Every hour sepsis goes undetected, a patient's survival odds fall by up to 7.6%. On 12 May 2026, the US Food and Drug Administration cleared the first AI system designed to close that window — before a doctor even suspects infection.

What TREWS Actually Does

Bayesian Health, a Johns Hopkins University spinout, received FDA 510(k) clearance for its Targeted Real-time Early Warning System (TREWS) — the first AI-powered device cleared to continuously monitor hospitalised patients for sepsis without any prior clinical suspicion as a trigger. Most existing early-warning tools wait for a clinician to flag concern before running an assessment. TREWS works the other way: it pulls from electronic health record (EHR) data — vitals, lab results, medications, procedures — and surfaces risk scores in real time, continuously, for every inpatient.

According to Bayesian Health's announcement, the system detects sepsis an average of 5.7 hours earlier than standard care — though MedTech Dive notes the company cites a broader range of nearly two to 48 hours faster than traditional methods, depending on patient presentation. The company also reports 82% sensitivity and an 89% provider adoption rate at sites where it has been deployed — both vendor-stated figures.

The Study Behind the Clearance

The FDA submission was anchored by a prospective study published in Nature Medicine in 2022, spanning 764,707 patient encounters across five hospitals — 17,538 of whom had confirmed sepsis. The headline finding: when clinicians responded to a TREWS alert within three hours, in-hospital sepsis mortality fell by 18%. Patients also showed lower organ failure rates and shorter stays.

764,707patient encounters in 2022 Nature Medicine study
18%reduction in in-hospital sepsis mortality
5.7 hrsaverage earlier detection vs. standard care (vendor-stated)
82%sensitivity for sepsis detection

Those are not modelled projections. They come from real clinical outcomes at real hospitals — which is precisely why the FDA's 510(k) pathway was applied here under the Breakthrough Device Designation Bayesian received in 2023.

Who Is Already Using It

Cleveland Clinic, MemorialCare (California), and the University of Rochester School of Medicine are confirmed deployment sites. These are not pilot programmes; the system is live in those institutions' clinical workflows.

CEO Suchi Saria, who founded Bayesian Health after losing someone she loved to sepsis, described the clearance in full: "FDA clearance is a critical milestone, and it's also the consequence of years spent validating that this fits into clinician workflows and helps them get ahead of deterioration instead of reacting to it." The 89% provider adoption figure lends some weight to that framing — though it remains a vendor-stated metric.

The Reimbursement Path That Makes This Stick

A cleared device without a payment route rarely scales fast. Bayesian's position here is unusually strong. The company is pursuing a New Technology Add-on Payment (NTAP) through the US Centers for Medicare and Medicaid Services (CMS), with the final CMS decision expected in early August 2026 and NTAP payments potentially beginning October 2026. If approved, hospitals would receive incremental reimbursement for each Medicare patient monitored — a direct financial incentive to deploy.

Sepsis costs US hospitals over US$50 billion annually. A tool that demonstrably cuts mortality and length of stay, with a payment mechanism attached, is a different commercial proposition than most AI health tools currently on the market.

The Broader Picture

TREWS is not alone in this space — Mednition's KATE AI and Isansys Lifecare's Patient Status Engine target overlapping use cases. What sets the Bayesian clearance apart is the scale of the prospective validation and the regulatory landmark: no other continuous AI sepsis monitor has FDA clearance. That distinction matters for procurement committees in US health systems, and increasingly for international hospital groups watching the US regulatory record before committing their own budgets.

Sepsis remains the leading cause of preventable death in US hospitals. Moving from reactive diagnosis to continuous, automated surveillance has been a clinical goal for years. TREWS getting cleared — and potentially reimbursed — is the clearest sign yet that this shift is no longer theoretical.