St. Luke’s uses AI, real-time monitoring to improve sepsis prevention and survival rates
St. Luke’s University Health Network is using a combination of artificial intelligence and continuous patient monitoring to improve detection and treatment of sepsis, one of the most life-threatening conditions faced by hospital patients.
The health system has implemented a system that tracks patient data around the clock, helping care teams identify early warning signs before conditions worsen.
“We utilize the Epic Sepsis Model – predictive AI – in the background, along with programming engineered by St. Luke’s, to monitor every patient every moment of the day for rising risk,” said Charles Sonday, St. Luke’s Associate Chief Medical Information Officer. “If a patient rises above a certain threshold … an alert goes to both the provider and the nurse to examine that patient and begin treatment.”
Officials said the approach combines technology with direct clinical response, allowing staff to intervene earlier and reduce the likelihood of severe complications or intensive care unit admissions.
Across the network, patient information including vital signs, lab results and clinical notes is continuously monitored and transmitted to a virtual response center staffed by registered nurses. Patients showing signs of deterioration are flagged for immediate attention.
The system has contributed to improved outcomes, including higher survival rates and fewer unexpected ICU transfers, according to the network.
“The sepsis bundle is a list of elements that focus on very timely metrics,” said Diana Tarone, St. Luke’s Senior Network Director of Quality. “We are checking to see if the patient is hypotensive and needs a vasopressor, what antibiotics are administered. It is a constant and consistent watching, monitoring and reassessing of the patient.”
At the virtual response center, clinicians monitor multiple patients simultaneously, prioritizing those at highest risk.
“It’s a very complex list of factors and there has to be a lot of communication between the physician team, the advanced practitioner team and the nursing team,” said Dr. Jennifer Axelband, St. Luke’s Associate Medical Director of Critical Care Education. “We use AI monitoring in the background to help us stay compliant with these bundles because statistics show that if you stay compliant with the metrics established by CMS, it decreases patient mortality by 30%.”
St. Luke’s reported sepsis treatment compliance rates of nearly 88% in its most recent fiscal year, compared with a national average of 64%, according to data from the Centers for Medicare and Medicaid Services.
