ABSTRACT The objective of this competing continuation (renewal) application is to determine whether modifications to a contingency management (CM) intervention improve outcomes and reduce costs in heavy drinkers with serious mental illness (SMI). Up to 46% of adults with SMI experience an alcohol use disorder in their lifetimes. Alcohol use contributes to high rates of homelessness, psychiatric hospitalization, HIV infection, cigarette smoking, and drug use in this population, for which CM is an especially promising treatment. In CM, patients receive tangible rewards for demonstrating drug abstinence. CM for alcohol use requires a biomarker that can detect alcohol use for more than 48 hours after consumption. As no such biomarker was available until recently, little research has investigated CM as a treatment for alcohol use disorders. In our initial funding period we found that the alcohol biomarker ethyl glucuronide (EtG) can detect drinking for up to 5 days when administered as part of a randomized 12-week trial of CM. Those randomized to EtG-based CM were 3 times more likely to submit alcohol-negative EtG tests than controls. CM participants also had lower levels of heavy drinking, stimulant drug use, and cigarette smoking than controls. However, CM was ineffective for participants with an average pre-treatment EtG level that indicated frequent, recent heavy drinking (EtG > 499 ng/mL). We propose to investigate whether 2 strategies – a) increasing reinforcer magnitude or b) reinforcing light drinking before reinforcing abstinence – can improve outcomes in heavy drinkers with SMI. While initial research indicates that these strategies are associated with improved outcomes in treatment-resistant drug users and cigarette smokers, no randomized trial has compared them, investigated them in alcohol users or adults with SMI, investigated their relative cost-effectiveness, or investigated modifiers of CM efficacy using a theoretical model. Therefore, we will compare the efficacy of these 2 approaches to the CM intervention implemented in the initial funding period in heavy drinkers with SMI. A total of 400 participants receiving treatment as usual at 2 treatment agencies will take part in a 4-week induction period. Participants (n=240) who attain a mean EtG > 499 ng/mL during the induction period will be randomized to either a) 4 months of standard-magnitude reinforcement CM for submitting alcohol-abstinent EtG samples (EtG < 100 ng/mL) (Usual CM), b) 4 months of high-magnitude CM for submitting alcohol-abstinent EtG samples (High-Magnitude CM), or c) 1 month of CM for submitting alcohol samples that indicate light drinking (EtG < 500 ng/mL), followed by 3 months of CM for submitting alcohol-abstinent EtG samples (Shaping CM). The primary outcome will be EtG-verified alcohol abstinence during the last 3 months of treatment (when all reinforcement is contingent on abstinence) and during 6 months of follow-up. We will also investigate group difference...