PROJECT SUMMARY: Many individuals with alcohol use disorder (AUD) receive no treatment, including evidence-based pharmacotherapies. Previous research on AUD medication use has primarily relied on claims data, and thus focuses on only a portion of the AUD medication treatment process (i.e., filled prescriptions). Using only claims data, it is not possible to distinguish the characteristics of individuals who do not receive a prescription from a clinician from those who receive one but do not fill it. It is also not possible to understand how medication prescriptions and fills fit into the trajectory of AUD treatment received. To design practice and policy interventions that will improve AUD medication use, clinicians and policy makers need to understand who “falls off” at each point in the treatment process. We propose to use a dataset from a large national insurer with linked claims and electronic health record (EHR) data for the period 2010-2023, as well as interviews of AUD providers, in a mixed methods study. The linked data will allow us to observe written prescriptions for each patient as well as whether or not the prescriptions were filled. The linked data will also allow us to incorporate clinically relevant information available in EHRs but not claims. Using these data, we will: 1) For individuals diagnosed with AUD, identify patient and provider characteristics associated with being prescribed an AUD medication and the time from index AUD encounter to prescription; 2) For individuals diagnosed with AUD who are prescribed an AUD medication, identify patient and prescriber characteristics associated with filling the prescription and with time to prescription fill; and 3) Identify trajectories of AUD care that incorporate information on prescriptions written for individuals diagnosed with AUD, and examine the association between different treatment trajectories and indicators of subsequent relapse and injury. The study's mixed methods approach, including semi-structured interviews of providers who treat AUD about the factors that influence their AUD medication prescribing decisions, will allow a more nuanced examination of AUD medication prescribing than quantitative analyses alone allow. The findings will help pinpoint where in the treatment process different types of people fall out of treatment and how different treatment trajectories are associated with relapse and injury, informing future intervention and policy design.