Abstract
This approach focused on identifying specific variables that predict the likelihood of readmission. It involved clinical, utilization, and demographic variables that are generally available on hospital computer abstract databases. The approach included a process for identifying and comparing individual variables with the highest risk of readmission. It also contained a procedure for assembling risk populations including combinations of variables. The approach demonstrated the potential for using risk analysis to maximize the focus of clinical management on patient outcomes while reducing the amount of resources required for this process.