New hospital software detects sepsis earlier and improves patient survival chances.
A new hospital tool designed by a research team at Johns Hopkins University could help doctors identify symptoms of sepsis much earlier, giving patients a better chance of survival. The sepsis detection system recently received clearance from the U.S. Food and Drug Administration (FDA) and is among the first medical systems of its kind, powered by artificial intelligence (AI) approved for patient care.
Sepsis is a dangerous condition that happens when the body has an extreme reaction to an infection. The immune system, which normally fights off germs, begins damaging healthy organs and tissues instead. If it’s not identified early on and treatment is delayed, sepsis can quickly become deadly. Doctors say fast treatment is one of the biggest factors in survival, but the condition is often difficult to recognize in its early stages because symptoms can look similar to many other illnesses.
Patients with sepsis may develop a fever, confusion, trouble breathing, fatigue, weakness, or a fast heart rate. These symptoms are common in hospitals for many reasons, which can make early diagnosis difficult even for experienced medical teams. In some cases, doctors may not suspect sepsis until it’s already spread throughout the patient’s body.
The new sepsis detection system was built to help solve that problem. Researchers created software that reviews electronic health records (EHRs) and patient information in real time. The system looks for warning signs that may point to sepsis before doctors or nurses notice obvious symptoms. Hospital staff then receive alerts so they can quickly examine the patient and begin treatment if needed.
According to The Johns Hopkins team, the technology has been able to detect sepsis between two and 48 hours earlier than older methods used in hospitals. That extra time may allow doctors to start antibiotics, fluids, and other treatments sooner, improving a patient’s chance for recovery.
The system is called the Targeted Real-Time Early Warning System. It was developed through federally funded research and later brought into hospital settings through the healthcare company Bayesian Health. Researchers reported that hospitals using the system saw sepsis fatalities drop by an estimated 20 percent.
The project was led by Suchi Saria, a professor at Johns Hopkins and director of the university’s AI and Healthcare Lab. Saria began working to bring the research into hospitals after losing her nephew to sepsis in 2017. Researchers involved in the project said the system works as an added layer of support for doctors rather than replacing medical judgment.
Several hospital systems across the U.S. have already started using this AI technology. These include the Cleveland Clinic, MemorialCare, and the University of Rochester School of Medicine. In addition to lower sepsis-related death rates, the hospitals testing the technology saw fewer medical complications, and shorter hospital stays for patients.
Medical experts involved in the project said one reason the system performs well is because it can handle large amounts of messy hospital data that can be difficult for humans to sort through quickly. Hospitals collect endless streams of information from lab tests, heart monitors, nurse notes, medications, and patient histories. The software constantly reviews that information and searches for patterns linked to sepsis.
FDA clearance may also help more hospitals adopt the system because facilities using the technology may qualify for Medicare and Medicaid reimbursement through a federal payment program supporting new medical tools. That financial support could make it easier for hospitals to add the system to patient care programs.
Sepsis continues to be one of the leading causes of death inside hospitals across the nation. More than 250,000 Americans die from the condition each year. Doctors have long searched for better ways to recognize the condition before patients become dangerously sick. Researchers believe this new technology could change how hospitals monitor patients at risk for severe infections. Instead of waiting until symptoms become obvious, medical teams may now receive warnings much earlier, giving patients faster care during one of the most dangerous medical emergencies treated in hospitals.
Sources:
Johns Hopkins researchers develop life-saving AI for early sepsis detection
Join the conversation!