use of an average barriers score inst ead of an absolute sum controls for variability across participants in the total number of recommendations provided by the psychologist. This average barriers variable, which is a continuous variable ranging from 0 to 1.4 ( M = .62, SD = .06), was then ent ered as the independent variable in a regression analysis with 4 month mean adherence as the outcome variable. Mean adherence scores were calculated for each participant by adding individual adherence ratings (rated on a 5point Likert scale) across all r ecommendations and dividing this sum by the total number of recommendations prescribed. This procedure produced a continuous variable ranging from 1.83 to 5 ( M = 3.81, SD = .11). Dividing the sum by total number of recommendations once again accounted fo r the variability in number of recommendations across participants. After accounting for child age at recruitment in Block 1, mean barriers was entered in Block 2 of a hierarchical regression analysis. Contrary to our hypothesis, mean barriers did not si gnificantly predict mean adherence at the 4 month follow up ( t = .208, p = .836; Table 311 ). While the initial model with child age as the only predictor of 4month adherence was significant ( R2 = .11, F = 5.07 , 63


p < .05), the addition of mean barriers did not result in a significant change in explained variance ( R2 = .11 , F = 2.50, p = .09). While this analysis indicates that the mean number of barriers reported across recommendations did not predict participants’ ov erall adherence across recommendation types (represented as mean adherence), it does not speak to whether or not adherence to a specific recommendation is predicted by the total number of perceived barriers associated with that recommendation. To further examine the relationship between adherence ratings and barriers, a bootstrapped ANOVA was conducted with adherence ratings set as the criterion variable and barriers as the primary predictor, with child age entered as a covariate. As with the barriers var iable, the distribution of adherence ratings varied significantly from the normal distribution, as it was highly negatively skewed due to a high frequency of ratings of 5. To account for violations of the normality assumption, bootstrapping was performed in this analysis. Unlike the previous analysis which did not show a relationship between mean adherence and mean barriers, a negative relationship emerged between individual adherence ratings and number of barriers, wherein parents who reported a greater n umber of barriers reported lower rates of adherence ( F = 13.21, p < .001). Once again, a follow up bootstrapped logistic regression analysis was conducted with number of barriers as the primary independent variable and a dichotomous adherence score as th e outcome variable. Child age was entered as a predictor in the first block, and total barriers was entered in the second block . To create the dichotomous adherence outcome variable, 4 month adherence ratings for each recommendation were recoded as 0 or 1 representing nonadherence or partial 64


adherence (recoded as 0 for ratings of 1 through 4) and full adherence (recoded as 1 for ratings of 5), respectively. The decision to use 5 as the cut point for full adherence was made because ratings of 5 on the 5 point Likert type scale were given for 55% of all recommendations, thus using this cut p oint divided the recommendations into two groups of similar size. Corroborating the previous ANOVA, total perceived barriers significantly predicted the dichotomous adherence score at the 4 month follow up ( B = 1.16, Wald test = 25.89, p < .001; Table 3 12 ). The model with only child age as a predictor had a 2 log likelihood value of 324. 27 . The addition of the total barriers variable led to a significant decrease in the model’s 2 log likelihood value, 2 (2, N = 258) = 35.59, p < .001. F ewer perceived barriers were associated with a greater likelihood of reporting full adherence to a specific recommendation. Hypothesis 1c analysis. To compare adherence ratings across recommendation types, a bootstrapped ANOVA was conducted with adherence ratings as the outcome variable and recommendation type as the grouping variable, controlling for child age as a covariate. Means and standard dev iations of adherence ratings for each recommendation type are presented in Table 3 6 . The lowest 4 month mean adherence was reported for psychological recommendations ( M = 3.31), followed by nonpsychological professional consultation ( M = 3.45), active s elf help ( M = 3.8 6 ), and school based recommendations ( M = 4.13). The highest mean adherence was for recommendations for physician consultation/services ( M = 4.47). In this analysis, adherence ratings differed significantly across recommendation types ( F = 5 . 45, p < .0 0 1). Once again, post hoc analyses were conducted by running the ANOVA without child age as a covariate. Based on results of these post hoc test s with Bonferroni 65


corrections, a dherence to psychological recommendations was significantly low er than adherence to school based recommendations ( p < .05) and physician consultation/services ( p < .01), but it did not differ significantly from other non psychological professional consultation/services ( p = 1.00) or active self help ( p = .49). Adherence to recommendations for physician consultation/services was significantly higher than adherence to recommendations for other non psychological professional consultation/services ( p < .05). Table 3 13 presents the 4month mean difference in adherence ratings for each pair of recommendation types. A bootstrapped logistic regression analysis was conducted once again as a follow up to this bootstrapped ANOVA, comparing the likelihood of ratings of ful l adherence versus partial adherence or nonadherence across recommendation types. For this analysis, the recoded dichotomous adherence variable was entered as the outcome variable, with child age entered in the model’s first block. For the categorical re commendation type variable, recommendations for psychological services was set as the reference category against which all other dummy coded recommendation type variables were compared. With recommendation type added to the logistic regression model, the 2 log likelihood decreased significantly from the model that included only child age ( Model 2 ( 5 , N = 267) = 20.87 , p < .01), indicating that the likelihood of belonging to the full adherence group varied significantly across recommendation types (Wald t est = 15.41, p < .01 ; Table 3 14). Recommendations for psychological services were significantly less likely to be associated with full adherence than school based recommendations ( B = 1.26, p <.01) and recommendations for physician 66


consultation/services ( B = 1.62, p < .01). These results were consistent with results of the ANOVA analysis using the original 5point adherence variable. Hypothesis 1d analysis. Results of previous analyses suggest that increased barriers are associated with lower rates of adherence overall. Additionally, both adherence ratings and the number of reported barriers vary significantly across recommendation types, with recommendations for psychological services predicting lower rates of adherence and more reported barriers to adherence. This next analysis was intended to demonstrate whether recommendation type and barriers interact to predict parent ratings of adherence. Table 3 15 present s means and standard deviations for adherence ratings separately by recommendation type at each level of the total barriers variable. A bootstrapped ANOVA was conducted to explore the role of perceived barriers as a moderator in the relationship between r ecommendation type and adherence. For this analysis, adherence at 4 months was compared across recommendation types and number of barriers. Additionally, a barriers by recommendation type interaction term was included in the model to detect any moderation effect between recommendation type and barriers. In addition to child age, mean adherence was also included as a covariate to control for each participant’s overall adherence across recommendation types. Results of the bootstrapped ANOVA revealed sign ificant main effects for mean adherence ( F = 78.12, p < .001) and total barriers ( F = 13.02, p < .001). Although earlier analyses indicated a significant relationship between recommendation type and adherence, the main effect of recommendation type was no longer significant with the addition of the recommendation type and barriers interaction term ( F = 1.71, p = .15). 67


There was a significant interaction between recommendation type and barriers ( F = 2.02, p < .03). To further explore this interaction effect, correlational analyses were conducted separately for each recommendation type examining the relationship between adherence ratings and barriers. While the Pearson correlation was negative for each recommendation type, indicating lower rates of adherence as barriers increased, this relationship was only significant for psychological recommendations ( r = .32, n = 42, p < .05; d = .68), school based recommendation ( r = .52, n = 69, p < .01; d = 1.22), and active self help ( r = .31, n = 67, p < .05; d = .65), but not for physician consultation/services ( r = .32, n = 34, p = .06 ; d = .68 ) or other nonpsychological professional consultation ( r = .2 6 , n = 46, p = .09 ; d = .53 ). It is notable that the magnitude of these correlations were all similar, despite differences in significance levels. Aim 2 Analyses As with Aim 1, the analyses below were conducted using mean barriers and mean adherence ratings for Time 3 (4month follow up), as well as parent ratings of symptom severity and impairment from Time 1 (initial contact). Hypothesis 2a analysis. To test the relationship between parent perceptions of child symptom severity/impairment and adherence, two hierarchical regres sion analyses were conducted using data from the Behavior Assessment System for Children, Second Edition (BASC2 PRS) and Perceived Impairment Scale (PIS) completed during the recruitment phase of the study (Time 1). For perceived symptom severity, scores on each of the four BASC2 PRS composite scales (Internalizing Problems, Externalizing Problems, Behavioral Symptoms Index, and Adaptive Skills) were entered as predictors 68


in a hierarchical regression (controlling for child age in Block 1) with Time 3 (4 month follow up) mean adherence scores as the primary dependent variable. A separate hierarchical regression was conducted with total scores from the PIS predicting mean adherence at the 4month follow up, once again controlling for child age in Block 1. Re sults of these analyses did not support the hypothesized relationship between symptom severity and adherence, as there was no significant relationship between symptom severity ( R2 = . 13, F (5, 38) = 1.10, p = .38 ; Table 3 16) or impairment ( R2 = .11 , F (2,42) = 2.51, p = .09; Table 317 ) as reported at the time of the initial assessment and mean adherence to recommendations at 4 months. Mean adherence scores remained nearly constant as BASC scores and PIS scores increased. Hypothesis 2b analysis. Alth ough 4 month adherence was not significantly related to initial ratings of symptom severity or impairment, the hypothesized role of barriers as a moderator in this relationship was nonetheless explored, as the addition of an independent variable can someti mes uncover relationships between other independent variables and the outcome that were not previously significant. The hypothesis that barriers moderate the relationship between parent ratings of child symptom severity/impairment and adherence was thus examined by creating an interaction term for each BASC2 PRS composite score (representing perceived symptom severity) and mean barriers. An interaction term was also computed for the total PIS score (perceived impairment) and mean barriers. These interact ion terms were subsequently entered in the third block of their respective hierarchical regressions, with child age entered in the first block and the severity or impairment scores entered in the second block along with perceived barriers. The barriers, c hild symptom severity, 69


and impairment variables were centered prior to computing the respective interaction terms to control for multicollinearity. In the first regression analysis, the addition of the four interaction terms consisting of each BASC2 PRS composite score and total barriers did not result in a significant increase in explained variance in the outcome variable ( R2 = .23 , F = .96, p = .50 ; Table 3 18). Similarly, the PIS and barriers interaction term was not a significant predictor of adherenc e in the second regression analysis ( R2 = .13, F = 1.49, p = .22 ; Table 319). Overall, adherence at 4 months was not predicted by measures of symptom severity and impairment, or their respective interactions with reported barriers. Aim 3 Analyses Hypothesis 3a analysis. The hypothesized positive relationship between parenting stress and mean adherence to recommendations was examined using a hierarchical regression analysis. Block 1 of the analysis included child age, and the Total Stress composit e score from the Parenting Stress Index (PSI) was then entered in Block 2, with mean adherence at 4 months serving as the dependent variable. Parenting stress, as rated by the PSI, explained almost no variance in mean adherence ratings at 4 months ( R2 cha nge = .00, F = 2.50, p = .10 ; Table 320 ). Hypothesis 3b analysis. The role of barriers as a moderator in this relationship was examined by creating an interaction term for the PSI Total S tress and mean barriers. After centering the interaction term to control for multicollinearity, it was then entered in the third block of a hierarchical regression, with child age entered in the first block and the PSI Total S tress and mean barriers variables entered in the second block. In this regression analysis, th e mean barriers by PSI Total Stress interaction did not 70


emerge as a significant predictor of mean adherence at 4 months ( R2 = . 14, F = 1.55, p = .21 ; Table 321 ). Aim 4 Analyses Hypothesis 4a analysis. One way ANOVAs were planned to assess the relationship between feedback modality and 1) percent recall of recommendations at Time 2 and Time 3 (calculated as the number of recommendations correctly recalled during the 2month and 4 month follow ups divided by the total number of recommendations gi ven), and 2) perceived understanding of recommendations at Time 2 and Time 3 (calculated as the sum of understanding scores across all recommendation types from the 2month and 4month follow ups divided by the total number of recommendations given). Part icipants were classified into one of two groups based on the type of feedback they received (written only or written + oral) and these groups were then compared on both outcome variables (parent recall and parent understanding) at 2 month and 4month follow ups. Means and standard deviations for percent recall and perceived understanding are presented separately by feedback group in Table 322 . Overall percent recall across both groups was 36% at 2 months and 37% at 4 months. Based on the results of the A NOVA, feedback groups did not differ with regard to the proportion of recommendations recalled at the 2month ( F = .10, p = .75) or 4month follow ups ( F = .41, p = .53). Mean understanding was marginally higher at both time points for the group that received both oral and written feedback (2 months: M = 4.93, 4 months: M = 4.98) than the group that only received written feedback (2 months: M = 4.83, 4 months: M = 4.83), but this difference was not significant at 2 months ( F = 1.75, p = .20) or 4 months ( F = 2.18, p = .15). 71


Hypothes 4b analysis. A one way ANOVA was used to examine the relationship between feedback modality (written only or written + oral) and mean adherence. For this analysis, adherence ratings from Time 2 (2month follow up) and Time 3 (4month follow up) were combined, allowing for inclusion of more participants. For participants who had data for both timeframes, only data from the 2month follow up was used. In this way, all adherence ratings used for this analysis represent the first adherence ratings obtained for each participant, eliminating any possible impact of the telephone interview itself on the adherence ratings included in this analysis. A variable representing the number of days elapsed between the initial assessment and the follow up interview was used as a covariate. While the mean adherence score was somewhat higher for the group that received both oral and written feedback ( M = 4.01) than for the group that only received written feedback ( M = 3.78), this difference was not significant ( F = 1.41 , p = .23 ; Table 322). Aim 5 Analyses To examine the relationship between contact at the 2 month follow up and adherence at the 4month follow up, a oneway ANOVA was conducted. Mean adherence at Time 3 (4month follow up) w as compared between Group A, which participated in the 2month follow up, and Group B, which did not participate in the 2 month follow up (Table 323 ) . The mean 4month adherence of group A ( M = 3.89, SD = .81) was slightly higher than Group B’s 4month adherence ( M = 3.72, SD = .69). However, this difference was not statistically significant ( F = .55, p = .46). 72


Aim 6 Analyses The relationship between adherence to recommendations and change in parent ratings of symptom severity and functioning was exami ned with several hierarchical regression analyses testing the hypothesis that greater parental adherence to recommendations would be associated with greater improvements in child symptoms and functioning. To examine the relationship between adherence and change in parent ratings of symptom severity, separate analyses were conducted for each of the four BASC2 PRS composite scales (Internalizing problems, Externalizing Problems, Behavioral Symptoms Index, and Adaptive Skills). For each composite scale, a ch ange score was calculated by subtracting scores obtained at the time of the initial assessment (Time 1) and ratings obtained at the 4month follow up (Time 3). For the Externalizing Problems, Internalizing Problems, and Behavioral Symptoms Index composite scales, these change scores were calculated as the simple difference between scores at Time 1 and scores at Time 3 (Time 1 minus Time 3). The opposite computation was completed for the Adaptive Skills composite (Time 3 minus Time 1) to ensure that a posi tive score on all change variables indicated improvement (higher scores on Adaptive Skills indicate better functioning). Change scores across scales ranged from an increase of 16 points to a decrease of 40 points, with change score means ranging from .96 for Adaptive Skills to 3.04 for Behavioral Symptoms Index. Substantial individual variability was also observed for these variables, with standard deviations ranging from 6.27 to 12. 14. C hange scores were then entered as the dependent variables in four separate hierarchical regression analyses with mean adherence as the primary predictor. For each analysis, child age was entered in Block 73


1, and the respective BASC2 PRS composite score at Time 1 w as entered in Block 2 to control for initial symptom severity. Mean adherence scores from the 4month follow up were entered in Block 3 as the primary predictor of interest. Prior to conducting the regression analyses, paired samples t tests were conducted for each BASC2PRS composite scale to determine whether there were significant changes in scores between time points. Table 3 4 shows the mean scores at Time 1 and Time 3, as well as the results of each t test. Across scales, scores were noted to improv e from the initial BASC2 PRS administration to the 4month follow up administration, but these differences were not statistically significant (all p’s > .05). There was no relationship between parent ratings of adherence at the 4month follow up and change scores on the BASC2 PRS composite scales: Externalizing Problems ( R2 = .36, t = .66, p = .51 ; Table 3 24), Internalizing Problems ( R2 = .51, t = .82, p = .42 ; Table 3 25 ), Behavioral Symptoms Index ( R2 = .41, t = .68, p = .51 ; Table 326 ), and Adapti ve Skills Index ( R2 = .25, t = .86, p = .40 ; Table 3 27 ) To examine the relationship between adherence to recommendations and change in parent ratings of impairment, change scores were calculated for the PIS by subtracting Time 3 scores from Time 1 scores. This produced a continuous change variable for which positive numbers indicated improvement in parent ratings of child impairment in functioning. The PIS change scores variable had a standard deviation of 7.97 and indicated that change scores ranged from an increase of 31 points to a decrease of 13 points. This change variable was then used as the outcome variable in a hierarchical regression analysis, and mean adherence was entered as the primary predictor. As in the previous analysis, child age was entered in Block 1. This was 74


followed by Time 1 PIS scores in Block 2 to control for baseline ratings of impairment. Mean adherence was then entered in Block 3. Similar to the BASC2 PRS scales, PIS total scores did not change significantly between Time 1 ( M = 40.86, SD = 8.48) and Time 3 ( M = 40.98, SD = 8.23), t = .10, p = .92. The regression analysis indicated there was no association between adherence ratings at 4 months and changes in PIS scores from Time 1 to Time 3 ( R2 = .3 5, t = .95, p = .35 ; Tabl e 328 ). 75


Table 3 1. Demographic characteristics — parent and c hild Parent Characteristic n (%) Child Characteristic n (%) Gender Gender Male 13 (14.4) Male 58 (64.4) Female 76 (84.4) Female 32 (35.6) Ethnicity Ethnicity Caucasian 72 (80) Caucasian 66 (73.3) African American 11 (12.2) African American 11 (12.2) Hispanic 2 (2.2) Hispanic 7 (7.8) Asian American 2 (2.2) Asian American 2 (2.2) Other 2 (2.2) Other 3 (3.3) Income Grade Level 6,000 10,000 4 (4.4) Preschool 2 (2.2) 11,000 20,000 11 (12.2) K 8 (8.9) 21,000 30,000 13 (14.4) First 14 (15.6) 31,000 40,000 6 (6.7) Second 13 (14.4) 41,000 50,000 14 (15.6) Third 6 (6.7) 51,000 60,000 4 (4.4) Fourth 7 (7.8) 31 (34.4) Fifth 5 (5.6) Sixth 4 (4.4) Marital Status Seventh 12 (13.3) Married 56 (62.2) Eighth 4 (4.4) Divorced/Separated 18 (20) Ninth 4 (4.4) Living Together 1 (1.1) Tenth 3 (3.3) Single 7 (7.8) Eleventh 1 (1.1) Widowed 1 (1.1) Twelfth 3 (3.3) Separated 5 (5.6) Relationship to child Parent 80 (92) Grandparent 4 (4) Other 3 (3) 76


Table 3 2. Mean child and caregiver age for whole sample and for groups A and B Full Sample Group A Group B ( n = 89) ( n = 59) ( n = 30) M SD M SD M SD t (87) p Child Age 10.02 3.58 9.85 3.72 10.37 3.32 .65 .52 Caregiver Age 40.18 9.59 41.57 9.94 37.29 8.27 1.97 .05 77


Table 3 3. Mean child and caregiver age for participants with and without follow up data Full Sample Follow up data obtained No follow up data obtained ( n = 89) ( n = 53) ( n = 36) M SD M SD M SD t (87) p Child Age 10.02 3.58 9.53 3.63 10.75 3.43 1.59 .76 Caregiver Age 40.18 9.59 38.60 7.95 42.48 11.31 2.08 .04 78


Table 3 4. Means and standard deviations for BASC2 PRS, PIS, and PSI Time 1 Time 3 Difference n M SD Range n M SD Range t (df) p Behavior Assessment System for Children (BASC2 PRS) Externalizing Symptoms 87 62.0 14.64 35 107 29 58.45 12.09 38 85 1.67 (27) .11 Internalizing Symptoms 87 59.7 15.28 35 98 29 56.83 12.10 38 81 1.17 (27) .25 Behavioral Symptoms Index 87 68.0 14.17 43 97 29 63.66 12.93 42 90 1.77 (27) .09 Adaptive Skills 87 35.2 9.74 18 59 29 37.76 9.38 13 58 .81 (27) .42 Perceived Impairment Scale (PIS) Total Score 87 41.6 7.89 21 59 43 40.98 8.23 25 65 .10 (42) .92 Parenting Stress Inventory (PSI) Total Stress 27 78.0 19.87 13 145 79


Table 3 5. Descriptive data for total barriers at 2 months and 4 months for participants with follow up data, total and by recommendation type. Total barriers Dichotomous scale (0, 1) Observed range n M (SD) n 1 (%) 2 month follow up All recommendation types ( n = 176) 119 .68 (.75) 0 3 92 (52.3) Psychological recommendations ( n = 29) 27 .93 (.80) 0 3 21 (72.4) School based recommendations ( n = 43) 29 .67 (.68) 0 2 24 (55.8) Physician consultation/services ( n = 24) 14 .58 (.78) 0 2 10 (41.7) Other non psychological professional consultation/services ( n = 31) 22 .71 (.78) 0 2 16 (51.6) Active self help ( n = 49) 27 .55 (.74) 0 3 21 (42.9) 4 month follow up All recommendation types ( n = 256) 156 .61(0.70) 0 3 130 (50.4) Psychological recommendations ( n = 40) 38 .90 (.76) 0 3 30 (71.4) School based recommendations ( n = 69) 34 .49 (.66) 0 3 29 (42.0) Physician consultation/services ( n = 34) 17 .50 (.57) 0 2 16 (47.1) Other non psychological professional consultation/services ( n = 46) 26 .57 (.69) 0 3 22 ( 47.8) Active self help ( n = 67) 41 .61 (.74) 0 3 33 (49.3) 1 Values for n represent the frequency with which at least one barrier was reported. 80


Table 3 6. Descriptive data for adherence ratings at 2 m onths and 4 months, total and by recommendation type. 5 point scale (1 5) Dichotomous scale (0, 1) Recommendation type n M (SD) Observed range n 1 (%) 2 month follow up All recommendation types 176 3.97 (1.40) 1 5 97 (55.1) Psychological recommendations 29 3.17 (1.60) 1 5 9 (31.0) School based recommendations 43 4.14 (1.13) 1 5 21 (48.8) Physician consultation/services 24 4.21 (1.53) 1 5 18 (75.0 Other non psychological professional consultation/services 31 4.26 (1.37) 1 5 23 (74.2) Active self help 49 4.00 (1.31) 1 5 26 (53.1) 4 month follow up All recommendation types 267 3.85 (1.51) 1 5 178 (66.7) Psychological recommendations 45 3.31 (1.65) 1 5 23 (44.2) School based recommendations 70 4.13 (1.41) 1 5 54 (77.1) Physician consultation/services 36 4.47 (1.06) 1 5 30 (83.3) Other non psychological professional consultation/services 47 3.45 (1.78) 1 5 28 (59.60) Active self help 69 3.86 (1.37) 1 5 43 (62.3) 1 Values for n represent the frequency of 5 point (full adherence) ratings 81


Table 3 7. Number of recommendations for participants with follow up data presented for total sample and by recommendation type . Recommendation type n (%) All recommendations 308 (100) Psychological services 52 (17) School based Recommendations 77 (25) Physician consultation/services 41 (13) Other non psychological professional consultation/services 56 (18) Active self help 82 (27) 82


Table 3 8. Frequencies and examples for each barrier type . Barrier type Examples n (%) Access problems Lack of provider/services in community, service not covered by insurance, inaccessibility due to travel distance or other transportation difficulties (i.e., no vehicle), no nearby providers 49 (32) Financial problems Expensive copays, no insurance coverage, inability to manage other expenses associated with services 25 (16) Competing time or scheduling demands Child extracurricular activities, child or parent employment, child attending other services, childcare needs 31 (20) Caregiver characteristics Caregiver attitudes and beliefs pertaining to healthcare providers or the specific recommendations received; disagreement regarding helpfulness or appropriateness of recommendation or results of the evaluation 19 (12) Child characteristics Child attitudes and beliefs; child refusal to participate; nature and severity of child's difficulties 26 (17) Knowledge problems Poor understanding of results and recommendations; not having access to the report (i.e., lost it, never received it) 4 (3) Note . % = percentage of barriers reported by participants for whom follow up data is available ( n = 53). 83


Table 3 9. Means, standard deviations, and m ean differences for total barriers. Recommendation type 1 2 3 4 5 M SD 1. Psychological recommendations — .41* .4 .34 .29 .9 .76 2. School based recommendations — .01 .07 .12 .49 .66 3. Physician consultation/services — .07 .11 .5 .57 4. Other nonpsychological professional consultation/services — .05 .57 .69 5. Active self help — .61 .74 Note . Mean difference scores for total barriers were obtained by subtracting the mean barriers of the recommendation type listed in the left vertical column by the mean barriers of the recommendation type numbered in the top horizontal row. *p < .05. ** p < .01. 84


Table 3 10. Recommendation type as predictor of dichotomous barriers variable. 4 month total barriers Variable Model 1 Model 2 B S.E. Wald statistic B S.E. Wald statistic Constant 1.33*** .40 11.30 .52 .51 1.05 Child age 0.14** .04 12.80 0.15*** .04 14.17 Recommendation type 10.65* RecType(1) 1.39* .44 5.14 RecType(2) 1.14* .50 3.38 RecType(3) .86 .47 6.26 RecType(4) 1.09* .44 1.05 2LL 346.08 332.39 2 13.65*** 25.26*** 2 11.61 Note . Recommendation type reference category is psychological services. RecType(1) = School based recommendations; RecType (2) = Physician consultation/services; RecType(3) = Other nonpsychological professional consultation/services; RecType(4) = Active self help. *p < .05. ** p < .01. *** p <.001. 8 5


Table 3 11. Mean barriers as predictor of 4 month mean adherence . 4 month mean adherence Model 1 Model 2 Variable B t B t Constant 4.48 14.12*** 4.5 13.58*** Child age .07 2.25 .07 2.01 Mean barriers .07 .21 R 2 .11 .11 F 5.07* 2.5 .001 2.57 Note . n = 45. *p < .05. *** p <.001. 86


Table 3 12. Total barriers as predictor of dichotomous adherence variable at 4 month follow up. 4 month dichotomous adherence Variable Model 1 Model 2 B S.E. Wald statistic B S.E. Wald statistic Constant 1.47*** .41 12.82 1.83*** .45 16.86 Child age .08* .04 4.13 .04 .04 .80 4 month total barriers 1.16 .23 25.89 2LL 324.27 292.86 . n = 258. *p < .05. *** p <.001. 87


Table 3 13. Means, standard deviations, and mean differences for adherence r atings. Recommendation type 1 2 3 4 5 M SD 1. Psychological recommendations — 0.82* 1.16** .14 .56 3.17 1.6 2. School based recommendations — .34 .68 .26 4.14 1.13 3. Physician consultation/services — 1.03* .6 4.21 1.53 4. Other nonpsychological professional consultation/services — .42 4.26 1.37 5. Active self help — 4.00 1.31 Note . Mean difference scores for adherence were obtained by subtracting the mean adherence of the recommendation type listed in the left vertical column by the mean adherence ratings of the recommendation type numbered in the top horizontal row. *p < .05. ** p < .01. 88


Table 3 14. Recommendation type as predictor of dichotomous adherence at 4 month follow up . 4 month dichotomous adherence Variable Model 1 Model 2 B S.E. Wald statistic B S.E. Wald statistic Constant 1.43*** .40 12.97 1.69*** .42 15.93 Child age 0.08* .04 3.97 0.10* .04 5.76 Recommendation type 15.41** RecType(1) 1.26* .42 9.00 RecType(2) 1.62* .55 8.88 RecType(3) .23 .43 .30 RecType(4) .53 .40 1.78 2LL 335.90 319.03 . n = 267. Recommendation type reference category is psychological services. RecType(1) = School based recommendations; RecType(2) = Physician consultation/services; RecType(3) = Other nonpsychological professional consultation/services; RecType(4) = Active self help. *p < .05. ** < .01. *** p <.001. 89


Table 3 15. 4 month mean adherence and standard deviations per recommendation type at each level of total barriers 4 month adherence 0 Barriers 1 Barrier 2 Barriers 3 Barriers Recommendation type n M SD n M SD n M SD n M SD Psychological recommendation 12 3.83 1.75 24 3.25 1.60 4 3.00 1.83 2 1.00 .00 School based recommendations 40 4.65 .98 25 3.60 1.61 3 2.33 .58 1 1.00


Table 3 16. BASC2 PRS composite scores as nonsignificant predictors of 4month adherence. 4 month mean adherence Model 1 Model 2 Variable B t B t Constant 4.44*** 13.59 4.08 3.64 Child age 0.06* 2.10 .06 2.02 BASC2 PRS Externalizing Symptoms .00 .38 BASC2 PRS Internalizing Symptoms .00 .54 BASC2 PRS Behavioral Symptoms Index .00 .14 BASC2 PRS Adaptive Skills .01 .80 R 2 .10 .13 F 4.40 1.10 .10 4.42 Note . n = 45. *p < .05. ** p < .01. *** p < .001. 91


Table 3 17. PIS total score as non significant predictor of 4 month adherence. 4 month mean adherence Model 1 Model 2 Variable B t B t Constant 4.48*** 14.12 4.35*** 7.05 Child age 0.07* 2.25 0.07* 2.24 PIS Total Score . .25 R 2 .11 .11 F 5.07* 2.51 .00 .06 Note . n = 45. *p < .05. ** p < .01. *** p < .001. 92


Table 3 18. BASC2 PRS composite scores, total barriers, and interaction terms as predictors of 4 month adherence. 4 month mean adherence Variable Model 1 B Model 2 B Model 3 B Constant 4.44*** 4.39** 5.80* Child age .06* .07* .06 BASC2 PRS Externalizing Symptoms .00 .01 BASC2 PRS Internalizing Symptoms .00 .03 BASC2 PRS Behavioral Symptoms Index .00 .04 BASC2 PRS Adaptive Skills .01 .00 4 months total barriers .04 .02 BASC2 PRS Externali zing Symptoms x total barriers .00 BASC2 PRS Internalizing Symptoms x total barriers .01 BASC2 PRS Behavioral Symptoms Index x total barriers .01 BASC2 PRS Adaptive Skills x total barriers .00 R 2 .09 .14 .23 F 4.28* .97 .96 .05 .09 .38 .94 Note . n = 43. Interaction terms were created by first centering each composite score variable to control for multicollinearity. *p < .05. ** p < .01. *** p < .001 . 93


Table 3 19. PIS total score, total barriers, and interaction terms as p redictors of 4 month adherence . 4 month mean adherence Variable Model 1 B Model 2 B Model 3 B Constant 4.48*** 4.36*** 5.00*** Child age 0.07* 0.07* 0.07* Total barriers .01 .23 Perceived Impairment Scale total score .00 .01 Perceived Impairment Scale x total barriers .01 R 2 .11 .11 .13 F 5.07* 1.66 1.49 .00 .02 .06 .98 Note. n = 45. The interaction term was created by first centering the PIS total score variable to control for multicollinearity. *p < .05. ** p < .01. *** p < .001. 94


Table 3 20. P SI T o tal S tress as non significant predictor of 4 month adherence . 4 month mean adherence Variable Model 1 B Model 2 B Constant 4.52* 4.53 Child age .07 .07 Parenting Stress Index Total Stress .00 R 2 .11 .11 F 5.12 2.50 .00 .00 Note. N = 42. *p < .05. ** p < .01. *** p < .001. 95


Table 3 21. PSI Total Stress, total barriers, and inter action terms as predictors of 4month adherence . 4 month mean adherence Variable Model 1 B Model 2 B Model 3 B Constant 4.52*** 4.51*** 5.29*** Child age 0.07* .07 .07 Parenting Stress Index Total Stress .00 .01 4 months total barriers .08 .01 Parenting Stress Index x total barriers .02 R 2 .11 .11 .14 F 5.12* 1.64 1.55 .00 .03 .03 1.83 Note. N = 45. The interaction term was created by first centering the PSI Total Stress variable to control for multicollinearity. *p < .05. ** p < .01. *** p < .001. 96


Table 3 22. Mean adherence at first follow up, percent recall, and mean understanding presented separately by feedback modality . Written feedback only Oral + written feedback Difference n M (SD) Observed range n M (SD) Observed range F (df) p Mean adherence at first follow up* 28 3.78 (.67) 2.3-5.0 21 4.01(.66) 2.8-5.0 1.41 (1) .23 Percent recall 2 months 12 .38 (.32) 16 .34 (.26) .10 (1) .75 4 months 27 .39 (.33) 18 .33 (.29) .41 (1) .53 Mean understanding* 2 months 12 4.83 (.25) 4.3-5.0 12 4.93 (.13) 4.5-4.7 1.75 (1) .20 4 months 27 4.83 (.36) 3.5-5.0 16 4.98 (.22) 4.7-5.0 2.18 (1) .15 * Range of possible scores for adherence and understanding variables is 1 – 5. Mean adherence and mean understanding were calculated by calculating the sum of adherence and understanding ratings, respectively, and dividing these sums by the total number of recommendations prescribed. 97


Table 3 23. Descriptive data for mean adherence by group and test of significance for group difference . Group A Group B Difference n M (SD) n M (SD) F (df) p Mean adherence across recommendation types 23 3.89 (.81) 22 3.72(.69) .55 (1) .46 98


Table 3 24. 4month adherence as nonsignificant predictor of BASC2PRS Externalizing Symptoms Index difference score . BASC2 PRS Externalizing Symptoms difference score Variable Model 1 B Model 2 B Model 3 B Constant 5.17 14.76* 8.55 Child age .24 .23 .33 BASC2 PRS time 1 Externalizing Symptoms .32 0.32* 4 months mean adherence 1.23 R 2 .01 .34 .36 F .34 6.56 4.42 .33 .01 12.64 .44 Note. N = 27. *p < .05. 99

PAGE 100

Table 3 25. 4month adherence as nonsignificant predictor of BASC2PRS Internalizing Symptoms Index difference score . BASC2 PRS Internalizing Symptoms difference score Variable Model 1 B Model 2 B Model 3 B Constant 14.54* 15.41 6.23 Child age 1.13** .89 1.04* BASC2 PRS time 1 Internalizing Symptoms 0.46*** 0.46*** 4 months mean adherence 1.89 R 2 .14 .49 .51 F 4.31 12.21 8.26 .35 .01 17.39 .67 Note. N = 27. *p < .05. ** p < .01. *** p < .001. 100

PAGE 101

Table 3 26. 4month adherence as nonsignificant predictor of BASC2PRS Behavioral Symptoms Index difference score . BASC2 PRS Behavioral Symptoms Index difference score Variable Model 1 B Model 2 B Model 3 B Constant 10.50* 11.09 4.75 Child age .71 .62 .73 BASC2 PRS time 1 Behavioral Symptoms Index 0.31** 0.31** 4 months mean adherence 1.28 R 2 .10 .40 .41 F 2.93 8.23** 5.52** .30 .01 12.27 .46 Note. N = 27. *p < .05. ** p < .01. 101

PAGE 102

Table 3 27. 4month adherence as nonsignificant predictor of BASC2PRS Adaptive Skills Index difference score . BASC2 PRS Adaptive Skills Index difference score Variable Model 1 B Model 2 B Model 3 B Constant 4.49 13.19* 18.67* Child age .34 .28 .39 BASC2 PRS Time 1 Adaptive Skills Index .25* .24* 4 months mean adherence 1.29 R 2 .05 .23 .25 F 1.29 3.64* 2.65 .18 .02 5.75 .75 Note. N = 27. *p < .05. 102

PAGE 103

Table 3 28. 4month adherence as nonsignificant predictor of Perceived Impairment Scale difference score . Perceived Impairment Scale difference score Variable Model 1 B Model 2 B Model 3 B Constant 5.63 13.59* 19.81* Child age .58 .62* .53 Perceived Impairment Scale Time 1 score .48*** .48*** 4 months mean adherence 1.42 R 2 .07 .34 .25 F 3.19 10.06*** 6.99** .18 .02 5.75 .75 Note. N = 27. ** p < .01. *** p < .001. 103

PAGE 104

CHAPTER 4 DISCUSSION Summary of Findings This study examined 1) predictors of parents’ and guardians’ adherence to recommendations following their child’s psychological evaluation, 2) the relationship between parental adherence and changes in ratings of child symptom severity and impairment, and 3) the relationship between feedback modality and subsequent recall of and adherence to recommendations. Among demographic factors, child age emerged as the only significant predictor of parental adherence. While this relationship was not found by MacNaughton and Rodrigue (2001) or Dreyer, O’Laughlin, Moore, and Milam (2010), it is likely that this is because these earlier study samples were limited to children aged 4 to 12 years and 5 to 13 years, respectively, while the present study included adolescen ts up to 17 years of age. The finding that older age is associated with lower rates of parental adherence is not surprising and likely attributable, at least in part, to the greater autonomy possessed by older youths as well as the greater number of factors competing for adolescents’ time (i.e., employment, socializing with peers, extracurricular activities, lengthy homework assignments). Additionally, school based interventions are likely more difficult to obtain for students in middle school and high sc hool where special education services are diminished due to a movement toward mainstreaming students. The role of child age may be examined further in future research by exploring differences in the relationship between adherence and barriers by developmental level. Corroborating the findings of MacNaughton and Rodrigue (2001), the present study found no relationship between parenting stress or child symptom 104

PAGE 105

severity/impairment at the time of the evaluation and subsequent parental adherence to recommendati ons. These findings are in contrast to those of Dreyer and colleagues (2010) who found a significant association between parenting stress and adherence in their sample of children referred to an AttentionDeficit/Hyperactivity Disorder (AD/HD) clinic. Th ese findings also differ from research examining predictors of adherence in pediatric health populations (Chisholm, Atkinson, Donaldson, Noyes, Payne, & Kelnar, 2006; Janicke, D.M., & Finney, J.W., 2003; Marhefka, Tepper, Brown, & Farley, 2006). One facto r that may be contributing to these discrepant findings is the heterogeneity of the samples obtained for the present study and for the MacNaughton and Rodrigue study as compared to the more specific populations (i.e., AD/HD, diabetes, asthma, HIV) examined in research demonstrating a significant relationship between parenting stress and adherence. Although the present study is not sufficiently powered to conduct this analysis, future research examining adherence in larger samples of heterogeneous clinical populations can examine differences in the relationship between parenting stress and adherence across different diagnostic groups. Understanding when parenting stress does or does not predict subsequent adherence can aid in the development of targeted int erventions to improve adherence. Additionally, rates of adherence or percent recall of recommendations did not differ significantly by feedback modality. Parents and guardians did not report higher rates of adherence when they received inperson oral feedback than when they received written feedback only. This was contrary to expectations, as individuals who receive oral feedback presumably have more opportunities to have their questions pertaining to the recommendations and the results of the evaluation addressed than 105

PAGE 106

individuals who simply receive a written report. A larger sample and randomized assignment to feedback modality is needed to better understand the role of feedback modality in predicting adherence rates. Additionally, the timing of feedback may also be an important factor in determining adherence. During the telephone interviews completed for this study, families often commented on their feelings of being overwhelmed immediately upon receiving the results of the evaluation, and these feelings may interfere with parents’ ability to fully process the information provided to them in the feedback session. Examining the possible impact of providing feedback at different time points (i.e., several days after submitting the written report ve rsus immediately after giving families the report; one feedback session versus two) is warranted. Furthermore, although there were no significant differences in adherence rates across providers for this sample, the person providing the feedback and the sty le of feedback used are also factors to be examined in future research as we continue to strive toward helping families to be more successful in following through with recommendations. Parental adherence was further examined as a predictor of change in chi ld symptom severity and impairment, with the expectation that higher rates of adherence to recommendations would correspond with greater improvements in symptoms and impairment as rated by parents and guardians four months after the receipt of the evaluati on results. Contrary to expectations, this relationship was not substantiated. Parents who reported higher rates of adherence did not also report a significant decrease in symptoms from the initial assessment to the 4month follow up as measured with the Behavior Assessment System for Children, Second Edition (BASC2106

PAGE 107

PRS), nor did they report a significant decrease in impairment as measured with the Perceived Impairment Scale (PIS). The failure to find a relationship between adherence and change in child symptom severity or impairment may be due in part to methodological limitations, including the small subset of participants who completed the follow up measures as well as the use of broadband measures to assess symptom severity and impairment. Broadband measures, such as the BASC2PRS and the PIS, measure problems across a broad range of domains and may not be specific enough to capture change in the particular areas of difficulty with which children presented and which are targeted by the interventions f amilies were encouraged to pursue. Further, there is a lack of standardized measures that have been validated to assess the impact of child difficulties on important areas of functioning in a meaningful way. The vast majority of measures that are available and widely utilized focus on symptom count and severity (i.e., intensity), without addressing the actual impact of these symptoms across functional domains (i.e., quality of life, academic performance, meaningful relationships, family cohesion). There is a need to develop measures that can be used to assess impairment within a heterogeneous population of individuals for whom impairment may reflect many different things. As in previous studies measuring predictors of adherence (i.e., MacNaughton & Rodri gue, 2001), adherence rates in the present study were once again the lowest for recommendations to psychological services. This was followed by non psychological professional consultation/services. The highest rates of adherence were reported for recommendations for physician consultation/services, with parents reporting nearly full adherence to these recommendations. Not surprisingly, perceived barriers once again 107

PAGE 108

emerged as the strongest predictor of lower rates of parental adherence to recommendations , corroborating previous findings (MacNaughton & Rodrigue, 2001; Nock & Photos, 2006; Patterson & Chamberlain, 1994). The average number of barriers reported per recommendation varied across recommendation types. The highest number of barriers per recomm endation was reported for recommendations for psychological services and active self help, while the lowest number of barriers per recommendation was noted for physician consultation/services, corresponding with the lowest and highest ratings of adherence, respectively. The strength of the negative relationship between barriers and adherence also varied across recommendation types, reaching statistical significance only for recommendations for psychological services, school based recommendations, and recom mendations for active self help. Medium effect sizes were noted in the relationship between barriers and adherence for psychological services and active self help recommendations, while a large effect size was obtained for school based recommendations. The higher effect size obtained for school based recommendations may reflect the fact that school based recommendations often require implementation by the child’s school and school personnel. While the parent plays an important role in sharing the evaluation results with the school and advocating for the child to ensure access to an appropriate education, schools ultimately have policies and procedures in place that may affect how quickly strategies designed to help address the child’s difficulties can be implemented. Specifically, Response to Intervention (RtI) is an approach that has largely replaced the traditional procedures for determining special education (ESE) eligibility. This system of assessment and intervention consists of four tiers, with hi gher 108

PAGE 109

level tiers representing more highly tailored and individualized interventions. Progressing through the RtI tiers can be timeconsuming and may prevent students with more subtle learning problems from receiving the types of ESE services that are some times included as school based recommendations in psychological reports. Furthermore, schools also have limited resources with which to serve large student populations. When parents are faced with barriers to adherence to school based recommendations, they may feel powerless to overcome these barriers, resulting from their own beliefs and assumptions or from the response they receive from the child’s school when they attempt to implement recommendations from an outside provider. The types of barriers m ost commonly reported by parents in the current study with regard to school based recommendations were access problems, which included reports of the school’s inability or unwillingness to implement interventions, accommodations, or other services recommended by the child’s psychologist. With regard to recommendations for psychological services, the relatively stronger relationship between barriers and adherence likely reflects the demandingness of following through with recommendations for psychological services, and families may be less able or willing to overcome barriers that interfere further with adherence. Adherence to recommendations for psychological services, such as participation in psychological interventions, can be taxing on a family’s time and financial resources, particularly when these services are not covered by insurance, and they require a level of commitment and effort over time that is not matched by the demandingness of adherence to recommendations for consultation and services with physicians and/or other nonpsychological professionals. In the case of physician consultation/services, 109

PAGE 110

appointments are typically no more frequent than once per month, with betweenvisit demands being limited to monitoring of symptoms and medication adm inistration. While other nonpsychological services, such as occupational, physical, and speech/language therapies often do require a weekly time commitment similar to that associated with psychological services, the role of the parent or guardian may be perceived as less arduous. Further, these non psychological professional services can often be provided within the school setting, reducing the impact of barriers commonly associated with accessing services in an outpatient setting. Similarly, recommendations that involve active self help can require substantial effort from the parent or guardian, often in the absence of external structure or support, thus perceived barriers may be particularly difficult to overcome. Notably, there was a different patte rn of reported barriers observed for these recommendations as compared to other types of recommendations. While access problems and competing time or scheduling demands were most frequently cited as barriers to adherence overall across recommendation types, the most frequent barriers to adherence noted for active self help recommendations were characteristics of the child, such as the child’s refusal to participate in self help strategies or activities or the severity of the child’s difficulties. Challen ges and Limitations There are several noteworthy challenges that were faced in conducting the present study as well as a number of important limitations that should be considered when examining these results. While the target number of 90 enrolled partici pants was achieved, the high rate of participants lost to follow up due in large part to challenges associated with contacting families by telephone resulted in a final sample that differed 110

PAGE 111

in certain ways from the initial sample. For instance, those parents and caregivers who completed follow up portions of the study were younger than parents and caregivers who did not participate in the study’s follow up interviews. Older caregivers may experience different types of barriers or they may follow through w ith recommendations at different rates than younger caregivers. Moreover, while the recruitment procedures for the current study resulted in a sample of participants that was diverse in terms of presenting problems and severity of symptoms, this sample ex cludes children and adolescents who have difficulties but have not been brought in for an assessment. Consequently, this sample may not be representative of all children who would benefit from a psychological evaluation. Moreover, unlike a previous study that examined adherence four weeks after the receipt of recommendations (MacNaughton & Rodrigue, 2001), the present study examined adherence 2 months and 4 months after families received the evaluation report. While this longer lag time was designed to gi ve families more time to complete recommendations and ultimately was intended to result in more accurate adherence ratings, the longer time period between the assessment and the follow up interviews may have also interfered with parents’ and caregivers’ ab ility to recall recommendations received as well as their level of adherence to these recommendations. Additionally, families were sometimes noted to have difficulty identifying the source of the recommendations they recalled, as their children may have participated in multiple evaluations across a variety of healthcare providers. The greater heterogeneity of the current study sample as compared to other previous studies (Dreyer et al, 2010; Human & Teglasi, 1993) is another noteworthy confound that could not be controlled for due to 111

PAGE 112

insufficient sample size. Participants varied considerably with regard to the providers and trainees conducting the evaluations, the number and thoroughness of the recommendations provided, and the amount of time lapsed between their evaluation and the receipt of the evaluation results. These factors may all have an impact on the degree to which families understand and follow through with recommendations they receive, as well as the types of recommendations to which they adhere. Another methodological limitation pertains to the nature of correlational research and retrospective data collection. When examining the relationship between perceived barriers and adherence, it is easy to assume that barriers influenced adherence. However, given that adherence ratings and barriers were obtained at the same time, it is also possible that parents and guardians were more likely to identify barriers retrospectively for recommendations with which they were non adherent than for recommendations they did complete, even if these differences in barriers were not actually present or perceived at the time that implementation was attempted. Furthermore, the independence of parental adherence to multiple recommendations is also questionable, pa rticularly when recommendations for psychological services are included, as therapists may help families to navigate other systems and can facilitate adherence to other types of recommendations as part of their therapeutic work. The possible influence of adherence to psychological recommendations on adherence to other recommendation types was not examined in this study but warrants further investigation. The use of single informants to obtain adherence ratings is an additional limitation of the present s tudy. Adherence ratings were not obtained from other 112

PAGE 113

caregivers or other collateral sources, such as medical and school records, due to feasibility and privacy concerns. Obtaining information from multiple sources may result in a more accurate assessment of the degree to which families follow through with recommendations. Similarly, the use of caregiver self report as the single source of adherence ratings is in itself a limitation, as the desire to be perceived positively by the interviewer or concerns about this information being shared with their evaluating psychologist may have influenced parents’ reports of adherence, resulting in a surprisingly high frequency of ratings of full adherence (55%). While the use of a 5point Likert scale was intended t o provide more variability in ratings of adherence and capture the fact that adherence is typically not an all or nothing construct, this scale is likely more appropriate for certain types of recommendations than for others, which may have contributed to d ifferences in ratings of adherence across recommendation types. For instance, full adherence to a recommendation for physician consultation/services may have been achieved with a one time visit to a physician’s office in which a possible medication trial was discussed, even if no medication trial was initiated. In contrast, full adherence to a recommendation for psychological intervention may have required attendance at weekly sessions and active engagement in therapy. It is easy to see how ratings of 5, representing full adherence, may be more easily assigned to the former recommendation than the latter, even in individuals who did initiate psychological intervention and who are participating in these services as recommended. Finally, statistical limit ations must also be highlighted, as these can highly influence the interpretability and generalizability of study results. Numerous separate analyses were conducted across the six aims of this study, without using alpha corrections to control 113

PAGE 114

for multiple analyses. Although Bonferroni corrections were applied for post hoc analyses when appropriate, significant p values that approach the conventional cutoff of .05 warrant replication to ensure that these results are not an artifact of multiple analyses. M oreover, because not all participants received recommendations of every type, analyses examining differences between recommendation types were conducted at the recommendation level, with each participant being represented multiple times in each analysis. Although controlling for mean adherence was designed to reduce the impact of lack of independence between individual adherence ratings, this is a noteworthy limitation that could only be addressed by using multilevel modeling techniques for which the present study was not sufficiently powered. Thus efforts were made to use additional analyses that did not violate independence assumptions to corroborate these other findings whenever possible. Clinical Implications and Future Directions The role of barriers as a consistent and prominent predictor of adherence to recommendations following psychological evaluations is one that cannot be ignored by psychologists. When examining results and creating a list of interventions and other strategies that are known to effectively address the child’s identified problems, the challenges to adherence that families are likely to face must be taken into consideration and used to help families prioritize the activities that are likely to have the most impact at the lowest cos t. The first step in addressing these barriers, however, is to be able to effectively identify them early on. Consequently, it is imperative that clinicians include a barriers assessment as part of their standard evaluation procedure (MacNaughton & Rodri gue, 2001). This information can then be used to structure the written report and 114

PAGE 115

oral feedback provided to families in such a way that 1) addresses reported barriers by providing alternatives and identifying available resources, and 2) emphasizes adherence to recommendations that are likely to be more feasible for families. One phase of identification should be at the point of the evaluation, in which a standardized barriers questionnaire can be incorporated into the evaluation or less formally as part of the clinical interview. As a next step, these barriers should be discussed jointly by the clinician and the family as part of the feedback session in order to determine whether there are resources that can be drawn upon for overcoming these challenges o r whether alternative interventions or strategies may be more appropriate for the child in light of these barriers. In some cases, barriers may be related to the parent or guardian’s attitudes or beliefs about the recommended interventions, and it is wise for these to be addressed early on to prevent them from interfering with adherence after the psychologist is no longer in contact with the family. Because barriers may not become evident to families until after they have attempted to follow through with recommendations, however, follow up telephone communication between the parent or guardian and the clinician several weeks after the report is received may be a good approach. This would allow the clinician an additional opportunity to intervene by helping families to identify additional resources of which they may not be aware, as well as by reprioritizing the list of recommendations in light of barriers that were encountered. If the ultimate goals of a psychological evaluation are to characterize the di fficulties exhibited by the child who is being assessed and to provide evidencebased recommendations that are likely to result in meaningful change 115

PAGE 116

for the child, then it follows that the identification and management of barriers to adherence is a necessary step toward meeting our clinical goals as psychologists. In order to justify the extra time and effort that would be required by the implementation of a formal barriers assessment and barriers intervention component in the evaluation and feedback proc edures, demonstrating the practical benefits of these added tasks in a measurable way will be necessary, particularly in a world of managed healthcare. This will require the development of a formalized barriers assessment tool, standardization of adherenc e measures, and more effective ways of demonstrating the impact of adherence on meaningful changes in children’s functioning. Although higher rates of adherence did not correspond with overall improvements in child functioning as measured by the broadband measures utilized in the present study, these findings may be an artifact of inadequate measures rather than a true indication that higher rates of adherence do not result in meaningful improvements in child functioning. Changes in scores on broadband measures that focus on symptom counts and frequencies to estimate severity do not appear to adequately capture the types of meaningful changes that interventions produce and that were described by families during their telephone interviews . Therefore, the d evelopment of more specific measures of impairment that show relations with adherence to recommendations and that can be administered to heterogeneous clinical groups i s imperative, as this would also help to justify allocating resources toward the goal of improving adherence. These measures should be focused on capturing real world indicators of impairment, such as the frequency of behavioral referrals in school or the presence/absence of meaningful peer relationships. The great 116

PAGE 117

challenge in developing s uch measures will be capturing functioning and impairment across a wide range of domains in a parsimonious manner. Another important consideration is the role of the psychologist as liaison, helping families to navigate the different organizations and syst ems within which they must function in order to address the difficulties identified in a psychological evaluation. In order to be most effective in this role as liaison, it is imperative that psychologists are continually educated and familiarized with these systems, understanding the procedures and challenges inherent to each system. For instance, in order for psychologists to provide meaningful and realistic school based recommendations to families, they must be familiar with the procedures used by spec ific school districts for determining eligibility for services. Additionally, using phrasing that is familiar to school personnel and that demonstrates understanding of how the school system functions is likely to result in greater receptivity by school personnel than using unfamiliar terminology or making recommendations that are not feasible within the system. By tailoring a report to its audiences and being cognizant of these audiences when developing our recommendations, we are likely to be more effec tive providers of psychological evaluation services. As an exploratory aim, this study examined the possible impact on adherence of contacting families for follow up several weeks after receipt of the evaluation results. To examine this, adherence ratings were compared across groups at the 4month follow up to determine whether higher rates of adherence were reported by participants who were contacted at 2 months versus those who did not receive the 2 month telephone contact. While group differences were minor and not statistically significant, it is important to 117

PAGE 118

note that improving adherence was not an intended function of the telephone interview, nor was the interview designed specifically to help families become more adherent. Anecdotally, however, qu alitative interview information obtained from parents and guardians suggests that reviewing recommendations with the principal investigator of this study prompted some families to revisit their child’s psychological evaluation results and consider pursuing services or self help activities they may not have explored initially following the evaluation. Additionally, families were given the opportunity to discuss barriers they faced and to consider possible strategies for overcoming these barriers. As the natural next step after learning more about predictors of adherence following a psychological evaluation, future research focused on improving adherence should consist of systematically examining both the feasibility and effectiveness of contacting families by telephone several weeks or months after the evaluation in order to review their adherence, identify barriers, and explore ways to overcome these barriers to increase adherence. For many families, it is not until a reevaluation takes place one or more years later that these recommendations and the barriers faced are revisited, and this translates into much time lost for children who may not be receiving the help they greatly need. This intervention can include a brief adherence and barriers assessment, as well as problem solving strategies and other interventions (i.e., motivational interviewing) to facilitate greater adherence. Determining the effectiveness of a standardized follow up telephone contact with families after an evaluation as one possible intervention for improving adherence is greatly encouraged. As is demonstrated in the present study, parents and guardians receive a diverse range of recommendations to address the problems identified in their child’s 118

PAGE 119

psychological evaluation. Successful completion of these recommendations requires access to and collaboration among a number of professionals and community organizations beyond the efforts of the individual parent and child. However, parents often find that navigating these systems effectiv ely and facilitating multidisciplinary cooperation can be a substantial challenge. Another area in which further research and program development is needed is in the establishment of systems and procedures designed to bridge communication between the diff erent entities that serve the diverse needs of children. Making services more accessible and less difficult for families to navigate would likely have a positive influence on adherence. 119

PAGE 120

APPENDIX DEMOGRAPHIC QUESTIONNAIRE Child Assessment Research Project Demographic Questionnaire Today’s Date: _________________________ For all questions on this form, please either write in or circle the correct information: Child’s name: ___________________________ Child’s date of birth : _____________________ Sex: male female Child’s current grade in school: ____________ Ethnicity: Asian American African American Caucasian Hispanic Other_____________ Assessment history: Has your child ever been ev aluated by a psychologist? Yes No If yes, please list below: Assessment Date Reason Psychologist/Clinic Name Please list any previous psychological or behavioral diagnoses your child has received: Treatment history: Has your child ever received treatment for psychological or behavioral problems? Yes No Treatment Reason Dates Has your child ever been on any prescription medications for psychological or behavioral problems? Yes No If yes, please complete the following: Medication Reason Dates Taken Has the child lived with you at least 50% of the time for the last 6 months? Yes No 120

PAGE 121

Demographic Questionnaire – Caregiver Information Your name : ______________________________ Date of birth : ______ _________________ Sex: male female Relationship to child: _________________ Cell # : __________________ Home #: __________________ Work #: ________________ _ __ Home Address: Ethnicity: Asian American African American Caucasian Hispanic Other_____________ Primary language s poken by you : _________________ Other l anguages spoken in the home: _________________ Marital S tatus: Complete the following for ALL PEOPLE living in child’s primary home , INCLUDING YOURSELF : Relationship to Chil d Age Highest grade completed Occupation YOU: Current yearly income of family whom this child LIVES WITH (circle one) : 3. Less than $5,000 5. $31,000 $40,000 4. $6,000 $10,000 6. $41,000 $50,000 5. $11,000 $20,000 7. $51,000 $60,000 6. $21,000 $30,000 8. $61,000 and above Please provide the name a nd contact information for one person who does not live with you and who m we can contact at follow up in the event that your p hone number or address has changed. Name:__________________________ Phone #___________ ____ ___ Relationship to you ____ ________ 1. Married 5. Living Together 5. Single 2. Divorced 6. Widowed 6. Separated 121

PAGE 122

LIST OF REFERENCES Abidin, R.R. (1995). Parenting stress index: Test manual . Charlotesville, VA: Pediatric Psychology Press. B arrett, P.M., Farrell, L., Pina, A.A., Peris, T.S., & Piacentini , J. (2008). Evidencebased psychosocial treatments for child and adolescent obsessivecompulsive disorder. Journal of Clinical Child & Adolescent Psychology, 37(1), 131 155. B rookmanFrazee, L., Gabayan, E.N., Haine, R.A., & Garland, A.F. (2008). Predicting frequency of treatment visits in community based youth psychotherapy. Psychological Services, 5 (2), 126138. C hambless , D.L., & Ollendick , T.H. (2001). Empirically supported psychological interventions: Controversies and evidence. Annual Review of Psychology, 52, 685716. C hambless, D.L., Sanderson, W., Shoham, V., Johnson, S., Pope, K., Crits Cristoph, P., Baker, M., Johnson, B., Woody, S., Sue, S., Beutler, L, Williams, D., & McCurry, S. (1996). An update on empirically validated therapies. The Clinical Psychologist, 49, 518. C hisholm, V., Atkinson, L., Donaldson, C., Noyes, K., Payne, A., & Kelnar, C. (2006). Predictors of treatment adherence in young children with type 1 diabetes. Journal of Advanced Nursing, 57(5), 482 493. C urry, K.T. & Hanson, W.E. (2010). National survey of psychologists’ test feedback training, supervision, and practice: A mixed methods study. Journal of Personality Assessment, 92 (4), 327336. Da vid Ferdon, C., & Kaslow, N.J. (2008). Evidencebased psychosocial treatments for child and adolescent depression. Journal of Clinical Child & Adolescent Psychology, 37(1), 62 104. D reyer, A.S., Laughlin, O., Moore, J., & Milam, Z. (2010). Parental adherence to clinical recommendations in an ADHD evaluation clinic. Journal of Clinical Psychology, 66(10), 11011120. E yberg, S.M., Nelson, M.M., & Boggs, S.R. (2008). Evidencebased psychosocial treatments for children and adolescents with disruptive behavior. Journal of Clinical Child & Adolescent Psychology, 37(1), 215237. Ga rcia, J.A., & Weisz, J.R. (2002). When youth mental health care stops: Therapeutic relationship problems and other reasons for ending youth outpatient treatment. Journal of Consulting and Clinical Psychology, 70(2), 439 443. 122

PAGE 123

Gass, C.S., & Brown, M.C. (1992). Neuropsychological test feedback to patients with brain dysfunction. Psychological Assessment, 4, 272 277. Geffk en , G.R., Keeley, M.L., Kellison, I., & Storch, E.A. (2006). Parental adherence to child psychologists’ recommendations from psychological testing. Professional Psychology: Research and Practice, 37(5), 499505. Ha yes, A.F. (n.d.) Indirect. SPSS, SAS, and Mplus Macros and Code. Retrieved on September 10 , 2011 from nd-m plusmacros-and-c ode.html. Hu man, M.T., & Teglasi, H. (1993). Parents’ satisfaction and compliance with recommendations following psychoeducational assessment of children. Journal of School Psychology, 31, 449 467. Jani cke, D.M., & Finney, J.W. (2003). Children’s primary health care services: social cognitive factors related to utilization. Journal of Pediatric Psychology, 28(8), 547558. Janz , N.K., & Becker, M.H. (1984). The health belief model: A decade later. Health Education Quarterly, 11(1), 1 47. Kaz din, A.E. (2000). Perceived barriers to treatment participation and treatment acceptability among antisocial children and their families. Journal of Child and Family Studies, 9(2), 157 174. Kaz din, A. E., Holland, L., & Crowley, M. (1997a). Family experience of barriers to treatment and premature termination from child therapy. Journal of Consulting and Clinical Psychology , 65, 453 – 463. Ka zdin, A.E., Holland, L., Crowley, M. & Breton, S. (1997b). Barriers to Treatment Participation Scale: Evaluation and validation in the context of child outpatient treatment. Journal of Child Psychology and Psychiatry, 38(8), 1051 1062. Kaz din, A.E., & Mazurick, J.L. (1994). Dropping out of child psychotherapy: Distinguishingearly and late dropouts over the course of treatment. Journal of Consulting and Clinical Psychology, 62, 1069 – 1074. Kaz din, A.E., & Wassell, G. (1999). Barriers to treatment participation and therapeutic change among children referred for conduct disorder. Journal of Clinical Child Psychology, 28, 160– 172. Kaz din, A.E., & Wassell , G. (2000). Predictors of barriers to treatment and therapeutic change in outpatient therapy for antisocial children and their families. Mental Health Services Research, 2 (1), 2740. 123

PAGE 124

Koroloff, N. M., Elliott, D.J., Koren, P.E., & Friesen, B.J. (1996 ). Linking low income families to children’s mental health services: An outcome study. Journal of Emotional and Behavioral Disorders, 4, 2 11. Kourany, R.F.C., Garber, J., & Tornusciolo, G. (1990). Improving first appointment attendance rates in chil d psychiatry outpatient clinics. Journal American Academy Child and Adolescent Psychiatry, 29, 657660. MacLean, L.M., Greenough, T., Jorgenson, V., & Couldwell, M. (1989). Getting through the front door: Improving initial appointment attendance at a mental health clinic. Canadian Journal of Community Mental Health, 8, 123 133. Maichenbaum, D., & Turk, D.C. (1987). Facilitating treatment adherence: A practitioner’s guidebook. New York: Plenum. Marhefka, S.L., Tepper, V.J., Brown, J.L., & Far ley, F.F. (2006) Caregiver psychosocial characteristics and children’s adherence to antiretroviral therapy. AIDS Patient Care and STDs, 20(6), 429437. Moore, T.R., & Symons, F.J. (2009). Adherence to behavioral and medical treatment recommendations by parents of children with Autism. Journal of Autism and Developmental Disorders, 39, 1173 1184. Moses, T. (2011). Parents’ conceptualization of adolescents’ mental health problems: Who adopts a psychiatric perspective and does it make a difference. Community Mental Health, 47, 6781. Nock, M.K. & Ferriter, C. (2005). Parent management of attendance and adherence in child and adolescent therapy: A conceptual and empirical review. Clinical Child and Family Psychology Review, 8(2), 149 166. No ck, M.K., Ferriter, C., & Holmberg, E. (2006). Parental beliefs about treatment credibility and effectiveness: Assessment and relation to subsequent treatment participation. Journal of Child and Family Studies, 16, 2738. Nock, M.K., & Photos, V. (2 006). Parent motivation to participate in treatment: Assessment and prediction of subsequent participation. Journal of Child and Family Studies, 15(3), 345 358. Ownby, R. L., & Wallbrown , F. (1986). Improving report writing in school psychology. In T. Kratochwill (Ed.), Advances in School Psychology, 5, 7 49. Pinto, A. (2003). Effect of treatment recommendation format on parental satisfaction and compliance. Dissertation Abstracts International: Section B: The Sciences and Engineering, 63(11 B), 5532. 124

PAGE 125

Pope, K.S. (1992). Responsibilities in providing psychological test feedback to clients. Psychological Assessment, 4 (3), 268 271. P reacher, K.J., & Hayes, A.F. (2008). Asymptomatic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behavior Research Methods, 40(3), 879891. P rochaska, J.O., DiClemente, C.C., & Norcross, J.C. (1992). In search of how people change: Applications to addictive behaviors. American Psychologist, 47(9), 11021114. R eynolds, C.R., & Kamphaus, K.W. (2004). Behavioral assessment system for children (2nd ed.). Circle Pines, MN: American Guidance Service. R ichardson, L.A. (2001). Seeking and obtaining mental health services: What do parents expect? Archives of Psychiatric Nursing, 15(5), 223231. R ogers, S.J., & Vismara, L.A. (2008). Evidencebased comprehensive treatments for early autism. Journal of Clinical Child & Adolescent Psychology, 37 (1), 8 38. R osenstock, I.M. (1974). Historical origins of the health belief model. Health Education Monographs, 2(4), 328 335. S herman, M.L., Barnum, D.D., BuhmanWiggs, A., & Nyberg, E. (2009). Clinical intake of child and adolescent consumers in a rural community mental health center: Does wait time predict attendance? Community Mental Health Journal, 45 , 78 84. S ilverman, W.K., Ortiz, C.D., Viswesvaran, C., Burns, B.J., Kolko, D.J., Putnam, F.W., & Amaya Jackson, L. (2008). Evidence based psychosocial treatments for children and adolescents exposed to traumatic events. Journal of Clinical Child & Adolescent Psychology, 37 (1), 156 183. S ilverman, W.K., Pina, A.A., & V iswesvaran, C. (2008). Evidence based psychosocial treatments for phobic and anxiety disorders in children and adolescents. Journal of Clinical Child and Adolescent Psychology, 37(1), 105130. T allent, N. (1993). Psychological report writing (4th ed.). Englewood Cliffs, NJ: Prentice Hall. T eagle, S.E. (2002). Parental problem recognition and child mental health service use. Mental Health Services Research, 4, 257 266. 125

PAGE 126

Tharinger, D.J., Hersh, B., Christopher, G.B., Finn, S.E., Wilkinson, A., & Tran, A. (2008). Assessment feedback with parents and preadolescent children: A collaborative approach. Professional Psychology: Research and Practice, 39(6), 600609. Waldron, H.B., & Turner, C.W. (2008). Evidencebased psychosocial treat ments for adolescent substance abuse. Journal of Clinical Child & Adolescent Psychology, 37(1), 238261. Wells, K.C., Epstein, J.N., Hinshaw, S.P., Conners, C.K., Klaric, J., Abikoff, H.B., et al. (2000). Parenting and family stress treatment outcomes in attention deficit hyperactivity disorder (ADHD): An empirical analysis in the MTA study. Journal of Abnormal Child Psychology, 28, 543 553. 126

PAGE 127

BIOGRAPHICAL SKETCH Eileen Matias Davis is originally from Hialeah, Florida where she graduated as class valedictorian in 2002 from Hialeah High School. In 2006 she earned a Bachelor of Arts degree in psychology from Harvard University in Cambridge, Massachusetts. During he r time as an undergraduate, Eileen worked at Harvard’s Laboratory for Developmental Studies and Clinical Research Laboratory. Upon graduating, Eileen took a position as a first grade teacher at Meadowlane Elementary School, a Title I primary school in South Florida. In 2008, s he entered the Clinical and Health Psychology doctoral program at the University of Florida with a concentration in clinical child psychology. As a graduate student at UF, she has worked under the mentorship of Brenda A. Wiens, Ph.D ., in the National Rural Behavioral Health Center. Eileen earned a Master of Science degree in 2010. In 20122013 she completed a one year internship at Texas Children’s Hospital in Houston, Texas in the areas of clinical child and pediatric psychology. She then went on to receive her Doctor of Philosophy degree in August 2014 after completing her doctoral dissertation. Eileen’s clinical training experiences range from psychological, neuropsychological, and developmental assessment of children and adolescents to parent management training, cognitive behavioral therapy, exposure with response prevention, and group therapy with youths and their families in the areas of clinical child psychology and pediatric psychology. Her research focuses on identifying predictors of adherence following child and adolescent assessments. 127