Thirty days are inadequate for assessing readmission following complex hepatopancreatobiliary procedures

Introduction: Early readmissions (30 days) have been used as a measure of health care quality. The purpose of our study was to evaluate patterns of readmission for a longer period (up to 2 years) following Hepatopancreatobiliary (HPB) surgery in the state of New York. Methods: The State Planning and Research Cooperative System database was utilized to identify patients undergoing complex HPB procedures between 2010 and 2012. Patients were followed for 2 years following surgery to identify all-cause readmissions. Factors for readmissions included patient demographics, comorbidities, perioperative complications, surgery type, and academic status. Multivariable generalized linear mixed models were performed to evaluate risk factors for readmissions. Results: There were 6207 complex HPB procedures with 1272 (20.49%) unplanned 30-day readmissions, 816 (13.15%) unplanned 31–90-day readmissions, 1678 (27.03%) unplanned 91-day to 1-year readmissions, and 1404 (22.62%) 1–2-year readmissions. After adjusting for other possible confounding factors, risk factors for 30-day readmissions include surgery type, as pancreatectomy and gallbladder patients are more likely to have a 30-day readmission than hepatectomy patients, facility type, as academic centers are more likely to have a readmission, male gender, presence of any comorbidity, and peri-operative complications. Risk factors for 31–90-day readmissions include race, insurance group, any comorbidity or any peri-operative complication, and 30-day readmissions. Risk factors for 91-day to 1-year readmissions include male gender, race, any comorbidity, 30-day readmissions, and 31–90 days’ readmissions. Risk factors for 1–2-year readmissions include presence of any comorbidity, and previous 91-day to 1-year readmissions. Conclusion: The 30-day readmission window is an inadequate, but predictive, measure of total readmission following complex HPB procedures.