To test hypotheses that gender, age, living situation, employment status, marital status, deterioration or history of depressive disorders are associated with SUD in these subjects, we used a T-test for age at interview, years of school, and deterioration. Fischer’s exact test was used for gender, depressive syndrome or episode, marital status, employment, and migration status. Yates’ correction test was used for relative onset of SC and SUD. Binary logistic regression was used to model the relationship between the dependent variable SUD and predictor variable living situation. To control for confounding effects of SUD and gender in sociodemographic variables such as marital status and employment, logistic regression methods were used for each outcome variable. In order to test whether country of residence was associated with different rates of SUD logistic regression methods were utilized.
