ABSTRACT Recovering from Alcohol Use Disorders (AUD) following treatment is difficult, with 43-83% relapsing post- discharge. These chronically high rates arise, in part, because of a gap in our knowledge about the co-evolution of in the moment risk mechanisms associated with relapse among those recovering from AUD, including Craving, Negative Affect, Stressors and Access to alcohol (CASA). These four factors co-evolve among important protective factors that buffer relapse risk (e.g., improved sober support and psychosocial function associated with post-treatment community AUD recovery programs), and context-based risk factors such as near-at-hand places where alcohol is available. The immediate goal of this project is to develop the means to assess and understand the impact of these environmental/contextual factors for recovering alcoholics, towards a longer term research objective of developing effective just-in-time interventions that mitigate the risk of relapse. In this project, we will leverage novel assessment methodologies via a responsive ecological momentary assessment (rEMA) system developed by our team. This app-based system allows researchers to incorporate geospatial, time- specific and event-specific criteria for enhanced monitoring of (and just-in-time delivery of interventions to) recovering patients in post-treatment settings. Consistent with PA-18-620 from NIAAA, this project aims to build foundational evidence and refine the data collection/analysis strategy needed to achieve our long term objective, via the following specific aims: Aim 1) Pilot the use of our rEMA platform to capture fine-grained longitudinal data on Craving, Affect, Stressors and Access to alcohol among 60 recovering alcoholics across 30 days, to evaluate CASA construct validity within rEMA data and assess the extent and impact of data missingness; Aim 2) Test incentive modalities that will inform us regarding best practices that maximize adherence to the rEMA protocol; Aim 3) Develop analytic strategies based machine-learning models of rEMA data, to forecast imminent risk and quantify causal pathways among CASA variables. To accomplish these aims, we will recruit and enroll adults (n=60; 50% female) in non-hospitalized residential AUD treatments in the Central Plains. Using smartphone assessment with GPS monitoring technology, we will collect longitudinal behavioral and experiential data from each participant over 30 days. Analysis of these data will advance our understanding of co-evolving contextual/social-environmental factors that impact the success/failure of the recovery effort, while establishing foundational evidence from which time/event-specific smartphone-based just-in-time interventions may be developed. The final phase of the project includes the creation of an advisory and dissemination board composed of local and national-level stakeholders, AUD treatment practitioners and other researchers that will guide the development of such an intervent...