Large-scale psychosocial interventions in complex emergencies call for a screening procedure to identify individuals at risk. To date there are no screening instruments that are developed within low- and middle-income countries and validated for that purpose. The present study assesses the cross-cultural validity of the brief, multi-informant and multi-indicator Child Psychosocial Distress Screener (CPDS). The CPDS data of total samples in targeted catchment areas of a psychosocial care program in four conflict-affected countries (Burundi n = 4193; Sri Lanka n = 2573; Indonesia n = 1624; Sudan n = 1629) were studied to examine the cross-cultural construct validity of the CPDS across settings. First, confirmatory factor analyses were done to determine the likelihood of pre-determined theory-based factor structures in each country sample. Second, multi-sample confirmatory factor analyses were done within each country sample to test measurement equivalence of the factor structure as a measure of construct validity. A 3-factor structure reflecting the theoretical premises of the instrument (e.g., child distress, child resilience and contextual factors) was found in the samples from Burundi, Sri Lanka and Indonesia, albeit with context specific deviations. The robustness of the 3-factor structure as an indicator of construct validity was confirmed within these three samples by means of multi-sample confirmatory factor-analyses. A 3-factor structure was not found in the Sudan sample. This study demonstrates the comparability of the assessment by the CPDS of the construct 'non-specific psychosocial distress' across three out of four countries. Robustness of the factor structure of the CPDS within different samples refers to the construct validity of the instrument. However, owing to context-specific deviations of inter-item relationships, the CPDS scores cannot be compared cross-culturally, a finding that confirms the need for attention to contextual factors when screening for non-specific psychosocial distress.
Read full abstract