Knowledge Bank

Cancer patients are at a high risk of hospital readmissions due to the complex nature of the disease. Currently, there is not a reliable predictive measure to assess hospital readmissions in cancer patients, but there have been attempts to determine readmissions based on other factors. The purpose of the study is to determine whether or not assigning a logistic model driven risk score to a patient at the time of discharge is an accurate predictor of readmissions, and to determine which additional factors influence readmissions. If assigning a risk score at the time of discharge is an accurate predictor of readmissions, it could allow other cancer designated hospitals to implement a similar predictive model. Predetermined variables were extracted from patient medical records within the Electronic Health Record (EHR), and these variables were compiled into a database where a predefined algorithm was implemented to calculate patient risk scores. The predicted likelihood of a readmission was then compared to the actual 30-day readmission status of the patients as a measure of the reliability of the risk score. After a data analysis was conducted, it was found that the readmission predictive model was accurately predicting readmissions based on a Chi-Square test (p< 0.001). The Kappa score for the agreement between actual readmission and the patient’s assignment to a high risk category was 0.125, which is a low Kappa score, and shows that there is work to be done to make the model better at predicting readmissions.