Toward a stepped care approach to treating problem drinkers: the predictive utility of within-treatment variables and therapist prognostic ratings

Abstract

Aims: Cost containment, a central issue in current health planning, encourages the use of brief interventions. Although brief interventions for problem drinkers have proved successful, a portion of such individuals do not change their alcohol use during treatment.

Design: Repeated measures design (pre-treatment, within-treatment and 6 months post-treatment).

Setting and participants: To identify individuals at risk for continued problem drinking, predictors of post-treatment drinking were examined for 212 problem drinkers who presented for treatment in an outpatient treatment clinic.

Intervention: All participants completed a brief cognitive behavioral motivational intervention.

Measurements: At the pre-treatment assessment demographic, drinking pattern, severity of dependence and other cognitive variables (e.g. self-efficacy, goal choice) were collected. Within-treatment, drinking pattern and cognitive variables such as self-efficacy and goal choice were again measured.

Findings: Regression analyses showed that therapist prognosis ratings contributed significantly to the prediction of outcome even when pre-treatment variables were controlled. However, when within-treatment variables were included in the prediction, variables such as within treatment drinking eliminated the predictive utility of therapist prognosis ratings. This pattern held for both percentage of days abstinent and drinks per drinking day at a 6-month follow-up.

Conclusions: It is suggested that a stepped care approach based on prediction models that include clients' within-treatment response can be applied to the treatment of problem drinkers who show little initial response to treatment.