Opportunistic Biomarkers of Frailty in Acute Cardiac Care

We and others have previously demonstrated the ability of frailty markers to predict mortality and major morbidity in older adults with heart disease. The unanswered question remains how to measure frailty in the emergency department or cardiac intensive care unit where urgent decisions must be made yet where the usual frailty questionnaires and physical tests are either unfeasible or unreliable. There is a timely need to improve the status quo of frailty assessment options with markers that are objective, reliable, practical, and adaptable to the logistical challenges posed by acutely ill patients. Thus, our objective is to use machine learning to extract frailty markers from readily available imaging tests, laboratory tests, and electrocardiograms; and demonstrate their value for the prediction of fatal and nonfatal adverse health events in hospitalized patients.