Continuous AI monitoring earns regulatory validation, marking a milestone for Bayesian’s real-time clinical intelligence platform and the new standard of proactive care it delivers.

NEW YORK, May 12, 2026 /PRNewswire/ — Bayesian Health today announced that its sepsis flagging device has received FDA 510(k) clearance—the first continuous AI sepsis monitor ever to do so. Early sepsis detection is a core module within Bayesian’s real-time clinical intelligence platform, which continuously monitors every patient, surfaces those who need attention, and gives clinicians the tools to set the right actions in motion.

The clearance, which builds on Bayesian’s prior FDA Breakthrough Device Designation, signals a shift in what health systems can expect from clinical AI: a validated, real-time layer of clinical intelligence clinicians trust and use at the bedside. EHRs remain the backbone of clinical documentation, but weren’t designed to continuously interpret patient data in real time. Bayesian builds on their foundation—adding always-on clinical reasoning that surfaces insights inside the workflows clinicians already use and helps them act with confidence.

A New Approach to Sepsis Detection

Real-time clinical intelligence requires a fundamentally different architecture: one that continuously reasons over evolving patient state across all data types, rather than triggering on episodic rules. It’s this architecture that underpins Bayesian’s platform, providing health systems with a single solution for early deterioration across multiple conditions like sepsis, respiratory failure, and cardiac events.

“Sepsis has been the focus of my work for more than a decade — a direction set, in part, by losing someone I loved to it. The work behind this clearance spans more than a decade: the deep research, the peer-reviewed validation, and the deployments that proved it works at the bedside. 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. That’s the bar clinical AI should be held to,” said Suchi Saria, PhD, CEO and Founder of Bayesian Health.

Why Sepsis is Often Caught Too Late

Sepsis is the clearest example of a broader challenge: clinical deterioration can emerge subtly and is often recognized too late. Sepsis is a leading cause of death in U.S. hospitals¹, and an expensive condition to treat, driving over $50 billion in hospital costs annually². Each hour of delayed treatment drops survival by as much as 7.6 percent³. Despite decades of investment, sepsis care remains hampered by inconsistent early recognition.

Unlike Bayesian’s solution, other FDA-cleared sepsis solutions require a clinician to suspect sepsis first, whether it’s a biomarker-based diagnostic blood test or a software risk score calculated using the results of the blood test. Continuous sepsis monitoring tools have historically sat outside FDA oversight, and clinicians tune them out: they fire constantly, often on patients already identified or who turn out to be fine. Bayesian’s platform is the first cleared solution to flag sepsis before clinical suspicion, running continuously across every hospitalized patient and surfacing deterioration in those who don’t yet look sick, at a precision clinicians trust.

“Catching sepsis before a clinician suspects it is a needle-in-a-haystack problem. Sepsis prevalence in acute care is just 2 to 4 percent, while traditional diagnostic tests run at 20 to 40 percent on patients already suspected of having it. Missing a single case is catastrophic, and that demands a level of precision most AI can’t meet, even tools that look promising in preliminary studies,” said Neri Cohen, MD, PhD, Head of Clinical Enterprise at Bayesian Health. “Few companies have done the rigorous data science and independent validation required to meet that bar. Fewer still have done the equally rigorous work that has to follow: building it into clinical workflows, earning clinician adoption at the bedside, and measuring outcomes long enough to know it’s still working. That’s what makes clinical AI deliver in practice”

Validated at Scale

When clinicians act on Bayesian’s alerts in time, sepsis patients are 18% less likely to die in the hospital. That finding comes from a 2022 peer-reviewed Nature Medicine study⁴ spanning 764,707 patient encounters (17,538 with sepsis) across five hospitals in both academic and community-based settings with 2,000+ providers using the software. This research showed accurate early detection (1 in 3 cases were physician confirmed) at high sensitivity (82%) and significant lead time (5.7 hours earlier), high provider adoption (89%) — a rate rarely seen in clinical AI, where most tools are ignored or turned off.

“Health systems have spent decades building the digital foundation of care through platforms like their EHR. What’s been missing is a real-time intelligence layer that can continuously interpret that data and help clinicians act earlier. We’ve built that infrastructure — not for a single condition, but across the conditions that drive the most preventable harm. Proactive, continuous monitoring saves lives. It should be the standard in every hospital,” Saria said.

“Deploying AI for sepsis detection across a health system like Cleveland Clinic requires rigorous evidence and seamless clinical integration,” said James Morrison, MD, Chair of the Cleveland Clinic Enterprise Sepsis Steering Committee. “Our focus is ensuring these tools deliver meaningful, actionable insights at the bedside where they matter most. This FDA clearance is an important milestone, validating the standard of safety and performance we demand for our clinicians and the patients we serve.”

“After evaluating multiple sepsis tools, our clinicians are finding that Bayesian provides early detection with a high rate of accuracy — a tool they can trust and use to diagnose and treat their patients,” said Gregg Nicandri, MD, Chief Digital and Information Officer at University of Rochester Medicine. “By enabling clinicians to get ahead of sepsis instead of chasing it, Bayesian is changing what we expect from clinical AI in this category.”

A Pathway to Reimbursement

FDA 510(k) clearance also positions Bayesian Health favorably to receive approval under the Centers for Medicare and Medicaid Services’ New Technology Add-on Payment (NTAP) program starting in October 2026. The final decision on NTAP approval will be issued early August. If approved as expected, this would give hospitals deploying the cleared sepsis flagging device access to potentially incremental payment for cases in which this solution is used, supporting both the clinical and financial case for proactive sepsis monitoring.

Bayesian’s Broader Platform

This clearance applies to Bayesian’s sepsis flagging device, part of the broader clinical intelligence platform designed to address the leading drivers of deterioration, mortality, length of stay, and ICU utilization—using a single, continuously learning system.

About Bayesian Health

Bayesian Health is the real-time clinical intelligence platform that helps health systems deliver proactive, high-reliability care. By continuously reading the full patient record to establish each patient’s baseline and detect meaningful change over time, Bayesian applies complex clinical reasoning to identify the patients who truly need attention and surface clear next steps. Built to work within leading EHR systems, the platform enhances existing clinical workflows by adding a continuous intelligence layer that helps teams intervene earlier and more effectively. Bayesian has partnered with leading health systems to make the long-promised shift from reactive to proactive care across key drivers of mortality, readmissions, and length of stay, hardwiring quality as a durable strategy rather than a series of disconnected initiatives.

Learn more at https://bayesianhealth.com.

View original content to download multimedia:https://www.prnewswire.com/news-releases/bayesian-health-receives-first-ever-fda-clearance-for-continuous-ai-sepsis-monitoring-302769190.html

SOURCE Bayesian Health