OBJECTIVE Emergency department (ED) cardioversion and discharge of atrial fibrillation (AF) is an evolving treatment. Emergency department cardioversion patients have few comorbidities, and their discharge directly from the ED leads to a sicker in-patient population of AF patients. This study examines whether the quality care markers, hospital charges (HC) and length of stay (LOS), negatively reflect the practice of ED cardioversion. METHODS Median HC and LOS were determined for 2 different quality assessment reporting models. In a standard model (SM), patients discharged from the ED were not included in any hospital statistics and only admitted, or observation patients were used to calculate the HC and LOS of AF patients. In an inclusive model (IM), patients discharged from the ED were also included in the hospital statistics but given the same LOS as observation patients. Differences across medians were analyzed using Wilcoxon rank sum tests. RESULTS A total of 312 patients were evaluated for AF over an 18-month period. Of these, 197 (62%) were admitted, 21 (7%) were placed in observation status, and 95 (31%) were discharged from the ED. Median values for LOS were 3 days (interquartile range [IQR], 1-5) for the SM and 1 day (IQR, 0-4) for the IM. Median values for HC were $33062 (IQR, $19267-$60614) for the SM and $20059 (IQR, $4249-$47195) for the IM. CONCLUSION Emergency department cardioversion selects out a less sick cohort of patients whose removal from a hospital's admission numbers negatively skews quality performance profiles.