Beyond the Individual Level: A Secondary Analysis of Social Ecological Determinants of Mental Health Among Refugee Adolescents

Background and Purpose: Adolescent refugees go through critical developmental stages while experiencing traumatic events and post-migration difficulties. The forced migration literature on their mental health shows that they are at high risk of experiencing psychological problems and the role of individual protective and risk factors has been widely investigated. However, researchers call for more holistic approaches that account for the role of family-, community-, and societal-level factors impacting the mental health of refugee adolescents. This study aimed to address this gap by investigating how much different levels of post-migration factors explain mental health symptoms.

Methods: This study is based on secondary data analysis of a cross-sectional data from the first wave of the Refugees in the German Educational System (ReGES) study in Germany. The analytic sample consisted of 1,694 refugee adolescents aged 14–16 residing in Germany, with corresponding parent data. The main outcome of the study was mental health symptoms, assessed using a composite score. Post-migration factors at different levels were included as potential predictors. Hierarchical multiple regression models were estimated with cluster-robust standard errors at the parent level.

Results: The results showed that different post-migration factors at different levels were significantly associated with mental health symptoms. First, a number of individual-level factors, such as German language competence (B = 0.061, SE = 0.014, p < .001), were associated with mental health symptoms. When family-level factors were included in the analysis, it was found that factors including living in shared accommodation (B = 0.323, SE = 0.142, p < .01) were significantly linked to poorer mental health outcomes, while parental support was a protective factor (B = −0.063, SE = 0.038, p < .05). At the community level, one of the factors included in the analysis, teacher support, was identified as a key protective factor (B = −0.073, SE = 0.025, p < .001). Lastly, both societal-level factors included in the analysis, which were problems experienced in the host country (B = 0.073, SE = 0.016, p < .001) and discrimination (sometimes: B = 0.301, SE = 0.140, p < .05; often: B = 0.561, SE = 0.309; very often: B = 0.820, SE = 0.449), explained additional variance in mental health symptoms. The final model, with all levels included, accounted for the largest proportion of variance in mental health symptoms (R² = .27, p < .001).

Conclusions and Implications: Multi-level models that include structural barriers, as well as individual- and family-level factors, explain more variance in mental health symptoms among refugee adolescents compared to each level alone. The study has implementations for social work profession as service providers as it shows that mental health interventions that target individual-level factors may not be sufficient to improve the well-being of refugees, as long as broader contextual factors are not addressed through interventions and policies that acknowledge the social ecology in which they live. These findings are in line with the social work Person-in-Environment (PIE) perspective that encourages consideration at micro, meso, and macro levels.