Mediators of focused psychosocial support interventions for children in low-resource humanitarian settings: analysis from an Individual Participant Dataset with 3,143 participants.

ABSTRACT

BACKGROUND: Research on psychosocial interventions has been focused on the effectiveness of psychosocial interventions on mental health outcomes, without exploring how interventions achieve beneficial effects. Identifying the potential pathways through which interventions work would potentially allow further strengthening of interventions by emphasizing specific components connected with such pathways. METHODS: We conducted a preplanned mediation analysis using individual participant data from a dataset of 11 randomized controlled trials (RCTs) which compared focused psychosocial support interventions versus control conditions for children living in low- and middle-income countries (LMICs) affected by humanitarian crises. Based on an ecological resilience framework, we hypothesized that (a) coping, (b) hope, (c) social support, and (d) functional impairment mediate the relationship between intervention and outcome PTSD symptoms. A systematic search on the Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE, PubMed, PyscARTICLES, Web of Science, and the main local LMICs databases was conducted up to August 2018. The hypotheses were tested by using individual participant data obtained from study authors of all the studies included in the systematic review. RESULTS: We included 3,143 children from 11 studies (100% of data from included studies), of which 1,877 from six studies contributed to the mediation analysis. Functional impairment was the strongest mediator for focused psychosocial interventions on PTSD (mediation coefficient -0.087, standard error 0.040). The estimated proportion of effect mediated by functional impairment, and adjusted for confounders, was 31%. CONCLUSIONS: Findings did not support the proposed mediation hypotheses for coping, hope, and social support. The mediation through functional impairment may represent unmeasured proxy measures or point to a broader mechanism that impacts self-efficacy and agency.