HCA saves more than 5,500 lives with sepsis monitoring algorithms

Nashville‐based HCA Healthcare has saved more than 5,500 lives using sepsis algorithms that monitor every patient in every hospital that’s been part of the health system for more than a year.

With more than 30 million patient encounters annually at HCA, the algorithmic system is being put through its paces and getting impressive results in combating this life-threatening condition that is the ninth leading cause of death in U.S. hospitals and third among all intensive care units.

“Using data science to examine the big data product of interoperable health information, we now have algorithms to do what no clinician can do—monitor labs and other data 24x7x365 for every patient in 164 HCA hospitals,” Jonathan Perlin, MD, HCA’s president of clinical services and chief medical officer, told a Senate committee on Wednesday.

Data scientists at HCA developed the Sepsis Prediction and Optimization of Therapy (SPOT) system that uses artificial intelligence and algorithms based on patient vital signs, labs, nursing reports, and other data to help identify sepsis about 18 hours earlier than the best clinicians.

Also See: EHRs help identify patients at greatest risk of dying from sepsis

“This system identifies patients with sepsis more accurately than the best clinicians and excludes patients without sepsis twice as accurately,” testified Perlin.

He told lawmakers that more than 5,500 patients would have “tragically succumbed” to the overwhelming infection that turns the body’s immune system against itself had it not been for the sepsis algorithms used at HCA hospitals.

“For every hour of delay in diagnosis, mortality increases by four to seven percent—time is life,” Perlin added.

Maternal sepsis accounts for 12.7 percent of pregnancy-related U.S. deaths each year. However, according to new research published in Anesthesia & Analgesia, better screening tools are needed to detect sepsis in pregnant women.

“Currently, there are no good ways to identify these women early,” says Melissa Bauer, DO, assistant professor in the Department of Anesthesiology at Michigan Medicine. “We're working on figuring out the best way to do that.”

In their study, Bauer and her colleagues at Michigan Medicine and seven U.S. academic medical centers and in Israel examined medical records to determine which screening tools are most effective.

Specifically, researchers assessed three screening tools commonly used over the past two decades to identify sepsis:

• The Systemic Inflammatory Response Syndrome (SIRS) criteria, in use from 1992 to 2016.
  
• The quick Sequential Organ Failure Assessment (qSOFA), recommended by the Society of Critical Care Medicine and others to replace SIRS in 2016.
  
• The Maternal Early Warning (MEW) criteria, designed to identify women at risk for a wide array of maternal complications, including pre-eclampsia, hemorrhage and sepsis.
  

According to the study, the SIRS tool had the highest sensitivity but the lowest specificity, while qSOFA had the lowest sensitivity and the highest specificity. At the same time, researchers discovered that qSOFA would miss about half of the sepsis patients they reviewed.


“In my opinion, it is better for a test to have a higher sensitivity so that anyone with sepsis is caught,” contends Bauer.

Nonetheless, she acknowledged that specificity is also important. "If you have poor specificity, you'll probably run into alarm fatigue, with caregivers constantly on alert for patients who don't have anything wrong," concluded Bauer. “There has to be a balance.”