Bayesian methods in preoperative risk assessment for cardiac surgery

Many strides have been made in the last decade to improve the accuracy of preoperative risk estimation, particular for cardiovascular surgery. It is our goal to estimate the preoperative risk associated with cardiac bypass surgery for patients in different risk categories. These risk categories are determined by the Parsonett model. The Parsonett model assigns a risk value to a range of risk factors consisting of patient attributes and disease parameters. Logistic modeling is applied to generate a comprehensive risk function. The database being utilized contains over 3,000 patients who have had cardiovascular surgery within the last 5 years. This thesis will utilize a database comprised of preoperative risk categories and their respective surgical outcomes in order to uniformly rate institutional and surgical performance.