Researchers at RWJ Barnabas Health and Rutgers Robert Wood Johnson Medical School found that an artificial intelligence (AI)-enabled early warning system helped identify hospitalized patients at risk of rapid clinical deterioration sooner, contributing to fewer deaths among high-risk patients.
Published in NEJM AIThis is a diary. New England Medical Journal Group (DOI: 10.1056/AIoa2500973), This study evaluated outcomes for 23,132 high-risk patients at 11 RWJBarnabas Health hospitals. After implementing an AI-powered early warning system, mortality rates among high-risk patients decreased from 23.1 percent to 18.6 percent. This represents an 18 percent reduction in the risk-adjusted probability of in-hospital death.
Hospitalized patients can deteriorate rapidly, often before any obvious warning signs become apparent. Researchers evaluated the Epic Deterioration Index (EDI), an AI-enabled tool that continuously analyzes information already recorded in electronic medical records, such as vital signs, test results, nursing assessments, and age, to identify patients at high risk for severe clinical decline. The system recalculates the risk score every 15 minutes and automatically alerts rapid response teams when a patient reaches the highest risk category.
Before evaluating the technology in this study, RWJBarnabas Health and Rutgers spent several years developing and implementing a system-wide approach to using EDI across the hospital. The health system first integrated the tool into the electronic medical record at Robert Wood Johnson University Hospital, an academic medical center, piloted and refined the method and timing of alert delivery, established automated notifications for rapid response teams, trained clinicians in its use, and continuously monitored its performance. Rutgers researchers then partnered with RWJBarnabas Health to evaluate the impact of this approach in real-world clinical settings and rapidly deploy the platform to 10 other hospitals.
Our goal was to identify patients early, before they reach the point where intervention becomes much more difficult. The deterioration index provides an earlier point in time. Getting patients into critical care sooner can change the course of patient outcomes. ”
Thomas Nahas, MD, vice president of medical informatics and intensivist at RWJBarnabas Health, assistant professor of medicine at Rutgers University Robert Wood Johnson Medical School, and lead author of the study
When a patient reaches the highest risk threshold, automatic notifications are sent directly to the hospital’s rapid response team, allowing critical care professionals to quickly assess the patient and determine if additional intervention is required. After implementation, rapid response team activity among high-risk patients increased from 25.3 percent to 37.5 percent of hospital stay.
This study evaluated outcomes for high-risk adult patients treated at academic medical centers, regional teaching hospitals, and community hospitals across the RWJBarnabas Health system. Despite the increase in rapid response assessments, transfers to intensive care units did not significantly increase, while mortality rates decreased significantly.
“Once a patient’s condition begins to deteriorate, every minute counts,” said Dr. Andy Anderson, chief medical quality officer at RWJBarnabas Health and co-author of the study. “This study shows that combining AI-enabled tools with experienced clinical teams can help us identify at-risk patients faster and provide the right care at the right time. These findings highlight the potential for innovation to improve quality, safety, and outcomes for the patients we serve.”
“This is the purpose of an integrated academic health system,” said Stephen P. O’Mahony, MD, FACP, senior vice president and chief medical information officer at RWJBarnabas Health and lead author of the study. “We combined Rutgers’ rigorous methodology with the operational reach of RWJBarnabas 11 hospitals. The mortality benefit was not created by the algorithm, but by the partnership around the algorithm.”
The researchers noted that the improvement in mortality likely resulted from a combination of factors, including staff education, increased clinical awareness, electronic medical record alerts, and automated rapid response team notifications working together as a system-wide coordinated approach. Because this study evaluated the Epic Deterioration Index, a tool already available within Epic, one of the most widely used electronic health record systems in the country, the results could have implications for hospitals across the country looking to improve patient outcomes. Researchers are now evaluating the next stage of this effort, which will focus on identifying patients with rapidly rising risk scores in the hope of allowing even earlier intervention.
sauce:
Reference magazines:
Nahas, T.A. Others. (2026) Implementing rapid response with AI—correlation with mortality. NEJM AI. DOI: 10.1056/AIoa2500973. https://ai.nejm.org/doi/10.1056/AIoa2500973
