UF Health Researchers Use Artificial Intelligence to Predict Surgical Complications
In an article published in the Journal of the American Medical Association in May 2022, UF Health researchers sought to address the role an artificial intelligence platform, MySurgeryRisk, can play to accurately predict surgical outcomes. This research focused on whether MySurgeryRisk would demonstrate stable performance between development and prospective validation phases, and the feasibility of the system to collect data and disseminate real-time information to physicians’ mobile devices.
Researchers, taking a collaborative, interdisciplinary approach, conducted a prognostic study of more than 74,000 surgical procedures involving approximately 58,000 adult patients from 2014 to 2020 using automated electronic health record, or EHR, data inputs and multiple algorithm models composed of 135 variables. Algorithms were constructed to identify patterns to predict post surgery complications, including sepsis, wound complications and cardiovascular and neurologic issues.
They found MySurgeryRisk accurately predicted postoperative complications using automated real-time EHR data and mobile device outputs. Good predictive performance of an AI system can benefit both doctors and patients by expanding surgical decision-making and decreasing complications.
This is one of several AI initiatives underway at UF Health and the University of Florida to leverage the advantages of AI for better health outcomes based upon GatorTron™, a first-of-its-kind technology designed to mine electronic health records in a new way. GatorTron™ is a collaboration between UF and NVIDIA, an AI partnership that began with a gift to UF of an NVIDIA DGX SuperPOD AI supercomputer, the nation’s most powerful university-owned artificial intelligence supercomputer.