New “life saving” artificial intelligence developed in Israel tells doctors which patients are most at risk of death or serious illness should they develop a blood infection.
Just by giving the information to doctors from routine blood tests when patients arrive in hospital, the technology will “prevent many deaths” as it will prompt medical teams to closely monitor those at highest risk, Prof. Noam Shomron told The Times of Israel.
He noted that blood infections are common, with an estimated 250,000 occurring annually in the US, and they are a leading cause of death, adding that this makes it an important issue to tackle.
The AI tool interprets patterns from standard blood tests that hospitals tend to run routinely on all new admissions, and uses them to assess danger levels in the case of blood infection.
It predicts responses with 82 percent accuracy, Shomron said, adding that while some other AI tools exist, his stands out for accuracy and ease of use, as it requires no special testing of patients.
The research behind the tool has only just been peer-reviewed, and an Israeli hospital has already agreed to deploy the tech, said Shomron, adding he hopes it will be the first of many. It requires no hardware, just to be installed on a hospital’s existing computer system.
Together with his students Yazeed Zoabi and Dan Lahav and other collaborators, Shomron has just published details of the AI in the journal Scientific Reports. The system will soon go live at Tel Aviv Sourasky Medical Center, he said.
“Hospitals simply can’t pay the same attention to every patient for every risk factor, and to function effectively you need to asses the level of danger that different conditions pose for different patients,” said Shomron. “Our method does exactly that.
“If you know how to direct resources effectively in hospitals, you can save a lot of lives, which is the expectation here,” he said.
Most of the time, the blood system is a sterile one, but infection with a bacterium or fungus can occur during surgery, or as the result of complications from other infections, such as pneumonia or meningitis. The body’s immunological response to the infection can cause sepsis or shock, dangerous conditions that have high mortality rates.
The researchers analyzed electronic medical records of 8,000 patients at Tel Aviv’s Ichilov Hospital who were found to be positive for blood infections. They fed an algorithm patients’ metrics for common blood tests like albumen levels before infection, and how the patient fared during their infection.
The algorithm then detected patterns between blood test metrics and response to blood infection. “Using artificial intelligence, the algorithm was able to find patterns that surprised us, parameters in the blood that we hadn’t even thought about taking into account,” stated Shomron.
“We already know the importance of directing resources well, for example in breast cancer, when we give increased screening to those at risk. This development will allow us to do the same for blood infections, which will help many people.”