Toward predicting completion of substance abuse treatment

Abstract

This investigation attempts to identify factors which influence whether or not someone is likely to drop out of a chemical dependency treatment program. Dropping out is defined as someone who leaves treatment against medical advice. 2The subjects were patients from a private, non-profit, medically based, residential program. Nine demographic characteristics were abstracted from the charts on file for the patients at the treatment center. Two groups of 45 patients each were selected from the inpatient population.One group, the Completed Treatment group, comprised patients who had completed the 28 day program. The second group, the AMA Discharge group, comprised patients who dropped out of treatment within the first 4 to 10 days. The demographic characteristics analyzed were gender, number of drugs used by the patient, drug preference, method of admission, treatment history, marital success, social status, dependents living at home, and education.A stepwise discriminant analysis using the SPSS-x statistical package was performed to select the best linear, weighted combination of variables which would produce the greatest discrimination between the two groups. The variables selected by the discriminant analysis were drug preference, treatment history, marital success, and social status. These four variables proved to be significant in accounting for the greatest proportion of total variance at a significance level of .0559.The discriminant function was used to classify the individual subjects according to their highest probability of membership into each group. By comparing actual and predicted group memberships, using the original sample, the validity of the discriminant function and discriminant Further research, studying the role of variables which represent the patient's interaction with the treatment program rather than variables which simply identify static characteristics of a subject, is needed to more fully identify how a treatment agency might facilitate treatment completion.