healthcare technology medical AI FDA clearance sepsis artificial intelligence in healthcare Epic Systems clinical decision support Bayesian Health Johns Hopkins alert fatigue in hospitals

Five years ago, sepsis prediction software suffered a major setback when an algorithm developed by Epic, a leading electronic health record company, failed to deliver in real-world hospital settings. The software, designed to alert physicians to potential sepsis cases—a life-threatening reaction to infection—promised to reduce the more than 350,000 annual sepsis-related deaths in the United States.

Despite strong performance metrics on paper, the algorithm proved ineffective in practice. It generated an overwhelming number of alerts, leading doctors to ignore or disable them entirely. This failure highlighted a critical gap between theoretical performance and practical usability in clinical environments.

Now, a new wave of sepsis prediction models is emerging. Epic has released an updated version of its algorithm, while startups are testing their models in health systems across the country. Additionally, a research team is exploring the use of large language models to analyze clinical notes for early signs of sepsis.

On Tuesday, Bayesian Health, a company with roots at Johns Hopkins University, announced that its sepsis flagging device has received clearance from the U.S. Food and Drug Administration (FDA). This marks a significant step forward in the development of reliable, clinically viable sepsis detection tools.

Source: STAT News
How America Became a Nation of Legalized Vices and What It Costs Us
← Previous

How America Became a Nation of Legalized Vices and What It Costs Us

 Is Calling Boneless Wings 'Wings' False Advertising? Federal Judge Says No
Next →

Is Calling Boneless Wings 'Wings' False Advertising? Federal Judge Say...