Background: In acute care surgery, predicting mortality is important to determine appropriate patient transfer to a regional emergency general surgery (EGS) center. We hypothesized that distance to a referral center and severity of illness (SOI) would be predictors of death.
Methods: We performed a retrospective analysis of a prospectively collected EGS registry from 2004 to 2008. The study population consisted of all patients discharged from the EGS service with an available home zip code in the registry. Study data included age, gender, length of stay (LOS), intensive care unit (ICU) LOS, distance between our facility and patient home zip code, and need for operative management. Systemic inflammatory response syndrome/sepsis/shock, peritonitis, perforation, and acute renal failure were used as SOI indicators. Mortality at discharge was the primary outcome. Patients were stratified by survival and compared using non-parametric statistical tests. Logistic regression assessed the simultaneous contribution of age, SOI, and distance to risk of death.
Results: A total of 3,439 patients met study criteria. Females slightly outnumbered males (1,813, 52.7%) with a median age of 47 years. The overall LOS was 6.4 days±9.3 days, and 2,331 (67.8%) of the patients underwent operation. Mean distance was 41.5 miles±51.2 miles (median, 22.2). Overall mortality was 2.7%. Increasing distance, age, and presence of SOI indicators were associated with mortality in univariable analyses. In multivariable logistic regression controlling for patient age and SOI, increasing distance in miles was related to increased mortality (odds ratio, 1.005; p<0.001). This odds ratio equates to a doubling in odds of death for each 132 miles between our center and the patient's home zip code.
Conclusion: Age, SOI, and distance from a regional referral center explain much of the variation in mortality and can be used for triage to regional EGS centers.