Predictors of treatment dropout for the single and dual-diagnosed client

Elizabeth Sullivan (2001)

The expectation of this study was that therapist variables, as opposed to treatment variables and client variables, were more strongly linked to outpatient treatment dropout. Client data was collected from the treatment files of 228 outpatients between the ages of 17 and 80. Seventy-six files of the following populaces were studied: single diagnosed (mental health or substance abuse) and dual diagnosed (mental health and substance abuse). A two-step logistic regression method was employed. This resulted in a correct classification of treatment dropouts 84% of the time. The most significant predictor of dropout was the treatment variable--number of visits. The second strongest predictors of single diagnosed and dual diagnosed populace dropout were these client variables: male, minority, and unemployed with a positive family history. The regression equation was able to correctly classify the following as predictors of treatment completion 79% of the time: female, non-minority, employed, with a negative family history of mental health or substance abuse treatment, with either no history, or a successful history of past treatment. The study's findings are likely to have significant implications for treatment providers in that it will provide an informational barometer of possible retention issues before they occur. These findings then increase the opportunity to create effective intervention plans with the possibility of reducing treatment dropout.