Early AI Tool Cuts Sepsis Deaths by Spotting Infection Hours Before Doctors

USA, BaltimoreSat May 16 2026
The U. S. Food and Drug Administration has cleared a new artificial‑intelligence system that can spot sepsis much faster than clinicians normally do. Developed by researchers at Johns Hopkins University, the software scans patients’ electronic health records and flags possible infection signs up to 48 hours earlier than traditional symptom checks. Sepsis, a dangerous immune reaction to infection, kills more than 350 000 Americans each year. Time is critical; each hour of delay lowers a patient’s chance of survival. Until now, doctors relied mainly on obvious signs like fever or confusion—clues that can be mistaken for other illnesses. Because the AI tool raises alarms before doctors even suspect sepsis, it gives teams a valuable head start. “Lead time changes outcomes, ” says the lab’s director, who lost her nephew to sepsis in 2017 and then set out to create a real‑world warning system. The product, called the Targeted Real‑Time Early Warning System, has already lowered mortality by nearly 20 % in dozens of U. S. hospitals that tested it.
The system was first rolled out in 2023 at major centers such as the Cleveland Clinic and the University of Rochester School of Medicine. Those pilots reported shorter hospital stays and fewer deaths among sepsis patients, demonstrating the tool’s practical benefits. Now that it has FDA approval, hospitals can also qualify for Medicare and Medicaid reimbursement for using the technology. Experts say few AI applications can navigate the messy data of everyday clinical practice and provide actionable advice that doctors trust. The clearance marks a shift in what is considered standard care for a condition linked to about one in three in‑hospital deaths. It also signals that decades of research at Johns Hopkins are finally reaching bedside patients, not just laboratory models. The FDA’s fast‑track program helped bring the system to market quickly, recognizing its potential to save lives in life‑threatening conditions. By giving physicians an extra “set of eyes and ears, ” the tool could become a routine part of sepsis management, improving outcomes for patients who otherwise might wait too long to be treated. For more details on sepsis symptoms and causes, resources are available through the Mayo Clinic.
https://localnews.ai/article/early-ai-tool-cuts-sepsis-deaths-by-spotting-infection-hours-before-doctors-45fd0b3a

actions