Mental Illness and/or Mental Health? Investigating Axioms of the

To view this page ensure that Adobe Flash Player version 9.0.124 or greater is installed.

Please login or register to make a comment!

...Description...... more. less.

levelofsatisfactionorahighlevelofpositiveaffect. Inaddition,the DSM 3IV 3TR AxisV(i.e.,globalassessmentof functioning[GAF]assessment)isacontinuousvariableapproach thatisprofferedinthisarticleasaseconddiagnosticapproachfor mentalhealth.Here,theitemsandscalesthatmeasureeachmental healthclusteraresummedtogetherandcodedinto10-pointrange categories.Unlikethecategoricaldiagnosis,thecontinuousassess- mentofmentalhealthdoesnotmirroranyspecific DSM 3III 3R mentalillness.Hedonicwell-beingobviouslyremainspartofthe assessmentofmentalhealth,butitisnotconsideredanessential symptomtohaveahighlevelofmentalhealth.Instead,high-level mentalhealthrequiresindividualstoreporthighlevelsonmost(or all)measuresofsubjectivewell-being,whereasthecategorical diagnosisofflourishingrequiresonlyhighlevelsonoverhalfof themeasures.Thecontinuousassessmentofmentalhealth,there- fore,providesamoreholisticassessmentofmentalhealththanthe categoricaldiagnosis,anditisusedheretoassesswhetherresults dependonthediagnosisortheassessment. Therealsoremainsskepticismregardingthescientificandap- pliedvalueofmeasuringandclassifyingindividualsintermsof theirmentalhealth.AccordingtoMechanic(1999), cAlthoughthe conceptofpositivementalhealthisoneworthkeepinginmind,it isnotveryhelpfulinclassifyingdifferentpersons,groups,or populations d(p.2).WhetherMechanic 9sclaimiswarrantedisan empiricalquestionthathasyettobestudied.Therefore,andas displayedinFigure1,Ialsoinvestigatethefunctioningaxiomof thecompletehealthmodelinthisstudy.Completementalhealth (i.e.,theabsenceofmentalillnessandpresenceofflourishing)is morefunctionalandadaptivethanmoderatementalhealthorpure languishing(i.e.,absenceofmentalillness).Inturn,purelanguish- ingisasdysfunctionalasanepisodeofpurementalillness(i.e., presenceofmentalillnessbutalsothepresenceofmoderate mentalhealthormaybeevenflourishing).Last,completemental 1 DatafromtheNationalComorbidityStudyindicatethatonehalfofthe economicallyviablepopulation(i.e.,ages15 354years)willremainfreeof mentalillnessoveritslifetimeandabout70%annuallydoesnotfitthe DSM 3IV 3TR criteriaformostmentalillnesses. 540 KEYES illness(i.e.,presenceofmentalillnessandabsenceofmental health 4languishing)shouldbemoredysfunctionalthanapure mentalillness.Thesepredictionsmayonlyapplytothemost commonmentaldisorders(e.g.,anxietyandmood)ratherthanthe lesscommonbutmoreseverepsychoticdisorders.Nonetheless, thefunctioningaxiompredictsthatwithintherealmofmental health,completelymentallyhealthyindividualswillfunctionbet- terthanmoderatelymentallyhealthyindividuals,whointurn shouldfunctionbetterthanlanguishingindividuals. Method Sample DataarefromtheMidlifeintheUnitedStates(MIDUS)surveycon- ductedbytheMacArthurFoundation 9sResearchNetworkonSuccessful MidlifeDevelopment.TheMIDUSisanationalprobabilitysample,drawn withrandomdigitdialingprocedures,thatconsistedofEnglish-speaking, noninstitutionalizedadults,age25 374years,whoresidedinthe48con- tiguousstatesandwhosehouseholdincludedatleastonetelephone.The firststageofthemultistagesamplingdesignselectedhouseholdswith equalprobabilityviatelephonenumbers.Disproportionatestratifiedsam- plingwasusedatthesecondstagetoselectrespondents.Thesamplewas stratifiedbyageandgender,withoversamplingofmenbetweentheages of65and74years.Workingnonhousehold(e.g.,business)numberswere eliminatedbydefinition,andworkingnumbersthatwereunsuccessfully contacted10timeswerealsoeliminated. Table1 CategoricalDiagnosisofMentalHealth(i.e.,Flourishing) DiagnosticcriteriaSymptomdescription Hedonia: requireshighlevelonatleastone symptomscale(Symptoms1or2) 1.Regularlycheerful,ingoodspirits,happy,calmand peaceful,satisfied,andfulloflife( positiveaffect past30days ) 2.Feelshappyorsatisfiedwithlifeoverallor domainsoflife( avowedhappinessoravowedlife satisfaction ) a Positivefunctioning: requireshighlevelon sixormoresymptomscales(Symptoms3 313) 3.Holdspositiveattitudestowardoneselfandpastlife andconcedesandacceptsvariedaspectsofself ( self-acceptance ) 4.Haspositiveattitudetowardotherswhile acknowledgingandacceptingpeople 9sdifferences andcomplexity( socialacceptance ) 5.Showsinsightintoownpotential,senseof development,andopentonewandchallenging experiences( personalgrowth ) 6.Believesthatpeople,socialgroups,andsociety havepotentialandcanevolveorgrowpositively ( socialactualization ) 7.Holdsgoalsandbeliefsthataffirmsenseof directioninlifeandfeelsthatlifehasapurpose andmeaning( purposeinlife ) 8.Feelsthatone 9slifeisusefultosocietyandthe outputofhisorherownactivitiesarevaluedbyor valuabletoothers( socialcontribution ) 9.Exhibitscapabilitytomanagecomplex environment,andcanchooseormanageandmold environmentstosuitneeds( environmentalmastery ) 10.Interestedinsocietyorsociallife;feelssocietyand cultureareintelligible,somewhatlogical, predictable,andmeaningful( socialcoherence ) 11.Exhibitsself-directionthatisoftenguidedbyhisor herownsociallyacceptedandconventionalinternal standardsandresistsunsavorysocialpressures ( autonomy ) 12.Haswarm,satisfying,trustingpersonalrelationships andiscapableofempathyandintimacy( positive relationswithothers ) 13.Hasasenseofbelongingtoacommunityand derivescomfortandsupportfromcommunity ( socialintegration ) a Lifedomainsmayincludeemploymentandmarriageorcloseinterpersonalrelationship(e.g.,parenting). Figure1. Predictionsofpsychosocialfunctioningfromthecompletestate modelofmentalhealth.Signsarereversedfornegativeoutcomes.Pure mentalillness any12-monthmentaldisorderwithoutlanguishing;pure languishing anylanguishingwithoutany12-monthmentaldisorder. 541 COMPLETESTATEMODELOFHEALTH TheMIDUSsurveycompliedwithInstitutionalReviewBoardstan- dards,andinterviewersreadastandardinformedconsentprotocolatthe beginningofthetelephoneinterview.Adultswhoagreedtoparticipate wereadministeredacomputer-assistedtelephoneinterviewthatlasted45 minonaverageandwerethenmailedtwoquestionnairebookletsthat requiredabout1.5hronaveragetocomplete.Allparticipantswereoffered $20andacopyofafinalstudymonographasincentivesforparticipation. Witharesponserateof70%forthetelephonephaseandaresponserateof 87%fortheself-administeredquestionnairephase,theoverallresponse ratewas61%withasamplesizeof3,032respondents.Fieldprocedures lastedapproximately13monthsandwerebegunin1994andconcludedin 1995. Descriptiveanalysesarebasedontheweightedsampletocorrectfor unequalprobabilitiesofhouseholdandwithinhouseholdrespondentse- lection.Thesampleweightpoststratifiesthesampletomatchthepropor- tionsofadultsaccordingtoage,gender,education,maritalstatus,race, residence(i.e.,metropolitanandnonmetropolitan),andregion(northeast, midwest,south,andwest)onthebasisoftheOctober1995Current PopulationSurvey(seeKeyesetal. 9s,2002,studyforthedemographic characteristicsoftheMIDUSsample).Findingswereunchangedby whetherthesamplewasweighted;alldescriptiveanalysespresentthe findingsbasedontheweightedsample.Thesamplingdesigninvolved somecomplexitiesthatcouldintroducedesigneffectsthatinflatestandard errorestimates.However,simulationswithjackkniferepeatedreplications (seeKish&Frankel 9s,1974,study)onanarrayofvariablesrevealedvery smallstandarderrorinflationofdesign-basedestimates,eliminatingthe needtoadjuststatisticaltestsfordesigneffectsinthesedata. Measures Mentalillness. TheMIDUSused DSM 3III 3R (AmericanPsychiatric Association,1987)criteriatodiagnosementalillness,whichwereopera- tionalizedbytheCompositeInternationalDiagnosticInterviewShortForm (CIDI-SF)scales(Kessler,Andrews,Mroczek,Ustun,&Wittchen,1998). StudieshaveshownthattheCIDI-SFhasexcellentdiagnosticsensitivity anddiagnosticspecificityascomparedwithdiagnosesbasedonthefull CIDIintheNationalComorbidityStudy(Kessler,DuPont,Berglund,& Wittchen,1999).Duringthetelephoneinterview,theCIDI-SFwasusedto assesswhetherrespondentsexhibitedsymptomsindicativeof(a)MDE,(b) generalizedanxietydisorder,(c)panicdisorder,and(d)alcoholdepen- denceduringthepast12months(notethatalldiagnosticcriteriausedin thisstudyconformtothecriteriaoutlinedinthe DSM 3IV 3TR ). Mentalhealth. Aspartoftheself-administeredquestionnaire,respon- dentsindicatedhowmuchofthetimeduringthepast30days 4 call, d cmost, d csome, d calittle, dor cnoneofthetime d 4theyfeltsixsymptoms ofpositiveaffect.Thepositiveaffectsymptomswere(a)cheerful,(b)in goodspirits,(c)extremelyhappy,(d)calmandpeaceful,(e)satisfied,and (f)fulloflife.Theinternalreliabilityofthepositiveaffectscalewas.91. Moreover,respondentswereaskedto cratetheirlifeoverallthesedays don ascalerangingfrom0( worstpossiblelifeoverall )to10( bestpossiblelife overall ). RespondentsalsocompletedRyff 9s(1989)scalesofpsychologicalwell- beingandKeyes 9s(1998)scalesofsocialwell-being.Thepsychological well-beingscalesreflecthowmuchindividualsarethrivingintheirprivate, personallives.Thescales,witharepresentativeiteminparentheses,areas follows:self-acceptance( cIlikemostpartsofmypersonality d),positive relationswithothers( cMaintainingcloserelationshipshasbeendifficult andfrustratingforme d),personalgrowth( cForme,lifehasbeena continualprocessoflearning,changing,andgrowth d),purposeinlife( cI sometimesfeelasifI 9vedoneallthereistodoinlife d),environmental mastery( cIamgoodatmanagingtheresponsibilitiesofdailylife d),and autonomy( cItendtobeinfluencedbypeoplewithstrongopinions d).The measuresofsocialwell-beingoperationalizehowmuchindividualssee themselvesthrivingintheirpublic,sociallife.Thescales,witharepre- sentativeiteminparentheses,areasfollows:social-acceptance( cPeopledo notcareaboutotherpeoples 9problems d),socialactualization( cSocietyis notimprovingforpeoplelikeme d),socialcontribution( cMydailyactiv- itiesdonotcreateanythingworthwhileformycommunity d),socialcoher- ence( cIcannotmakesenseofwhat 9sgoingonintheworld d),andsocial integration( cIfeelclosetootherpeopleinmycommunity d). Eachscaleofpositivefunctioning(i.e.,psychologicalandsocial)con- sistedofthreeitemswitharelativebalanceofpositiveandnegativeitems andwereself-administered.Respondentsindicatedwhetheranitemde- scribedhowtheyfunctionedonascalerangingfrom1( stronglyagree )to 7( stronglydisagree ).Negativeitemswerereversecoded.Thethree-item scalesofpsychologicalwell-beinghaveshownmodestinternalconsistency (Ryff&Keyes,1995),whereasthelarger,20-itemscaleshaveshown excellentinternalconsistencyandconstructvalidity(seeRyff 9s,1989, study).Confirmatoryfactoranalysesalsoconfirmedtheproposedsix- factorstructureofpsychologicalwell-being(Ryff&Keyes,1995).The internalconsistencyofthecombined18itemsofpsychologicalwell-being inthisstudywas.81.Similarly,thethree-itemscalesofsocialwell-being haveshownmodest-to-excellentinternalconsistency,andthelargeritem scalesofsocialwellbeinghaveexhibitedgoodinternalconsistencyand constructvalidity(seeKeyes 9s,1998,study).Confirmatoryfactoranalyses supportedtheproposedfive-factorstructureofsocialwell-being(Keyes, 1998).Theinternalconsistencyofthecombineditemsofsocialwell-being inthecurrentstudywas.81.Table2presentsthebivariatecorrelationsand descriptivestatisticsofthecontinuous(i.e.,symptomcount)measuresof mentaldisorderandthesummedscaleofeachtypeofsubjective well-being. Table2 BivariateCorrelationsandDescriptiveStatistics Variable1234567 1.No.depressionsymptoms1.00.33.32.07 .33 .26 .16 2.No.generalizedanxietysymptoms1.00.26.02 a .19 .22 .15 3.No.panicattacksymptoms1.00.08 .22 .19 .14 4.No.alcoholdependencesymptoms1.00 .15 .08 .03 a 5.Summedscale,emotionalwell-being1.00.54.36 6.Summedscale,psychologicalwell-being1.00.53 7.Summedscale,socialwell-being1.00 Range0 370 3100 360 371 31512 3425 335 M 0.840.210.370.5311.0032.9022.10 SD 2.001.101.101.202.204.804.70 Note. Correlationsaresignificantat p \x2 .001(two-tailed). a Statisticallynonsignificant. 542 KEYES Beforeapplyingtheproposeddiagnosticcriteria,Idividedeachscaleof emotional,psychological,andsocialwell-beingbythenumberofconstit- uentitems,standardized,andIcomputedtertiles.Thestatisticaltertile definesthethresholdforhigh-level(i.e.,uppertertile)andlow-level(i.e., lowertertile)subjectivewell-being,becausetherearenootherunambig- uousthresholds.Thus,likementaldisorder,mentalhealthcanbeviewedas adeviationfromusualfunctioning,whichcanbeoperationalizedbythe statisticalaverageorbyanidealstandard(e.g.,Mechanic,1999).Inthis study,thetertileoperationalizesflourishingasabove-averagefunctioning andlanguishingasbelow-averagefunctioning. Psychosocialfunctioning. Respondentsindicatedwhethertheirhealth limitedthem calot, d csome, d calittle, dor cnotatall dfromdoinganyof nineinstrumentalactivitiesofdailylife.Theactivitiesincludedliftingand carryinggroceries,bathingordressingoneself,climbingseveralflightsof stairs,bending(kneelingorstooping),walkingmorethan1mile,walking severalblocks,walkingoneblock,performingvigorousactivity(e.g., running,liftingheavyobjects),andperformingmoderateactivity(e.g., bowlingorvacuuming).Theinternalconsistencyofthelimitationof activityofdailylivingscalewas.91. Respondentsindicatedthenumberofmissedandcutbackworkdays duringthepast30days.Specifically,ofthepast30days,respondentswere asked cHowmanydayswereyoutotallyunabletogotoworkorcarryout yournormalhouseholdworkactivitiesbecauseofyourphysicalhealthor mentalhealth? dSubsequently,ofthepast30days,andasidefromthose daystheyweretotallyunabletowork,respondentswereasked, cHow manyoftheotherdaysdidyouhavetocutbackonworkorhowmuchyou gotdonebecauseofyourphysicalhealthormentalhealth? dFollow-up questionsinquiredwhethertheworkcutbacksandlostdaysofworkwere duetophysicalhealth,tomentalhealth,ortoacombinationofmentaland physicalhealth.Inthisstudy,Ifocusedonlyonworkcutbacksandlost daysduetomentalhealth. Respondentsalsowereaskedwhetherthey cagreed dor cdisagreed dwith thefollowingstatements: cIoftenfeelhelplessindealingwiththeprob- lemsoflife dand cThereislittleIcandotochangetheimportantthingsin mylife. dAnalysesfocusedontheproportionthatagreedwiththehelp- lessnessstatements.Tomeasuregoalformation,Iaskedrespondents whetherthestatements, cIknowwhatIwantoutoflife dand cIfindit helpfultosetgoalsforthenearfuture, ddescribedthem calot, d csome, d ca little, dor cnotatall. dTwostatementsmeasuredrespondents 9resilience. First, cWhenfacedwithabadsituation,IdowhatIcantochangeitforthe better. dSecond, cIfindIusuallylearnsomethingmeaningfulfroma difficultsituation. dRespondentsindicatedwhethereachstatementde- scribedthem calot, d csome, d calittle, dor cnotatall. dLast,intimacywas measuredbyaskingrespondentstoindicatewhethertheir(a)spouseor partner(orotherfamilymembers)and(b)friends creallycare(s)about them. dRespondentsindicatedwhethereachstatementdescribedthem ca lot, d csome, d calittle, dor cnotatall. dInthisstudy,Ifocusedanalyseson theproportionthatsaidthatthegoalformation,resilience,andintimacy questionsdescribedtheirsituation calot. d Results Table3reportstheindicesoffitofconfirmatoryfactormodels thattestdifferenttheoriesofthelatentstructureofthemeasuresof mentalhealthandmentalillness.Threescalesservedasindicators ofthelatentconstructofmentalhealth:thesummedscalesthat reflectedemotionalwell-being(i.e.,satisfactionpluspositiveaf- fect),psychologicalwell-being(i.e.,sixscalesofpsychological well-beingsummedtogether),andsocialwell-being(i.e.,thefive scalesofsocialwell-beingsummedtogether).Foursummarymea- suresservedasindicatorsofthelatentconstructofmentalillness: thenumberofsymptomsofMDE,generalizedanxiety,panic disorder,andalcoholdependence.Theindependencemodelisa baselineinwhichtherewereasmanylatentconstructsasmeasures ofmentalhealthandillness(i.e.,eachmeasurereflectsaninde- pendentlatentfactor).Thechi-squarestatisticanddescriptivefit indiceswereverylarge,indicatingthatthetheoryofindependence isuntenable. Bycomparison,andconsistentwiththepsychiatricmodelof mentalhealth,Model2positsthatallmeasuresarecausedbya single,bipolarlatentdimension.Thechi-squarestatisticandde- scriptivefitindicesweremarkedlyimprovedforthesingle-factor modelrelativetotheindependencemodel.Moreover,thechi- squarecontrastoftheindependenceandsingle-factormodelsre- vealedahighlystatisticallysignificantreductionofchi-square, suggestingthatthesingle-factor(psychiatric)modelwasamore tenablemodelthantheindependencemodel.However,stemming fromthecompletestatemodelofhealth,Model3 4whichposits thatthemeasuresofmentalhealthandmentalillnessreflecttwo distinct,butuncorrelated,unipolarfactors 4alsohadmarkedly improvedfitindices.Thechi-squarecontrastoftheindependence andtwo-factor(orthogonal)modelrevealedahighlystatistically significantreductionofchi-square,suggestingthatittoowasa moretenablemodelthantheindependencemodel. Thefinalmodelfullyteststhestructuralaxiomofthecomplete healthmodel,thatis,measuresofmentalhealthandillnesses Table3 MaximumLikelihoodEstimationofConfirmatoryFactorModelsofTheoriesoftheLatent StructureofContinuousMeasuresofMentalHealthandDSM 3III 3RMentalDisorders Latentstructuremodel 2 df AGFICNRMSEAAIC difference 2 / df difference 1.Independence3,388.721.6025.5.223,402.7 2.Singleaxis582.614.88150.8.12687.7400.9 1 32 * 3.Twoaxes,orthogonal557.314.91157.7.11543.6404.5 1 33 * 4.Twoaxes,oblique a 162.413.97511.7.06193.5394.9 3 34 * Note.N 2,997.Mentalillnessmeasuresincludedmeasuresofthenumberofsymptomsoffourmental disorders:panicdisorder,majordepressiveepisode,generalizedanxiety,andalcoholdependence;mentalhealth measuresincludedthesummedscalesofemotionalwell-being,psychologicalwell-being,andsocialwell-being. AGFI adjustedgoodnessoffitindex;CN critical N; RMSEA root-mean-squareerrorofapproximation; AIC Akaikeinformationcriterion. a Thecorrelation(phi)betweenlatentfactorsis .53. * p \x2 .001. 543 COMPLETESTATEMODELOFHEALTH constituteseparatelatentfactors,andthetwolatentfactorsare correlated.Thechi-squarecontrastofthetwo-factor(orthogonal) Model3andthetwo-factor(oblique)Model4revealedahighly statisticallysignificantreductionofchi-square,suggestingthatthe correlatedtwo-factormodelwasthemosttenablemodelofthe structureofmentalhealthandillness.Moreover,thefitindicesfor Model4suggestedthatitwasanexcellentfittingmodeltothese data.Theadjustedgoodness-of-fitindexwas.97,thecritical N was twiceaslargeastherecommendedcutpointof200,andthe root-mean-squareerrorofapproximationandAkaikeinformation criterionweresmallerthanforModel3.Thecorrelationbetween thelatentfactorofmentalillnessandmentalhealthwas 3.53.The standardizedloadingsofallmentalhealthindicatorsontheirlatent factorwere.60orhigher,andthestandardizedloadingsofthe mentalillnessmeasuresontheirlatentfactorwere.50orhigher, exceptforalcoholdependence,whichwas.12. 2 Insum,data stronglysupportthestructuralaxiomhypothesis;thetheorythat themeasuresofmentalhealthandmentalillnessconstitutesepa- rate,correlatedaxesprovidesthebestfittingmodeltothesedata. Table4reportstheprevalenceofmentalhealthasdiagnosed categoricallyandassessedcontinuously.Only18%fitthecategor- icaldiagnosisofflourishing.Moreover,only9.6%fitintothe highestrangeofthecontinuousassessmentofmentalhealth(i.e., 50 360),whichwasthefourthmostprevalentrangeofmental healthscores.Rather,mostAmericans,roughly6in10according tothecategoricaldiagnosis,weremoderatelymentallyhealthy. Similarly,themostprevalentrangeofmentalhealthscoreswasthe approximatemidpointofthecontinuousassessment(i.e.,30.0 3 39.9).AlthoughfewAmericanswerelanguishing 4about17% accordingtothecategoricaldiagnosis 4itisimportanttonotethat nearlyasmanyAmericanswerementallyunhealthy(i.e.,languish- ing)aswerementallyhealthy(i.e.,flourishing).Similarly,al- thoughlessthan1%fellintothelowestrangeofthecontinuous assessment(i.e.,0.0 39.9),acombinedtotalofjustover20%fell belowtheapproximatemidpoint,withcontinuousscoresat29.9or lower. Table4alsoreportstheassociationofthefour DSM 3III 3R mentaldisordersandtheircomorbiditywiththementalhealth diagnoses.WhetherIusedthecategoricaldiagnosisorcontinuous assessment,thepatternofresultsisunequivocal:Flourishingin- dividualswereatthelowestriskofanyofthefour12-month mentalillnessesortheircomorbidity.Conversely,andbecause causalitymayoperateintheotherdirection,individualswithany ofthefourmentaldisorderswereataverylowriskofflourishing. Moreover,therelationshipofmentalillnessandmentalhealth representsagradient;theprevalenceofeachmentalillnessde- creasedaslevelofmentalhealthincreased.Forexample 4and 2 Thefindingsreportedhereincludealcoholdependenceasanindicator ofmentalillnessbecausereestimationofthetwo-factor(oblique)model withoutalcoholdependencedidnotchangeanyofthedescriptivefit indicesandonlychangedthelatentfactorcorrelationbyasinglepoint(i.e., 3.52). Table4 PrevalenceandComorbidityof12-monthDSM 3III 3RMentalDisordersWithMentalHealthDiagnosis(SampleWeighted) CategoricaldiagnosisContinuousassessment Total Mentally unhealthy ( languishing ) Moderately mentally healthy Mentally healthy ( flourishing )0.0 39.910.0 319.920.0 329.930.0 339.940.0 349.950.0 360.0 a n 5111,974547131094771,108962283 Mentaldisorder%16.965.118.00.43.716.237.532.69.6 Majordepressiveepisode n 14325426949111697412422 %28.013.44.869.245.023.315.37.74.214.1 Generalizedanxiety disorder n 5441252134306298 %10.62.20.441.719.37.12.70.60.73.3 Panicdisorder n 8011954256072403204 %15.76.30.930.822.912.66.54.21.16.9 Alcoholdependence n 4813415293887545194 %9.46.82.715.48.38.07.85.61.86.6 Comorbidity b n 8410747325671243193 %16.45.40.753.829.411.76.42.51.16.5 Note. Fortheseparatetestsoftheassociationofeachmentalhealthdiagnosis(categoricalandalsocontinuous)witheachmentalillnessdiagnosis, p \x2 .002(two-tailed)forallchi-squaretests. a Allseparatescalesofeachtypeofsubjectivewell-beingweresummedtogethertoformanoverallscore,andtheoverallscoreswererecodedtoreflect atotalscorerangingfrom0tothehighestscore,whichwas60.8;becauseonlythreescoresexceeded60,andnotevenbyafullpoint,thosethreescores weretopcodedintothe50 360category. b Twoormore(maximum 4)mentaldisordersduringthepast12months. 544 KEYES focusingonthecategoricaldiagnosisofmentalhealth 428%of languishingand13%ofmoderatelymentallyhealthyindividuals (comparedwith5%offlourishingadults)hadMDE.Flourishing individualswereover5timeslesslikelythanlanguisherstohave MDE.Flourishingindividualsalsowere23timeslesslikelythan languisherstohavehadcomorbidmentaldisordersoverthe12- monthperiod. Table5containsthecross-tabulationofwhetherindividualshad anyofthementaldisordersbythecategoricalmentalhealth diagnosistoobtaintheestimatesofthestatesofcompletemental health.Theprevalenceofcompletementalhealthwasnearly 17.0%,whereasjustover50.0%weremoderatelymentally healthy.Nearly10.0%hadaformofpurelanguishing,whichis languishingwithoutanyofthefourmentaldisorders.Purelan- guishersalsoreportedanaverageoflessthanonesymptomofany ofthementalillnesses,suggestingitisnotasubclinicalformof anyofthefourmentaldisorders.Ofthe22.9%ofadultswhohad anyofthefourmentalillnesses,7.0%werealsolanguishing, whereas15.9%hadapureformofmentalillness,meaningthis lattergrouphadeithermoderatementalhealthorwasflourishing. Thefunctioningaxiompredictsthatcompletementalhealthis the summumbonum ofpsychosocialfunctioning,suggestingitis moreadaptivethanmoderatementalhealthorpurelanguishing. FindingsinTable5primarilysupportthefunctioningaxiomofthe completementalhealthmodel.Occupationalandpsychosocial malfunctioningwaslowest,whereaspositivefunctioningwas greatest,amongadultswithcompletementalhealth.Healthlimi- tationsofactivitiesofdailyliving,anyhalf-daycutbackandany lossofadayofwork,andperceivedhelplessnesswereleastlikely amongadultswithcompletementalhealth.Moreover,goals,re- Table5 PsychosocialFunctioningbyCategoricalDiagnosisofCompleteMentalHealth(SampleWeighted) Mentalillness andlanguishing Puremental illness Pure languishing Moderately mentallyhealthy Completely mentallyhealthy n 2144842971,535503 Psychosocialfunctioning%7.015.99.850.616.6 Healthlimits 4anyofnine Activitiesofdailyliving(alot) n 731289834961 %34.326.433.022.712.1 Anyworkcutbackduringpast30days n 691043210617 %34.522.311.37.13.4 AnylostworkdayduringPast30days n 4963228919 %23.213.27.75.83.8 Helplessness Feelhelplessdealingwithlife 9sproblems n 17216119038327 %80.434.064.025.45.4 Cannotchangeimportantthingsinlife n 999612938142 %46.720.243.625.48.5 Goals Knowswhatwantsoutoflife n 5219560599342 %24.340.720.339.768.0 Findsithelpfultosetgoalfornearfuture n 3116336467267 %14.534.012.130.853.4 Resilience Changesbadsituationsforbetter n 5827273855397 %27.256.825.056.579.2 Learnsfromdifficultsituations n 7028492762346 %32.759.531.150.569.8 Intimacy Feelsreallycaredforbyspouse,partner, orfamily n 1574222081,374486 %73.487.470.089.596.8 Feelsreallycaredforbyafriend n 5824753729336 %27.251.818.148.567.3 Note. Fortheseparatetestsoftheassociationofthementalhealthdiagnosiswitheachpsychosocialfunctioningvariable, p \x2 .001(two-tailed)forall chi-squaretests. 545 COMPLETESTATEMODELOFHEALTH silience,andintimacyweregreatestamongadultswithcomplete mentalhealth.Indeed,completelymentallyhealthyadultsgener- allyfunctionedbetterthanmoderatelymentallyhealthyadults;in turn,moderatelymentallyhealthyadultsfunctionedbetterthan adultswithpurelanguishing. Incontrast,occupationalandpsychosocialmalfunctioningwas greatest,whereaspositivefunctioningwaslowest,amongadults witha12-monthmentalillnessontopoflanguishing.Health limitationsofactivitiesofdailyliving,anyhalf-daycutbackand anylossofanentiredayofwork,andperceivedhelplessnesswere greatestamongadultswithamentalillnessaswellaslanguishing. Goals,resilience,andintimacyalsowerelowestamongadultswho hadamentalillnessinadditiontolanguishing,who,aspredicted, functionedworsethanadultswithapurementalillness. Thecontrastinfunctioningbetweenpurelanguishingandpure mentalillnessesgenerallysupportsthefunctioningaxiom.In2of the11indices,individualswithapurementalillnessfunctioned worsethanindividualswithpurelanguishing.Namely,ahigher proportionofadultswithpurementalillness(comparedwithpure languishingindividuals)hadmoreworkcutbacksorlostmoredays ofwork.In9ofthe11indices,purelanguishingwasmore dysfunctionalthanpurementalillness.Forexample,onethirdof purelanguishers,comparedwithjustoveronequarterofindivid- ualswithpurementalillness,reportedanyhealthlimitationsof dailyactivities.Twiceasmanyadultswithpurementalillness (43%)asadultswithpurelanguishing(21%)saidthat cknowing whattheywantoutoflife ddescribedthemalot.Twiceasmany adultswithpurementalillness(33%)asadultswithpurelanguish- ing(15%)saidthat citishelpfultosetgoalsforthenearfuture d describedtheirsituationalot.Twiceasmanyadultswithpure mentalillness(59%)asadultswithpurelanguishing(26%)said that, cWhenfacedwithabadsituation,IdowhatIcantochange itforthebetter ddescribedthemalot.Nearly3timesasmany adultswithpurementalillness(54%)aswithpurelanguishing (18%)saidthata cfriendthatreallycaresaboutthem ddescribed theirsituationalot. Insum,completelymentallyhealthyadultsexhibitedthefewest healthlimitationsofdailyactivities,littleornoworkcutbacksor missedworkdays,andhighlevelsofpsychosocialfunctioning. Completementalhealthwasthesummumbonumoffunctioning, betterthantherelativeabsenceofmentalhealth 4thatis,moderate mentalhealthand,especially,purelanguishing.Inturn,languish- ingwasasbadas,andsometimesworsethan,thepresenceofa purementalillness.Last,completementalillness 4thatis,the absenceofmentalhealthqualanguishingandthepresenceofa mentaldisorder 4wasmostdysfunctional,evenmoredysfunc- tionalthanapurementalillness,atleastasmeasuredbysome indicators. Discussion MentalillnessessuchasMDEareasocietalburdenandare projectedtobecomemoreprevalentandburdensomebytheyear 2020(Murray&Lopez,1996).Assuch,thereislittleskepticism aboutthesocietalvalueofgreatersupportformentalillness research.Ratherthanexpandthescopeofbasicandapplied researchtoincludementalhealth(Gladis,Gosch,Dishuk,&Crits- Christoph,1999),treatmentandpreventionofmentalillnessap- pearatfirstglancetobemoreurgentpublichealthissues.How- ever,argumentsfororagainstthestudyofpositiveorcomplete mentalhealtharebasedonuntestedassumptions,twoofwhich wereinvestigatedinthisarticle. First,mentalhealthispresumedtobetheoppositeofmental illness;thus,theabsenceofmentalillnessequalsthepresenceof mentalhealth.Underthisassumption,ifsocietycaneffectively treatmentalillness,thenmoreindividualswillbecomementally healthy.Thecurrentstudyconfirmsempiricallythatmentalhealth andmentalillnessarenotoppositeendsofasinglecontinuum; rather,theyconstitutedistinctbutcorrelatedaxesthatsuggestthat mentalhealthshouldbeviewedasacompletestate. 3 Thus,the absenceofmentalillnessdoesnotequalthepresenceofmental health.Thestructureofmentalhealthasdistinctfrommental illnesswasimpliedbytheintroductionofAxisV(Luborsky,1962) andtheGAFscaletomeasureAxisV(Endicott,Spitzer,Fleiss,& Cohen,1976).However,AxisVremainsunderusedrelativeto other DSM 3IV 3TR axes,perhapsbecausetheGAFexhibitsrela- tivelypoorreliabilityandvalidity(Goldman,Skodol,&Lave, 1992;Roy-Byrne,Dagadakis,Unutzer,&Ries,1996),andpsy- chologistsandpsychiatristrateitsusefulnessasverylow(Frazee, Chicota,Templer,&Arikawa,2003). Second,itiscommonlyassumedthatclassifyingandmonitoring thementalhealthstatusofindividuals,groups,orpopulationsis worthless.Individualsfreeofmentalillnessareassumedtobe homogenous,functioningaboutthesameandmarkedlybetterthan mentallyillindividuals.However,thediagnosisandmeasurement ofmentalhealth 4howevercrudeandpreliminarysomemaythink oftheschemepresentedhere 4hasprovidedsomeinvaluable information.First,relativelyfewadults(i.e.,about2in10)who werefreeofanyofthefour12-monthmentaldisorderscouldbe classifiedasflourishingorcompletelymentallyhealthy.Almostas manyadultswerementallyunhealthy(i.e.,languishing)aswere mentallyhealthy(i.e.,flourishing),andmostadultsweremoder- atelymentallyhealthy.Second,diagnoseslessthanflourishing wereassociatedwithgreaterlevelsofdysfunctionsintermsof workreductions,healthlimitations,andpsychosocialfunctioning. Moreover,purelanguishingwasasdysfunctional(sometimes more)thanpurementalillness(althoughthepurementalillness wasassociatedwithgreaterworkproblems).Mentalillnesswhen combinedwithlanguishingwasmarkedlyworsethanapureform ofmentalillness. Inaddition,arecentstudyconfirmedthehypothesisthatthe completementalhealthdiagnosticstateswereindependentrisk factorsforcardiovasculardisease(CVD;Keyes,2004).Inthis study,Ifocusedonthecombinationofthecategoricaldiagnosisof mentalhealthwithMDE,becausethelatterhasbeenshowntobe ariskfactorforheartandarterialdiseases.Theunadjustedprev- alenceofanyCVDwas8%amongcompletelymentallyhealthy adults,comparedwith12%ofadultswithmoderatementalhealth, 12%ofadultswithpurelanguishing,and13%ofadultswithpure depression.Amongadultswhowerelanguishingandhadanepi- sodeofmajordepression,theprevalenceofanyCVDwas19%.In multivariateanalyses,completelymentallyhealthyadultshadthe 3 Thelatentfactorsofmentalhealthandmentalillnesscorrelatedat 3.53, indicatingthatonequarterofvariancebetweencommonmeasuresof mentalillnessandmentalhealth(i.e.,subjectivewell-being)isshared variance. 546 KEYES lowestriskforCVD.Anythinglessthancompletementalhealth 4 especiallyforpostmenopausalwomen(forreasonshypothesized andexplainedinthefullstudyreport) 4translatedintoelevated riskforCVDthatwascomparablewiththeriskassociatedwiththe knownriskfactorsofdiabetes,smoking,andlackofexercise. Thereareseveralpotentiallimitationstothecurrentstudy.First, theMIDUSmaynotprovidethemostrigoroustestofthestructural axiom.Onlyfourmentaldisordersweremeasured,andfuture researchshouldincludeawiderassessmentofmentaldisordersto fullyinvestigatethestructuralaxiom.Second,thesubjectivewell- beingscalesareself-reportmeasuresandpossiblyreflectabias towardwesternizedculturesanddevelopednations.Futurere- searchshouldincludereportsonthetarget 9swell-beingfrom others(e.g.,expertsorfamily)andinvestigatewhetherthestruc- tureofsubjectivewell-being(i.e.,hedoniaandeudaimonia),and dimensionstherein,arereplicatedinothercultures.Thesparse amountofcomparativeresearchinthisareasuggeststhatthe structureofsubjectivewell-beingislikelytobethesame,although thenumberandkindofeudaimonicdimensionsislikelytovaryby country(seeKeyes&Ryff 9s,2003,study). Third,theproposeddiagnosticcriteriaandthevalidityofthe diagnosesrequirefurtherwork.Thisstudyandpreviousresearch (Keyes,2002)adoptedacombinationofstatistical(e.g.,tertile cutpoints)andrational(e.g.,useof DSM 3IV 3TR numberofsymp- tomstomeetcondition)diagnosticcriteria.Althoughthisstudy suggeststhatfewadultsareflourishing,changesinthediagnostic criteriawouldobviouslychangethepointprevalenceestimatesand someconclusions.Italsoremainsanempiricalquestionwhethera categoricaltaxonoracontinuumbestrepresentsthelatentstruc- tureofmentalhealthasmeasuredinthisstudy.Futureresearch shouldalsoexploreadditionalcriteriaofmentalhealth(e.g.,a durationcriterion)andalternativemodelsofmentalhealthaswell asinvestigatetheconstructvalidityofanydiagnosesagainstexpert evaluations. 4 Findingsalsosuggestdirectionsforfutureresearchregarding thecausalnexusofstatesofmentalhealthwithmentaldisorders. Inparticular,islanguishingadiathesisfor,andisflourishinga protectivefactoragainst,theonsetandrecurrenceofmentalill- ness?Conceptually,onecanthinkofmentalhealthasthecontin- uumatthetopofthecliffwheremostindividualsreside.Flour- ishingindividualsareatthehealthiestandthereforefarthest distancefromtheedgeofthiscliff;languishingplacesindividuals veryneartheedgeofthecliff.Hence,languishingmayactasa diathesisthatisactivatedbystressorsthatpushindividualsoffthe cliffandintomentalillness. Last,findingsoftheassociationofcompletementalhealthwith lowhelplessnessandhighgoals,resilience,andintimacysuggests thatextanttalktherapiesmaybeusefulforpromotingflourishing aswellastreatingmentalillness.Prominenttheoriesoftheetiol- ogy,andtherefore,treatmentofdepressionhavefocusedon cognitive 3behavioralprocesses(e.g.,Hollon,Thase,&Markow- itz,2002).Thatis,mentalillnessessuchasdepressionresultfrom eitherorbothdistortedthoughtandbehavioralpatternsorthe inabilitytoformandmaintainpositiveinterpersonalattachments. Yearsofresearchonlearnedhelplessness(Seligman,1975)has consistentlyshownthatperceptionsofone 9shelplessnessand inabilitytochangebadsituationsleadtodistressanddepression, whereasframingoflifeintermsofgoalsisassociatedwithhigh levelsofsubjectivewell-being(Emmons,2003).Researchon resiliencehasshownthatmentalhealthdependsonanindividual 9s abilitytomentallycopewith,transform,andfindmeaningful lessonsfromthestressorsandlife 9schallenges(Ryff&Singer, 2003).Last,aplethoraofstudieshaveindicatedthatinterpersonal relationshipsthatsatisfyneedsforbelongingnessandsuccorare instrumentaltomentalhealth(Harlow,1958;Reis&Gable,2003). Insum,mentalhealthhasbeenstudiedfortoolongasmerely theabsenceofmentalillness.Itistimetoheedtheadviceofthe historicJointCommissiononMentalIllnessandHealth(see Gurin,Veroff,&Feld,1960;Jahoda,1958),whichrecommended thestudyandpromotionofmentalhealthaswellasthestudyand treatmentofmentalillness.Mentalhealthisclearlysomething positive,andanythinglessthanflourishingappearstofallshortof healthyfunctioninginlife. 4 Itisunclearwhowouldbethemostexpertjudgeofanotherperson 9s mentalhealth,namelywhetheritwouldbeapsychiatrist,psychotherapist, along-termcompanion,ateacherorcoach,orsomeoneelse. References AmericanPsychiatricAssociation.(1987). Diagnosticandstatisticalman- ualofmentaldisorders (3rded.,rev.).Washington,DC:Author. AmericanPsychiatricAssociation.(2000). Diagnosticandstatisticalman- ualofmentaldisorders (4thed.,textrev.).Washington,DC:Author. Bradburn,N.M.(1969). Thestructureofpsychologicalwell-being. Chi- cago:Aldine. Emmons,R.A.(2003).Personalgoals,lifemeaning,andvirtue:Well springsofapositivelife.InC.L.M.Keyes&J.Haidt(Eds.), Flour- ishing:Positivepsychologyandthelifewell-lived (pp.105 3128).Wash- ington,DC:AmericanPsychologicalAssociation. Endicott,J.,Spitzer,R.L.,Fleiss,J.L.,&Cohen,J.(1976).TheGlobal AssessmentScale:Aprocedureformeasuringoverallseverityofpsy- chiatricdisturbance. ArchivesofGeneralPsychiatry,33, 766 3771. Frazee,J.C.,Chicota,C.L.,Templer,D.I.,&Arikawa,H.(2003).The usefulnessoftheAxisVdiagnosis:Opinionsofhealthcareprofession- als. JournalofNervousandMentalDisease,191, 692 3694. Garrison,C.Z.,Schluchter,M.D.,Schoenbach,V.J.,&Kaplan,B.K. (1989).Epidemiologyofdepressivesymptomsinyoungadolescents. JournaloftheAmericanAcademyofChildandAdolescentPsychiatry, 28, 343 3351. Gladis,M.M.,Gosch,E.A.,Dishuk,N.M.,&Crits-Christoph,P.(1999). Qualityoflife:Expandingthescopeofclinicalsignificance. Journalof ConsultingandClinicalPsychology,67, 320 3331. Goldman,H.H.,Skodol,A.E.,&Lave,T.R.(1992).RevisingAxisVfor DSM 3IV :Areviewofmeasuresofsocialfunctioning. AmericanJournal ofPsychiatry,149, 1148 31156. Greenberg,P.E.,Stiglin,L.E.,Finkelstein,S.N.,&Berndt,E.R.(1993). Theeconomicburdenofdepressionin1990. JournalofClinicalPsy- chiatry,54, 405 3418. Gurin,G.,Veroff,J.,&Feld,S.(1960). Americansviewtheirmental health. NewYork:BasicBooks. Harlow,H.(1958).Thenatureoflove. AmericanPsychologist,13, 573 3 685. Headey,B.W.,Kelley,J.,&Wearing,A.J.(1993).Dimensionsofmental health:Lifesatisfaction,positiveaffect,anxiety,anddepression. Social IndicatorsResearch,29, 63 382. Hollon,S.D.,Thase,M.E.,&Markowitz,J.C.(2002).Treatmentand preventionofdepression. PsychologicalScienceinthePublicInterest,2, 39 376. Jahoda,M.(1958). Currentconceptsofpositivementalhealth. NewYork: BasicBooks. 547 COMPLETESTATEMODELOFHEALTH Kessler,R.C.,Andrews,G.,Mroczek,D.,Ustun,B.,&Wittchen,H.-U. (1998).TheWorldHealthOrganizationCompositeInternationalDiag- nosticInterviewShortForm(CIDI-SF). InternationalJournalofMeth- odsinPsychiatricResearch,7, 171 3185. Kessler,R.C.,DuPont,R.L.,Berglund,P.,&Wittchen,H.-U.(1999). Impairmentinpureandcomorbidgeneralizedanxietydisorderand majordepressionat12monthsintwonationalsurveys. American JournalofPsychiatry,156, 1915 31923. Kessler,R.C.,McGonagle,K.A.,Zhao,S.,Nelson,C.B.,Hughes,M., Eshleman,S.,etal.(1994).Lifetimeand12-monthprevalenceof DSM 3 III 3R psychiatricdisordersintheUnitedStates:ResultsfromtheNa- tionalComorbiditySurvey. ArchivesofGeneralPsychiatry,51, 8 319. Kessler,R.C.,&Zhao,S.(1999).Theprevalenceofmentalillness.In A.V.Horwitz&T.L.Scheid(Eds.), Ahandbookforthestudyofmental health:Socialcontexts,theories,andsystems (pp.58 378).NewYork: CambridgeUniversityPress. Keyes,C.L.M.(1998).Socialwell-being. SocialPsychologyQuarterly, 61, 121 3140. Keyes,C.L.M.(2002).Thementalhealthcontinuum:Fromlanguishing toflourishinginlife. JournalofHealthandSocialBehavior,43, 207 3 222. Keyes,C.L.M.(2003a).Completementalhealth:Anagendaforthe21st century.InC.L.M.Keyes&J.Haidt(Eds.), Flourishing:Positive Psychologyandthelifewell-lived (pp.293 3312).Washington,DC: AmericanPsychologicalAssociation. Keyes,C.L.M.(2003b).Promotingalifeworthliving:Humandevelop- mentfromthevantagepointsofmentalillnessandmentalhealth.In R.M.Lerner,F.Jacobs,&D.Wertlieb(Eds.), Promotingpositivechild, adolescent,andfamilydevelopment:Ahandbookofprogramandpolicy innovations (Vol.4,pp.257 3274).ThousandOaks,CA:Sage. Keyes,C.L.M.(2004).Thenexusofcardiovasculardiseaseanddepres- sionrevisited:Thecompletementalhealthperspectiveandthemoder- atingroleofageandgender. AgingandMentalHealth,8, 266 3274. Keyes,C.L.M.,&Lopez,S.J.(2002).Towardascienceofmentalhealth: Positivedirectionsindiagnosisandinterventions.InC.R.Snyder&S.J. Lopez(Eds.), Handbookofpositivepsychology (pp.45 359).NewYork: OxfordUniversityPress. Keyes,C.L.M.,&Ryff,C.D.(2003).Somatizationandmentalhealth:A comparativestudyoftheidiomofdistresshypothesis. SocialScience andMedicine,57, 1833 31845. Keyes,C.L.M.,Shmotkin,D.,&Ryff,C.D.(2002).Optimizingwell- being:Theempiricalencounteroftwotraditions. JournalofPersonality andSocialPsychology,82, 1007 31022. Keyes,C.L.M.,&Waterman,M.B.(2003).Dimensionsofwell-being andmentalhealthinadulthood.InM.Bornstein,L.Davidson,C.L.M. Keyes,&K.Moore(Eds.), Well-being:Positivedevelopmentthrough- outthelifecourse (pp.477 3497).Mahwah,NJ:Erlbaum. Kish,L.,&Frankel,M.R.(1974).Inferencefromcomplexsamples. JournaloftheRoyalStatisticalSociety,36, 1 337. Luborsky,L.(1962).Clinicians 9judgmentsofmentalhealth. Archivesof GeneralPsychiatry,7, 407 3417. McGregor,I.,&Little,B.R.(1998).Personalprojects,happiness,and meaning:Ondoingwellandbeingyourself. JournalofPersonalityand SocialPsychology,74, 494 3512. Mechanic,D.(1999). Mentalhealthandsocialpolicy:Theemergenceof managedcare. NeedhamHeights,MA:Allyn&Bacon. Murray,C.J.L.,&Lopez,A.D.(Eds.).(1996). Theglobalburdenof disease:Acomprehensiveassessmentofmortalityanddisabilityfrom diseases,injuries,andriskfactorsin1990andprojectedto2020. Cambridge,MA:HarvardSchoolofPublicHealth. Regier,D.A.,Narrow,W.E.,Rae,D.S.,Manderscheid,R.W.,Locke, B.Z.,&Goodwin,F.K.(1993).ThedefactoU.S.mentalandaddictive disordersservicesystem:Epidemiologiccatchmentareaprospective 1-yearprevalenceratesofdisordersandservices. ArchivesofGeneral Psychiatry,50, 85 394. Reis,H.T.,&Gable,S.L.(2003).Towardapositivepsychologyof relationships.InC.L.M.Keyes&J.Haidt(Eds.), Flourishing:Positive psychologyandthelifewell-lived (pp.129 3159).Washington,DC: AmericanPsychologicalAssociation. Roy-Byrne,P.,Dagadakis,C.,Unutzer,J.,&Ries,R.(1996).Evidencefor limitedvalidityoftherevisedglobalassessmentoffunctioningscale. PsychiatricServices,47, 864 3866. Ryan,R.M.,&Deci,E.L.(2001).Onhappinessandhumanpotentials:A reviewofresearchonhedonicandeudaimonicwell-being. Annual ReviewofPsychology,52, 141 3166. Ryff,C.D.(1989).Happinessiseverything,orisit?Explorationsonthe meaningofpsychologicalwell-being. JournalofPersonalityandSocial Psychology,57, 1069 31081. Ryff,C.D.,&Keyes,C.L.M.(1995).Thestructureofpsychological well-beingrevisited. JournalofPersonalityandSocialPsychology,69, 719 3727. Ryff,C.D.,&Singer,B.(1998).Humanhealth:Newdirectionsforthe nextmillennium. PsychologicalInquiry,9, 69 385. Ryff,C.D.,&Singer,B.(2003).Flourishingunderfire:Resilienceasa prototypeofchallengedthriving.InC.L.M.Keyes&J.Haidt(Eds.), Flourishing:Positivepsychologyandthelifewell-lived (pp.15 336). Washington,DC:AmericanPsychologicalAssociation. Sartorius,N.(2001).Theeconomicandsocialburdenofdepression. JournalofClinicalPsychiatry,62, 8 311. Seligman,M.E.P.(1975). Helplessness:Ondepression,development,and death. SanFrancisco:Freeman. Sigerist,H.E.(1941). Medicineandhumanwelfare. NewHaven,CT:Yale UniversityPress. Smith,M.B.(1959).Researchstrategiestowardaconceptionofpositive mentalhealth. AmericanPsychologist,14, 673 3681. Tellegen,A.,Watson,D.,&Clark,L.A.(1999).Onthedimensionaland hierarchicalstructureofaffect. PsychologicalScience,10, 297 3303. U.S.PublicHealthService.(1998). Suicide:AreportoftheSurgeon General. Rockville,MD:Author. U.S.PublicHealthService.(1999). Mentalhealth:AreportoftheSurgeon General. Rockville,MD:Author. Waterman,A.S.(1993).Twoconceptionsofhappiness:Contrastsof personalexpressiveness(eudaimonia)andhedonicenjoyment. Journal ofPersonalityandSocialPsychology,64, 678 3691. Watson,D.,&Tellegen,A.(1985).Towardaconsensualstructureof mood. PsychologicalBulletin,98, 219 3235. WorldHealthOrganization.(1948).WorldHealthOrganizationConstitu- tion.In BasicDocuments. Geneva,Switzerland:Author. ReceivedFebruary11,2004 RevisionreceivedNovember6,2004 AcceptedNovember10,2004 548 KEYES

less