The differentiation between depressive and anxious adolescent females and controls by behavioural self-rating scales

The differentiation between depressive and anxious adolescent females and controls by behavioural self-rating scales
of 9
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
  Research report The differentiation between depressive and anxious adolescent females andcontrols by behavioural self-rating scales Eva Henje Blom a, ⁎ , Jan-Olov Larsson b , Eva Serlachius a,b , Martin Ingvar a a Department of Clinical Neuroscience, and Osher Center for Integrative Medicine, Karolinska Institutet, Sweden b Department of Woman and Child Health, Karolinska Institutet, Sweden a r t i c l e i n f o a b s t r a c t  Article history: Received 18 May 2009Received in revised form 7 July 2009Accepted 7 July 2009Available online 20 August 2009 Background:  This studyaimed to validate the ability of frequently used self-assessment scales inSwedish child and adolescent psychiatric practice to differentiate between adolescent girls withmanifest anxiety disorders and depression from those with less severe symptoms. Methods:  The receiver-operating characteristic (ROC) curve wascalculatedfor Beck's DepressionInventory(BDI),Beck'sAnxietyInventory(BAI),HospitalAnxietyandDepressionScale(HAD),theemotionalsubscale(SDQ-em),theimpactscoreandthetotaldif  fi cultiesscoreoftheStrengthsandDif  fi culties Questionnaire and Sense of Coherence (SOC) in a sample of 73 adolescent, femalepatients, diagnosed with one or several anxiety disorders and/or depression. ROC was alsocalculated for 66 age-matched controls. Results:  SOC and the SDQ-em showed the best ability to differentiate cases of anxiety disordersand/ or depression from non-cases. SOC and SDQ-em had an equivalent ability to differentiatedepression from non-cases compared to the specialised scales for depression, BDI and HAD-dep.SOC and SDQ-emwere signi fi cantlybetterindifferentiatingcasesofanxietyfrom non-casesthanthe specialised scales BAI and HAD-anx. Selection bias and several forms to  fi ll in can havein fl uenced the result. Conclusions:  SOCandSDQ-emseemedtobevalidtoolsforidentifyinggirlswithanxietydisordersand depression. This is of clinical importance since self-reported symptoms of anxiety anddepression show a major increase in adolescent girls and methods to identify those in need of treatment are needed.© 2009 Elsevier B.V. All rights reserved. Keywords: Adolescent femalesBehavioural self-rating scalesAnxiety disordersMajor depressionDifferentiating cases/ non-casesReceiver-operating characteristic curve 1. Introduction A systematic evaluation of the increasing prevalence of self-reported symptoms of anxiety and depression in adoles-centsisimportant.Theprevalenceofself-reportedsymptomsof anxiety in Swedish adolescents has increased more than threetimesduringthelasttwodecadesandwas33.1%infemalesand11.9%inmalesin2007(Statistics-Sweden,2007).Internationalstudies reveal great discrepancies in the prevalence of anxietydisorders, with one-year prevalence ranging from 2.8 to 18.4%in 15 – 18 year olds (Feehan et al.,1994; Fergusson et al.,1993;Goodwin et al., 2004; Lewinsohn et al., 1998; McGee et al.,1990). International studies also show that lifetime prevalenceof depression ranges from 9 to 20% at the end of adolescence(Costello et al., 2003; Fergusson et al., 2005; Lewinsohn et al.,1998; Reinherz et al.,1993). In Sweden, the cost of depressionhas doubled in the past eight years, making it a major publichealth concern (Sobocki et al., 2007).Psychiatric self-assessment scales aimed at measuringsymptoms of anxiety and depression can serve diversepurposes.Someinstrumentshavebeendesignedforsystema-tic screening, whereby adolescents with psychiatric symp-toms are identi fi ed at an early stage so that methods forpreventionofpsychiatricdiseaseprogressioncanbeprovided.  Journal of Affective Disorders 122 (2010) 232 – 240 ⁎  Corresponding author. Department of Clinical Neuroscience and OsherCenter for Integrative Medicine, Karolinska Institute, MR Research Center,N8:00, Karolinska University Hospital, 17176 Stockholm, Sweden. Tel.: +46738019212; fax: +46 8 51773266. E-mail address:  eva.henjeblom@ki.se (E. Henje Blom).0165-0327/$  –  see front matter © 2009 Elsevier B.V. All rights reserved.doi:10.1016/j.jad.2009.07.006 Contents lists available at ScienceDirect  Journal of Affective Disorders  journal homepage: www.elsevier.com/locate/jad  This is of importance, since previous research has shown thatmild psychiatric symptoms increase the risk of severepsychiatric disease and suicidal thoughts, minor/sub thresh-old depression, or dysthymia increase the risk of furthersevere psychiatric disease (Fergusson et al., 2005; Woodwardand Fergusson, 2001).It isa majoradvantage in clinicalpractice tohave accesstovalidated instruments to discriminate caseness from non-caseness.Byidentifyingspeci fi cthresholdscores,thescreeninginstruments can be used for such discriminant purposes. Theinstrumentscanthenserveasanaidtoschoolhealthservicesorgeneral practitioners in identifying adolescents who should bereferredforpsychiatricassessmentandinpotentiallyhelpingtoimprovetheef  fi cacyofintakeroutinesatopenpsychiatricunitsfor children and adolescents.If psychometric requirements for diagnostic indices areful fi lled, specialised instruments can also assist in meeting theclinical challenge of differentiating between anxiety disordersand depression and differentiating between different anxietydisorders, thus serving as a complement to clinical assessmentand diagnostic interviews. The instruments can also beused inresearch and clinical practice for assessing the severity of symptoms, or for measuring the outcome of treatment.There are several specialised instruments designed toidentify symptoms of either depression or anxiety. In thisstudy, we chose instruments which have been translated intoSwedish and are most frequently used in Swedish child andadolescent psychiatric clinical practice and mental healthservices. Beck's Depression Inventory (BDI) is a frequentlyused rating scale in adolescent psychiatry for screening andmeasuring symptoms of depression over a two-week periodprior to the assessment. A modi fi ed version, BDI-II, has beendeveloped and shown good psychometric properties in nonclinical adolescent samples (Osman et al., 2008). Likewise,Beck's Anxiety Inventory (BAI) is often used in adolescentresearch and clinical practice, with good psychometricproperties reported (Osman et al., 2002). The HospitalAnxiety and Depression Scale consists of both a depressionand an anxiety subscale (HAD-dep, HAD-anx) and wassrcinally created to detect states of depression and anxietyinthesettingofahospitalmedicaloutpatientclinic(ZigmondandSnaith,1983).Inareviewofthe747identi fi edpapersthatused HAD, the instrument was found to perform well inassessing the symptom severity and caseness of anxietydisorders and depression in both somatic, psychiatric andprimarycare patients and in the general population (Bjellandet al., 2002).In addition to the specialised instruments, we alsoincluded The Strengths and Dif  fi culties Questionnaire (SDQ)andTheSenseofCoherence(SOC)sincepreviousresearchhasshown that the accuracy of general purpose behaviour check-lists is high (Aebi et al., 2009; Rey and Morris-Yates,1992).SDQ is widely used internationally as a screening instru-ment for psychiatric symptoms in children and adolescents(Goodman et al., 2004) and currently an agreed coremeasuring instrument in the international Child and Adoles-cent Mental Health (CAMH) Outcome Research Consortium(CORC). SDQ covers emotional symptoms, hyperactivity,conduct and peer problems, but the emotional scale (SDQ-em) does not differentiate between symptoms of anxiety anddepression.Thescoresofeachsubscalecanbeaddedtoatotaldif  fi cultiesscore(SDQ-tot).Thereisalsoasubscalemeasuringpro-social behaviour and a special impact scale (SDQ-imp).Since there are three different versions of the SDQ, it ispossible to cross-check information from parents, teachersandself-reports.SDQwasoriginallydesignedforchildrenandadolescents up to 16 years of age but has been used for 17 – 19 year olds in Finland and Norway (Koskelainen et al., 2001;Van Roy et al., 2006). The Swedish self-report version of SDQ for adolescents has been shown to have acceptable psycho-metric properties (Lundh et al., 2008; Svedin and Priebe,2008).SOC is of special interest, as it represents a salutogenticmodel which focuses on factors that support human healthand well-being rather than on factors that cause disease. Nopsychiatric symptoms are measured but rather the globalorientation to one's inner and outer environments, which ishypothesized to be a signi fi cant determinant of location andmovement on the health ease-disease continuum. It does notrefer to a speci fi c type of coping strategy, but to factors whichare the basis for successfully coping with stressors. The SOCscale is considered to be a reliable, valid, and cross-culturallyapplicable instrument (Antonovsky and Sagy, 1986; Erikssonand Lindstrom, 2005), which is strongly related to perceivedhealth, especially mental health (Eriksson and Lindstrom,2006).In adolescents, there seem to be manifest psychologicaldifferencesbasedongenderwhichcontributetovulnerabilitytoanxietydisordersanddepression.Onesuchdifferenceisforgirls to exhibit a greater tendency to engage in ruminativethinking and increased interpersonal sensitivity (Breslauet al., 1995; Nolen-Hoeksema et al., 1999). A dramaticpredominance of mood disorders in females emerges at theonset of puberty (Angold et al., 1998). Psychiatric self-assessment scales are used for both boys and girls withouttaking gender into consideration. Our study did not aim tosolve this issue and only focused on girls in order to limitvariability in the sample.The over-all aim of this study was to rank the ability of established psychiatric self-assessment scales to differentiatebetween cases of affective disorders (one or several anxietydisorders and/or depression) and non-cases in adolescentgirls.Wealsowantedto fi ndthemosteffectiveinstrumentstospeci fi callyidentifycasesofanxietydisordersanddepression,and to identify cut-off scores for each scale. 2. Methods  2.1. Samples The clinical sample consisted of adolescent girls ( n =73)withameanage16.8 years(range14.5 – 18.4 years)whowerepsychiatricpatientsandhadaprimarydiagnosis,validatedbyDAWBA of major depression and/or one or several anxietydisorders (general anxiety syndrome, social phobia, speci fi cphobia, panic disorder, separation anxiety, post-traumaticstress disorder). Patients with severe autism or psychoticsymptoms were not included in the study. The subjects hadongoing treatment contact (median duration 11 months) atoneof13openpsychiatricclinicsforchildrenandadolescentssituatedin thecenterof Stockholm, itssuburbsandin smallertownsnearby.The fi rstauthorinformedthestaffattheclinics 233 E. Henje Blom et al. /  Journal of Affective Disorders 122 (2010) 232 –  240  (doctors, psychologists and social workers) about the studyandthestaffthenaskedtheirpatientsaboutparticipationandalso gave them written information. From what have beenreported from the staff 85% of the informed patientsparticipated. The reason to decline participation was fear of blood sampling (taken for other reasons) or parentaldisagreement.Clinical assessment and DAWBA-diagnosis (Developmentand Wellbeing Assessment) were used to establish casestatus. Eight subjects were excluded from the clinical samplebecause the DAWBA was incomplete or could not con fi rmdiagnosis of major depression or clinical anxiety.The control sample consisted of adolescent girls ( n =66),with a mean age of 16.5 years (range 15.9 – 17.7 years)recruited from one high school in a small rural town, one inStockholm city, one in an af  fl uent northernsuburb and one ina less af  fl uent southern suburb with a large immigrantpopulation. The  fi rst author met the students in theirclassrooms and gave them oral and written informationabout the study. About 80% of the informed studentsparticipated, the participation ratio being similar for allschools. The main reasons for declining to participate werefear of blood sampling and reluctance to miss school-hours.An informed consent form was signed by the subject andat least one parent, and all subjects received two cinematickets after completing the measurements. The study wasapproved by the Regional Ethics Committee at KarolinskaInstitutet. None of the authors reported any con fl ict of interest.  2.2. Instruments The subjects of the clinical sample completed the DAWBAand their forms at the open units whereas the controlscompleted their forms at the of  fi ces of the school nurse. Anendeavourwasmadetocollectalldataononeoccasion,butinsome cases, new appointments had to be made to completeall stages of the procedure. This involved testing one subjectat a time under the supervision of one or two assistants. Arandomly-sorted pack of forms was handed to each subject,who was then free to choose which order to complete them. Developmentand WellbeingAssessment(DAWBA) isasemi-structured diagnostic interview designed to generate ICD-10and DSM-IV psychiatric diagnoses on 5 – 17 year olds. DAWBAhas consistently generated sensible estimates of prevalenceand association with risk factors (Ford et al., 2003). Whencompared to clinical diagnoses, DAWBA diagnoses supportgood validity (Alyahri and Goodman, 2006; Goodman et al.,2000a; Mullick and Goodman, 2005). The DAWBA inter-viewers worked from computerized versions of the test inwhich all information had been collated by software thatpredicted the likely diagnosis. Two of the authors indepen-dently rated the computer-generated summary sheets anddiagnoses forms to decide whether or not to accept thecomputer diagnoses or lack of diagnoses. Hospital Anxiety and Depression Scale (HAD)  includesseven questions about depression and seven about anxiety(Zigmond and Snaith, 1983). Each question scores 0 – 3points. A total score of 8 – 10 points indicates mild to mode-rate symptoms; above 10 indicates a clinically signi fi cantcondition. Beck Depression Inventory (BDI)  consists of 21 items ratedona4-pointscaleandyieldsatotalscorebysummationoftheratings for the individual items (Beck et al., 1961). The totalscore is divided into four categories: no depression (0 – 9 p),mild depression (10 – 16 p), moderate depression (17 – 29 p)and severe depression (30 – 63 p). When this study wasdesigned, the BDI-II was not yet validated for the Swedishversion while BDI-A1 is used in this study. Beck Anxiety Inventory (BAI)  contains 21 items assessingthe degree to which the respondent has been affected by thephysical or cognitive symptoms of anxiety during the pastweek (Beck et al., 1988). BAI items are also meant to re fl ectpanic attack symptoms. The total score is divided into fourclinical categories: low degree of anxiety (0 – 9 p), middle tomoderate level of anxiety (10 – 18 p), moderate to severeanxiety (19 – 29 p) and severe anxiety (30 – 63 p). Strengths and Dif   fi culties Questionnaire(SDQ) ismadeupof 25 statements regarding psychological attributes and beha-viours, forming  fi ve subscales: hyperactivity/inattention,emotional symptoms, conduct problems, peer problems andpro-social behaviours (Goodman, 2001; Goodman et al.,2000b). Thepro-socialscalemeasuressocialcapacityincludinghelpfulness, generosityand empathy. The subscales score 0 – 10points and the total dif  fi culty scale scores 0 – 40 points. In thisstudy, only the self-report versionwas used. Sense of Coherence (SOC)  contains 29 items measuringsalutogenticfactors(Antonovsky,1993;AntonovskyandSagy,1986). Every item is rated on a 7-point scale giving amaximum score of 203. In adult populations, the averagehas been estimated to be 140. High scores indicate a goodsense of coherence.  2.3. Statistical analyses ROC-analyses were used to assess the diagnostic accuracyof the different tests and to compare the usefulness of thetests. A ROC-curve could be estimated by calculating thesensitivityandspeci fi cityofeach test atevery possible cut-off point and plotting sensitivity against speci fi city. The accuracyof a test could be estimated by the area under the ROC-curve(ROC-auc). ROC-aucs as a measure for predicting diagnosiscould be interpreted as follows: poor (0.50 – 0.70); moderateto fair (0.70 – 0.80); good (0.80 – 0.90), and excellent (0.90 – 1.00) (Ferdinand, 2008).For each scale based on the ROC report, the criterion valuecorresponding with the highest accuracy (minimal falsenegative and false positive results) was chosen as the cut-off score. ROC-aucs for different tests were compared with  t  -tests. Probability levels of 0.05 or less were consideredsigni fi cant and con fi dence intervals of 95% were reported.Analyses were done in Statistica 8.0 (www.statsoft.com) orStata 10 (www.stata.com). 3. Results  3.1. Sample characteristics 16.7% of the controls were recruited from the small ruraltown,21.2%fromtheaf  fl uentnorthernsuburb,24.2%fromthecity center, and 37.9% from the southern suburb. Parentunemployment and single-parent status did not differ 234  E. Henje Blom et al. /  Journal of Affective Disorders 122 (2010) 232 –  240  signi fi cantly between the clinical sample and controls, butparents of non-Swedish origin were less frequent in theclinical sample (Table 1).The median duration of treatment contact was 11 months.The DAWBA interview concluded that 19.2% of the subjectsful fi lled the criteria for major depression only, 32.9% for onlyone or several anxiety disorders, and 47.9% received thecombined diagnosis of both depression and one or severalanxiety disorders. The representation of different anxietydisorders varied and co-morbidity of several anxiety dis-orders was common. Of the patients diagnosed with anxietydisorders, 49.3% had only one anxiety syndrome, 26.0% hadtwo, and 4.1% had three or more anxiety disorders. Of the 73patients in the clinical sample 27 had co-morbid psychiatricdiagnosis in addition to anxiety disorders and/or depression.Of theses 27 patients 25 had one psychiatric diagnosis andtwo patients had two psychiatric diagnoses in addition to theanxiety disorder and/or depression. The clinical sampleincluded two patients with conduct disorders, three withAspergers syndrome, two with OCD, one with anorexia, ninewith bulimia and twelve with eating disorders UNS. Thisgroup did not show extreme scores on any of the assessmentscales. On the contrary they scored lower that the group withboth anxiety disorder and depression (data not shown). SDQ results for the present controls compared to other Scandina-vian community samples of adolescent girls show that theresults of SDQ-tot and SDQ-em were similar (Table 2).  3.2. Descriptive results for each instrument by the diagnosis groups and controls The mean scores of each scale and subscalewas comparedbetween the clinical sample and controls. All scales andsubscales showed signi fi cantly higher scores in the clinicalsample. The largest difference was found in SDQ-em (adj  Z  − 7.68), SDQ-imp (adj  Z  − .64) and SOC (adj  Z   7.24)  p N 0.0001(Table 3).When studying the clinical sample, all the scales (SOC,SDQ-tot, SDQ-em, SDQ-imp, BDI and HAD-anx and HAD-dep)showed that subjects with only anxiety disorders had thelowest score in each scale, subjects with only depression hadhigher scores, and subjects showing co-morbidity with bothanxiety disorders and depression had the highest scores(Table 4).  3.3. Discriminant ability of the instruments 3.3.1. Total clinical sample versus controls According to ROC-auc, the scales which were mostaccurate in discriminating between cases of anxiety disordersand/ or depression and non-cases were, in order of accuracy:SOC (0.89), SDQ-em (0.88), BDI (0.86), SDQ-imp (0.86), andSDQ-tot (0.84). The ROC-curve for the Sense of Coherencescale discriminating cases of anxiety disorders and/ordepression from non-cases is shown in Fig.1.There were no statistically signi fi cant differences betweenSOC and the other four scales. For each scale, cut-off scoresbased on the ROC report with the criterion value correspond-ing to the highest accuracy were calculated to differentiatebetween cases of anxiety disorders and/ or depression andnon-cases in adolescent girls (Table 5a).  3.3.2. Patients with depression versus controls The scales with best discriminant accuracy for cases of depression were: SOC (0.93) and BDI (0.89), SDQ-em (0.87),SDQ-tot (0.85) and HAD-dep (0.81). Compared with SOC, theBAI scale was statistically signi fi cantly less accurate in iden-tifying cases of depression (Table 5b).  3.3.3. Patients with one or several anxiety disorders versuscontrols The ability to discriminate between cases of anxietydisorders and non-cases was best for SOC (.84), SDQ-em(0.83), (SDQ-imp (0.77) and SDQ-tot (0.75). The specialisedanxiety scales were less accurate: BAI (0.69) and HAD-anx(0.71). Both SOC and SDQ-em were signi fi cantly better thanthe specialised anxiety scales BAI (0.69) and HAD-anx (0.71)in differentiating between cases of anxiety and non-cases(Table 5c).  3.3.4. Patients with both one or several anxiety disorders anddepression versus controls The co-morbid cases of both anxiety disorders anddepression were best differentiated with BDI (0.94), SOC(0.93), SDQ-em (0.92) and SDQ-imp (0.92) and no statisti-callysigni fi cantdifferenceswerefoundbetweenBDIandeachof the scales SOC, SDQ-em and SDQ-imp (Table 5d). 4. Discussion The main  fi nding of this study was that, of the assessedscales SOC and SDQ-em exhibited abilities in the range good-excellentindifferentiatingbetweencasesofanxietydisordersand/or depression from non-cases. SOC and SDQ-em also hadequivalent excellent ability to differentiate between cases of depression and non-cases compared to the specialised scalesfor depression BDI and HAD-dep. This compares to previousstudies investigating the discriminant properties of the ChildBehavioural Check List (CBCL) for adolescent depression(showing a ROC-auc of 0.78) and of the Youth Self Report  Table 1 Socio-demographic data in the clinical sample and in controls.ClinicalsampleControls  Chi  2 b N  =73  N  =66Parents' occupational statusBoth parents employed 56 49At least one parent unemployed 17 (23.3%) 16 (24.2%) –  1 missing 0.03 (ns)Family situationLiving with two parents 43 48Living with single parent 24 (32.9%) 18 (27.3%)Living at boarding school 3 a – 3 missing  –  0.78 (ns)Parents ethnic backgroundAt least one parent Swedish 66 50Both parents with non-Swedish background4 (6.1%) 16 (24.2%)3 missing  –  9.3 ⁎⁎ a Living at boarding school counted as living with two parents. b Ns=non-signi fi cant. ⁎⁎  p b 0.01.235 E. Henje Blom et al. /  Journal of Affective Disorders 122 (2010) 232 –  240  (YSR)(witha ROC-aucforadolescentgirlsof0.73(Ferdinand,2008; Reyand Morris-Yates,1991,1992)). A similar indicationfordiagnosticvalidityofdepressionhasalsobeenfoundintherecently developed version of YSR, and the DSM AFF scale,with a ROC-auc of 0.907 (Aebi et al., 2009) and 0.91(Ferdinand, 2008) in comparable samples.SOC and SDQ-em also showed good ability in differentiat-ing between cases of anxiety disorders and non-casescompared with the specialised scales BAI and HAD-anx. TheFerdinand study (Ferdinand, 2008) showed a ROC-auc of 0.65 – 0.70forCBCLindiscriminatingbetweencasesofanxietydisorders and non-cases and 0.64 – 0.76 for YSR.It is interesting that the SOC instrument evidences suchaccuracyin discriminatingbetweenpatientsandnonpatientsin this age-group since the items are not concerned withpsychiatric symptoms, but with ways of coping with innerand outer stressors. A review of 458 studies showed that theSOC is strongly related to perceived mental health (Erikssonand Lindstrom, 2007). The stronger the SOC, the better theperceived health in general, at least for those with an initiallyhigh SOC. This relationship was manifested in study popula-tions regardless of age, sex, ethnicity, nationality, and studydesign. The SOC seems to be able to predict health and is animportant contributor for the development and maintenanceof people's health (Eriksson and Lindstrom, 2006). Anto-novsky, who developed SOC, considered it to measure traitrather than state. The theoretical model of the SOC constructpostulates that a person's SOC is stabilized by the end of young adulthood, showing only minor  fl uctuations except inthe case of major changes in patterns of life experiences(Antonovsky, 1993). It was later shown that SOC exhibits acertain degree of stability in middle to late adolescence andthat relationships between SOC and healthy adaptation, maybe evident inyoungerage-groups thanpreviouslyanticipated(Torsheim et al., 2001). Test – retest correlations in adultsshow that SOC is comparatively stable after 10 years, but notas stable as Antonovsky initially assumed (Eriksson andLindstrom, 2005). Further studies are needed to investigateSOC as an outcome measure for psychiatric treatment inadolescents, as improvement of SOC scores in this age-groupmay predict future mental health to a better degree thanpsychiatric assessment scales.The discriminant power of SDQ-em and SDQ-tot wasalso an interesting  fi nding. This  fi nding is similar to those of a recent study conducted by Goodman, who reports thatSDQ seems to be a genuinely dimensional measure of childmental health (Goodman and Goodman, 2009). The proper-ties of the self-report subscales of SDQ have been analysedby latent trait theory and the RASCH model focusing on theoperating characteristics of the items, revealing the SDQ-emto be highly reliable compared to the other subscales(Hagquist, 2007). The positive properties of SDQ-em are inline with our  fi ndings of a good discriminant power. In theSvedin-study, the inter-correlation between SDQ-em andSDQ-tot was high (0.73), indicating that the two subscalesmeasure the same symptoms, and possibly explaining whySDQ-tot had unexpectedly high discriminant power foremotional disorders. The same study showed item-total  Table 2 SDQ means and standard deviations for the present controls compared to other Scandinavian community samples of adolescent girls.Psychiatric self-assessment scalePresentstudySwedish communitysample — LundhSwedish communitysample — SvedinNorwegian communitysample — RonningNorwegian communitysample — Van RoyFinnish communitysample — Koskelainen15 – 17 yearsold15 years old 17 – 19 years old 15 – 16 years old 16 – 19 years old 15 – 17 years N  =66  N  =498  N  =530  N  =290  N  =ca 3500  N  =ca 350SDQ-tot 10.18 (6.16) 10.40 (5.15) 11.54 (4.93) 10.59 (5.51) 11.3 (5.2) 11.8 (1.7)SDQ-em 3.73 (2.38) 3.17 (2.14) 3.95 (2.48) 3.01 (2.35) 3.6 (2.3) 3.6 (2.2)SDQ-imp 0.86 (1.71)  – – –  1.7 (2.0)  –  Table 3 Comparison of means and standard deviations of each scale and subscale inthe clinical sample and controls.Psychiatric self-assessment scaleClinical sample Controls Adjusted  Z  a Mean (SD)  N  b Mean (SD)  N  b BDI 25.10 (12.01) 67 9.85 (8.50) 66  − 7.05 ⁎⁎⁎ BAI 22.81 (11.00) 70 13.38 (9.69) 66  − 5.44 ⁎⁎⁎ HAD-dep 8.61 (4.31) 70 3.72 (3.44) 64  − 6.35 ⁎⁎⁎ HAD-anx 11.21 (4.00) 69 6.54 (4.27) 64  − 5.88 ⁎⁎⁎ SDQ-tot 18.25 (5.36) 73 10.18 (6.16) 66  − 6.70 ⁎⁎⁎ SDQ- em 7.018 (1.69) 73 3.73 (2.38) 66  − 7.68 ⁎⁎⁎ SDQ-imp 3.93 (2.43) 73 0.86 (1.71) 66  − 7.64 ⁎⁎⁎ SOC 95.8 (20.48) 69 137.1 (26.9) 50 7.24 ⁎⁎⁎ a Mann Whitney  U  -test (similar results when  t  -test for independentsamples was used). b Due to incomplete forms numbers are reduced compared to srcinalsamples. ⁎⁎⁎  p b 0.001.  Table 4 Mean values of each scale and subscales in the subgroups of the clinicalsample: anxiety disorder only, depression only and co-morbidity with bothdepression and anxiety.Psychiatric self-assessment scaleAnxiety disorderonlyDepressiononlyCombineddiagnoses F  -valueMean (SD)  n  Mean (SD)  n  Mean (SD)  n BDI 15.5 (9.1) 22 24.1 (8.0) 13 32.1 (10.6) 32 19.2 ⁎⁎⁎ BAI 18.0 (8.1) 23 17.0 (6.4) 12 28.0 (11.6) 35 9.8 ⁎⁎⁎ HAD-dep 9.2 (3.2) 23 11.1 (2.9) 12 12.6 (4.3) 34 6.0 ⁎⁎ HAD-anx 5.8 (3.2) 23 8.5 (4.3) 12 10.5 (4.0) 35 10.8 ⁎⁎⁎ SDQ-tot 15.5 (5.7) 24 18.5 (4.1) 14 20.0 (4.9) 35 5.8 ⁎⁎ SDQ-em 6.3 (1.7) 24 6.9 (1.3) 14 7.9 (1.5) 35 8.2 ⁎⁎ SDQ-imp 3.1 (2.5) 24 3.5 (2.2) 14 4.7 (2.3) 35 3.3 ⁎ SOC 107.7 (20.7) 24 95.8 (9.7) 12 87.1 (19.2) 33 8.5 ⁎⁎ Differences between groups were calculated by one-way ANOVA. ⁎  p b 0.05. ⁎⁎  p b 0.01. ⁎⁎⁎  p b 0.001.236  E. Henje Blom et al. /  Journal of Affective Disorders 122 (2010) 232 –  240
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!