Artificial Neural Network Analysis Examining Substance Use Problems Co-Occurring with Anxiety and Depressive Disorders Among Adults Receiving Mental Health Treatment

Abstract

Objective: The co-occurrence of anxiety disorders, depressive disorders, and substance use problems was examined. Methods: The Mental Health Client-Level Data dataset was used to conduct logistic regression models and an artificial neural network analysis. Logistic regression analyses were conducted among adults with anxiety (n = 547,473) or depressive disorders (n = 1,610,601) as their primary diagnosis who received treatment in a community mental health center. The artificial neural network analysis was conducted with the entire sample (N = 2,158,074). Results: Approximately 30% of the sample had co-occurring high-risk substance use or substance use disorder. Characteristics including region of treatment receipt, age, education, gender, race and ethnicity, and the presence of co-occurring anxiety and depressive disorders were associated with the co-occurring high-risk substance use or a substance use disorder. Conclusions: Findings from this study highlight the importance of mental health facilities to screen for and provide integrated treatment for co-occurring disorders.