Seeking a Second Opinion: Use of the Same versus a Different Hospital for Patients Requiring Recurrent Emergency Department Visits

Abstract

Implementation of the Centers for Medicare and Medicare Services Hospital Readmission Reduction Program in 2012 fueled research regarding recurrent hospital utilization. This research has primarily focused on inpatient utilization and has been limited by use of hospital-level datasets which do not allow for tracking of patients across institutions and use of large administrative datasets which lack detailed information about individual patient visits. Under-estimation of event rates is a common limitation of single institution or hospital system-based research. Rates of same versus different emergency department (ED) utilization, however, have not yet been documented. The primary objective of this study is to determine the proportion of 30-day ED revisits occurring at a differnt hospital as the index visit. We hypothesized that 20% of 30-day ED revisits occur at a different hospital than the index ED visit. This is a retrospective single-state cohort study using data describing all ED treat-and-release visits and 30-day return ED visits in Florida (FL) from July, 1 2010 - June 30, 2011. Data were obtained from the State Emergency Department Database (SEDD) and the State Inpatient Database (SID) of the Agency for Healthcare Research and Quality Healthcare Cost and Utilization Project. All ED treat-and-release visits were followed forward to identify the next ED visit within 30 days and the hospital where that visit took place. ED treat-and-release without a subsequent 30-day visit were excluded from analysis. The primary outcome of the percent of ED revisits occurring at a different hospital as compared to the prior ED visit, per day, was modeled using beta regression. There were 1,156,441 ED discharges that had a subsequent 30-day ED revisit, with 33% of these return visits (386,368) occurring at a different hospital. The proportion of ED revisits occurring at a different hospital changed significantly in the first 72 hours after an index ED visit as compared to revisits occurring after 72 hours (z=-24.3, p<0.001). In unadjusted analysis, 64.5% of re-visits that occurred on the same day (day 0) as the index visit occurred at a different hospital than the index visit, 38.3% of re-visits on day 1 occurred at a different hospital, and 26.7% of revisits on day 2 occurred at a different hospital. After day 3, the proportion of ED revisits occurring at a different hospital than the index visit remained relatively constant at 30-31%. (Figure). This study found that a substantial proportion of ED revisits occur at different hospitals than the site of the index ED discharges, especially in the first 24-48 hours after an ED discharge. These findings highlight the importance of performing research regarding hospital utilization at a level larger than a single institution, as at least 30% of return visits are likely to be missed within a single institution. In addition, they emphasize the important of identifying effective methods to facilitate communication about patient care across institutions to maximize the delivery of safe and efficient emergency care.