A study examined through machine learning thousands of medical files, and made it possible to predict risk factors for the development of severe morbidity and mortality when infected with a resistant bacteria or fungus
Blood infections are one of the world's leading causes of morbidity and mortality, therefore it is of great importance to identify the risk factors for the development of severe morbidity and mortality already at the time of infection with a bacterium or fungus. Researchers from Tel Aviv University trained software using an artificial intelligence method to study electronic medical records of patients at Ichilov Hospital in Tel Aviv who were found to be positive for blood infection. After studying the data and disease history of each patient, the software automatically identified risk factors in the medical files with an accuracy of 82%. According to the researchers, the new development could be used as an early warning system for doctors, by ranking the patients according to the risk of worsening the disease.
predict the chance of life-threatening
For the most part, our circulatory system is a sterile system, but infection with bacteria or fungus is possible during surgery or other surgical intervention, or as a result of a complication of other infections such as pneumonia and meningitis. The diagnosis of the infection is made by taking a blood sample and transferring it to a growth medium for bacteria and fungi, when the immune response of the infection can cause sepsis and a state of shock, which are dangerous conditions with high mortality rates.
Behind the groundbreaking research, which can save many lives, are the students Yazid Zoevi and Dan Lahav from Prof. Noam Shomron's laboratory From the Sackler Faculty of Medicine, in collaboration with Dr. Ahuva Weiss Mailik, head of the Center for Artificial Intelligence at the Ichilov Hospital, Prof. Amos Adler and Dr. Orli Kehat. The results of the study were published in the journal Scientific Reports.
"We worked on the medical files of 7,889 patients hospitalized at Ichilov Hospital between 2014 and 2020 who were found to be positive for blood infections, during hospitalization and up to 30 days after, whether the patient died or not," explains Prof. Noam Shomron. "We fed the medical files into software that operates on the basis of artificial intelligence, and we wanted to see if the artificial intelligence would recognize data patterns in the medical files that would allow us to automatically predict which patients would develop severe morbidity and even death as a result of the infection."
After the training, the artificial intelligence developed at Tel Aviv University reached an accuracy level of 82% in predicting the course of the disease, even ignoring obvious factors such as the age of the patients and the number of hospitalizations. After the researchers entered the patient's data, the algorithm knew how to predict the course of the disease, and the meaning is that in the future it will be possible to rank patients in terms of the danger posed to their health ahead of time.
"Through the use of artificial intelligence, the algorithm was able to find patterns that surprised us and parameters in the blood that we did not even think to consider," says Prof. Shomron. "We are now working with the medical teams to understand how the information can be used to rank the patients in terms of the severity of the infection. The trick is to use the prediction of the software to direct the doctors to the patients who are at the highest risk."
"Ramot - the commercialization company of Tel Aviv University, believes in the ability of this groundbreaking technology to contribute to reducing costs for the hospitalization systems. This is an example of effective collaboration between university researchers and hospitals, which improves the quality of medical care in Israel and around the world. We at Ramot will continue to lead and promote this type of collaboration," says Keren Perimor Cohen, CEO of Ramot.
AdvD on the subject of the science website:
