PROJECT SUMMARY Stroke occurs commonly in older patients and is a leading cause of long-term disability. Disability and death from stroke can be reduced with timely reperfusion interventions (i.e., intravenous thrombolysis and endovascular thrombectomy) and high-quality stroke center care. Unfortunately, patients from minority or underserved populations have less access to reperfusion interventions and acute stroke expertise. This is largely because stroke centers – the hospitals best equipped to care for stroke patients – are not geographically distributed to match where patients live and many patients initially present to smaller, non- stroke center hospitals. Valuable work has been done to improve the capabilities of that first, smaller hospital, for example through the expansion of telestroke. However, the most effective interventions for stroke will require transfer. Endovascular thrombectomy is only available at hospitals with advanced capabilities, and there are significant geographic and racial/ethnic disparities in access to this highly efficacious intervention. An optimized system of interhospital stroke patient transfers could be a solution to more equitable access to high-quality stroke care. While many factors contributing to disparities in access and outcomes are relatively fixed (e.g., patient demographics or hospital locations), the process of interhospital patient transfers is dynamic and subject to intervention. The decision to transfer, and the timeliness and destination of that transfer decision can be influenced toward improved outcomes tomorrow. Yet there are many factors influencing patient transfer decisions. In this mixed-methods study, we use statewide, all-payer claims data to study the network of over 340 California hospitals connected through stroke patient transfer to understand the relationship between transfer decisions and disparities in patient access and outcomes. We will apply methods from network science, which are ideally suited to characterize and study the dynamic, multi-level, interdependent structure of the stroke patient transfer network. We will develop and apply new methods in network community detection to identify clusters of hospitals closely connected through patient transfer and to study how the racial/ethnic composition of hospital clusters is associated with stroke-related capabilities in the cluster and how clusters change over time (Aim 1). We will quantify patient-level racial and ethnic disparities in stroke access and outcomes and identify factors at the hospital cluster-level associated with disparities (Aim 2). Then, upon identifying clusters of hospitals that are positive and negative deviants (as outliers in achieving high levels of access and low disparities), we will use qualitative methods to identify scalable strategies to reduce inequalities in stroke access and outcomes (Aim 3). We believe that that these solutions will be valuable to health system leaders and policy makers aiming...