Project Summary The broad goal of the proposed work is to conduct a randomized controlled trial of a specialized computer- delivered cognitive-behavioral therapy (CBT) to supplement standard alcohol use disorder (AUD) treatment in patients with a co-occurring anxiety disorder (“comorbidity”). Comorbidity is both common in AUD treatment patients (up to 50%) and confers a substantial increase in the risk of a return to drinking in the months following treatment. Because research shows that simply adding a standard psychiatric treatment does not substantially improve the AUD outcomes of comorbid individuals, we developed a CBT-based intervention aimed at disrupting the positive-feedback loop (“vicious cycle”; VC) of mutually aggravating negative affect and drinking behavior/urges (the “VC-CBT”). In an RCT, AUD treatment patients who received the therapist- delivered VC-CBT demonstrated significantly improved alcohol use outcomes as compared to those who received a standard anxiety treatment. Unfortunately, most community-based AUD treatment programs do not have clinical staff with the specialized training and technical expertise needed to deliver the VC-CBT. To help bridge this “research-to-practice” gap, we went on to develop a fully autonomous and interactive computer- delivered version of the VC-CBT and have demonstrated its functionality in AUD patients. Now, we propose to test the clinical efficacy of the computer-delivered VC-CBT, as well as the mechanisms and processes by which it is hypothesized to work. Aim I is a randomized controlled trial comparing the computer-delivered VC- CBT to an intensity-matched computer-delivered active control intervention that focuses on healthy lifestyles. 256 individuals in residential AUD treatment who have a comorbid anxiety disorder will receive either the VC- CBT or the active control intervention to obtain 200 cases that complete a 1-, 4- and 8-month follow-up. We predict the VC-CBT group will demonstrate superior alcohol-related outcomes at follow-up relative to the control group. Aim II evaluates the extent to which the computer-delivered VC-CBT selectively imparts the skills and knowledge targeted and whether they convey (mediate) the interventions therapeutic effect. This entails a formal series of “causal steps” analyses of the associations of: treatmentskills/knowledge; skills/knowledge outcomes; and, treatmentoutcomes with vs. without statistically controlling the effect of skills/knowledge. Aim III will test the theoretically-derived prediction that the computer-delivered VC-CBT moderates (i.e., weakens) the association between levels of real-time negative affect and drinking behavior/ urge. This will be accomplished by analyzing a series of twice-daily ecological momentary assessments (EMAs) that participants record in their natural environment for the 7 days prior to each of the three follow-up assessments. The impact of this work would be to provide a scalable and inexpensive means of impro...