288 Characteristics of Repeat Emergency Department Users at a University Medical Center: Frequent Emergency Department Utilization Is Associated With Higher Rates of 30-day Inpatient Readmission

Study Objectives: Repeat emergency department (ED) users are common in academic and community hospitals alike. We conducted this study to characterize repeat ED users, according to frequency of repeat encounters. Methods: We performed a retrospective analysis of all ED visits over a 6-month period using an electronic database. Repeat ED users (REDUs) were defined as having 2 or more ED visits within the study period. We categorized patients as non-REDUs (single visit), low REDUs (2-3 visits), moderate REDUs (4-9 visits) and high REDUs (> 9 visits). Comparisons were made to identify differences between REDUs and non-REDUs according to several demographic, medical and visit-related characteristics. Admission and 30-day readmission rates were calculated. Statistical analysis was performed using Chi-Square tests for categorical data and Fisher's exact test for small sample sizes. Continuous data were analyzed using one way ANOVA and Student's t-tests. Results: REDUs accounted for 19.7% of all individual ED patients and 39.7% of visits. Low REDUs accounted for 28.9% of visits averaging 2.2 visits per patient, moderate REDUs accounted for 8.9% of visits averaging 4.9 visits per patient and high REDUs accounted for 1.9% of visits averaging 16.9 visits per patient. REDUs were more likely to be African American, living within close proximity to the hospital, publically insured, indigent and chronically ill. High REDUs were more likely to receive Medicare coverage (42.9%) and had significantly greater odds of having psychiatric comorbidites. Of note, low and moderate REDUs demonstrated no significant difference in the rate of inpatient admission per visit relative to that of non-REDUs (14.2%, 14.9% and 15.1% of ED visits, respectively), while high REDUs had significantly lower admission rates per visit (4.6%, p < 0.001). When admitted, all REDU groups demonstrated significantly higher rates of 30-day readmission (low- 25.8%, moderate- 42.1% and high- 68.2% of inpatient visits, p values < 0.05) in comparison to non-REDUs (6.7%). High REDUs had significantly shorter lengths of stay (average 3.86 days, p < 0.05) compared with non-REDUs at 6.06 days. Conclusion: REDUs tend to have more chronic illness, greater numbers of comorbidities and higher 2-year mortality than non-REDUs. Admission rates by visit were similar for all categories except for high REDUs, indicating that ED visits by low to moderate REDUs are justified as necessary. While high REDUs incur fewer and shorter inpatient admissions, they exhibit the highest rates of 30-day readmission. Federal programs designed to penalize providers for 30-day readmissions may be unreasonable for all patient populations.