Perspective: Limiting Antimicrobial Resistance with Artificial Intelligence/Machine Learning

Le résumé fourni par la source

The author traces his experience with the application of computers in clinical microbiology over the past 60 years, specifically in directing clinicians to treat bacterial infections diagnosed by the laboratory and the antibacterial agent(s) that could be used to treat those infections. Appropriate use of antibiotics will result in reduced antimicrobial resistance, which is increasing worldwide. An early form of AI, Mycin (1976), a system based on rules provided by experts designed to propose antibiotic regimens for central nervous system infections, was never applied due to the limitations in the number of rules that could be incorporated into the clinical workflow. Machine learning (ML) was developed to overcome the limitations of expert systems. Several variables that influence the outcome bacteria/drug interaction, such as the source of the infection, absence of antimicrobial resistance markers, patients' health profile, and the historical susceptibility within the hospital and the local area are incorporated in the proposed comprehensive AI/ML program. The role of AI in the discovery of new antimicrobial agents is also addressed.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.