Benchmarking Low Volume / High Risk Trauma Procedures to Improve Surgical Outcomes and Improve Readiness Using Machine-Learning Algorithms

In May 2019, the Defense Health Board (DHB) released a report finding that the Military Health System (MHS) was endangering patient and warfighter lives by not adequately training for, measuring, and benchmarking the outcomes of low-volume, high-risk procedures. More than 30 years of healthcare research demonstrate that patient outcomes for procedures are significantly better in facilities that routinely complete them. One analysis by US News and World Report found that facilities in the top quintile of volume of heart failure cases had mortality rates 23% lower than facilities with lower patient volumes. Alexandria Health has developed a novel machine learning (ML) -based benchmarking methodology to drive better outcomes, reduce healthcare costs, and improve the patient experience. The Predictive OpTimal Trees in Emergency Surgery Risk (POTTER) Calculator uses AI-developed decision trees to sort patients into distinct and highly relevant cohorts, allowing medical professionals to compare outcomes across identical patient groups, showing how individual physicians, hospitals and systems compare against each other and industry standards.  Alexandria Health has partnered with MIT to train AI/ML-powered algorithms and develop medical expert-guided decision trees to assess surgical outcomes in trauma and emergency surgeries for the DHA’s Joint Trauma System and the 59th Medical Wing.