A. M. Ingraham1, S. Fernandes-Taylor1, J. Schumacher1, X. Wang1, M. Saucke1, C. C. Greenberg1 1University Of Wisconsin,Department Of Surgery,Madison, WI, USA
Introduction: Transferred emergency general surgery (EGS) patients constitute a highly vulnerable, acutely ill population. Guidelines to facilitate timely, appropriate EGS transfers are lacking. We determined patient- and hospital-level factors associated with interhospital EGS transfers. Determining these factors represents a critical first step in identifying who necessitates transfer and minimizing delays to definitive care.
Methods: Adult EGS patients (defined by American Association for the Surgery of Trauma ICD-9 diagnosis codes) were identified within the 2008-2013 Nationwide Inpatient Sample (NIS) (n=17,175,450). Patient- and hospital-level factors were examined as predictors of transfer to another acute care hospital with a multivariate proportional cause-specific hazards model. Because patients may succumb to death or discharge to other locations rather than transfer, a competing risks analysis considering the NIS design assessed the effect of risk factors for transfer. In addition to variables in the Table, the model included patient-level characteristics (sex, age, race, insurance, patient income based upon zip code, Charlson Comorbidity Index [CCI], EGS diagnosis, procedures performed, day of admission) and two hospital-level factors (total number of discharges and region).
Results: 1.8% of patients were transferred (n=317,357). Transferred patients were on average 62 years old and most commonly had Medicare (52.9% [n=167,921]), private (26.7% [n=84,851]), or Medicaid insurance (10.7% [n=34,020]). 67.8% were white. The most common EGS diagnoses among transferred patients were related to hepatopancreatobiliary (n=90,734 [28.6%]) and upper gastrointestinal tract (n=59,958 [18.9%]) conditions. Most transferred patients (n=269,215 [84.8%]) did not have a procedure prior to transfer. Transfer was more likely if patients were in small or medium versus large facilities, government versus private facilities, and rural or urban non-teaching versus urban teaching facilities (Table). Patient-level factors associated with transfer included male sex (Hazard Ratio [HR] 1.09 [95% Confidence Intervals (CI) 1.07-1.11]), CCI (HR CCI of 2 1.01 [95% CI 1.04-1.11] and HR CCI of 3 1.17 [95% CI 1.13-1.20]), and admission on a Saturday or Sunday (HR 1.04 [95% CI 1.02-1.06]).
Conclusion: We identified patient- and hospital-level characteristics of EGS transfers to another acute care hospital. Hospital-level characteristics more strongly predicted the need for transfer than patient-related factors. Consideration of these factors by providers at non-tertiary centers as they care for EGS patients in the context of the resources and capabilities of their local institutions may reduce time to definitive care and improve patient outcomes.
