
HCA estimates 5,500 lives have been saved because of artificial intelligence now used to detect sepsis, a deadly infection common in hospitals. The Nashville-based hospital chain’s top doctor highlighted the technological advance during a recent congressional hearing.
Sepsis essentially turns the body’s immune system against itself, often resulting in organ failure. And
it’s a leading cause of death in intensive care units. While common, the condition is difficult to catch early.
So HCA has been using size to its advantage, harnessing artificial intelligence to analyze real-time patient data like vital signs and nursing reports. The centralized system can piece together a diagnosis 20 hours before the best physician, according to HCA’s estimates. And each hour increases the survival rate between 4 and 7 percent.
“While we haven’t yet done a formal financial assessment of how less care and shorter hospitalization generates lower cost, we can tell you this, we’re saving lives through this big data and predecessor strategy,” HCA chief medical officer Jonathan Perlin said.
Perlin
spoke in late November to the Senate Health Committee, chaired by Tennessee Republican Lamar Alexander. He’s
regularly spotlighted Nashville-based health care firms and experts in a running series of hearings on health care cost.
HCA has named its technology “Sepsis Prediction and Optimization of Therapy,” or SPOT. And Perlin says choosing an acronym that spells out a common name for a dog was no coincidence.
“It really does act as our sepsis sniffer,” Perlin said
in a press release. “The whole point is for it to sniff smoke and put the ‘fire’ out before it becomes catastrophic.”
Hospitals around the country are exploring new ways to fight sepsis, including several big studies. One tests a cocktail of viatamins and steroids.
And a large academic study has become controversial for putting patients at potentially more risk as doctors test higher and lower doses of medication.
But speed of response is widely seen as the key, and several hospitals are working on projects similar to HCA’s.
At Harborview Medical Center in Seattle, the
hospital tracks abnormal vital signs and requires nurses to explain any figures that might otherwise be an early hint at sepsis, like high or low temperature, elevated breathing, high blood pressure and high white blood cell counts. No single symptom indicates sepsis, but the automated system looks for patterns.
“Sepsis is a really frustrating disease,” Harborview pulmonologist
David Carlbom told NPR. “There’s no blood test for sepsis. … There’s nothing you can look at under the microscope and say, ‘This is sepsis.’ ”
HCA is trying to become the national leader in setting a new standard of care for sepsis.
“We’ll publish our findings. We’ll share our algorithms,” Perlin tells WPLN. “We would also say, though, that the real magic is not just in the computer science and the scale, it’s in marrying the computer science with the clinical workflow.”
But Perlin says it’s not enough for other hospitals to collect and analyze their data in the same way. Doctors and nurses also have to act on it immediately to catch sepsis before it’s too late.
