Many risk stratification algorithms have been developed to predict readmission after cardiac procedures and acute cardiac syndromes. However, these algorithms are not often used in actual clinical practice. Some of the few risk stratification algorithms that are routinely used in clinical practice predict different clinical outcomes such as stroke or death. One could speculate that algorithms that predict stroke or death associated with specific clinical decisions are more important to patients and caregivers than algorithms that predict readmission. However, even these algorithms that predict risk for events other than readmission have limited discrimination. For example, the median c-statistic (the ability to discriminate between patients who will have and will not have events) of the CHADS2-VASc is 0.66 (0.61–0.69) and the STS mortality score is 0.70 (0.64–0.76) [ ]. Readmission risk algorithms, on the other hand, could be especially useful from a health systems perspective, allowing institutions to target more resources at higher risk patients to reduce readmission and healthcare utilization. One of many factors that limit the use of readmission risk algorithms in clinical practice is lack of discrimination. In that setting, machine learning (ML) methods have the potential to improve our ability to predict who will experience readmission after a procedure.