Bayesian Health gets FDA clearance for sepsis monitor

Sepsis, when a patient’s body responds improperly to an infection, is a major challenge for providers. It’s the third leading cause of death in US hospitals, with 1.7 million people affected every year. It also costs healthcare $38 billion annually, according to a 2023 study across several universities.

New York startup Bayesian Health is looking to address this challenge with—you guessed it—AI. The company announced it received FDA 510(k) clearance for its continuous sepsis monitor, which uses AI to flag the condition, on May 12.

This is a step toward widespread adoption, though Bayesian already has a huge advantage on its side: the power of Cleveland Clinic. The 23-hospital health system has invested in and supported the company since its development (which we first reported on back in December).

Now with FDA approval, Bayesian is entering its next stage of growth.

“Ultimately, we can only produce outcomes when clinicians trust the system and act on it,” founder Suchi Saria told Healthcare Brew. “You can think of the FDA as a very meaningful and critical level for building trust.”

How it works. Usually, providers order sepsis tests after a patient starts showing symptoms like fever, confusion, and shortness of breath, Saria said. But research shows that early identification can make a big difference in saving lives and reducing harm overall. The University of Kansas Health System, for example, dropped its sepsis mortality rate from 49% to 5% after it started tracking data, the Association of American Medical Colleges reported in 2023.

“We’re running in the background, we flag the case within workflow, and then we transparently summarize why this patient was caught,” Saria said.

Bayesian’s platform pulls patient data, like lab results, medical records, and vitals to track sepsis risk. It then provides clinicians with a course of action to treat the patient. The technology was tested at Cleveland Clinic as well as Baltimore-based Johns Hopkins, which spun out Bayesian, in 2023.

“Sepsis is a complex condition that can be challenging to diagnose,” James Morrison, chair of Cleveland Clinic’s enterprise sepsis steering committee, told us via email. “We believe sepsis detection is one way advanced technology can meaningfully support clinicians and improve patient outcomes.”

Asher Perzigian, global and Americas clinical and care delivery practice lead at consultancy Accenture, said the challenge in implementing this kind of tool is not in the model itself, but rather how it fits into care overall.

“If clinicians are overwhelmed by alerts, lack confidence in the recommendations, or can’t easily act on the information provided, even highly accurate systems may struggle to deliver their intended value,” he told us in an email.

The exact dollar investment into Bayesian hasn’t been disclosed, though Saria said testing the continuous sepsis monitor within Cleveland Clinic has helped prove its usefulness.

“What this FDA clearance is doing is validating decades of work that has gone into proving that a solution works not just on paper but at the bedside,” Saria said.