Evaluating the effects of Antimicrobial abuse

This research will utilize the 2018-2021 SOAR dataset, analyzing bacterial isolates from community-acquired respiratory tract infections (CA-RTIs) using machine learning (ML) methodologies. We will extract minimum inhibitory concentration data and apply ML algorithms—such as random forests, Extreme Gradient Boosting, artificial neural networks, logistic regression, clustering, and predictive neural networks—to uncover hidden resistance patterns, identify phenotypic clusters, and model complex interactions across CLSI, EUCAST, and PK/PD breakpoints. To generalize these findings to Zambia, the ML models will be contextualized using Zambia’s local epidemiological variables and antibiotic formularies. By training on regional SOAR data and applying geographic adaptation techniques, the models will generate localized, predictive AMR risk profiles relevant to the Zambian clinical setting. Applying these ML-driven insights to Zambia will significantly: Help improve patient outcomes: Predictive ML models can forecast resistance probability at the point of care, enabling clinicians to bypass ineffective empiric therapies and rapidly select targeted treatments, thereby reducing mortality and treatment failures. Strengthen stewardship: By identifying key drivers of resistance and high-risk microbial clusters, ML provides data-driven guardrails to optimize antibiotic prescriptions and precisely restrict the misuse of reserve drugs. Inform public health practice: ML forecasting algorithms can project emerging resistance trends, allowing Zambian health authorities to proactively update treatment guidelines and essential drug lists before resistance becomes endemic. Strengthen health systems: Integrating ML into AMR surveillance transitions Zambia’s health infrastructure from reactive reporting to predictive analytics, enhancing laboratory capacity and building data-driven resilience against infectious diseases. Ultimately, this research leverages advanced ML to transform global surveillance data into proactive, localized strategies for combating AMR in Zambia.