Cleveland Clinic researchers have developed an artificial intelligence (AI) model to determine the best combination and timeline for prescribing drugs to treat bacterial infections based on the speed at which bacteria grow under certain perturbations. This model, led by Dr. Jacob Scott and his team, was published in PNAS and aims to address the issue of antibiotic resistance due to overuse of antibiotics, potentially leading to more deaths from antibiotic-resistant infections than cancer by 2050.
Antibiotics have significantly increased the average lifespans in the US by almost ten years by lowering fatality rates of minor health issues. However, the overuse of antibiotics has led to bacteria developing mutations that resist treatment. The concept of antibiotic cycling, where healthcare providers rotate between different antibiotics over specific time periods, has shown promise in effectively treating diseases by minimizing resistance and maximizing susceptibility to antibiotics.
One challenge with antibiotic cycling is determining the best strategy, as there is currently no standardized approach between hospitals for which antibiotic to use, for how long, and in what order. Dr. Weaver and Dr. Maltas utilized computer models to predict how bacteria’s resistance to one antibiotic could make it weaker to another, aiming to identify drug cycling regimens that can minimize resistance and maximize susceptibility.
By applying reinforcement learning, which allows a computer to learn from its mistakes and successes to determine the best strategy to complete a task, the research team’s AI model was able to figure out the most efficient antibiotic cycling plans to treat multiple strains of E. coli and prevent drug resistance. This study is among the first to apply reinforcement learning to antibiotic cycling regimens, highlighting the potential of AI in supporting complex decision-making in antibiotic treatment schedules.
Dr. Weaver and his team envision their AI model not only managing individual patient infections but also informing how hospitals treat infections on a larger scale. They are also working to expand the application of their model beyond bacterial infections to other deadly diseases that may develop treatment resistance, such as drug-resistant cancers. This research showcases the potential of AI in optimizing treatment regimens and addressing challenges in bacterial infections and other medical conditions that evolve treatment resistance.
