Open Access

Background Dropout from healthcare interventions can negatively affect patients and healthcare providers through impaired trust in the healthcare system and ineffective use of resources. Research on this topic is still largely missing on refugees and asylum seekers. The current study aimed to characterize predictors for dropout in the Mental Health in Refugees and Asylum Seekers (MEHIRA) study, one of the largest multicentered controlled trials investigating the effectiveness and cost-effectiveness of a nationwide stepped and collaborative care model. Methods Predictors were multiply imputed and selected for descriptive modelling using backward elimination. The final variable set was entered into logistic regression. Results The overall dropout rate was 41,7%. Dropout was higher in participants in group therapy (p = 0.001; OR = 10.7), with larger satisfaction with social relationships (p = 0.017; OR = 1.87), with difficulties in maintaining personal relationships (p = 0.005; OR = 4.27), and with higher depressive symptoms (p = 0.029; OR = 1.05). Participants living in refugee accommodation (p = 0.040; OR = 0.45), with a change in social status (p = 0.008; OR = 0.67) and with conduct (p = 0.020; OR = 0.24) and emotional problems (p = 0.013; OR = 0.31) were significantly less likely to drop out of treatment. Conclusion Overall, the outcomes of this study suggest that predictors assessing social relationships, social status, and living conditions should be considered as topics of psychological treatment to increase adherence and as predictors for future research studies (including treatment type).

Adolescent refugees and asylum seekers (ARAS) are highly vulnerable to mental health problems. Stepped care models (SCM) and culturally sensitive therapies offer promising treatment approaches to effectively provide necessary medical and psychological support. To our knowledge, we were the first to investigate whether a culturally sensitive SCM will reduce symptoms of depression and PTSD in ARAS more effectively and efficiently than treatment as usual (TAU). We conducted a multicentric, randomized, controlled and rater-blinded trial across Germany with ARAS between the ages of 14 to 21 years. Participants (N = 158) were stratified by their level of depressive symptom severity and then equally randomized to either SCM or TAU. Depending on their severity level, SCM participants were allocated to tailored interventions. Symptom changes were assessed for depression (PHQ) and PTSD (CATS) at four time points, with the primary end point at post-intervention after 12 weeks. Based on an intention-to-treat sample, we used a linear mixed model approach for the main statistical analyses. Further evaluations included cost–utility analyses, sensitivity analyses, follow-up-analyses, response and remission rates and subgroup analysis. We found a significant reduction of PHQ (d = 0.52) and CATS (d = 0.27) scores in both groups. However, there was no significant difference between SCM and TAU. Cost–utility analyses indicated that SCM generated greater cost–utility when measured as quality-adjusted life years compared to TAU. Subgroup analysis revealed different effects for the SCM interventions depending on the outcome measure. Although culturally sensitive, SCMs did not prove to be more effective in symptom change and represent a more cost-effective treatment alternative for mentally burdened ARAS. Our research contributes to the optimization of clinical productivity and the improvement of therapeutic care for ARAS. Disorder-specific interventions should be further investigated.