Risk factors for 30-day readmission after cardiac surgery : a multisite Victorian study

Background: In international settings up to 40% of patients are readmitted within 30-days of cardiac surgery. In Australia >12,000 adults undergo cardiac surgery annually. Evidence describing rates and reasons for readmission is sparse. The aim of this study was to determine incidence and predictors for 30-day readmission after cardiac surgery in three Victorian hospitals. Methods: Site-specific data, prospectively collected for contribution to the Australian and New Zealand Society of Cardiac and Thoracic Surgeons database, from a major public, major private, and regional Victorian centre were analysed. From2011 to 2016 data were collected on 9,860 patients. Readmission was defined as an inpatient within 30 days from surgery. Patients presenting to the emergency department, admitted to short-stay, or transferred to rehabilitation were not classified as a readmission. Results: A total of 1,004 (10.2%) patients were readmitted within 30 days. Eighty-seven (5.3%) were readmitted for more than one reason, mostly related to surgery (n = 764, 7.8%). Univariate analysis of demographic characteristics, cardiac risk factors, operative variables, and early postoperative complications revealed 23 factors associated with readmission. A direct entry logistic regression model retained preoperative immunosuppression (OR 1.58, 95% CI 1.13–2.19), atrial fibrillation (OR 1.32, 95% CI 1.11–1.57), and New York Heart Association Class 3 (OR 1.27, 95% CI 1.09–1.48), and postoperative red blood cell transfusion (OR 1.23, 95% CI 1.07–1.41) and serum creatinine (OR 1.00, 95% CI 1.00–1.01) as predictors of readmission (all p≤0.002). Despite excellent goodness-of-fit (Hosmer-Lemeshow 0.98) the model had poor explanatory value (Nagelkerke R2 2.7%). Conclusion: Patient and process factors require investigation to determine impact on readmission rates.