Researchers at the University of British Columbia have found that artificial intelligence can detect sepsis early.
Sepsis is responsible for at least one in five deaths worldwide, including severe cases of COVID-19, according to a UBC news release on Tuesday.
But experts warn that early detection of the life-threatening condition is difficult.
UBC researchers say that because sepsis is defined as the body’s passive response to infection and has a variety of symptoms — including fever, fatigue, hyperventilation and rapid heart rate — it is often first manifested by other illnesses. Maybe.
Arjun Baghela, a UBC graduate student, says, “This new technique dissects the passive immune responses involved in sepsis like never before, providing new insights into the biological processes involved in any type of sepsis, including COVID-19.” also includes.” Analysis .
“People don’t know much about sepsis, but in 2020, the number of fatal sepsis deaths is likely to be much higher than one in five, because all people who die from COVID-19 are actually Died of sepsis.”
According to the Centers for Disease Control, infections that can lead to sepsis usually begin in the skin, lungs, urinary tract or gastrointestinal tract.
It can take anywhere between 24 and 48 hours for a health care provider to confirm a patient has sepsis. But for every hour of delay in providing treatment, the risk of death increases by 7.6 percent, which highlights the need for rapid detection, the UBC researchers say.
“Typically, a patient comes into the emergency room feeling critically ill, with a set of symptoms that are unsuccessfully non-specific,” says UBC professor Dr. Bob Hancock.
“The doctor sees that patient if they have a bunch of symptoms and says, ‘This is a patient who could have sepsis,’ but only if they have some certainty can they start treating them right away. It’s a ‘look and see’ game for the first 24 to 48 hours.”
Hancock says it’s imperative to know if a patient has sepsis as soon as possible, given that it’s so common and widespread antimicrobial resistance is a risk if antibiotics are used more than necessary.
For the study, published in the journal EBioMedicine, researchers examined a total of 348 patients across four continents. The scientists say they were able to confirm their findings by re-examining two other larger studies for a total of 1,062 patients.
The research found that severe sepsis can be detected when a patient first comes to medical care.
“Using machine learning, also known as artificial intelligence, the researchers were able to identify the set of genes that predict whether a patient will get severe sepsis, and in five different ways (subtypes/ endotype) in which sepsis manifests itself,” the study reads.
Researchers say the discovery will eventually lead to testing that will allow health care providers to identify the body’s passive response to infection and determine appropriate treatment.
“This is important because the two subtypes are associated with a very high risk of severe sepsis and death,” reads the study. “These biomarkers also worked in the ICU, where it was shown that one endotype was particularly lethal, with a mortality rate of 26 percent.”
Researchers said the technology to measure gene expression already exists in hospitals, and can be done within just two hours of being admitted to the emergency room.