Access to primary health care in the U.S. is an ongoing problem for poorly insured patients, leading many to seek care in discontinuous ways through the Emergency Department (ED). While most ED visits are appropriate, many may be "prevented" by timely and potentially less costly ambulatory care. Several definitions of potentially avoidable ED visits have therefore been developed as measures of primary care access, quality, and cost. Particularly for publicly-insured patients, many healthcare organizations use individual case management, triggered by “preventable” ED visits, to attempt better patient outcomes and lower costs. But this actuarial approach often ignores the contextual determinants of health. We therefore propose using geographic information systems to augment individual risk assessment with spatial epidemiology to identify the characteristics of patient and neighborhood “hotspots” of risk for potentially avoidable ED utilization. Our study will determine: (1) what patient characteristics, including neighborhood, are associated with risk of potentially avoidable ED visits; (2) among patients with potentially avoidable ED visits, what patient characteristics, including neighborhood, are associated with the risk of higher total cost of care for those visits; (3) what neighborhood characteristics are associated with risk of having potentially avoidable ED visits, and (4) given 1-3, what are the particular patient demographics, neighborhoods, and conditions that, on qualitative assessment, appear most likely to benefit from patient-centered educational, primary care, or alternative healthcare resource interventions to reduce potentially avoidable ED visits and their costs? Partnering with Regions Hospital administration, we will use the information generated to identify patient and neighborhood center interventions to achieve the triple aim of better quality care and lower costs through improvements in public health.