Incidence and Predictors of 30-day Readmission of Patients in Patients Hospitalized with Acute Pancreatitis: An Analysis of National Readmission Database

Abstract

Introduction: Few studies have evaluated the relationship between patient characteristics in acute pancreatitis and risk for hospital readmission within 30 days. We aimed to study patient, hospital and society-level predictors of readmission after hospitalizations for acute pancreatitis. Methods: We performed a retrospective analysis using the National Readmission Database (NRD) in adult patients (Age > 18 years) discharged with a principle diagnosis of acute pancreatitis from January 2013 - December 2013. Index stays were re identified from January to November to allow for a 30-day readmission window. Index stays required “live” discharge status and non-missing length of stay. Clinical variables were based on secondary diagnoses at the time of first admission. Logistic regression was used to determine the predictive value of the selected variables. Statistical analyses were performed using SAS (Version 9.4, The SAS institute, Cary, NC). Results: 243,816 index acute pancreatitis discharges were identified during the study period, with 39,623 readmissions within 30 days (16.2%). Odds of readmission increased after hospitalizations complicated by bile duct obstruction {OR 1.26 (1.15 - 1.39)}, while they decreased in patients who had had cholecystectomy {OR 0.46 (0.42 - 0.51)} during the index admission. Readmissions also were less likely with advancing age {45-64yrs: OR = 0.90(0.85-0.95); 65-84yrs: OR = 0.53(0.49-0.58); >85yrs: OR = 0.47(0.41-0.53)} and females gender {OR 0.95 (0.91-0.99)}. Discharges associated with private insurance had a lower likelihood of readmission compared to Medicare {OR 0.57 (0.53-0.61)}, and there was no difference between Medicare and Medicaid {OR 1.00 (0.93-1.07)}. Rising Charlson comorbidity scores correlated with a higher odds of readmission {1.08 (1.03 - 1.14) for CCS of 1 and OR 1.44 (1.36 - 1.51) for scores >1}. In addition, readmissions of patients discharged to facilities (LTC) {OR 1.36 (1.25 - 1.49)} were higher compared to routine home discharges. Conclusion: In this exploratory nationwide database study, 16.2 % of acute pancreatitis discharges were readmitted within 30 days. Importantly, cholecystectomy during the index admission was associated with a 50% less chance of readmission, supporting the current guidelines for performing early cholecystectomy in mild biliary pancreatitis. Higher Charlson comorbidity scores and discharge to LTC facility were also associated with higher readmission rates.Figure 1