Abstract
BACKGROUND: In Sub Saharan Africa, there has been limited research on instruments to identify specific mental disorders in children in conflict-affected settings. This study evaluates the psychometric properties of three self-report scales for child mental disorder in order to inform an emerging child mental health programme in post-conflict Burundi. METHODS: Trained lay interviewers administered local language versions of three self-report scales, the Depression Self-Rating Scale (DSRS), the Child PSTD Symptom Scale (CPSS) and the Screen for Child Anxiety Related Emotional Disorders (SCARED-41), to a sample of 65 primary school children in Burundi. The test scores were compared with an external 'gold standard' criterion: the outcomes of a comprehensive semistructured clinical psychiatric interview for children according the DSM-IV criteria (the Schedule for Affective Disorders and Schizophrenia for School-Age Children - K-SADS-PL). RESULTS: The DSRS has an area under the curve (AUC) of 0.85 with a confidence interval (c.i.) of 0.73-0.97. With a cut-off point of 19, the sensitivity was 0.64, and the specificity was 0.88. For the CPSS, with a cut-off point of 26, the AUC was 0.78 (c.i.: 0.62-0.95) with a sensitivity of 0.71 and a specificity of 0.83. The AUC for the SCARED-41, with a cut-off point of 44, was 0.69 (c.i.: 0.54-0.84) with a sensitivity of 0.55 and a specificity of 0.90. CONCLUSIONS: The DSRS and CPSS showed good utility in detecting depressive disorder and posttraumatic stress disorder in Burundian children, but cut-off points had to be put considerably higher than in western norm populations. The psychometric properties of the SCARED-41 to identify anxiety disorders were less strong. The DSRS and CPSS have acceptable properties, and they could be used in clinical practice as part of a two-stage screening procedure in public mental health programmes in Burundi and in similar cultural and linguistic settings in the African Great Lakes region.