Multiple predictors of dropout from alcoholism treatment

A common problem in treating alcoholics is the high dropout rate. Many studies have identified individual factors associated with dropout, eg, poor motivation and previous dropout. We believe the present study reports the first major effort to use multivariate analyses to predict dropout in a large (792), one-year follow-up study of alcoholics, and examines the possibility that medical and nonmedical treatments lead to differential dropout rates. A multiple classification analysis technique showed that treatment variables as opposed to client characteristics were the best predictors of dropout. Patients remaining in treatment were more likely to have a variety of medical interventions, eg, medication and medical assessment, than those who dropped out. Results were similar to studies using other techniques and have interesting implications for the treatment of alcoholics, raising questions about current trends toward nonmedical treatment of alcoholism.