Predicting treatment dropout from a therapeutic community for polydrug abusers



Journal Title

Journal ISSN

Volume Title


Texas Tech University


Residential therapeutic communities are effective but costly treatment alternatives for substance abuse. Attrition from this form of substance abuse treatment has been historically high. The cost of treatment for a therapeutic community may be reduced by identifying early dropouts from treatment. Research has suggested that the Minnesota Multiphasic Personality Inventory (MMPI) and certain demographic variables can be used to identify those patients at risk for early treatment dropout from substance abuse treatment programs, including therapeutic communities. The predictive ability of psychological and demographic variables can be enhanced by using a multivariate statistical approach rather than using t-tests to compare pre-treatment data.

This study used the three validity and ten clinical scales of the MMPI plus two demographic variables in determining a discriminant function that separated treatment dropouts from treatment stayers from a traditional residential therapeutic community for polydrug abusers. The two demographic variables were: length of substance abuse before treatment and employment status. The discriminant function was able to successfully separate the original sample into completers and dropouts. Dropouts reported more psychopathology on the MMPI; they had a longer history of substance before treatment; and they had poor employment histories before treatment. An examination of the correlations of the predictor variables with the discriminant function revealed that scales F, 4(PD), 7(PT), 8(SC), and 9(MA) were the more important variables in the discriminant function. Original cases and new cases from a cross-validation sample were classified better than the expected rate of 30% for stayers and 70% for dropouts. Participants in this study were from Alternative House, the therapeutic community component of the Alternative Drug Abuse Program in Houston, Texas. The participants were predominately young, white male probationers who were stipulated to treatment by the district criminal courts.