End-of-treatment abstinence self-efficacy, behavioral processes of change, and posttreatment drinking outcomes in Project MATCH.

Abstract

This study evaluated whether alcohol abstinence self-efficacy at the end of alcohol treatment was moderated by utilization of behavioral processes of change (coping activities used during a behavior change attempt). It was hypothesized that self-efficacy would be differentially important in predicting posttreatment drinking outcomes depending on the level of behavioral processes, such that the relation between self-efficacy and outcomes would be stronger for individuals who reported low process use. Analyses were also estimated with end-of-treatment abstinence included as a covariate. Data were analyzed from alcohol-dependent individuals in both treatment arms of Project MATCH (Matching Alcoholism Treatments to Client Heterogeneity; N = 1,328), a large alcohol treatment study. Self-efficacy was moderated by behavioral process use in predicting drinking frequency 6 and 12 months posttreatment and drinking quantity 6 months posttreatment such that self-efficacy was more strongly related to posttreatment drinking when low levels of processes were reported than high levels, but interactions were attenuated when end-of-treatment abstinence was controlled for. Significant quadratic relations between end-of-treatment self-efficacy and 6- and 12-month posttreatment drinking quantity and frequency were found (p < .001, ƒ² = 0.02-0.03), such that self-efficacy most robustly predicted outcomes when high. These effects remained significant when end-of-treatment abstinence was included as a covariate. Findings highlight the complex nature of self-efficacy's relation with drinking outcomes. Although the interaction between self-efficacy and behavioral processes was attenuated when end-of-treatment abstinence was controlled for, the quadratic effect of self-efficacy on outcomes remained significant. The pattern of these effects did not support the idea of "overconfidence" as a negative indicator.