Introduction: With the majority of U.S. hospitals not having primary percutaneous coronary intervention (pPCI) capabilities, the time spent at transferring emergency departments (EDs) is predictive of clinical outcomes for patients with ST-elevation myocardial infarction (STEMI). Compounding the challenges of delivering timely emergency care are the known delays caused by ED crowding. However, the association of ED crowding with timeliness for patients with STEMI is unknown. We sought to examine the relationship between ED crowding and time spent at transferring EDs for patients with STEMI.
Methods: We analyzed the Centers for Medicare and Medicaid Services (CMS) quality data. The outcome was time spent at a transferring ED (i.e., door-in-door-out [DIDO]), was CMS measure OP-3b for hospitals with ≥10 acute myocardial infarction (AMI) cases requiring transfer (i.e., STEMI) annually: Time to Transfer an AMI Patient for Acute Coronary Intervention. We used four CMS ED timeliness measures as surrogate measures of ED crowding: admitted length of stay (LOS), discharged LOS, boarding time, and waiting time. We analyzed bivariate associations between DIDO and ED timeliness measures. We used a linear multivariable regression to evaluate the contribution of hospital characteristics (academic, trauma, rural, ED volume) to DIDO.
Results: Data were available for 405 out of 4,129 hospitals for the CMS DIDO measure. These facilities were primarily non-academic (99.0%), non-trauma centers (65.4%), and in urban locations (68.5%). Median DIDO was 54.0 minutes (IQR 42.0,68.0). Increased DIDO time was associated with longer admitted LOS and boarding times. After adjusting for hospital characteristics, a one-minute increase in ED LOS at transferring facilities was associated with DIDO (coefficient, 0.084 [95% CI [0.049,0.119]]; p<0.001). This translates into a five-minute increase in DIDO for every one-hour increase in ED LOS for admitted patients.
Conclusion: Among patients with STEMI presenting to U.S. EDs, we found that ED crowding has a small but operationally insignificant effect on time spent at the transferring ED.