Psychometric properties of the Turkish version of the Clinician-Administered PTSD Scale for Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (Turkish CAPS-5)

Abstract

Background: In the subsequent revision of Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5; American Psychiatric Association, 2013) symptoms of diagnostic criteria for post-traumatic stress disorder (PTSD) are defined in four clusters and the number of PTSD symptoms was expanded to 20. The Clinician-Administered PTSD Scale (CAPS) is the most widely used structured clinical interview and recognized as the golden standard in PTSD diagnosis. The final revision of the clinical interview form as the CAPS for DSM-5 (CAPS-5) was advanced in line with the recent revisions in DSM-5 with regards to the PTSD definition. The aim of this study was to examine the psychometric properties of the Turkish version of CAPS-5 in clinical samples and healthy controls. Methods: In the present study, 30 inpatients with PTSD and 30 inpatients with major depressive disorder consecutively presented to the Psychiatry Outpatient Clinic Yüzüncü Yıl University Research Hospital, and 30 healthy controls were enrolled. All participants were included if only they reported an index trauma in the Life Events Checklist for DSM-5 (LEC-5) that bothered them during the past month. Subjects were administered a socio-demographic questionnaire, the Dissociative Experiences Scale (DES), Beck Depression Inventory (BDI), Beck Anxiety Inventory (BAI) along with the LEC-5, CAPS-5 and PTSD Checklist for DSM-5 (PCL-5). We used confirmatory factor analysis to compare a structured clinical interview (CAPS-5) and a self-report measure, the PCL-5 and to examine DSM-5 implied four-symptom clusters and several factor structures proposed in the literature to understand which model best represents the latent factor structure of PSTD symptoms. Using multivariate analysis of covariance, concurrent validity of both self-report and structured clinical interview was evaluated. Receiver operating characteristics (ROC) curve was utilized to obtain an optimal cut-off value of the PCL-5 scores in order to use in demarcating cases with non-cases. Results: Even though DSM-5 implied four-factor model adequately fit to either data collected using self-report or clinician-administered measures of PTSD, the latent structure of PTSD symptoms measured by either CAPS-5 or PCL-5 were best represented by six-factor Externalizing Behaviors model, particularly compared to seven-factor Hybrid model. In comparison to depressive and control groups, PTSD patients reported greater scores on the PCL-5, DES, BDI, and BAI and McNemar χ2 values between two applications with two weeks interval were unsubstantial. Additionally, PTSD patients exhibited greater symptom endorsement on B, C, D, E, F, G symptom clusters and dissociative subtype than depressive patients and controls. Using signal detection analysis, a significant area under the curve (AUC) was calculated for the PCL-5 (AUC = 0.87 p < 0.001 asymptotic 95% Confidence Interval = 0.798–0.942). The PCL-5 had excellent diagnostic utility with 0.90 sensitivity and 0.80 specificity on a cut-off score ≥47. Conclusion: Turkish versions of the CAPS-5 and PCL-5 are demonstrated to have very good psychometric properties. Implications regarding the findings are discussed.

Keywords

• PTSD Checklist for DSM-5
• factor structure
• concurrent validity
• pathological dissociation
• depression