St. Luke’s uses AI system to predict patient decline, cut cardiac arrests and ICU transfers
St. Luke’s University Health Network is using artificial intelligence across all of its hospitals to monitor patients around the clock and intervene before medical conditions worsen — an approach that has significantly reduced cardiac arrests and emergency transfers to intensive care.
All medical-surgical units at St. Luke’s employ Epic’s Deterioration Index, a proprietary AI tool that analyzes real-time data such as heart rate, blood pressure, breathing patterns, lab results, patient age and nursing documentation across the network’s 16 campuses. The system continuously calculates a patient’s risk of decline and relays that information to the St. Luke’s Virtual Response Center.
There, registered nurses monitor the AI-generated scores and alert bedside nurses, nursing supervisors or the rapid response team when a patient’s condition appears to be deteriorating.
Earlier this year, the Hospital and Healthsystem Association of Pennsylvania recognized St. Luke’s for its use of the system, awarding the network a 2025 “In Safe Hands” Achievement Award for measurable improvements in patient safety and outcomes.
The predictive model has helped drive a 34% reduction in cardiac arrests and a 12% decrease in rapid response transfers to the ICU, according to the network. Survival rates increased from 24% to 36%, and the technology helped avoid unnecessary ICU transfers, saving an estimated $1.75 million in 2024.
“Humans could essentially calculate these risk scores,” said Julie Tanhauser, Information Technology Strategic Planner at St. Luke’s, “but we can’t continuously do it every second for every patient among hundreds of patients on our 16 campuses. That’s your AI piece.”
The tool has become fully integrated into the workflow for physicians, nurses and advanced practitioners, allowing clinicians to respond quickly when a patient’s score reaches critical thresholds.
St. Luke’s leaders say the initiative showcases how machine learning can be responsibly incorporated into hospital care, strengthening clinical judgment while enhancing patient safety. Over the past three fiscal years, the program has reduced unexpected ICU transfers, lowered cardiac arrest rates and improved overall patient survival.
