Temptation to drink as a predictor of drinking outcomes following psychosocial treatment for alcohol dependence

Background: Alcohol craving, defined as the subjective experience of an urge or desire to use alcohol, has been identified in numerous settings as a significant predictor of alcohol use and alcohol relapse following treatment for alcohol use disorders. Yet, numerous limitations to the conceptualization and measurement of drinking temptation have led many researchers to question whether self-reported drinking temptation is a useful construct for evaluating treatments for alcohol use disorders.

Methods: Secondary analyses of data from Project MATCH, a multisite randomized clinical trial, were conducted to examine the association between a single-item measure of self-reported "temptation to drink" and drinking outcomes. The first goal was to determine whether temptation to drink changed during the course of treatment for alcohol dependence. The second goal was to assess the predictive validity of temptation to drink, assessed during the fourth session of treatment, as a predictor of past 30-day drinking rates and past 90-day drinking-related consequences at 1 and 3 years following treatment.

Results: The temptation to drink decreased significantly during treatment, and self-reported temptation to drink during the fourth session of treatment was significantly associated with numerous drinking outcomes (including quantity, frequency, and consequences) at 1 year posttreatment (R = 0.04 to 0.11) and number of drinks per drinking day at 3 years following treatment (R = 0.02). A dichotomous measure of temptation to drink (not at all tempted vs. all other levels of temptation) had greater sensitivity as a predictor of drinking outcomes at 1 and 3 years posttreatment than alternative drinking measures (e.g., any drinking, any heavy drinking days) assessed during treatment.

Conclusions: A single-item measure of temptation to drink was a reasonable predictor of short- and long-term drinking outcomes following treatment and comparable to commonly used measures of drinking outcomes for alcohol clinical trials.