Psychometric properties of the Turkish version of the Patient Health Questionnaire-Somatic, Anxiety, and Depressive Symptoms

Objective: In this study, we examined the psychometric properties of the Turkish version of the Patient Health Questionnaire Somatic, Anxiety, and Depressive Symptoms (PHQ-SADS) developed by Kroenke, Spitzer, and Williams. Method: This study sample consisted of 200 outpatients admitted to Erenkoy Mental Health Training and Research Hospital, Erenkoy Physiotherapy and Rehabilitation Hospital, and Kartal Training and Research Hospital and 240 graduate students in Karadeniz Technical University. Participants were administered the 90-item Symptoms Checklist (SCL-90R) and the 7-item Whiteley Index, along with the PHQ-SADS. A month later, the PHQ-SADS was readministered to 60 of the students. To investigate the internal consistency of the scale and its subscales, corrected item-total correlations were examined to establish the effect on the Cronbach coefficients and internal consistency of each item of the subscales. Test-retest correlations were also analyzed for reliability. Factorial structure was investigated using principal component analysis. The validity of distinguishing congruent and specific groups was also investigated for validity. Results: Total scores on the scale showed an adequate test-retest consistency (r = 0.54, 0.52, and 0.76, respectively). All items showed adequate correlations (r > 0.26) in the test-retest analysis. Cronbach a values were 0.86 (control), 0.93 (patient), and 0.92 (total) on the test of internal consistency. When the questions were analyzed individually, the item-total correlation for item 7 of the PHQ somatization subscale was found to be inadequate in the control group. Exploratory factorial analysis and varimax rotation results showed that the scale provided a 4-factor structure. In the validity analysis, a significant difference between the patient and the control group mean values was determined. The SCL-90R, 7-item Whiteley Index, and SCL-90R somatization subscales were found to be sufficiently related to the number of symptoms to establish criterion-related validity. Conclusion: Findings with respect to internal consistency, test-retest consistency, item-total correlation, factorial structure, distinguishing validity for specific groups, and criterion-related validity for the PHQ-SADS show that the scale is acceptable in terms of validity and reliability for the Turkish population. (C) 2012 Elsevier Inc. All rights reserved.