Intermountain Health is participating in a national effort to use artificial intelligence and machine learning to develop better ways of treating three of the most common conditions that land people in the intensive care unit.

Pneumonia, sepsis and acute respiratory distress syndrome, also called ARDS, together kill hundreds of thousands of people in the United States each year. During the COVID-19 pandemic, numerous people who were hospitalized in critical condition ultimately died from these conditions.

Through a $51.6 million grant from the National Institutes of Health, researchers from 22 hospitals across the country, in collaboration with a clinical coordinating center at Vanderbilt University Medical Center, will identify what makes survivors of these critical illnesses different, allowing them to find targeted treatments specific to individual people.

Intermountain Health is leading one of six groups of health care systems that make up the APS Phenotyping Consortium, which plans to bring together many of the nation’s experts in phenotyping critical care illnesses. Other sites include the University of California San Francisco, University of Colorado, University of Michigan, University of Pennsylvania and Vanderbilt University Medical Center.

“For 20 years, the scientific community has been searching for effective treatments for ARDS and sepsis patients, without much success,” said Dr. Samual Brown, investigator of the Intermountain-led InterACT Center with the consortium and vice president for research at Intermountain Health.

Part of the problem, Brown said, has been imprecise treatments where it’s assumed everyone is the same and will respond to the same therapies.

“It’s as if we’ve been looking at apples and oranges and watermelons, and treating everyone as if they were apples,” he said. “We need to be more precise and bring true precision medicine to these critically ill patients.”

There are many different causes for and severities of ARDS, pneumonia and sepsis. Emerging evidence suggests that ARDS, pneumonia and sepsis likely represent many distinct disorders with different underlying causes and disease mechanisms.

Brown said each of the three illnesses can be devastating for the patient who ends up in the ICU.

“Intensive care units are complicated places,” Brown said. “The reality is kind of a horror show for the people who live it. They are in a strange state of suspended animation where they ache, are sore and in pain, sometimes short of breath, their sleep-wake cycles are thrown out of whack. People often hallucinate and have waking nightmares.”

Brown said they are there for life-saving reasons, of course, but it’s brutal, and as many as two-thirds of people are severely traumatized once they’re released from the hospital.

“These conditions can kill 30% to 40% of those afflicted and survivors have a hard, slow recovery,” Brown said.

Researchers will study 5,000 patients by collecting information on them through blood and genetic tests and tracking their progress and outcomes. Then, using advanced technologies like advanced analytics, AI and machine learning, they will identify patient types and actionable targets within each type that would then be suitable for further, more specific clinical trials.

Brown said that this is a radically different way of approaching patients with ARDS, sepsis and pneumonia and will lead to a new treatment paradigm.

Newsletter