St. Luke’s Uses AI in Hospitals to Head Off Cardiac Arrest and Other Problems

When you or a loved one is admitted to the hospital’s medical-surgery unit (med-surg), St. Luke’s University Health Network has an award-winning extra layer of care. 

All med-surg units at St. Luke’s use a proprietary artificial intelligence (AI) tool to predict, identify and rapidly address the needs of its admitted or under-observation patients to reduce unanticipated transfers to the ICU – and save lives. 

It’s like a superhero sidekick: AI that watches over people in the hospital every second of every day, awake or sleeping. 

This tool looks at metrics such as heart rate, blood pressure, breathing, lab work, patient age and nursing documentation in real time at all 16 St. Luke’s campuses.  

The information is transmitted to the St. Luke’s Virtual Response Center, which is staffed by a team of registered nurses who, depending upon a patient’s deterioration score, can alert the patient’s bedside nurse, a nursing supervisor, or in the more rapidly deteriorating patients, alert St. Luke’s rapid response team (RRT) to identify and treat patients early, before they require transfer to the ICU or have a sudden cardiac arrest in the med-surg unit. 

Earlier this year, St. Luke's utilization of the AI tool called Epic’s Deterioration Index was recognized by The Hospital and Healthsystem Association of Pennsylvania (HAP) with a 2025 “In Safe Hands” Achievement Award” for the initiative that provided significant measurable gains in operational, quality and financial metrics while reducing cardiac arrest events, overall morbidity and mortality. 

This proactive approach led to a 34% reduction in cardiac arrests and a 12% drop in rapid response transfers to the ICU, with survival rates climbing from 24% to 36%.   

Beyond clinical outcomes, the Deterioration Index predictive model has helped to deliver measurable budgetary impact, saving $1.75 million in 2024 by avoiding unnecessary ICU transfers. 

“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 Deterioration Index transforms patient safety through predictive analytics.  

Today, the initiative is deeply embedded in St. Luke’s workflow for physicians, nurses and advanced practitioners so they are alerted immediately when a patient's Deterioration Index score crosses critical thresholds before something goes critically wrong, allowing providers and nurses a chance to intervene faster with the proper treatment. 

As part of the Network’s broader AI strategy, the Deterioration Index exemplifies how machine learning can be responsibly integrated into care delivery, improving outcomes while still preserving and enhancing clinician judgment. 

This initiative has enhanced patient care outcomes by improving survival rates, diminishing unexpected ICU transfers, and lowering code blue (cardiac arrests) incidents over a period of three fiscal years.