Addictive Behaviors, Vol. 5, pp. 259-264, 1980 Printed in the USA. All rights reserved. 03064603/80/030259-06$02.00/O Copyright 0 1980 Pergamon Press Ltd zyxwvutsrqponm PREDICTORS OF TREATMENT OUTCOME FOR ALCOHOLICS TREATED BY INDIVIDUALIZED BEHAViOR THERAPY * STEPHEN A. MAISTO Research Triangle Institute MARK B. SOBELL and LINDA C. SOBELL Addiction Research Foundation Toronto, Ontario, Canada Abstract-Predictors of abstinent and controlled drinking outcomes were investigated for 69 male alcoholics who had participated in a study of individualized behavior therapy for alco- holics. Percent of days abstinent 19-24 mos following treatment and a measure reflecting con- trol over drinking (Controlled Drinking Index) during that same interval were regressed on 3 posttreatment, 3 within-treatment and 7 pretreatment variables. These analyses suggested that drinking outcomes stabilize between 1 and 2 years following treatment and revealed that drink- ing behavior during the 6 mos preceding the criterion outcome interval was the best predictor of both outcomes. While a strong relationship was found between a controlled drinking treat- ment goal and controlled drinking outcomes, treatment goal was not related to abstinent out- comes. Total pretreatment alcohol-related hospitalizations were negatively related to controlled drinking outcomes. The contributions of these findings to an understanding of Client X Treat- ment X Outcome interactions are discussed. Accruing evidence shows that successful treatment of alcohol problems does not necessarily require complete abstinence from alcohol (Lloyd & Salzburg, 1975; Pattison, Sobell & Sobell, 1977; Sobell, 1978). Given recent findings, improvement might be more appropriately defined as a zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA reduction in alcohol problems-a reduction in the consequences that are associated with the consumption of alcohol or even abstaining from alcohol. The multiple demonstrations that non- abstinent treatment goals can be attained by some clients indicate a need to empirically deter- mine which treatment goals and methods are more likely to be effective with which clients (Kissin, 1977; Miller &Caddy, 1977; Nathan & Bridell, 1977; Sobell, 1978). Ideally, matching clients with goals could be accomplished by randomly assigning different populations of clients to various treatment modalities; however, practical and ethical considera- tions have made such research difficult. Unfortunately, clinical judgment is presently the major method by which client-treatment matching is most often accomplished. However, there is a wide variation among clinicians and researchers regarding which clients are most appropriate for nonabstinent treatment goals. For example, opinions range from suggestions that nonabstinent goals may have greater value for clients who have experienced fewer psychosocial impairments from alcohol or who have been less physically debilitated by alcohol (Miller, 1976), to specula- tion that these methods are most appropriately used as a last resort with chronic alcoholics who have histories of recalcitrant attitudes toward treatment and repeated treatment failures (Ewing & Rouse, 1973; Nathan & Briddell, 1977). *Preparation of this paper was supported, in part, by grants ST32 AA07072 and lR18 AA00486 from the National Institute on Alcohol Abuse and Alcoholism. Requests for reprints should be sent to: Mark Sobell, Head of Sociobehavioural Treatment Research, Clinical Institute, Addiction Research Foundation, 33 Russell Street, Toronto, Ontario, M5S 2Sl Canada. 259