Johns Hopkins
FDA clears first AI tool to flag sepsis before clinicians can
Clinical Takeaway: Hospitals using the tool may gain hours of lead time on sepsis detection, with real-world deployments showing reductions in mortality and length of stay. FDA clearance also unlocks CMS reimbursement, which should reduce the financial barrier to adoption.
Sepsis kills an estimated 250,000 patients in the U.S. each year and is involved in roughly one in three in-hospital deaths, in part because its early signs overlap with many other conditions. Every hour of delayed treatment reduces the odds of survival, but existing screening tools generally activate only after a clinician has already begun to suspect the diagnosis. The Targeted Real-Time Early Warning System, developed at Johns Hopkins and commercialized by Bayesian Health, is designed to flag patients before that concern even forms.
The system integrates with the electronic health record and runs clinical AI continuously against the full stream of inpatient data, including vitals, labs, medications, and clinician notes. In real-world deployment across dozens of U.S. hospitals, it has detected sepsis cases roughly 2 to 48 hours earlier than traditional methods, with a corresponding 18% reduction in sepsis mortality. Sites that piloted the tool under the FDA's Breakthrough Designation, including Cleveland Clinic, MemorialCare, and the University of Rochester School of Medicine, also reported reductions in morbidity and length of stay.
The clearance is the first FDA approval for an AI tool that performs pre-suspicion sepsis screening, a category distinct from the screening checklists and rule-based alerts already embedded in many EHR systems. Those tools generally rely on a clinician to enter or confirm sepsis-relevant findings; this system is designed to surface the signal earlier, before a workup has been initiated. Hospitals that deploy it can seek payment under the CMS New Technology Add-on Payment (NTAP) program, which reimburses inpatient use of newly cleared technologies and is intended to offset deployment costs.
"Pre-suspicion screening is what creates lead time, and lead time is what changes outcomes in sepsis," said lead researcher Suchi Saria, PhD, professor and director of the AI & Healthcare Lab at Johns Hopkins University, who began translating her lab's research into a clinical system after losing her nephew to sepsis in 2017. "Once a clinician already suspects sepsis, the clock has been running, often for hours or even days. No other cleared test or device monitors for sepsis prior to clinician suspicion."
Source: Johns Hopkins University. 2026 May 12. FDA approves early warning system for sepsis