An Israeli healthcare provider says it has significantly boosted doctors’ ability to prescribe the right antibiotics, by introducing an artificial intelligence tool.
Antibiotic resistance is rife, and means that many patients aren’t helped by the initial antibiotic given to them, and need to switch medicines.
This not only harms patients by slowing down treatment; in the bigger picture, it also reinforces a vicious cycle: The more antibiotics people are exposed to, the more antibiotic resistance grows.
Maccabi Healthcare Services is now reporting a 35 percent reduction in the need to change antibiotics among urinary tract infection patients who were prescribed with the help of a new artificial intelligence tool deployed earlier this year, and developed at the Technion-Israel Institute of Technology.
Dr. Shira Greenfield, director of medical informatics at Maccabi, said that the tool provides a high-tech solution to “a global problem that all healthcare entities are working to solve.”
She added, “The significance of administering personalized antibiotic treatment is that it lowers the risk of antibiotic resistance developing.”
Dr. Idan Yelin of the Technion Faculty of Biology, told The Times of Israel that the AI tool started as a research project, conducted with Prof. Roy Kishony. They published peer-reviewed research on the idea in 2019, but for two years it was stuck in the conceptual stage, and not deployed in a real-world setting.
The theory was sound, he said, namely that if an algorithm was fed data on antibiotics that succeeded or failed, along with information on patient age, gender, physical condition and a range of other parameters, it would help predict the best antibiotics for other patients.
Over recent months, Maccabi deployed it on the first sample of patients, namely those suffering from urinary tract infections. It is now expected to deploy the algorithm for other infections — and other healthcare organizations are expected to make use of the algorithm too.
“We took this into a real life setting, and were happy to see that it is reducing the rate of antibiotic mismatches to patients by 35%,” Yelin said.
“For patients, this makes treatment shorter and more effective. And in general, as antibiotic resistance grows and given it’s a major concern, we want to reduce number of frequency of mismatches, treatment failures, and amounts of antibiotics being used, and this algorithm can help us to do so.”