CMcNabandGMeadows - PDF document

CMcNabandGMeadowsIntroductionAssessingm
CMcNabandGMeadowsIntroductionAssessingmentalhealthcareneedsinprimarycareOneinthreepatientsattendingAustraliangeneralpracticehavecommonformsofdepressive,anxietyorsomatoformdisorders,and10–25%havedepres-sivedisorderssevereenoughtojustifyevidence-basedTreatmentofdepression,incontrast,accountsforlessthan4%ofconsultationsinAustraliangeneralpractice.2,3Reducingthisdisparitymaybehelpedbydevelopingstructuredscreeningandassessmentinstrumentssuitableforprimarycare,andeffortsinthisdirectionhavetypicallyinvolveddevelopmentofbriefsymptomscales.Theutilityofsymptomdetectionislimited,how-ever,unlessitisthenfollowedbythedevelopmentofasharedunderstandingofneeds,basedonwhichanappropriateclinicalinterventioncanbedelivered.Giventhatsymptommeasuresdonotexaminethedegreetowhichparticularmentalhealthneedsareperceived,theassessmentofperceivedneedmayfacilitateboththedevelopmentofthissharedunderstandingandtheprovisionofsuitableclinicalExistingneedsassessmenttoolsareinappropriateforgeneralroutineusewithinprimarycare.Suchmeasures,whichincludetheCardinalNeedsSched-ule,theMedicalResearchCouncil(MRC)NeedsforCareAssessmentSchedule,theCamberwellAssess-mentofNeed,andtheBangorAssessmentofNeedProfiletypicallyassumepre-existingmentalhealthdifficulties,and/orarecompletedbycliniciansratherthanconsumers.Theneedsassessedwithinthemincludebothmentalhealthandrelatedneeds,suchasdomesticskillsandaccesstotransportandamen-ities;giventhediversityofpotentialreportedneeds,thesemeasuresarelengthy(e.g.CardinalNeedsSched-ule:62min;CamberwellAssessmentofNeed:25min;BangorAssessmentofNeedProfile:20min).TheTwo-WayCommunicationChecklist(2-COM)isarecentlydevelopedmeasureofperceivedneedinpatientswithschizophreniathatrepresentsasub-stantialimprovementintermsofbrevity;however,itcontinuestotakeapproximately13minutesto11,12EvidencefromtheAustralianNationalSurveyofMentalHealthandWellbeing(NSMHWB),thelargestepidemiologicalsurveyconductedinAustraliatodate,hasprovidedsomeindicationsastothelikelyutilityofdirectenquiryintoperceivedneedformentalhealthcareinthegeneralpopulation.Theneedsassessmentinstrumentdevelopedforthatsurvey,thePerceivedNeedforCareQuestionnaire(PNCQ),assessedfivedomainsofperceivedneed:information,medication,counsellingorpsycho-therapy,socialinterventionsandskillstraining.Itcategorisedeachoftheseasbeingatalevelofnoneed,unmetneed,partiallymetneed,orfullymetneed;whereneedwasunmetorpartiallymet,itassessedbarrierstocare.Formalpsychometriceval-uationpriortoadoptionforthesurveysupportedconstructvalidityandshowedittohaveacceptableThePNCQiscurrentlybeingusedwithinseveralservicesresearchprojectsinAustralia,NorthAmerica,andAfrica,eitherasacomputer-assistedintervieworinitspaperformusedbytrainedinterviewers.Itisabriefandsimplemeasureinitscomputer-assistedform,typicallytakinglessthantwominutestoThepaperformofthePNCQ,incon-trast,isa17-pageinstrumentandrequiresconsider-abletraininginadministeringtheskiprules;itsroutineuseinprimarycaresettingsisthereforeBeyondblue,theAustraliannationaldepressioninitiative,commissionedthedevelopmentofabriefversionofthePNCQaspartofitssubstantialpro-grammeofresearchintoment

alhealthinprimarycare.Theusewithinprimar
alhealthinprimarycare.Theusewithinprimarycarescreeningandassessmentofabriefperceivedmentalhealthneedmeasure,inadditiontosymptomscales,waspro-posedasameanstoassistintheroutinemonitoringofclinicalcourseandserviceresponseinbothre-searchandclinicalsettings,particularlywithrespecttoexamininglevelsofmetandunmetneedandbarrierstocare.ThispaperreportsthedevelopmentoftheGeneral-practiceUsers’Perceived-needInventory(‘GUPI’),whichrepresentsabriefpaperformofthePNCQ.AstudyisreportedbasedinprimarycarethatprovidessomeempiricalassessmentofthePNCQintermsofuser-centredness,feasibility,utility,andpsychometricproperties,asguidedbypublishedcriteriaforeval-uatingsuchmeasures.WhiletheGUPIisnotcon-sideredsolelyasascreenforpsychiatricproblems,theutilityofthemeasurewouldbesupportedbyevidenceofatleastsomeassociationbetweentheGUPIcategoriesandthepresenceofmentalhealthproblems(criterionvalidity);test–retestreliabilityisalsoconsidered.TheGeneral-practiceUsers’Perceived-needInventory(‘GUPI’)DevelopmentThedevelopmentphaseinvolvedtriallingaseriesofdraftversionsinconsultationwithgeneralprac-titioners(GPs),mentalhealthcareprofessionals,andprimarymentalhealthcareconsumers,resultingintheGUPI(seeAppendix1).TheGUPIenquiresintodomainsofperceivedneedcommonwiththoseassessedbythePNCQ,andalsointobarrierstoreceivingcarewhereaneedisidentified.Thispaperconcentratesonpropertiesoftheperceivedneedcomponentofthemeasure.StudyprocedureandsettingThreegeneralpracticesinMelbournetookpartinthestudy.Formsoutliningthestudyweredisplayedatreception,andreceptionistsinvitedconsecutiveattendeestoparticipateinthestudy.ConsenterswereapproachedbyresearchersinwaitingroomsbeforetheirGPconsultation;questionnaireswereadministeredinquietroomswithinthegeneralpractice,inthepresenceoftheresearcher.ThestudywasapprovedbytheUniversityofMelbourneHu-manResearchEthicsCommittee.Attime1,participantswereadministeredabatteryofpencil-and-papertests,including:ademographicsquestionnaire(assessingage,gender,employmentstatus,occupation,highereducationallevel,maritalstatus,andnumberofafeedbackquestionnaire,exploringconsumerperceptionsoftheGUPI’sacceptabilityandutilitymeasuresoflikelymentalhealthproblemschosenforbrevityandpracticality,inordertoexaminecriterionvaliditytheSomaticandPsychologicalHealthReport(‘SPHERE’),a12-itemquestionnaireassessingpsychologicalandsomaticsymptomsindicatingcommonmentaldisordersingeneralpracticethreeitemsoftheShort-FormHealthSurvey(‘SF-8’),canvassinggeneralhealth(scoredonasix-pointLikertscale),androlelimitationsresult-ingfromphysicalandemotionalhealthprob-lems(scoredonfive-pointLikertscales)theGUPI.Anaverageofsevendayslater(time2)(standarddeviation(sd)=0.74,range=6–10),participantsweretelephonedforadministrationofthetime2Concurrentcriterion-relatedvaliditywasexam-inedusingSPHERE‘caseness’,andcontinuousscoresonthethreeitemstakenfromtheSF-8HealthSurvey.Test–retestreliabilitywasexaminedusingGUPIresponsesattimes1and2.ResultsOne-hundredandninety-sevenGPattendeeswereapproached,and122agreedtoparticipateinthestudy(62%);83oftheseparticipantswereavailableforretest.Participants’meanagewas45years(sd=16.12),a

nd77%ofthesamplewasfemale.Themodalemploy
nd77%ofthesamplewasfemale.Themodalemploymentstatuswasthatofpensioner(33.6%ofthesample),followedbyworkingfull-time(24.6%).UtilityoftheGUPI:theresultsfromthefeedbackquestionnaireThelargemajorityofparticipantsratedtheGUPIaseasytounderstand(92.6%ofrespondersresponded‘agree’or‘stronglyagree’withthisstatement),andeasytocomplete(95.3%responded‘agree’or‘stronglyagree’).Amajoritydescribeditasusefulforthemtocomplete(62.1%responded‘agree’or‘stronglyagree’).ParticipantsgenerallyregardedtheGUPIaspossiblyhelpfulincommunicatingconcernstheymighthavetoGPs(45.2%responded‘agree’or‘stronglyagree’,27.9%responded‘neitheragreenordisagree’),andasimilarnumberreportedthattheywouldbepreparedtocompletetheGUPIeverytimetheyvisitaGP(43.2%responding‘agree’or‘stronglyagree’,26%reporting‘neitheragreenorDescriptivestatisticsofGUPIresponsesFifty-ninepercentofparticipantsindicatedsomeformofneedontheGUPIateithertime1ortime2.Chi-squareanalysisdidnotrevealanyassociationbetweendemographicvariablesandperceivedneed.OfparticipantswhoreportedneedontheGUPI,themodalbarrierreportedtoaccessingmeansofmeet-ingneedwasapreferencetomanageoneself(38.3%attime1).Attime1,reportingofperceivedneedoneachitemwasasfollows:item1,50%;item2,45.1%;item3,36.9%;item4,11.5%;anditem5,25.4%.CMcNabandGMeadowsReliabilityandvalidityoftheGUPIStabilityofreportingofneedFifty-ninepercentofthesamplereportedneedattime1,while35.2%reportedneedattime2,astatisticallysignificantchange(=83,McNemartest).Thosewhoreportedneedattime1andtime2weresignificantlymorelikelytoreportaneedformedicationattime1thanthosewhoreportedneedattime1andnotattime2(90.7%vs.(1)=12.12,0.001).Testsfordifferenceinreportingofperceivedneedacrosstime1andtime2werenon-significantforfourofthefiveGUPIitems,suggestingstabilityintheseitemsmeasures0.55;counselling:1.00;socialinterventions:1.00;skillstraining:0.73).Responsestothemedicationitemweresignificantlydifferentovertime(0.001);participants’report-ingofneedwithrespecttomedicationfellfrom45.0%attime1to37.3%attime2.SPHEREresultsandcriterionvalidationReportofbothsomaticandpsychologicalsymp-tomsbeyondthresholdontheSPHERE(‘level1caseness’)hasthehighestlevelofoveralleffi-17,19thispaperthereforeconsidersonlySPHERElevel1caseness.Table1demonstratesthesensitivityandspecificityofeachitem,andtheGUPIasawhole.Items1,2,and3appearhighinsensitivityandmoderateinspecificity;therefore,thosewhomeetcriteriaforlevel1casenessaccord-ingtotheSPHEREarehighlylikelytoreportaneedonitems1,2,or3,andthosewhodonotreachthiscasenessthresholdontheSPHEREarequiteunlikelytoreportaneedontheseGUPIitems.Items4and5appearlowinsensitivityandhighinspecificity;hence,thereappearsalowprobabilitythatindivid-ualswillreportperceivedneedontheGUPI,givenSPHEREcaseness,butahighprobabilitythatpartici-pantswillnotreportGUPIneediftheydonotreportSPHEREcaseness.Items4and5,therefore,appeartotargetindividualswhodonotachieveSPHEREthreshold.OverallefficiencyfortheGUPIis63.1%withrespecttoSPHERElevel1caseness.Multivariateanalysesofvariancerevealedsignifi-cantrelationshipsbetweentime1perceivedneedandSFHealthSurveyscores,withthosereportingneedscoringhigheronthegeneral((1,

116)=13.09,0.001),somatic((1,116)=6.03,0
116)=13.09,0.001),somatic((1,116)=6.03,0.05),andpsychological((1,116)=34.75,0.001)healthLogisticregressionwasconductedtoexaminethedegreetowhichGUPIitems,singlyandincombin-ation,predicted‘level1’SPHEREcaseness.NosingleitemsignificantlypredictedSPHEREcaseness.Prin-cipalcomponentsanalysissuggestedafirstfactoroftheitems1,2,and3,andasecondfactoroftheremainingtwoitems.Whenitemsloadingonthefirstandsecondfactorsrespectivelywereenteredintoasequentiallogisticregression,thefirstblocksignificantlypredictedSPHERElevel1caseness(3)=23.96,0.001);thesecondblockdidnotcontributesignificantlyabovethis((2)=0.66,TheseanalysessuggestedthatthefirstthreeitemsoftheGUPIalonemaybevalid.UsingthisshortformoftheGUPI(‘GUPI-SF’),comparisonswithcriterionvariableswerelargelyunchanged.ScoresontheSF-HealthitemswereassociatedwithGUPI-SFperceivedneed(item1:(1,116)=14.04,0.001;item2:(1,116)=6.64,0.05;item3:(1,116)=0.001).‘Level1’SPHEREcasenesswasrelatedtoperceivedneedontheGUPI-SF((1)=15.05,Fisher’s0.001,oddsratio=4.92,confi-denceintervals(CIs)=2.13–11.36);participantsreportingperceivedneedontheGUPI-SFweremorelikelytoreportlevel1SPHEREcasenessthanthosewithoutneed.Minimaldifferencesinsensi-tivityandspecificitybetweentheGUPIandtheGUPI-SFemerged(GUPI-SFsensitivity:0.78,CIs0.64–0.88;specificity:0.58,CIs0.47–0.69).OverallefficiencyscoresweremarginallyhigherfortheGUPI-SFforSPHERElevel1caseness(65.57%).Table1SensitivityandspecificityoftheGUPI,item-by-itemandoverallItem1Item2Item3Item4Item5OverallLevel1casenessSensitivity(CI)0.76(0.62–0.86)(0.46–0.73)(0.27–0.55)Specificity(CI)0.65(0.66–0.85)(0.73–0.90)TheGeneral-practiceUsers’Perceived-needInventory(‘GUPI’)DiscussionUtilityoftheGUPIParticipantsoverwhelminglyfoundthequestion-naireeasytounderstandandcomplete.Amajorityreporteditasusefulandpotentiallyhelpfulincom-municatingconcerns.Fortheminoritywhodidnot,thebrevityandclarityoftheinstrumentmeanthatitpresentslittleinthewayofresponseburden.ItispossiblethattheframingofthepresentationoftheGUPIasaresearchmeasureinthisinstancediminisheditsfacevalidity,andthatpatientsmaybemorelikelytoseeitasausefultoolwereitpresentedwithinthecontextofnormalclinicalcare.LevelsofreportedneedontheGUPILevelsofreportedperceivedneedontheGUPIappeartobemuchhigherthanthosegenerallydemonstratedinepidemiologicalsurveys(59%ontheGUPI,versus13.8%intheAustralianNationalSurveyforMentalHealthandWellbeing;19.4%intheUSNationalCo-MorbiditySurvey;11.7%inanOntariostudy;22.4%intheChristchurchpsychi-atricepidemiologystudy;7.3%inaFinnishnationalwithcross-nationaldifferencesbeinginpartatleastexplainedbydifferencesininstrumen-Thathigherlevelsofperceivedneedwerereportedinthisstudyisunsurprising,asassessmentofperceivedneedislikelytoyieldhigherpro-portionsofneedwhenindividualsarehelpseeking,astheyaretosomedegreewhentheyattendgeneralpractice.Onlyonestudytodatehasexaminedmentalhealthneedingeneralpracticeattendees,butitisnotcomparablewiththisstudyasitdefined‘need’accordingtoobjectivesymptomatologyTimetrendsinGUPIresponsesReliabilityanalysesgenerallysuggestedthatpartici-pantswerelesslikelytoreportmentalhealthneedsattime2thantime1.Thismightb

eregressiontowardthemean.Alternatively,t
eregressiontowardthemean.Alternatively,theGPconsultationmayhaveservedthroughprovisionofinformationoraprofessionalopiniontoeitherallowtheneedtobemetortoreassurethepatientsothatneedwasnolongerperceived.CompletionoftheGUPImightinsomecaseshavepromptedpatientstoraisetheirmentalhealthconcernswiththeGP.Perceivedneedformedicationwasmorefre-quentlyendorsedforthosewhoseperceivedneedswerestablethanthosewhoseneedswerenotstable.ThissuggeststhatthoseconsistentlyreportingneedontheGUPIovertimemaybeadifferentsubsamplethanthosewhodonot,intermsofduration,severityofpathologyand/ordistress.Onesubsamplereport-ingneedattime1mayhaveacontinuing,stableneed,andperceivetheirdifficultiestobemore‘bio-logical’or‘severe’;theothermayhaveaperceivedneedthatislesssevereand,therefore,morelikelytorespondtoreassuranceorbrieftreatment,ortonaturallyremitovertime.ValidityoftheGUPICriterion-relatedconcurrentvalidityoftheGUPIwasgenerallysupported.Perceivedneedwasassociatedwithpoorergeneralhealthandagreaterlevelofdisabilityduetobothphysicalandemotionaldiffi-culties,asassessedbytheitemstakenfromtheSF-8.ParticipantswhoreportedneedontheGUPIgener-allydemonstratedlevel1casenessontheSPHERE.Thissupportstheappropriatenessofassessingper-ceivedneedwithinprimarycare,notonlythroughthevalueofprovidinganopportunityforGPstoraisementalhealthconcernsintheconsultation,butalsobecauseofthestronglinkbetweenperceivedneedandtheseinstrumentsasproxiesforcaseness.AgreementbetweendifferentitemsovertimewasparticularlyhighbetweenthefirstthreeitemsoftheGUPI,tappinginformationregardingemotionalproblems,medicationforemotionalproblems,andcounselling;andbetweenthefinaltwoitems,tap-pingpracticalissuesandsocialskills.Intuitively,thefirstthreeitemsoftheGUPItapamore‘psychologi-cal’componentofmentalhealthcare(i.e.recog-nitionofpsychologicaldifficultiesandpursuitofdifferenttreatmentoptions);thefinaltwoitemsfocusonmoresociallyorientedfeaturesofneed.Thishigherlevelofagreementbetweenthefirstthreeitemssupportsthisintuition.McNemar’sstat-istic,factoranalysis,logisticregression,andsensi-tivityandspecificityanalysisfurthersuggestedthattheGUPImayperformadequatelyinapsycho-metricsense,evenifitems4and5aredeleted;psychometricpropertiesoftheGUPI-SFweregener-allysimilartotheGUPI.Itappears,therefore,thatareformulatedGUPI,containingonlythefirstthreeitems,mayhaveadvantagesofbrevityofcom-pletiontimeandsimplicityofadministration,with-outsacrificingpsychometricqualities.However,itisimportanttobearinmindthatneedssampledbyitems4and5willbeclinicallysignificantforsomepeople,andthattodeletetheseitemsmayreducetheinstrument’scontentvalidity.Thebrieferformmaybeconsideredparticularlysuitableinpackagesofinstrumentationwheresymptomorfunctionalquestionsarealsobeingincluded,buteveninitsCMcNabandGMeadows‘full’form,theGUPIisashortinstrumentforpatientstocomplete.Whilesensitivityishigh,specificityisonlymod-erate.Mostindividualswhomeetthethresholdforpsychiatriccasenessreportaperceivedmentalhealthneed,butasubstantialnumberoftheindi-vidualswhodonotmeetpsychiatriccasenessdoreportaneed.However,wewouldsuggestthatthesefeaturesoftheGUPIdonotrenderitredundant.I

tisarguablymostimportanttoensurehigherle
tisarguablymostimportanttoensurehigherlevelsofsensitivity–tobeinclusive,andtoreduce‘falsenegatives’–ratherthanspecificityintheinstanceofassessmentofmentalhealthneeds.Thismayparticularlybethecaseinthegeneralpracticecon-text,giventhattruepositivesmayattimesbepar-ticularlydifficulttodetect.Recentfindingssuggestthathigherlevelsofinsightintotheexistenceofamentalhealthproblemareassociatedwithamoreaccuratereportingofneed.Giventhatthosewithlowerinsightarepresumablylesslikelytodisclosesymptomsaswellastoreportneed,detectionofdifficultiesinthiscaseisleftentirelytotheprac-titioner.Whilementalhealthprofessionalshavebeenfoundtoidentifyneedsevenintheabsenceoftheirpresentationbythepatient,thismaybeamorechallengingataskfortheGP,giventheoftentime-limitedandpressurednatureofthepatient–professionalrelationship.Itishenceimportantthatanymeasureofperceivedmentalhealthneedishighinsensitivity,andtheGUPIfulfilsthisTheGUPI’smoderatespecificitymightbeseenasproblematiciftheaimofassessingperceivedneedisitsuseasaproxyforpsychiatriccaseness.Theidentificationofperceivedneed,however,goesbe-yondthis.Perceivedneedsassessmentmayallowfacilitationofthemeetingofpatientneeds,arguablyatleastpartoftheprofessional–patientrelationship,evenifitisnottheprofessional(inthiscase,theGP)whodirectlymeetstheneed.Further,perceivedmentalhealthneedidentificationmayassistintheearlyidentificationofsubthresholddifficulties,whichifaccompaniedbyatimelyinterventionmaypre-venttheemergenceofmoreseriousmorbidity.LimitationsofthestudyThisstudyhassomeacknowledgedmethodologicallimitations.Firstly,itwasnotpossibletodeterminewhetherstudyrefusersdifferedfromparticipants.Relatedly,completionratewasonlymoderate,with62%ofthoseapproachedagreeingtoparticipate.Itispossiblethatrefusalwasaresultoftheveryfactofexperiencingaperceivedneedformentalhealth-care,and/orobjectivesymptomatology.However,thosewhoexperiencedmentalhealthdifficultiesmayhavebeenmorelikelytoparticipatethanthosewhodidnot,althoughthisappearsunlikelyasthestudywasnotintroducedasresearchspecifictomentalhealthissues.Ineithercase,thestudy’sfindingsmaybelimitedingeneralisability.ItispossiblethatcompletionratesoftheGUPIwouldbehigherifthiswereframedasanaspectofroutineclinicalcareratherthanasaresearchproject;futureresearchfocusingontheuseoftheGUPIaspartofnormalclinicalpracticemayservetofurtherclarifythegeneralisabilityofthesefindings.Additionally,theGUPI’sadministrationdifferedovertime:inpersonattime1,andbytelephoneattime2.Whilethisprocedureprobablysecuredagreatersamplesizeatfollow-up,itmayhaveledtoclientsbeingmorereluctanttodiscloseperceivedneedverbally,incontrasttothelesspersonalques-tionnaireformatoftime1.Briefcriterionvalidationmeasureswereselectedtoreduceresponseburden;theappropriatenessofthesemeasuresasproxiesforfullpsychiatricinterviewsisnotsettled.Astest–retestreliabilityanalysisfocusedontheGUPI,weareunabletodeterminewhetherreductioninperceivedneedreportedontheGUPIovertimeisassociatedwithareductionin‘actual’needoneithertheSPHEREortheSF.Giventhatthisstudyisaprimarycarestudy,italsodoesnotallowexaminationofperceivedmentalhealthneedinthosepeoplewhodonot

presenttogeneralpractice,approximately20
presenttogeneralpractice,approximately20%inacalendaryear.WehavenotestablishedwhetheradministrationoftheGUPIresultsinconsumersraisingissueswiththeirGPandtherebyhavingneedsmet,orhavingperceivedneedsextinguishedviareassurance.Finally,samplesizeconstrainedusfromexaminingthedegreetowhichsensitivityandspecificityoftheGUPIdiffersacrossconsistentandinconsistentre-sponders;itispossiblethatthemeasure’ssensitivityandspecificitywouldbegreaterinasampleofconsistentresponders.Ifneedsassessmentresultsinnoimprovementinthepatient’scareoroutcome,itsutilitywouldseemquestionable.ThemethodologyofthisstudydidnotallowexaminationofthisissuewithrespecttotheGUPI,butexistingevidencesuggeststhatpro-vidinghealthprofessionalswithinformationre-gardingapatient’sperceivedneedmayimprovepatientoutcome,oratleastbenefittheclinicalrela-tionship.Twouncontrolledtrialshavefoundpro-vidinghealthprofessionalswithinformationaboutpatients’perceivedneedsimprovedpatientout-27,28Assessmentofperceivedneedhasalsobeenassociatedwithanimprovementinpatient-reportedqualityofpatient–doctorcommunication,andchangesinpatientmanagement.Inonerecentrandomisedcontrolledtrial,feedbackofstandard-isedneedsassessmentdidnotimprovetheoutcomeTheGeneral-practiceUsers’Perceived-needInventory(‘GUPI’)ofolderpeopleattendingpsychiatricdayhospi-afurtherclusterrandomisedcontrolledtrialinadultswithseverementalillnesssimilarlyfoundnoclinicalimprovementbutdiddetectanincreaseinpatientlevelsofsatisfaction.Inbothofthesestudies,thecomparisongroupincludedpatientswhoreceivedregularneedsassessmentconductedbymultidisciplinaryhealthprofessionals,togetherwiththepatientandtheirfamily;thismayhavelimitedtheabilitytodetectbetween-groupdiffer-SuggestionsforfutureresearchWesuggestthatfurtherresearchontheGUPIisdesirabletobuildonthefindingsofthisstudy.AstudyinageneralpopulationsettingwithoutGPconsultationinterventionmaymoreadequatelyestimatereliability.Morecomplexstudydesignsinprimarycarecouldestablishtheutilityoftheinstru-mentinclinicalplanninganddecisionmaking.SuchdesignscouldalsoexplorethekeyquestionofthepossibleroleoftheGUPIininfluencingthetherapeuticcontentandoutcomeofconsultations,andexaminemorefine-grainedissuessuchaswhereandhowitcanbebestusedwithinconsultation.CurrentresearchutilisingtheGUPImaybeabletoprovidefurtherinformationonitsperformance,particularlywhenusedinparallelwithothersymp-tomscalesthanthoseusedinthisstudy.TheGUPIisagenerallywell-receivedmeasure.Con-currentcriterion-relatedvaliditywasstronglysup-ported.Perceivedneedsratesreducedoverabriefintervaloftime.ThismighthavebeenduetotheroleofaninterveningGPconsultationinreducingperceivedneedbyallowinganopportunityforreassurance;GUPIadministrationmayhavecon-tributedtothis.Fromthepointofviewofpsycho-metrics,theuseofaGUPIshorterformwithonlythree‘psychological/psychiatric’itemsreceivedIneitherfive-itemorthree-itemforms,theGUPIgoesasubstantialwaytowardsmeetingrelevantevaluativecriteriaassetoutinEvansetal’sMentalHealthNeedsAssessmentCriticalAppraisalCheck-Onthebasisofthesefindings,theuseoftheGUPIissuggestedasapotentiallyusefulmeasurewithinprimarycareresearchsettings.Suchfurthe

ruse,particularlyincombinationwithsympto
ruse,particularlyincombinationwithsymptommeasures,wouldallowfurtherexplorationofitspsychometricproperties.Meanwhile,practitionersmayconsidertargetedclinicaluseoftheGUPI,andfurtherresearchmaysupportmoreextensiveadop-tionwithinclinicalsettings.ThefullGUPIisbriefenoughtobeacceptableincombinationwithshortsymptommeasures,andtheuseoftheGUPI-SFcanbeconsideredwhereextremebrevityisdesired.Thisprojectwasfundedby‘Beyondblue:thenationaldepressioninitiative’.ThanksareextendedtoIreneBobevskiforparticipationindraftingoftheinstru-ment,toIreneBobevski,GillianPlant,andDaniGoldfortheirinvolvementindatacollection;tothegeneralpractitionersfromtheNorthWestdivisionwhotookpartindiscussionssupportingdevelop-mentandwhofacilitatedrecruitment,particularlyDrJohnHodgson,DrBobLong,andDrJohnStanton;andfinallytotheconsumerswhogenerouslygaveoftheirtimetotakepart.1HickieIB,DavenportTA,NaismithSLetal.SPHERE:Anationaldepressionproject.MedicalJournalof2Bridges-WebbC,BrittH,MilesDAetal.MorbidityandtreatmentingeneralpracticeinAustralia1990–MedicalJournalofAustralia3BrittHandMillerGC.TheBEACHstudyofgeneralMedicalJournalofAustralia4GoldbergDandWilliamsP.AUser’sGuidetotheGeneralHealthQuestionnaire.Windsor:NFER-Nelson,5HickieIB,DavenportTAandRicciCS.ScreeningfordepressioningeneralpracticeandrelatedmedicalMedicalJournalofAustralia6MarshallM,HoggLI,GathDHetal.TheCardinalNeedsSchedule:amodifiedversionoftheMRCNeedsforCareAssessmentSchedule.7BebbingtonPE,BrewinCR,MarsdenLetalMeasuringtheneedforpsychiatrictreatmentinthegeneralpopulation:ThecommunityversionoftheMRCNeedsforCareAssessment.8BrewinCR,WingJK,MangenSPetal.Principlesandpracticeofmeasuringneedinthelong-termmentallyill:theMRCNeedsforCareAssessment.PsychologicalMedicine9PhelanM,SladeM,ThornicroftGetal.TheCamberwellAssessmentofNeed:thevalidityandreliabilityofaninstrumenttoassesstheneedsofpeoplewithseverementalillness.BritishJournalof10CarterMF,CrosbyC,GeertshuisSetal.Developingreliabilityinclient-centredmentalhealthneedsJournalofMentalHealthCMcNabandGMeadows11vanOsJ,AltamuraAC,BobesJetal.2-COM:aninstrumenttofacilitatepatient–professionalcom-municationinroutineclinicalpractice.ActaPsych-iatricaScandinavica12vanOsJ,AltamuraAC,BobesJetal.EvaluationoftheTwo-WayCommunicationChecklistasaclini-calintervention.BritishJournalofPsychiatry13MeadowsG,FosseyE,HarveyCetal.Theassessmentofperceivedneed.In:AndrewsG,HendersonSUnmetNeedinPsychiatry:problems,resources,.Cambridge:CambridgeUniversityPress,2000,pp.390–8.14MeadowsG,BurgessP,BobevskiIetal.Perceivedneedformentalhealthcare:influencesofdiagnosis,demographyanddisability.PsychologicalMedicine15MeadowsG,HarveyC,FosseyEetal.Assessingperceivedneedformentalhealthcareinacom-munitysurvey:DevelopmentofthePerceivedNeedforCareQuestionnaire(PNCQ).SocialPsychiatryandPsychiatricEpidemiology16EvansS,GreenhalghJandConnellyJ.Selectingamentalhealthneedsassessmentscale:Guidanceonthecriticalappraisalofstandardizedmeasures.JournalofEvaluationinClinicalPractice2000;6:379–93.17HickieIB,DavenportTA,Hadzi-PavlovicDetalDevelopmentofasimplescreeningtoolforcom-monmentaldisordersingeneralpractice.JournalofAustralia18

WareJE,KosinskiM,DeweyJEetalAManualforUs
WareJE,KosinskiM,DeweyJEetalAManualforUsersoftheSF-8HealthSurvey.Lincoln,RI:Quality-MetricIncorporated,2001.19ClarkeDMandMcKenzieDP.Anexaminationoftheefficiencyofthe12-itemSPHEREquestionnaireasascreeninginstrumentforcommonmentaldisordersinprimarycare.AustralianandNewZealandJournalofPsychiatry2003;37:236–9.20MeadowsG,BurgessP,FosseyEetal.Perceivedneedformentalhealthcare,findingsfromtheAustralianNationalSurveyofMentalHealthandWell-being.PsychologicalMedicine21KatzSJ,KesslerRC,FrankRGetal.TheuseofoutpatientmentalhealthservicesintheUnitedStatesandOntario:theimpactofmentalmorbidityandperceivedneedforcare.AmericanJournalofPublicHealth22HornblowA,BushellJA,WellsJEetal.Christchurchpsychiatricepidemiologicalstudy:useofmentalhealthservices.NewZealandMedicalJournal23LehtinenV,JoukamaaM,JyrkinenEetal.NeedformentalhealthservicesoftheadultpopulationinFinland:resultsfromtheMiniFinlandHealthSur-ActaPsychiatricaScandinavica24BoardmanJ,HenshawCandWillmottS.NeedsformentalhealthtreatmentamonggeneralpracticeBritishJournalofPsychiatry25CarterMF.Therelationshipofaself-reportedas-sessmentofneedinmentalillnesstoinsight.ofMentalHealth26MeadowsG,LiawT,BurgessPetal.Australiangeneralpracticeandthemeetingofneedsformen-talhealthcare.SocialPsychiatryandPsychiatricEpi-27MarshallM,LockwoodA,GreenGetal.Systematicassessmentsofneedandcareplanninginseverementalillness:Clusterrandomisedcontrolledtrial.BritishJournalofPsychiatry28O’LearyDandWebbM.Theneedsforcareassess-ment:alongitudinalapproach.PsychiatricBulletin29AshayeOA,LivingstonGandOrrellMW.Doesstandardizedneedsassessmentimprovetheout-comeofpsychiatricdayhospitalcareforolderpeople?Arandomizedcontrolledtrial.AgingandMentalHealth30DepartmentofHealth(UK).EffectiveCareCo-ordi-nationinMentalHealthServices:modernisingthecareprogrammeapproach–apolicybooklet.DepartmentofHealth,1999.CONFLICTSOFINTERESTADDRESSFORCORRESPONDENCEProfessorGrahamMeadows,Director,SouthernMentalHealthAdultResearch,TrainingandEvalu-ationCentre,DepartmentofPsychologicalMedi-cine,MonashUniversity,POBox956,DandenongVic3175,Australia.Phone:+61(0)395541941;Accepted???????TheGeneral-practiceUsers’Perceived-needInventory(‘GUPI’)TypeofhelpIwouldlikemyGPtodiscussthiskindofhelpwithmeIdon’tneedtodiscussthiskindofhelp.Iamalreadygettingthiskindofhelp,(eitherfrommyGPorsomewhereelse)InformationaboutemotionalproblemsorgettingtreatmentfortheseproblemsMedicationortabletstohelpyouwithemotionalproblems***Counselling;includinganykindofhelptotalkthroughyourproblems***Helptosortoutpracticalissuessuchashousingormoneyproblems***Helptoimproveyourabilitytowork,tocareforyourself,touseyourtimeortomeetpeople***Appendix1:TheGeneral-practiceUsers’Perceived-needInventory(GUPI)ThesequestionsaskwhetheryouwouldlikeyourGeneralPractitionertodiscusswithyouanyofthefollowingkindsofhelp,forcommonemotionalproblemssuchasfeelingdepressedoranxious.YourGPmightoffertohelpyouinthisway,oryoumightpreferyourGPtosuggestanalternativesourceofhelp.Pleasefirstcarefullyreadthelistofthreechoices,oneatthetopofeachofthecolumns,thenfillinonecirclelikethisineachrow,fortheoptionwhichbestappliestoyou.Haveanyofth

efollowingreasonsstoppedyouinthelastfew
efollowingreasonsstoppedyouinthelastfewweeks,fromgettinganyofthesekindsofhelp,orfromgettingasmuchhelpasyoumayhaveneeded.FillinanycirclesthatapplytoyouNotapplicable,Ihaven’tneededanyofthesekindsofhelp......................................................................Ipreferredtomanagemyself.......................................................................................................Ididn’tthinkanythingwouldhelp..................................................................................................Ididn’tknowwheretogethelp......................................................................................................Iwasafraidtoaskofhelporwhatotherswouldthinkofme.....................................................................Icouldn’taffordthemoney........................................................................................................Iaskedbutdidn’tgethelp.........................................................................................................ThankyouforyourhelpInternationalresearchTheGeneral-practiceUsers’Perceived-needInventory(‘GUPI’):abriefgeneralpracticetooltoassistinbringingmentalhealthcareneedstoprofessionalCatharineMcNabBALLB(Hons)MSc(Hons)ResearchFellowGrahamMeadowsMDMPhilMBChBMRCP(UK)MRCPsychFRANZCPDepartmentofPsychologicalMedicine,MonashUniversity,Melbourne,AustraliaTheGeneral-practiceUsers’Perceived-needInventory(‘GUPI’)isapracticalinstrumenttoidentifyperceivedneedformentalhealthcareingeneralpractice.EmpiricalfindingsreportedhereexploretheutilityandacceptabilityoftheDesignandsettingCriterionvalidityandtest–retest-reliabilitystudiesinmetropolitangeneralpracticesinMelbourne,Australia.One-hundredandtwenty-twoat-tendeesatgeneralpracticesexaminedcross-sectionally;83examinedlongitudinally.MainoutcomemeasuresPerformanceoftheGUPIagainstthe12-itemSomaticandPsycho-logicalHealthReport(SPHERE)questionnaireandgeneral,emotionalandphysicalill-healthitemsfromtheShortFormHealthSurvey.PerceivedneeddeclaredthroughtheGUPIwaspositivelyassociatedwithcombinedpsychologicalandsomaticSPHEREcaseness,andhigherscoresonHealthSurveyitems.Sensi-tivityofthemeasureforboththeseproxiesofpsychiatriccasenesswasgood.Asubsetofthreeitemshasperformanceinpsychometrictermslargelyequivalenttothefive-itemversion.Withinareliabilitystudy,whereageneralpractitionerconsultationoccurredwithinthetimeframeofthetestandretestdesign,overallperceivedneedreducedinfrequencyovertime.StabilitythroughtimewasassociatedwithperceivedneedforPerceivedneedasascertainedbytheGUPIisassociatedwithpoorergeneralhealth,physicalandemotionaldifficulties,andlikelypsychiatriccaseness.Theinstrumentmaybeuse-fulinthree-orfive-itemforms,thelatterallowingforultra-briefadministrationincombinationwithsymptomand/ordisabilitymeasures.Reassur-anceortreatmentmayaccountforadownwardtrendofperceivedneedthroughtime.Thisinstru-mentextendstherangeofbriefmentalhealthcareneedsassessmentinstrumentsavailableforgen-eralpracticeuse.Keywords:mentaldisorders,mentalhealthser-vices,needsassessment,primaryhealthcare,PrimaryCareMentalHealth2005RadcliffePublishing