Predicting attrition in long-term residential substance use disorder treatment: A modifiable risk factors perspective.

Although numerous factors are associated with attrition in substance use disorder (SUD) treatment, many are unmodifiable and therefore difficult to target in efforts to improve treatment outcomes. The current study sought to identify the strongest and most modifiable predictors of attrition in long-term residential SUD treatment from myriad characteristics associated with treatment termination. Archival data were examined for 2,069 adults (74% male; 38% non-Hispanic White) who entered a long-term residential SUD treatment facility between January 2010 and June 2016. Program staff recorded clients' demographic, situational, substance use, and intake data at admission; discharge data were recorded at termination. To increase the likelihood our results were clinically meaningful, we randomly split our sample, ran 2 5-step hierarchical logistic regressions, and cross-validated our results. Across samples, we found younger age, having less than a high school education (Step 1), unstable living arrangements (Step 2), greater prior month use of primary substances, less prior month use of alcohol, and prior year needle use preceding treatment (Step 4), and longer recommended length of stay in treatment (Step 5) predicted attrition. To improve long-term residential SUD treatment completion, we propose treatment adaptations begin with the most modifiable predictors of attrition. Accordingly, the current data indicate initial focus should be placed on refurbishing the process through which recommended treatment durations are approached by providers. Subsequent focus should be placed on modifiable factors that present greater systemic challenges, followed by those that are unmodifiable but can be indirectly targeted by interventions tailored to specific underrepresented groups. (PsycInfo Database Record (c) 2020 APA, all rights reserved).