A new algorithm created by researchers at Mount Sinai Hospital, New York, has learned to identify subtle changes in electrocardiograms (ECGs) to predict whether a patient is developing heart failure.