A Machine Learning Early Warning System for Intraoperative Hypotension

This randomized trial evaluated depth and duration of intraoperative hypotension during noncardiac surgery, comparing a machine learning early warning system to standard care. Hypotension was defined as a MAP <65 mm Hg for one minute or more. The early warning system significantly reduced the degree and duration of hypotension compared to controls. Control patients suffered two adverse events resulting in death (7%), whereas none occured in the intervention group.