Predictors of hospital transfer and associated risks of mortality in acute pancreatitis

Background: There is limited research in prognosticators of hospital transfer in acute pancreatitis (AP). Hence, we sought to determine the predictors of hospital transfer from small/medium-sized hospitals and outcomes following transfer to large acute-care hospitals.

Methods: Using the 2010-2013 Nationwide Inpatient Sample (NIS), patients ≥18 years of age with a primary diagnosis of AP were identified. Hospital size was classified using standard NIS Definitions. Multivariable analyses were performed for predictors of "transfer-out" from small/medium-sized hospitals and mortality in large acute-care hospitals.

Results: Among 381,818 patients admitted with AP to small/medium-sized hospitals, 13,947 (4%) were transferred out to another acute-care hospital. Multivariable analysis revealed that older patients (OR = 1.04; 95%CI 1.03-1.06), men (OR = 1.15; 95%CI 1.06-1.24), lower income quartiles (OR = 1.54; 95%CI 1.35-1.76), admission to a non-teaching hospital (OR = 3.38; 95%CI 3.00-3.80), gallstone pancreatitis (OR = 3.32; 95%CI 2.90-3.79), pancreatic surgery (OR = 3.14; 95%CI 1.76-5.58), and severe AP (OR = 3.07; 95%CI 2.78-3.38) were predictors of "transfer-out". ERCP (OR = 0.53; 95%CI 0.43-0.66) and cholecystectomy (OR = 0.14; 95%CI 0.12-0.18) were associated with decreased odds of "transfer-out". Among 507,619 patients admitted with AP to large hospitals, 31,058 (6.1%) were "transferred-in" from other hospitals. The mortality rate for patients "transferred-in" was higher than those directly admitted (2.54% vs. 0.91%, p < 0.001). Multivariable analysis revealed that being "transferred-in" from other hospitals was an independent predictor of mortality (OR = 1.47; 95% CI 1.22-1.77).

Conclusions: Patients with AP transferred into large acute-care hospitals had a higher mortality than those directly admitted likely secondary to more severe disease. Early implementation of published clinical guidelines, triage, and prompt transfer of high-risk patients may potentially offset these negative outcomes.