Abstract Background For nearly 51 million persons in the United States who lack health care insurance, the emergency department (ED) functions as a safety net where no patient is denied care based on ability to pay, and much public rhetoric has characterized ED utilization by uninsured patients. We estimated national ED utilization by uninsured patients and compared uninsured and insured ED patients in terms of demographics, diagnostic testing, disposition and final diagnoses. Methods We analyzed data from the National Hospital Ambulatory Medical Care Survey (2006-2009) stratified by insurance status. Demographic data, diagnoses, testing, and procedures performed in the ED were tabulated for each visit. Weighted percentages provided by National Hospital Ambulatory Medical Care Survey were used to estimate national rates for each variable, and multivariate models were constructed for predicting testing, procedures, and admission. Results The 135085 ED visits represent 475 million patients visits, of which 78.9 million (16.6%) were uninsured. Compared with insured patients, uninsured patients were more often male (51.1% vs 44.3%) and younger (age 18-44 years, 66.2% vs 35.4%). Uninsured patients had lower rates of circulatory/cardiovascular (7.5% vs 4.1%) and respiratory diagnoses (14.6% vs 11.8%). Uninsured patients had fewer diagnostic tests and procedures and fewer hospital admissions than those with insurance. In our multivariate models, insurance status was predictive of testing and procedures but not hospital admission. Conclusions Uninsured patients account for approximately 20 million or 1 in 6 ED visits annually in the United States and have differences in demographics, diagnoses, and ED utilization patterns from those with insurance.