Methods: We analyzed data from two waves of the National Epidemiologic Survey on Alcohol and Related Conditions. The sample consisted of 2,631 adults who met criteria for DSM-IV alcohol abuse/dependence at Wave 1 (2001-2002) and who had complete data on drinking behavior at Wave 2 (2004-2005). Outcomes included no alcohol use (referent), moderated drinking, or ongoing AUD. We used chi-square tests to compare group characteristics and multivariate logistic regression to model outcomes, accounting for the complex survey design with stratum, cluster, and sampling weights.
Results: Few participants (5%) reported abstinence at Wave 2. Seeking help for drinking was associated with abstinence in the full sample, but only small proportions reported obtaining treatment (2%), attending 12-step groups (3%), or seeking other, non-specialist help (4%) between survey waves. Among the majority that continued drinking, approximately equivalent proportions reported moderated drinking and no longer met AUD diagnostic criteria (48%) or had ongoing AUD (47%). All racial/ethnic minorities had lower odds of any continued drinking, while sexual minorities and those reporting concurrent cannabis use had higher odds of any continued drinking. Differentiating drinking groups, men and those with only public insurance had higher odds of ongoing AUD (adjusted odds ratio [aOR] 1.80; 95% confidence interval [CI] 1.44, 2.27; and OR 1.80; 95% CI 1.24, 2.60, respectively). Three factors were associated with lower odds of moderated drinking: greater AUD severity (OR .95; 95% CI .92, .98), seeking help for drinking between survey waves (OR .38; 95% CI .21, .66), and rural residence (OR .61; 95% CI .43, .88).
Discussion: Although abstinence is the least risky outcome, it is relatively rare. Following harm reduction principles, which call for maximizing the chances of moderated drinking and decreasing the risk of ongoing AUD, our epidemiologic profiles provide a foundation for future work. Results suggest priority populations for targeted interventions (e.g., men; rural populations) and potential intervention targets (e.g., cannabis co-use). Further research is needed to elaborate the mechanisms leading to moderated drinking after AUD.