Klinik Psikofarmakoloji Bülteni-Bulletin of Clinical Psychopharmacology ISSN: 1017-7833 (Print) 1302-9657 (Online) Journal homepage: https://www.tandfonline.com/loi/tbcp20 Comparative Validity and Reliability Study of The QIDS-SR16 in Turkish and American College Student Samples Haluk Mergen (Associate Professor of Family Medicine), Ira H. Bernstein (Professor of Clinical Sciences), Vedide Tavli (Professor of Paediatric Cardiology), Kurtulus Ongel (Associate Professor of Family Medicine), Talat Tavli (Professor of Cardiology) & Seref Tan (Associate Professor of Faculty of Education) To cite this article: Haluk Mergen (Associate Professor of Family Medicine), Ira H. Bernstein (Professor of Clinical Sciences), Vedide Tavli (Professor of Paediatric Cardiology), Kurtulus Ongel (Associate Professor of Family Medicine), Talat Tavli (Professor of Cardiology) & Seref Tan (Associate Professor of Faculty of Education) (2011) Comparative Validity and Reliability Study of The QIDS-SR16 in Turkish and American College Student Samples, Klinik Psikofarmakoloji Bülteni- Bulletin of Clinical Psychopharmacology, 21:4, 289-301, DOI: 10.5455/bcp.20110223124825 To link to this article: https://doi.org/10.5455/bcp.20110223124825 © 2011 Taylor and Francis Group, LLC Published online: 08 Nov 2016. Submit your article to this journal Article views: 207 View related articles Citing articles: 1 View citing articles Full Terms & Conditions of access and use can be found at https://www.tandfonline.com/action/journalInformation?journalCode=tbcp21 Araştırmalar / Original Papers DOI: 10.5455/BCP.20110223124825 Comparative Validity and Reliability Study of The QIDS-SR16 in Turkish and American College Student Samples Haluk Mergen1, Ira H. Bernstein2, Vedide Tavli3, Kurtulus Ongel4, Talat Tavli5, Seref Tan6 ÖZET: ABS TRACT: Türk ve Amerikalı üniversite öğrencilerinde hızlı Comparative validity and reliability study of depresif belirti envanteri-özbildirim formu’nun the QIDS-SR16 in Turkish and American college (HDBE16-ÖF) karşılaştırmalı olarak geçerlik ve student samples güvenirliği Objective: To evaluate the validity and reliability of the Amaç: Öğrenci ağırlıklı Türk örneklemine uygulanan Quick Inventory of Depressive Symptomatology, self- Türkçe’ye çevrilmiş 16 maddelik Hızlı Depresif Belirti reported version, in a Turkish student sample (QIDS-SR16-T) Envanteri-Özbildirim Formu’nun (HDBE16-ÖF): a) Amerikalı by comparing it to (a) the American version (QIDS-SR16- üniversite öğrencilerine uygulanan orijinal Amerikan ver- US) and (b) the Turkish version of the Beck Depression siyonu (QIDS-SR16-US) ve b) aynı Türk öğrenci örneklemin- Inventory (BDI-II-T). de Beck Depresyon Envanteri-II (BDI-II) ile karşılaştırılarak Materials and Methods: Slightly modified versions of geçerlik ve güvenirliğinin ortaya konması amaçlanmıştır. the QIDS-SR16-T, and the BDI-II-T were administered to Çalışmamız Türkiye ve Amerika Birleşik Devletleri arasında 626 outpatients at the Uludağ University campus-based yapılan bir kültürlerarası geçerlilik çalışmasıdır. family health center. The QIDS-SR16-US was administered Metod: Uludağ Üniversitesi yerleşkesi Aile Sağlığı to 584 respondents at an American university. SAS and Merkezi’ne ayaktan başvuran öğrenci ağırlıklı 626 hasta- MPlus were used to provide descriptive statistics, classical ya; www.ids-qids.org adresinden ulaşılabilen ve kısmen exploratory factor analysis, and item response theory modifiye edilerek Türkçe’ye çevrilen HDBE16-ÖF ve BDI-II analyses (in the form of a multiple group confirmatory testleri uygulandı. Ayrıca Güneybatı Teksas Üniversitesi’nde factor analysis). HDBE16-ÖF envanterinin İngilizce orijinal versiyonu olan Results: The internal consistency (Cronbach α) of the QIDS-SR16-US 584 öğrenciye uygulanmıştır. Betimleyici ista- QIDS-SR16-T was 0.77. Both QIDS-SR16 versions were tistik, klasik açıklayıcı faktör analizi ve madde tepki kuramı unidimensional, but the BDI-II-T was not. The mean QIDS- analizleri, SAS ve MPlus istatistik programları ile yapılmıştır. SR16-T and QIDS-SR16-US item-total correlations were Bulgular: Türk deneklerin ortalama yaşı 21,1±2,16 (stan- similar. The correlation between the QIDS-SR16-T and BDI- dart sapma) olup %67,8’i kadındı. Türk öğrencilerin aile II-T was 0.72 (.90 when disattenuated). Multiple-group içi depresyon öyküsü: annede %29, babada %8, kardeşte confirmatory factor analysis suggested that the QIDS- 1 %14, kendisinde %16 ve akrabada %5 olarak bulundu. SR16-T and QIDS-SR16-US had the same factor loadings but Associate Professor of Family Medicine, Amerikalı deneklerden 225 olguda hiç yaş belirtilmemiş different intercepts. This reflects group differences in level Uludağ University Family Health Center, Bursa-Turkey haldeyken ortalama yaş 20.0±3,5 (standart sapma) ve tüm of depression, perhaps because the Turkish respondents, 2Professor of Clinical Sciences, University of deneklerin %63,6’sı kadın olarak saptandı. HDBE16-ÖF’nın unlike their US counterparts, were seen in a medical context Texas Southwestern Medical Center, Dallas, USA 3 madde ortalaması 6,94±4,85 (standart sapma) bulundu. where illness-related depression is more prevalent. Scores Professor of Paediatric Cardiology, Yeditepe HDBE16-ÖF’ün iç tutarlılık katsayısı (Cronbach α) 0,78 idi ve on the QIDS-SR16-T and the BDI-II-T were also equated. University, İstanbul-Turkey 4Associate Professor of Family Medicine, ortalama madde-toplam korrelasyon katayısı 0,47 (0,33- Discussion: The QIDS-SR16-T has good psychometric Tepecik Research & Training Hospital, 0,61) bulundu. QIDS-SR16-US’nin kaşılaştırılabilir madde properties and convergent validity with the BDI-II-T. Its İzmir-Turkey 5 ortalaması 6,09±3,76, Cronbach α 0,74, madde-toplam use is recommended when a self-reported instrument is Professor of Cardiology, Celal Bayar korrelasyon katsayısı 0,43 (0,24-0,54) olarak bulundu. appropriate. University, Manisa-Turkey6Associate Professor of Faculty of Education, Hem HDBE16-ÖF hem de QIDS-SR16-US tek boyutlu iken BDI- Uludağ University, Bursa-Turkey II tek boyutlu olarak bulunmadı. HDBE16-ÖF’ün ve QIDS- Key words: Major depressive episode, screening scale, SR16-US’un madde-total korelasyon ortalaması birbirine validity, reliability Ya zışm a Ad re si / Addr ess rep rint re qu ests to: benzerdi. BDI-II ile HDBE16-ÖF arasındaki korelasyon katsa- Haluk Mergen, M.D., A.S., Uludağ Universitesi Aile Sağlığı Merkezi, Görükle, Bursa-Turkey yısı 0.72 bulundu, bu değer disattenüe edildiğinde 0.90’a Bulletin of Clinical Psychopharmacology 2011;21(4):289-301 çıkmaktaydı. Çoklu grup doğrulayıcı faktör analizi HDBE16- Telefon / Phone: +90-532-441-9651 ÖF ve QIDS-SR16-US’un aynı faktör yüküne sahip olduğu farklı değişik eşiklerinin olduğunu ortaya çıkarmıştır. Bu Elektr o nik pos ta ad re si / E-mai l add ress:haluk.mergen@gmail.com durum depresyon düzeyinde grup farklılıklarını ortaya koymaktadır. Türk deneklerin, Amerikalı deneklerden farklı Gönderme tarihi / Date of submission: olarak daha fazla depresyon geçirdikleri söylenebilir. Ayrıca 01 Şubat 2011 / February 01, 2011 HDBE16-ÖF ve BDI-II’nın skorları birbirlerine eşitlenmiştir. Kabul tarihi / Date of acceptance: Tartışma: HDBE16-ÖF’ün, hem Türkiye hem de Amerika 23 Şubat 2011 / February 23, 2011 Birleşik Devletleri’nde depresyon tanısı için çok sık kulla- nılan BDI-II testi gibi iyi psikometrik özellikleri ve yapısal geçerliliği olduğu saptanmıştır. Pek çok ortamda HDBE - Bağıntı beyanı:16 H.M., I.H.B., V.T., K.Ö., T.T., S.T.: Yazarlar bu ÖF’ün kullanılmasını önermekteyiz. makale ile ilgili olarak herhangi bir çıkar çatışması bildirmemişlerdir. Anahtar sözcükler: Majör depresif epizod, tarama testi, geçerlilik, güvenilirlik Declaration of interest:H.M., I.H.B., V.T., K.Ö., T.T., S.T.: The authors reported noconflict of interest related to this Kli nik Psikofarmakoloji Bülteni 2011;21(4):289-301 article. Klinik Psikofarmakoloji Bülteni, Cilt: 21, Sayı: 4, 2011 / Bulletin of Clinical Psychopharmacology, Vol: 21, N.: 4, 2011 - www.psikofarmakoloji.org 289 Comparative validity and reliability study of the QIDS-SR16 in Turkish and American college student samples INTRODUCTION few minor changes were made to accommodate actual Turkish usage, e.g., references to weight in pounds was A major depressive episode (MDE) must last at least replaced by weight in kilograms. The specific translation two weeks and involve five of nine core symptoms: (a) of the QIDS-SR16-T used in this study is available from the sleep disturbance, (b) sad mood, (c) change in appetite first author. The QIDS-SR16 requires minimal training to and/or weight, (d) difficulty in concentration and decision administer and is freely available for use (7). The 16 items making, (e) negative self view, (f) thoughts of death or span the nine core symptom domains as defined by the text suicide, (g) loss of general interest, (h) reduced energy revision of version IV of the American Psychiatric level, and (i) restlessness or agitation. One of the symptoms Association’s Diagnostic and Statistical Manual of Mental must be (b) or (g) (1). Estimates of lifetime MDE Disorders (DSM-IV-TR) (8). These items are sleep prevalence in different countries r ange from 5 to 17% disturbance (items 1-4), depressed (sad) mood (item 5), with an average estimate of 12%. Various prevalence change in appetite or weight (items 6-9), concentration/ estimates obtained in Turkey include 8.4% (2), 26.2% (3), decision making (item 10), self view (item 11), suicidal and 39.4% (4). ideation (item 12), interest (item 13), energy/fatigue (item Because of the high prevalence of depression, accurate, 14), and psychomotor agitation/retardation (items 15 and time-efficient measurement of depressive symptom 16) (8). The determination of remission or partial remission severity is of great importance in conducting cost-efficient is based on the DSM-IV-TR, which recommends that all clinical trials (5). Self-reports are useful to both clinicians nine diagnostic criterion symptoms be assessed (5,8). The and researchers who wish to monitor treatment outcomes responses for each item range from zero to three, with zero in a time- and cost-effective manner. indicating the absence of that symptom in the past week, One way to evaluate the translation of a test is to and the remaining categories defining mild, moderate, and compare the results from that sample to one obtained from severe presence of the symptom in question in that time a sample that is fluent in the original language. In the period. The scoring scheme involves adding the scores for present case, we had samples of Turkish and American the nine symptom domains to yield a total score that college students. These samples are similar in terms of consequently ranges from 0 to 27. QIDS-SR16-US scores cognitive abilities, but as will be noted, are not similar in are commonly interpreted as follows: no depression (0-5), level of depression. Another aspect of the evaluation is to mild depression (6-10), moderate depression (11-15), see how well the measure correlates with an accepted severe depression (16-20), and very severe depression measure, both of which are available as translations, i.e., (≥21) (6); however, the Turkish version of the QIDS-SR16, its convergent validity. In the present case, the Beck QIDS-SR16-T, has not been evaluated for its reliability and Depression Inventory-II (BDI-II) is one such instrument. validity. The purpose of this study was to perform this This study evaluated the validity and reliability studies of evaluation. the QIDS-SR16-T by comparing our results obtained in our The inventory used for comparison, was the Beck sample to the American version, the QIDS-SR16-US. and Depression Inventory-II, a revision of the original scale by correlating the QIDS-SR16-T version with the BDI- published in 1996 by Beck et al (9,10). The Turkish II-T. version, the Beck Depression Inventory-II-Turkish (BDI- II-T), was validated by Tegin (11). The main advantages METHODS of the QIDS-SR16 are its brevity and survey of the nine major psychiatric symptoms as listed in the DSM-IV-TR. The 16 item self-reported version of the Quick In comparison, the BDI-II includes 21 items, the Hamilton Inventory of Depressive Symptomatology (QIDS-SR16) depression scale has 17, 21, 24, 28, or 31 items and the was first described by Rush et al. (2003) (6) and also exists Zung Self-Depression Scale has 20 items. The QIDS-SR16 in clinician rated and interactive voice formats (6). The also provides: 1) equivalent weightings (0-3) for each QIDS-SR16-T was obtained from the www.ids-qids.org symptom item, 2) clearly stated anchors that estimate the website. The translation and back-translation of the QIDS- frequency and severity of symptoms; and 3) matched SR16 was done by a translation team into 30 languages. A clinician and patient ratings between the scale’s clinician 290 Klinik Psikofarmakoloji Bülteni, Cilt: 21, Sayı: 4, 2011 / Bulletin of Clinical Psychopharmacology, Vol: 21, N.: 4, 2011 - www.psikofarmakoloji.org H. Mergen, I. H. Bernstein, V. Tavli, K. Ongel, T. Tavli, S. Tan and self-reported forms (6,12). columns and 583 rows for the QIDS-SR16-US, and 21 The Turkish sample consisted of students, who columns and 626 rows for the BDI-T. Thus, the number of presented consecutively to the outpatient clinic of the columns equaled the number of items on the actual test and university. On the other hand, the US sample was composed the number of rows equaled the number of participants in of students taking a course in introductory psychology at a the relevant sample. Each of the resulting 150 matrices university in the Southwestern United States. Since they was then subjected to principal component analysis and did not present or seek help at a clinic, this sample was the eigenvalues averaged within each of the three sets. A unselected with regard to psychological problems and thus test was considered unidimensional if (a) its first eigenvalue probably was less depressed as a group than the Turkish exceeded the average of the simulated eigenvalues for that sample. One case was eliminated from each sample due to test and (b) all subsequent obtained eigenvalues were errors in filling out the answer sheets. smaller than the corresponding simulated eigenvalues. The Turkish study conformed to the Helsinki Because the essential features of the exploratory analysis declaration requirements and had approval from the were also present in the item response theory analysis and University ethics committee. The Turkish students gave the latter will be discussed starting in the next paragraph, informed consent and voluntarily participated in the study. the former will not be presented except to note that loadings They completed the QIDS-SR16-T and BDI-II-T in 7-10 on the first (only) factor ranged from .45 (item 6, suicidal minutes. Out of 670 patients’ questionnaires, 44 ideation) to .74 (item 2, sad mood). All of these analyses questionnaires were eliminated because of problems such used SAS 9.2 as failure to complete the two tests. The gathering of the Because we were not comparing the Turkish sample 584 QIDS-SR16-US questionnaires met similar ethical with that of another nationality, the item response theory guidelines and was approved by the University Institutional analysis for the BDI-II-T follows the standard two Review Board. One US questionnaire was eliminated due parameter logistic polytomous format (17,18). Thus, the to a coding error. Thus, the Turkish and US samples probability of an affirmative response is given by consisted of 626 and 583 respondents, respectively. P(θ ) 1/ij = 1 + eai(θ-b ) . In this equation, ai describes the 21 item ij slopes (factor loadings) and bij describes the 63 (21 items Statistical Analyses x 3 dichotomous criteria) intercepts (locations) of the functions. These dichotomies are 0 vs. 1, 2, or 3 (no Descriptive analyses, classical test theory, exploratory pathology vs. at least some pathology), 0 or 1 vs. 2 or 3 (no factor analysis and item response theory analyses were or mild pathology vs. moderate or severe pathology), and used. An item response analysis is a form of confirmatory 0, 1, or 2 vs. 3 (no pathology, mild pathology, or moderate factor analysis that tests proposed item structures. A single pathology vs. severe pathology). Finally, θ denotes the group analysis was performed in the case of BDI-T, and magnitude of the latent variable (depression in this case). multiple group (Turkish vs. US) analyses in the case of the In order to make the equations estimable, the value of ai is QIDS-SR16. The descriptive analyses were evaluated using the same for all three intercepts so they form a set of three χ2 tests for discrete variables and t-tests for continuous (in the present case) parallel functions. The larger ai is, the variables. The classical test theory analyses generated item steeper the slope of the functions, and the more means, item standard deviations, item-total correlations discriminating they are with regard to θ. Though it comes (rit), scale means, and scale standard deviations. from a different theoretical perspective, ai serves a similar Exploratory factor analysis was used to evaluate the role to the rit of classical test theory, and the values of ai dimensionality of the three tests. This analysis is important and rit are typically correlated over items. The bij describes for its own sake, but is also important for the item response the tendency to choose the higher category of each analyses as well. Parallel analysis (13-16) was used to dichotomy. The higher its value, the more likely the lower decide upon the number of factors (dimensionality). The alternative is chosen and, if at 0, the intercept is said to be version used in this study consisted of generating a series at threshold, as the probabilities of the two alternatives are of 50 matrices of random normal deviates. Each matrix each .5. The values of bij serve a role similar to that of the had 9 columns and 626 rows for the QIDS-SR16-T, 9 item means in classical test theory, but there are three such Klinik Psikofarmakoloji Bülteni, Cilt: 21, Sayı: 4, 2011 / Bulletin of Clinical Psychopharmacology, Vol: 21, N.: 4, 2011 - www.psikofarmakoloji.org 291 Comparative validity and reliability study of the QIDS-SR16 in Turkish and American college student samples values/item in the present cast per set of functions vs. item letting ai vary freely, so it assumed equal intercepts in the one mean. Both ai and bij are scaled in a z-score metric. two samples. Model 3 constrained both ai and bij to The θ also has a factor variance, which was set at 1.0 to equality, thus assuming both equal slopes and equal define the unit of measurement. intercepts, but allowed the group means to differ. This The two versions of the QIDS-SR16 generate a series of assumes that any difference between nationalities is multiple group models, which evaluate differences proportional to the factor loadings and also assumes equal between the Turkish and US samples, thus complicating locations. Finally, model 4 constrained ai and bij to equality the basic Samejima model. The series incorporate various while constraining the group means to 0, i.e., effectively nested comparisons as described below. Two models are treating the groups as identical. Model 1 was evaluated by nested when one is made more specific than the other in a testing its overall fit in terms of χ2 whereas the remaining particular way. In the present case, this meant comparing models were evaluated by means of the Bentler-Satorra χ2 the fit of one model where certain parameters in the difference test (19). The latter is needed because weighted Turkish and US model were allowed to differ vs. a model least squares estimation was used rather than maximum in which they were required to be the same. The difference likelihood estimation, given the categorical nature of the in fit between the two models can be tested to see if the responses, so simple differences in χ2 could not be used to equality constraint is tenable vs. whether the two groups test the national differences. Models 2a, 2b, and 3 were require separate sets of parameters. There are a total of five tested against model 1, and model 4 was tested against models, described more fully in the next paragraph as model 3. However, even well-fitting models may generate models 1, 2a, 2b, 3, and 4. Each model generates five sets a significant χ2 because of incidental issues like of estimates/group. By analogy to the BDI-II-T, there are nonnormality. As a result, descriptive measures of fit have 9 slopes which describe the strengths of relation between assumed greater importance in evaluating models. In the nine domains (symptoms) and (θ) overall depression. particular, it is common to require the confirmatory fit Following standard terminology, these are denoted ai and index (CFI) to exceed .95 and the root mean square error there are 27 locations which describe the tendencies to of approximation (RMSEA) to be less than .05, and this choose the more pathological of the particular dichotomy. strategy was emphasized in this study. In general, if a These are denoted bij where i denoted the 9 domains and j particular model is unacceptable, further constraints on denoted the 3 dichotomies per domain as described with that model are also unacceptable. As a results, if model 3 the analysis of the BDI. There is also a group mean for θ can be rejected, so can model 4, though all results will be for each nationality, which was set at 0 in all but one of the presented for completeness. nested comparisons to be presented. Fourth, there is a The chosen model also provides a test information variance of θ for each nationality, which was always 1.0. function (TIF) for each of the three tests. This describes Finally, there is a residual variance associated with each of the sensitivity of θ to change as a function of its level. The the nine QIDS-SR16 domains. These factor variances were TIF serves a similar purpose to coefficient α in classical all fixed at 1.0, and the residual variances were all fixed at test theory but is a function of θ instead of a constant. 1.0, though the results of additional analyses, not reported, Finally, we used the procedure described in (20) to equate allowing the factor variances to differ from 1.0, did not the QIDS-SR16-T to the BDI-II-T. This involves finding differ materially from those to be presented. the expected a posteriori value of θ for each raw score. The first of the series of QIDS-SR16 models, model 1, Scores on the two tests are considered equated if they have allowed the values of ai and bij to differ for the Turkish and the same or highly similar values of θ. This assumption is US groups, i.e., the models were fit separately to the two met because the same participants took both tests. Mplus nationalities. Four additional models introduced various was used for these item response theory analyses. constraints. Model 2a constrained ai to equality between groups while letting values of bij vary freely. Thus, the RESULTS model assumed equal slopes (discriminations, factor loadings) in the Turkish and US samples. Conversely, The mean age of the Turkish participants was 21.1±2.16 model 2b constrained bij to equality within groups while years and 67.8% were female. The demographic data of 292 Klinik Psikofarmakoloji Bülteni, Cilt: 21, Sayı: 4, 2011 / Bulletin of Clinical Psychopharmacology, Vol: 21, N.: 4, 2011 - www.psikofarmakoloji.org H. Mergen, I. H. Bernstein, V. Tavli, K. Ongel, T. Tavli, S. Tan Tab le 1: Sociodemographic and medical properties of the these gender differences ranged from 2.09 to 4.79, and the Turkish subjects. associated p values ranged from .023 to .0001. In addition, N % greater irritability was associated with higher monthly Age income, t= 2.05, p= .041. Thoughts of death or suicide <20 years old 253 40.3 20-24 years old 354 56.4 were more frequent in students with a family history of >24 years old 21 3.3 depression, t= 2.86, p= .004 and in older students, t= 2.34, Education Illiterate 6 1 p= .019. Low energy level was also more common among Primary school 3 0.5 those with a family history of depression, t= 2.70, p= High school 18 2.9 .007. University 601 95.7 Income <500TL 243 38.7 CTT Analysis 501-1000TL 218 34.7 1001-2000TL 138 22.0 >2000TL 29 4.6 The scale mean (sd) of the QIDS-SR16-T was 6.94±4.85. Work The internal consistency reliability (Cronbach’s α) was Does not work 523 83.3 Does work 105 16.7 0.78, and the mean rit was 0.47 with a range of rit values Dwelling from 0.33 to 0.61. The comparable mean (sd) of the QIDS- Village 86 13.7 Town 88 14 SR16-US was 6.09±3.76. Cronbach’s α was .74, the mean City 454 72.3 rit was .43, and the range of rit was from .24 to 54. Number of sibling Table 2 contains the QIDS-SR16-T and QIDS-SR16-US No sibling 66 10.5 1 sibling 194 30.9 item statistics. The QIDS-SR16-T mean was significantly 2 siblings 135 21.5 higher than the QIDS-SR16-US mean, t= 3.35, p< .0001. 3 siblings 122 19.4 >3 siblings 111 17.7 Table 3 contains the comparable BDI-II-T statistics. Family Depression History Within the Turkish sample, the correlation between the N/A 347 55.3 QIDS-SR16 and the BDI21 was .75. Disattenuating by No 224 35.7 Yes 57 9.1 dividing this obtained correlation by the square root of the Note: percentages for a given variable may not add to exactly 100.0 because of product of the alpha coefficients (21) indicates that the rounding error. correlation between the underlying traits measured by the two scales is .90. Thus, they are highly similar, but not outpatients are illustrated in Table 1. Some of the students identical, measures of depressive symptomatology. had a family history of depression as well: mother 29%, father 8%, sibling 14%, self 16%, and relative 5%. The mean age of the American subjects was 20.0±3.5 years Tab le 2: QIDS-SR16 item and scale statistics: item means, with 225 missing observations, and 63.6% were female. item standard deviations, item-total correlations (rit), sample sizes (N), raw Cronbach’s α, scale means and scale standard No other demographic data were available. The Turkish deviations for Turkish and US samples subjects were slightly, but significantly older than the Turkey US American subjects, t (987) = 4.35, p < .0001. However, the Domain Mean Std rit Mean Std rit percentages of males and females were the same in the two 1 1.65 .94 .37 1.65 .80 .24 ethnic groups, χ2 (1) = 1.84, p = 0.17. 2 .77 .89 .61 .63 .68 .54 Both QIDS-SR -T and BDI-II-T scores were higher 3 .86 .93 .39 .91 .88 .3416 4 .92 .86 .61 .57 .70 .49 in women than men (t=2.97, p=0.03 and t=2.61, p=0.009). 5 .76 1.23 .49 .43 .80 .39 QIDS-SR -T scores related to a family history of 6 .13 .40 .33 .23 .54 .4116 7 .43 .74 .44 .38 .67 .45 depression (t=2.08, p=0.038). The incidence of depressive 8 .51 .73 .51 .55 .69 .49 symptomatology was significantly greater among females 9 .89 1.12 .52 .75 .75 .52 than males in six domains: (a) sleep disorders, (b) weight N 626 583 Raw α .78 .74 problems, (c) irritability, (d) sadness, (e) concentration/ Scale Mean 6.94 6.09 decision making, and (f) energy level. The t values for Scale Std 4.85 3.76 Klinik Psikofarmakoloji Bülteni, Cilt: 21, Sayı: 4, 2011 / Bulletin of Clinical Psychopharmacology, Vol: 21, N.: 4, 2011 - www.psikofarmakoloji.org 293 Comparative validity and reliability study of the QIDS-SR16 in Turkish and American college student samples Tab le 3: BDI-II-T item and scale statistics: item means, item were less than their simulated counterparts). Following a standard deviations, item-total correlations (rit), sample size promax rotation, the two Beck factors correlated .42. (N), raw Cronbach’s α, scale mean and scale standard deviation Equally important, the first factor accounted for Item Mean Std rit considerably more variance (22%) than the second (8%), 1 .40 .68 .63 so the latter is fairly minor. This second factor is also 2 .34 .53 .49 3 .26 .60 .44 difficult to interpret as was defined by items number 16 4 .50 .60 .62 (sleep problems), 18 (appetite), 19 (weight loss), 20 5 .60 .61 .60 (hypochondriasis), and 21 (interest in sexuality). These 6 .42 .73 .50 7 .33 .59 .65 results indicate the assumption that a scale measures only 8 .49 .67 .57 one dimension, which is relevant to the item response 9 .09 .38 .37 10 .50 .97 .45 analysis of the next section, is met for both versions of the 11 .55 .80 .49 QIDS-SR16, but the Beck analysis should be interpreted 12 .46 .71 .62 with some caution because of the presence of a minor 13 .59 .78 .65 14 .23 .63 .45 second factor. 15 .54 .64 .55 16 .47 .68 .49 17 .33 .64 .51 Main IRT Analyses 18 .24 .52 .42 19 .09 .33 .23 Table 4 contains the fit statistics for the various multiple 20 .28 .57 .32 21 .34 .73 .37 group models. The model χ2 for the baseline model 1 N 620 testing the two forms of the QIDS-SR16, which fit Raw α .89 St. Alpha .89 parameters separately to the two groups, was 133.68 and is Scale Mean 8.03 significant beyond the .0001 level. However, the CFI for Scale Std 7.61 model 1 was .981 and the RMSEA was .049 so the fit of this baseline model is acceptable in descriptive terms. The Exploratory Factor Analysis (Dimensionality) difference χ2 comparing models 1 and 2a was a significant 19.56 on 9 df, p < .05, but the CFI and RMSEA were The QIDS-SR16-T and the QIDS-SR16-US were both identical to that observed with model 1. Conversely, the χ 2 unidimensional by the parallel analysis criterion. comparing models 1 and 2b was 208.44 on 27 df, and the Specifically, the first eigenvalues of these two tests were CFI and RMSEA were .926 and .079. Model 2b can be 3.39 and 3.11 vs. 1.19 and 1.19 for the simulated data, and rejected by these criteria. Accordingly model 2a was the second eigenvalues were .92 and 1.02 vs. 1.13 and 1.13 tentatively accepted, i.e., it was assumed that QIDS-SR16-T for the simulated data. Additional obtained eigenvalues and QIDS-SR16-US have the same slopes (relations of θ to were all less than their simulated counterparts. However, the nine domains) but different thresholds (levels) in the the BDI-II-T was two dimensional as its first three two groups. The poor fit of model 2b makes testing of eigenvalues were 6.79, 1.46, and 1.10 vs. 1.34, 1.28, and model 3 irrelevant which, in turn, also makes testing model 1.24 for the simulated data (all subsequent real eigenvalues 4 irrelevant. Tab le 4: Fits of the multiple group models Model Intercepts Slopes Means χ² df p RMSEA CFI 1 Free Free Constrained 133.68 54 .00 .05 .98 2a Free Constrained Constrained 19.56 9 .02 .04 .98 2b Constrained Free Constrained 208.44 27 .00 .08 .93 3 Constrained Constrained Free 24.97 35 .00 .08 .92 4 Constrained Constrained Constrained 14. 1 .00 .08 .91 Note: RMSEA – root-mean square error of approximation, and CFI = comparative fit index. The χ2 for model 1 tests the overall fit of the model whereas the χ2 for the remaining models tests the difference in fit (model 1 is used to test models 2a, 2b, and3, and model 3 is used to test model 2). As a form of weighted least squares was used in testing, the Satorra-Bentler adjustment was employed in model testing. 294 Klinik Psikofarmakoloji Bülteni, Cilt: 21, Sayı: 4, 2011 / Bulletin of Clinical Psychopharmacology, Vol: 21, N.: 4, 2011 - www.psikofarmakoloji.org H. Mergen, I. H. Bernstein, V. Tavli, K. Ongel, T. Tavli, S. Tan Figure 2: Common domain slopes Turkish and US groups. Note that contrary to what was obtained in the classical analysis of the total tests scores, QIDS-SR16-US 0 vs., 1-2 thresholds were actually lower than QIDS-SR16-T 0 vs., 1-3 thresholds, implying that American respondents were more willing to report a pathological category than Turkish respondents. However, the remaining two comparisons were in line with the classical test theory analysis; QIDS-SR16-T thresholds were lower than QIDS-SR16-US thresholds 8 of 9 times in each case, implying the Turkish participants were the more willing to endorse the moderate and severe categories of depressive pathology.. Figure 2 contains the common values of the slope. As is usually the case, these slopes parallel the item-total correlations of the classical test analysis. Domain 2 (sad mood) has the highest value (is most discriminating) and domains 1 and 3 the lowest (are least discriminating). The remaining domains are closer to domain 2 than domains 1 and 3. Figures 1 and 2 thus describe the main features of the QIDS-SR16 item structures. The BDI’s structure is of lesser interest and does not meet the item response theory model’s assumption that a single dimension underlies the data so it will not be presented. However, it is available upon request from the authors. Figure 1: Domain thresholds for the three criteria (0 vs. 1-3 in the top panel, 0 and 1 vs. 2 and 3 in the middle panel, and 0-2 vs. 3 in the bottom panel), separately for the Turkish and US Test Information Functions groups. Figure 3 contains the test information functions (TIF) Figure 1 contains the thresholds for the three criteria (0 for the Turkish QIDS-SR16, the QIDS-SR16-US and the vs. 1-3 in the top panel, 0-1 vs. 2- 3 in the middle panel, BDI-II-T. As noted, these represent the change in the and 0-2 vs. 3 in the bottom panel), separately for the respective measure per change in depression (θ generically). Klinik Psikofarmakoloji Bülteni, Cilt: 21, Sayı: 4, 2011 / Bulletin of Clinical Psychopharmacology, Vol: 21, N.: 4, 2011 - www.psikofarmakoloji.org 295 Comparative validity and reliability study of the QIDS-SR16 in Turkish and American college student samples Tabl e 5: Results of equating the QIDS-SR16-T to the BDI-II-T QIDS-SR16-T BDI-II-T Score EAP Score EAP 0 -1.4 0 -1.3 1 -1.1 1 -1.0 2 -.9 1 -1.0 3 -.6 2 -.7 4 -.4 3 -.5 5 -.2 4 -.3 6 .1 5 .0 6 .2 7 .4 7 .4 8 .6 8 .7 9 .7 10 .9 9 1.0 11 1.0 Figure 3: Test information functions (TIF) for the Turkish and US 12 1.1 versions of the QIDS-SR16 and for the BDI-II-T 10 1.2 13 1.2 14 1.4 11 1.5 15 1.5 16 1.6 The points to note here are: (a) the high degree of similarity 12 1.7 17 1.7 between the two versions of the QIDS-SR16, (b) the fairly 18 1.8 substantial similarity among all three measures below the 13 1.9 19 1.9 20 2.0 latent variable mean (θ= 0), (c) the tendency of the BDI- 14 2.1 21 2.1 II-T to be most discriminating past this point, and (d) the 22 2.1 23 2.2 fall-off for the BDI-II-T at the positive tail (θ > 2). 15 2.3 24 2.3 25 2.4 Test Equating 16 2.5 26 2.517-18 2.6 27 2.6 19 2.7 28 2.7 Table 5 contains the results of equating the QIDS- 20 2.8 29 2.8 21 2.9 30 2.9 SR16-T and BDI-II-T. Thus, a raw score of 8 on the QIDS- 22 3.0 31 3.0 SR16-T equates to a raw score of 9 on the BDI-II-T because 23 3.1 32 3.1 both lead to an estimated θ of .7 on the normally distributed 24 3.2 33 3.225 3.3 34 3.3 depression scale. Not all values equate exactly so expected 35 3.4 a posteriori (EAP) values within ± .1 θ units were accepted 36 3.5 37 3.6 as matching, e.g., a QIDS-SR16-T score of 4 was treated as 26 3.7 38 3.7 matched to a BDI-II-T score of 3 as they produced EAP 39 3.7 values of -.4 and -.5 respectively. 40 3.827 3.8 41 3.8 42-45 3.9 DISCUSSION ³ 46 4.0 Note: EAP = Expected a posteriori value of θ (depression) A major finding was the essential equivalence of the loadings (trace line slopes) of the domains in the two values of coefficient α reflect the sample variances, as is samples, meaning that each domain measured depression true in the various studies. This is in part due to the fact that to the same extent in the two cases. These values are similar we did not have many severely depressed patients in our to those previously obtained from the QIDS-SR16-US, e.g. sample whereas other studies tend to run the gamut of (5, 22-29). Thus, the scale is unidimensional. Sad mood depression. The difference in intercepts (levels) is more relates most strongly to overall depression and suicidal difficult to interpret as it may reflect either the fact that the ideation relates least strongly, as noted in Fig. 2, and as Turkish respondents were recruited from a medical setting reported in various studies conducted on US samples. The where depression is perhaps more common and the 296 Klinik Psikofarmakoloji Bülteni, Cilt: 21, Sayı: 4, 2011 / Bulletin of Clinical Psychopharmacology, Vol: 21, N.: 4, 2011 - www.psikofarmakoloji.org H. Mergen, I. H. Bernstein, V. Tavli, K. Ongel, T. Tavli, S. Tan American respondents were not, or other sample low education public sector sample using, as here, both characteristics or differences due to translation. At this classical test theory and item response theory analyses. point, the former seems the more reasonable alternative. Overall, the self-reported and clinical versions of the QIDS The second major finding was the high correlation among were similar in their psychometric properties. Similarly, Turkish respondents between the BDI-II-T and the QIDS- Rush et al. (5), found the clinical and self-reported versions SR16-T, which is strong evidence for the latter’s convergent of the QIDS to compare well to the Hamilton Rating Scale validity. However, one difference between the two is that for Depression, which is perhaps the most widely used the BDI-II-T contains at least two dimensions, whereas the measure to evaluate depressive symptomatology in the QIDS-SR16-T is unidimensional, which is perhaps United States. Likewise, Doraiswamy, Bernstein Rush, et responsible for the lack of perfect correlation. This al. (27) found the Montgomery-Åsberg Depression Rating difference may also help explain the difference in TIF past Scale (MADRS), QIDS-C16, and QIDS-SR16 to perform the group mean on θ, i.e. mean depression level. similarly in an elderly population, where the MADRS is There is one important point regarding our test perhaps the most widely used measure in Europe. The α equating. We treated the QIDS-SR16-T and the BDI-II-T as coefficients ranged from 0.85 to 0.89. Moreover, Bernstein two self-descriptive measures of depressive et al. (31) also found the QIDS scales effective in the symptomatology with neither serving as a “gold standard” evaluation of patients with bipolar disorder. In addition, all to define depression for the other. This led to equating the of these papers found the QIDS scales to be unidimensional. two tests in terms of common inferred (θ) values. This is Other studies noted that QIDS scales are similar in quite different from the ROC approach used by Bilgel and reliability to comparable instruments used to evaluate Bayram (30) which does accept the BDI-II-T as a “gold depressive symptomatology (6,9,32-33). standard”. We have used ROC analysis in some earlier Brown et al. studied the QIDS-SR16, IDS-SR30, HRSD17 studies, e.g., (25) but the criterion to define depression was and Mini Asthma Quality of Life Questionnaire in a structured clinical interview and not a self-report of asthmatic patients at treatment exit because of the highly symptomatology. We accept that the interview is closer to co-occurrence of asthma with depression (34). Cronbach α a “gold standard” than is another self-reported measure. values were highest (0.95) for the IDS-SR30; because of its Lamoureux et al. (7) have provided another example of greater length. These values were 0.87 for the QIDS-SR16 a study for which ROC analysis was appropriate. They and the HRSD17. QIDS-SR16 and HRSD17 total scores are studied 155 heterogeneous primary care outpatients, highly correlated (r=0.85) as are QIDS-SR16 and IDS-SR30 similar to our validity study. They used both the clinician- scores (r=0.97). All three scales used in the Brown et al. rated and self-reported QIDS scales, which they compared study, showed comparable sensitivity to symptom change. with the results of the Structured Clinical Interview for Bernstein, Rush, Yonkers et al. (23) had postpartum DSM Disorders (SCID). They reported an area under the patients and non-postpartum female controls take the curve of 0.82. The value of Cronbach’s α was 0.86, which QIDS-SR16-US. Both groups showed low energy level and is somewhat greater than our value, perhaps reflecting restlessness/agitation. However, the non-postpartum differences in sample variability. They suggested a total group reported greater sad mood, suicidal ideation, and score cutoff of 13-14 for moderate depression, which reduced interest. Conversely, the postpartum group yields a sensitivity of 76.5% and specificity of 81.8%. exhibited psychomotor symptoms (restlessness/agitation) They emphasized the need for screening of MDE in and impaired concentration/decision-making. Carmody et primary care, which could substantially improve patient al. (35) compared the QIDS-SR16-US to the Montgomery outcomes, particularly when combined with efforts to Åsberg Depression Rating Scale and used the Orlando et promote adequate treatment and follow-up. al. (31) procedure to equate the two sets of scores. Several other studies illustrate the wide range of In summary, the QIDS-SR16-US has been used settings, in which the QIDS-SR16-US has been applied. successfully in a wide variety of settings, and the QIDS- These are important in this context given the similarity of SR16-T appears sufficiently similar to suggest its use in a the two versions of the QIDS-SR16. Bernstein, Rush, variety of settings to screen for depression, including Carmody et al. (23) studied a low income and relatively primary care settings. Klinik Psikofarmakoloji Bülteni, Cilt: 21, Sayı: 4, 2011 / Bulletin of Clinical Psychopharmacology, Vol: 21, N.: 4, 2011 - www.psikofarmakoloji.org 297 Comparative validity and reliability study of the QIDS-SR16 in Turkish and American college student samples Limitations monitor therapeutic effects. However, this was not the goal of this study. In addition, sensitivity to change was Because the study sample primarily consists of evaluated in the QIDS-SR16-US and found at least adequate university students seen at an outpatient clinical setting, it (36). Given the similarity of the two versions, this is would be of use to apply it to a more general Turkish important, albeit indirect evidence. population, especially an academic one that is similar to Although the details of the item responses analysis of the present American sample. A second important the Beck are less important than those of the QIDS-SR16, limitation is that we did not have repeated test data to the statement, expressed earlier regarding the need to evaluate stability and sensitivity to change of the Turkish interpret the results with caution due to the presence of a QIDS-SR16, which would be necessary if the scale were to minor second factor, is useful to note. References: 1. Sadock BJ, Sadock VA. Kaplan and Sadock’s Synopsis of Pscyhiatry: 12. Rush AJ, Trivedi MH, Carmody TJ, Ibrahim HM, Markowitz JC, Behavioral Sciences/Clinical Psychiatry (10th edition), Philadelphia: Keitner GI, et al. Self-reported depressive symptom measures: Wolters Kluwer, Lippincott Williams&Wilkins; 2007. p.527-8. sensitivity to detecting change in a randomized, controlled trial of chronically depressed, nonpsychotic outpatients. 2. Mergen H, Öngel K. Factors associated with depression among Neuropsychopharmacology 2005; 30(2):405-16. Turkish faculty of education freshmen by Beck Depression Inventory–II-Turkish [Dejavniki depresivnosti pri študentih prvega 13. Horn JL. An empirical comparison of various methods for estimating letnika pedagoške fakultete z vprašalnikom BDI-II-T]. Zdrav Vestn common factor scores. Educ Psychol Meas1965; 25: 313-22. 2009; 78:548-54. 14. Humphreys LG, Ilgen D. Note on a criterion for the number of 3. Bostancı M, Ozdel O, Oguzhanoglu NK, Ozdel L, Ergin A, Ergin common factors. Educ Psychol Meas1969; 29: 571-578. N, et al. Depressive Symptomatology Among University Students in Denizli, Turkey: Prevalence and Sociodemographic Correlates. 15. Humphreys LG, Montanelli RGJr. An investigation of the parallel Croat Med J 2005; 46(1):96-100. analysis criterion for determining the number of common factors. Multivariate Behav Res 1975; 10(2): 193-206. 4. Mergen H, Erdoğmuş Mergen B, Tan Ş, Öngel K. Evaluating The Depression and Related Factors Among the Students of the 16. Montanelli RGJr, Humphreys LG. Latent roots of random data Faculty of Education at Celal Bayar University. The New Journal correlation matrices with squared multiple correlations on the of Medicine 2008; 25:169-174. diagonal: a Monte Carlo study. Psychometrika 1976; 41(3): 341-48. 5. Rush AJ, Bernstein IH, Trivedi MH, Carmody TJ, Wisniewski S, 17. Samejima, F. Estimation of latent ability using a response pattern of Mundt JC, et al. An evaluation of the quick inventory of depressive graded scores. Psychometric Monograph 1969:(Suppl. 17):S1-S100 symptomatology and the hamilton rating scale for depression: a sequenced treatment alternatives to relieve depression trial report. 18. Samejima, F. Graded response model. In: van LindenWJ, Hambleton Biol Psychiatry 2006; 59(6):493-501. RK, editors. Handbook of modern item response theory. New York: Springer-Verlag; 1997. p. 85-100. 6. Rush AJ, Trivedi MH, Ibrahim HM, Carmody TJ, Arnow B, Klein DN, et al. The 16-item Quick Inventory of Depressive 19. Satorra A, Bentler PM. A scaled difference chi-square test statistic Symptomatology (QIDS) Clinician Rating (QIDS-C) and Self- for moment structure analysis. Psychometrika 2001; 66(4):507-14 Report (QIDS-SR): A psychometric evaluation in patients with chronic major depression. Biological Psychiatry 2003;54(5):573-83. 20. Orlando M, Sherbourne, CD, Thissen, D. Summed-score linking using item response theory: application to depression measurement. 7. Lamoureux BE, Linardatos E, Fresco DM, Bartko D, Logue E, Milo Psychol Assess 2000; 12(3): 354-9. L. Using the QIDS-SR16 to identify major depressive disorder in primary care medical patients. Behav Ther 2010; 41(3):423-31. 21. Nunnally JC, Bernstein IH, Psychometric Theory (3rd Ed.), New York: McGraw Hill; 1994. 8. American Psychiatric Association. Diagnostic and Statistical Manual for Mental Disorders (DSM-IV), Fourth Edition. Washington, DC: 22. Carmody TJ, Rush, AJ, Bernstein, IH, Brannan S, Husain MM, American Psychiatric Association, 1994. Trivedi MH.. Making clinicians lives easier: Guidance on use of the QIDS self-report in place of the MADRS. J Affect Disord 2006; 9. Beck AT, Steer RA, Brown GK. Beck Depression Inventory-Second 95(1-3): 115-8. Edition Manual. San Antonio, TX: The Psychological Corporation, 1996. 23. Bernstein IH, Rush AJ, Carmody TJ, Woo A, Trivedi MH. Clinical 10. Beck AT, Ward C, Mendelson M, Mock J, Erbaugh J. An inventory vs. self-report versions of the quick inventory of depressive for measuring depression. Arch Gen Psychiatry 1961; 4: 561-71. symptomatology in a public sector sample. J Psychiatr Res 2007; 41(3-4):239-46. 11. Tegin B. Depresyonda bilişsel bozukluklar Beck modeline göre bir inceleme Yayınlanmamış doktora tezi. [Cognitive Malfunctions 24. Bernstein IH, Rush AJ, Yonkers K, Carmody TJ, Woo A, McConnell in Depression. Survey according to Beck Model, unpublished K, et al. Symptom features of postpartum depression: are they dissertation]. Hacettepe Üniversitesi, Psikoloji Bölümü; 1980. distinct? Depress Anxiety 2008; 25(1):20-6. 298 Klinik Psikofarmakoloji Bülteni, Cilt: 21, Sayı: 4, 2011 / Bulletin of Clinical Psychopharmacology, Vol: 21, N.: 4, 2011 - www.psikofarmakoloji.org H. Mergen, I. H. Bernstein, V. Tavli, K. Ongel, T. Tavli, S. Tan 25. Bernstein IH, Wendt B, Nasr SJ, Rush AJ. Screening for Major 31. Bernstein IH, Rush AJ, Suppes T, Trivedi MH, Woo A, Kyutoku Depression in Private Practice. J Psychiatr Pract 2009; 15(2): 87-94. Y, et al. A psychometric evaluation of the clinician-rated Quick Inventory of Depressive Symptomatology (QIDS-C16) in patients 26. Bernstein IH, Rush AJ, Trivedi MH, Hughes CW, Macleod L, with bipolar disorder. Int J Methods Psychiatr Res 2009; 18(2):138- Witte BP, et al. Psychometric properties of the Quick Inventory 46. of Depressive Symptomatology in adolescents. Int J Methods Psychiatr Res 2010; 19(4):185-94. 32. Pignone MP, Gaynes BN, Rushton JL, Burchell CM, Orleans CT, Mulrow CD, et al. Screening for depression in adults: A summary 27. Doraiswamy PM, Bernstein IH, Rush AJ, Kyutoku Y, Carmody of the evidence for the U.S. Preventative Service Task Force. Ann TJ, Macleod L, et al. Diagnostic utility of the Quick Inventory of Intern Med 2002; 136(10):765-76. Depressive Symptomatology (QIDS-C16 and QIDS-SR16) in the elderly. Acta Psychiatr Scand 2010; 122(3):226-34. 33. Trivedi MH. Tools and strategies for ongoing assessment of depression: a measurement-based approach to remission. J Clin 28. Bernstein I H, Rush A J, Suppes T, Kyotoku Y, Warden D. The Psychiatry 2009;70 (Suppl. 6):26-31. Quick Inventory of Depressive Symptomatology (Clinician and Self-Report Versions) in Patients with Bipolar Disorder. CNS 34. Brown ES, Murray M, Carmody TJ, Kennard BD, Hughes CW, Khan Spectr 2010; 15(6):367-73. DA, et al. The Quick Inventory of Depressive Symptomatology- Self-report: a psychometric evaluation in patients with asthma and 29. Bernstein IH, Rush AJ, Stegman D, Macleod L, Witte B, Trivedi major depressive disorder. Ann Allergy Asthma Immunol 2008; MH. A Comparison of the QIDS-C16, QIDS-SR16, and MADRS in 100(5):433-8. an Adult Outpatient Clinical Sample. CNS Spectr 2010; 15(7): 458- 68. 35. Carmody TJ, Rush AJ, Bernstein IH, Brannan S, Husain MM, Trivedi MH. Making clinicians lives easier: guidance on use of the 30. Bilgel N, Bayram N. Turkish Version of the Depression Anxiety QIDS self-report in place of the MADRS. J Affect Disord 2006; Stress Scale (DASS-42): Psychometric Properties. [Depresyon 95(1-3):115-8. Anksiyete Stres Ölçeğinin (DASS-42) Türkçeye Uyarlanmış Şeklinin Psikometrik Özellikleri]. Archives of Neuropsychiatry- 36. Rush AJ, Bernstein IH, Trivedi MH, Carmody TJ, Wisniewski Nöropsikiyatri Arşivi Dergisi 2010; 47(2):118-26. S, Mundt JC, et al. An evaluation of the Quick Inventory of Depressive Symptomatology and the Hamilton Rating Scale for Depression: a Sequenced Treatment Alternatives to Relieve Depression trial report. Biol Psychiatry 2006;59(6): 493-501. Klinik Psikofarmakoloji Bülteni, Cilt: 21, Sayı: 4, 2011 / Bulletin of Clinical Psychopharmacology, Vol: 21, N.: 4, 2011 - www.psikofarmakoloji.org 299 Comparative validity and reliability study of the QIDS-SR16 in Turkish and American college student samples KISA DEPRESİF BELİRTİ ENVANTERİ ÖZBİLDİRİM FORMU (hasta tarafından doldurulacak) Son 7 gün boyunca, sizi en iyi tanımlayan seçeneği daire içine alınız. 1. Uykuya dalma: 0 Uykuya dalmam hiçbir zaman 30 dakikayı aşmıyordu. 1 Bu sürenin yarısından azında, uykuya dalmam en az 30 dk. sürüyordu. 2 Bu sürenin yarısından çoğunda, uykuya dalmam en az 30 dk sürüyordu. 3 Bu sürenin yarısından çoğunda, uykuya dalmam 60 dakikadan uzun sürüyordu. 2. Gece boyunca uyku: 0 Gece uyanmıyordum. 1 Her gece kısa sürelerle birkaç kez uyanarak, huzursuz ve hafif uyuyordum. 2 Gecede en az bir kez uyanıyordum, ancak kolayca tekrar uyuyordum. 3 Bu sürenin yarısından çoğunda, gece boyu birden fazla uyanıyordum ve 20 dakika ya da daha uzun süre uyanık kalıyordum. 3. Çok erken uyanma: 0 Bu sürenin çoğunda, kalkmam gereken zamandan en fazla 30 dakika önce uyanıyordum. 1 Bu sürenin yarısından çoğunda, kalkmam gerekenden 30 dk.dan uzun bir süre öncesinde uyanıyordum. 2 Hemen her zaman, gerekenden en az bir saat önce uyanıyordum, ancak sonuçta tekrar uyuyordum. 3 Gerekenden an az bir saat önce uyanıyordum ve bir daha uyuyamıyordum. 4. Çok fazla uyuma: 0 Gün içinde uyuklamaksızın, gecede en fazla 7/8 saat uyuyordum. 1 Gündüz uyuklamalar da dahil olmak üzere 24 saat boyunca, en fazla 10 saat uyuyordum. 2 Gündüz uyuklamalar da dahil olmak üzere 24 saat boyunca, en fazla 12 saat uyuyordum. 3 24 saat boyunca uyuklamalar da dahil olmak üzere, 12 saatten fazla uyuyordum. 5. Keder hissi: 0 Kederli hissetmiyordum. 1 Bu sürenin yarısından azında kederli hissediyordum. 2 Bu sürenin yarısından çoğunda kederli hissediyordum. 3 Bu sürenin hemen hepsinde kederli hissediyordum. 6. İştah azalması: 0 İştahımda her zamankine göre değişiklik olmadı. 1 Her zamankinden daha az miktar ya da sıklıkta yiyordum. 2 Her zamankinden belirgin olarak daha az ve kendimi zorlayarak yiyordum. 3 24 saat içinde nadiren ve yalnızca kendimi çok zorlayarak ya da başkalarının zorlaması ile yiyordum. 7. İştah artması: 0 İştahımda her zamankine göre değişiklik olmadı. 1 Her zamankinden daha sık yeme ihtiyacı duyuyordum. 2 Düzenli olarak, her zamankine göre daha sık ve/veya daha fazla miktarda yiyordum. 3 Hem öğünlerde hem de öğün aralarında aşırı yeme isteği duyuyordum. 8. Kilo verme (son iki hafta içerisinde): 0 Kilomda bir değişiklik olmadı. 1 Hafif bir kilo kaybım olduğunu hissediyorum. 2 1 kilogram ya da daha fazla verdim. 3 2,5 kilogram ya da daha fazla kilo kaybettim. 9. Kilo alma (son iki hafta içinde): 0 Kilomda bir değişiklik olmadı. 1 Hafif kilo aldığımı hissediyorum. 2 1 kilogram ya da daha fazla kilo aldım. 3 2,5 kilogram ya da daha fazla kilo aldım. 300 Klinik Psikofarmakoloji Bülteni, Cilt: 21, Sayı: 4, 2011 / Bulletin of Clinical Psychopharmacology, Vol: 21, N.: 4, 2011 - www.psikofarmakoloji.org H. Mergen, I. H. Bernstein, V. Tavli, K. Ongel, T. Tavli, S. Tan 10. Konsantrasyon (karar verme): 0 Her zamanki konsantrasyon ve karar verme yeteneğimde bir değişiklik yok. 1 Ara sıra kararsız olduğumu ya da dikkatimin dağıldığını hissediyorum. 2 Çoğunlukla dikkatimi toplamak ya da karar vermek bir çaba gösteriyorum. 3 Okumak için yeterince konsantre olamıyorum ya da basit kararları bile alamıyorum. 11. Kendime bakışım: 0 Kendimi diğerleri kadar değerli ve hak sahibi görüyorum. 1 Her zamankinden daha fazla kendimi suçluyorum. 2 Diğerleri için sorun kaynağı olduğuma büyük ölçüde inanıyorum. 3 Sürekli kendimdeki küçük ya da önemli eksiklikleri düşünüyorum. 12. Ölüm ya da intihar düşünceleri: 0 Ölüm ya da intiharı düşünmüyorum. 1 Hayatın boş olduğunu ya da yaşamaya değip değmeyeceğini düşünüyorum. 2 Haftada birkaç kez birkaç dakika boyunca intihar ya da ölümü düşünüyorum. 3 Günde birkaç kez intihar ya da ölümü bazı ayrıntılarıyla düşünüyorum ya da intihar için özgün planlar yaptım ya da yaşamıma son vermeyi denedim. 13. Genel ilgi: 0 Diğer insanlar ye genel aktivitelere ilgim, her zamankinden farklı değil. 1 Diğer insanlar ye genel aktivitelere ilgimin daha az olduğunu fark ediyorum. 2 Önceki aktivitelerimin yalnızca bir ya da ikisine ilgimin sürdüğünü fark ettim. 3 Önceki aktivitelerime hemen hemen hiç ilgim kalmadı. 14. Enerji düzeyi: 0 Her zamanki enerji düzeyimde bir değişiklik yok. 1 Her zamankinden daha kolay yoruluyorum. 2 Olağan günlük aktivitelerime başlamak ya da bitirmek için büyük çaba göstermem gerekiyor (alışveriş, ev işleri, yemek yapma ve işe gitme gibi). 3 Enerjim olmadığı için olağan günlük aktivitelerimin çoğunu yapamıyorum. 15. Yavaşlama hissi: 0 Her zamanki olağan hızımda düşünüp, konuşup hareket ediyorum. 1 Daha yavaş düşündüğümü ya da sesimin düzeyinin donuk olduğunu fark ediyorum. 2 Soruların çoğuna yanıt vermem birkaç saniye gerektiriyor ve düşüncemin yavaşladığına eminim. 3 Sıklıkla aşırı çaba harcamadan sorulara yanıt veremiyorum. 16. Huzursuzluk hissi: 0 Huzursuz hissetmiyorum. 1 Sık sık huzursuzluk hissediyorum, ellerimi ovuşturuyor ya da oturma biçimimi değiştiriyorum. 2 Hareket etme isteği duyuyorum ye çok huzursuzum. 3 Zaman zaman oturarak bekleyemiyorum ve dolaşma ihtiyacı duyuyorum. Puanlamak için: 1. Uyku ile ilgili 4 maddeden (1-4) en yüksek puanı seçiniz _____ 2. Madde 5 _____ 3. İştah ile ilgili 4 maddeden (6-9) en yüksek puanı seçiniz _____ 4. Madde 10 _____ 5. Madde 11 _____ 6. Madde 12 _____ 7. Madde 13 _____ 8. Madde 14 _____ 9. Psikomotor durumla ilgili 2 maddeden (15-16) en yüksek puanı seçiniz _____ Toplam puan: (0-27) _____ Klinik Psikofarmakoloji Bülteni, Cilt: 21, Sayı: 4, 2011 / Bulletin of Clinical Psychopharmacology, Vol: 21, N.: 4, 2011 - www.psikofarmakoloji.org 301