REFORM SC HOLARS - SMART PRESCRIBING

And it makes it harder to break the groupthink and confirmation bias that too often pervades the policy world, limiting the quality of decisions and the range of ideas that are considered. [...] AI is already being applied in the microbiology laboratory to support the diagnosis of infection, the identification and quantification of micro-organisms, and the analysis of antimicrobial susceptibility. [...] Current AI technologies have been developed to focus on clinical outcomes such as the prediction of sepsis in critical care, the diagnosis of TB or surgical site infection, the prediction of virological success of antiretroviral therapy, and the selection of an antibiotic regimen. [...] One review has identified 60 unique AI-based tools designed to support decision making for clinical diagnosis and management of infection.31 AI was used to support the diagnosis of infection, the early detection or stratification of sepsis, the prediction of response to antimicrobial therapy, the presence of antibiotic resistance, and the choice of antibiotic regimen. [...] 3.2.2 Improving end-user engagement Development of AI-based decision support tools requires end-user engagement to ensure that design of their interface and interpretability of their recommendations are acceptable.49 The early engagement of prescribers facilitates user-friendly design and can ensure that the tool is deployed and used at the correct place in the antibiotic decision making pathway.