The literature suggests readmission rates within 30 days of discharge after coronary artery bypass grafting (CABG) are high (10.8-20.9%). As part of value-based purchasing, within the next two years Medicare is expected to begin using clinical outcomes, like readmission, as metrics for determining reimbursement.

Although prior studies examined factors associated with readmission after CABG, none have developed a predictive model that identifies individuals at risk. Such a model would be instrumental in appropriately focusing resources to prevent readmission.

A univariate analysis of institutional data (demographic/preoperative/intraoperative/postoperative specifics) from all patients undergoing cardiac surgery between March 2003 and September 2010 was performed to determine factors related to 30-day readmission. Chi-squared and independent t-test were used for categorical and continuous variables, respectively. A logistic regression was performed with variables having tests with p-value ≤ 0.02.

Independent predictors of readmission were the presence preoperatively of congestive heart failure, chronic lung disease, body mass index ≥ 40, and time on cardiopulmonary bypass. A strong correlation between predicted and actual length of stay was uncovered, enhancing the ability to identify those at increased risk.

From the data, a mathematical model was developed to serve as a function for predicting probability of readmission.

By using patient-specific parameters for individual risk assessment for early readmission, a predictive model will provide the advantage of foresight when arranging postoperative services and determining follow up.

The ultimate goal is quality improvement by reducing the readmission rate from preventable processes. The model will help in achieving this and consequently allow for maximum Medicare reimbursement.