BACKGROUND: Little has been written about the geographic basis of emergency department (ED) visits. OBJECTIVE: The objective of this study is to describe the impact of geography on ED visits. METHODS: A retrospective analysis was conducted of ED visits during a 1-year period at a single institution using spatial interaction analysis that models the pattern of flow between a series of origins (census block groups) and a destination (ED). Patients were assigned to census block groups based upon their verified home address. The study hospital is the only Level I trauma, pediatric, and tertiary referral center in the area. There are 11 other hospitals with EDs within a 40-mile radius. Each patient visit within this radius, including repeat visits, was included. Patients with an invalid home address, a post office box address, or those who lived outside a 40-mile radius were excluded. ED visits per 100 population were calculated for each census block group. RESULTS: There were 98,584 (95%) visits by 63,524 patients that met study inclusion criteria. Visit rates decreased with increasing distance from the ED (p < 0.0001). Nineteen percent of patients lived within 2 miles, 48% within 4 miles, and 92% within 12 miles of the ED. The Connecticut border, 7 miles south of the ED (p < 0.0001), the Connecticut River, 1 mile west of the ED (p < 0.0001), and the presence of a competing ED within 1 mile (p < 0.0001) negatively impacted block group ED visit rates. Travel distance was related to the percentage of visits that were high acuity (p < 0.0001), daytime (p < 0.01), or resulted in admission (p < 0.0001). CONCLUSIONS: Geography and travel distance significantly impact ED visits. Copyright © 2011 Elsevier Inc. All rights reserved.