Search

Abstract Background During conflict, children and adolescents are at increased risk of mental health problems and in particular, anxiety and depression. However, mental health screening in conflict settings is problematic and carries risk making the need for fast, easy-to-administer, screening instruments paramount. The shortened version of the Revised Child Anxiety and Depression Scale (RCADS-25) is one method of rapidly assessing anxiety and depressive symptoms in youths. This self-report questionnaire demonstrates good internal consistency and diagnostic capacity in clinical and non-clinical populations. Nevertheless, few studies have tested the psychometric properties of translated versions of the RCADS-25 limiting its applicability worldwide. Objectives To expand the reach and utility of the RCADS-25, the present study sought to develop an Arabic version of the instrument (RCADS25-Arabic) and to explore its reliability and underlying factor structure. In light of changes to DSM classification, the effects of removing indicator variables for obsessive-compulsive disorder on the psychometrics of the RCADS25-Arabic were also explored. Method The scale was back translated into Modern Standard Arabic and administered to 250 Arabic speaking schoolchildren between 8 and 15 years of age in Syria. Mean and standard deviation were used to characterise the sample and summarize scores. The reliability and factor structure of the RCADS25-Arabic was explored using confirmatory factor analysis. Results Females were 127 and mean age was 12.11 ± SD 2.35. Males scored lower on anxiety (M 15.05 SD ± 8.0, t(248) = − 3.15, p = .003, d = 0.39) and internalizing factors (M 26.1 SD ± 13.1, t(248) = − 2.36, p = .0160, d = 0.31) with no statistical gender difference recorded for depression (t(248) = − 1.27, p = .202). Fit statistics were good for two- and one-factor solutions (χ2/df = 1.65, RMSEA 0.051, CFI .91, TLI .90 and χ2/df = 1.64 and RMSEA 0.051, CFI .91 and TLI .89 respectively). DIFFTEST showed no significant difference between models (χ2 diff (1) = 0.03, p