OBJECTIVES: ED use for non-emergencies contributes to crowding, increased healthcare costs, and, potentially, poor clinical outcomes.  Efforts to decrease inappropriate ED use have focused on insured individuals.  The purpose of this retrospective cohort study was to characterize frequent ED use among un- and underinsured individuals in central Texas. METHODS: Data were obtained from the ICare system, which includes information for > 800,000 individuals and > 5 million encounters within 24 Central Texas healthcare providers who arrange for or provide care for uninsured or underinsured individuals.  Persons who received care from an ICare-participating organization during calendar year 2007 were included in these analyses.  Frequent ED use was defined as at least 6 ED visits within either a calendar quarter or any contiguous three-month period.  Linear regression was used to estimate the relation between patient characteristics and total ED visits among frequent users. RESULTS: There were 216,364 ED visits in 2007; 128,538 individuals had at least 1 ED visit and 0.7% (n = 892) were considered frequent ED users.  Frequent users were mainly female (55.6%), and Caucasian (55.8%).  Hispanics and African-Americans accounted for 14.2% and 17.3% of frequent users, respectively. The regression model accounted for 12.1% of variability in the outcome.  Total clinic visits (beta = 0.05, p 0.037, 95% CI 0.003 – 0.094), inpatient admissions (beta = 1.35, p <0.001, 95% CI 0.976 – 1.716) and having any mental health diagnosis in any ED visit (beta = 3.278, p <0.001. 95% CI 2.060 – 4.496) were positively and statistically significantly associated with the number of ED visits among frequent users, after adjusting for age, sex, and race. CONCLUSIONS: In 2007, frequent users accounted for <1% of ED users, and 5.4% of total ED use.  A study of these frequent users may help identify opportunities for intervention to help underlying causes for these frequent visits.