Validation of a mental health assessment in an African conflict population

We studied the validity of the assessment of posttraumatic stress disorder (PTSD) and depression within the context of an epidemiological mental health survey among war-affected adolescents and young adults in northern Uganda. Local language versions of the Posttraumatic Diagnostic Scale (PDS) and the Depression section of the Hopkins Symptom Checklist (DHSCL) were administered by trained local interviewers. Correlations with probable predictor variables (i.e., trauma exposure), outcomes (e.g., impaired functioning), and local idioms of distress (i.e., spirit possession) were determined to estimate criterion-related construct validity. To assess convergent validity, expert clinicians reinterviewed a subsample using structured interviews (the Clinician Administered PTSD Scale [CAPS] and the Mini International Neuropsychiatric Interview [MINI]). Depression and PTSD symptoms as assessed by the local interviewers correlated with the context variables as predicted. After optimizing the scoring algorithm, we found good agreement between the PDS-based diagnoses and expert diagnoses. However, the concordance for depression diagnoses was not satisfactory. Results show that mental health assessments in African languages can produce reliable and valid data but that caution is warranted in the unevaluated transfer of cutoff scores and scoring algorithms.