Risk score for unplanned vascular readmissions

Abstract

Vascular surgery patients have high readmission rates, and identification of high-risk groups that may be amenable to targeted interventions is an important strategy for readmission prevention. This study aimed to determine predictors of unplanned readmission and develop a risk score for predicting readmissions after vascular surgery. The National Surgical Quality Improvement Program database for 2011 was queried for major vascular surgical procedures. The primary end point was unplanned 30-day readmissions. The data were randomly split into two-thirds for development and one-third for validation. Multivariable logistic regression was used to create and validate a point score system to predict unplanned readmissions. Overall, 24,929 patients were included, with 2507 readmissions (10.1%). A point-based scoring system was developed with the use of factors predictive for readmission, including procedure type; discharge destination; race; non-elective presentation; pulmonary, renal, and cardiac comorbidities; diabetes; steroid use; hypoalbuminemia; anemia; venothromboembolism before discharge; graft failure before discharge; and bleeding disorder. The point score stratified patients into 3 groups: low risk (0-3 points) with a readmission rate of 5.4%, moderate risk (4-7 points) with a readmission rate of 8.6%, and high risk (≥ 8 points) with a readmission rate of 16.4%. The model had a C-statistic = 0.67. Through the use of patient, operative, and predischarge events, this novel vascular surgery-specific readmission score accurately identified patients at high risk for 30-day unplanned readmission. This model could help direct discharge and home health care resources to patients at high risk, ultimately reducing readmissions and improving efficiency.