Data Challenge: Broad Spectrum Disperse Probabilistic Model to guide AMR stewardship using Mathematical and Spatial epidemics

This proposal presents a Broad Spectrum Disperse Probabilistic Model that utilises the Pfizer ATLAS dataset to generate actionable predictions regarding antimicrobial resistance (AMR) across various geographical locations and time periods. The model is designed to function as a surveillance decision-support tool, bridging static data and immediate stewardship requirements by employing advanced probabilistic algorithms and spatial-temporal analytics. The model aims to quantify the evolving resistance trends categorised by region, pathogen, and antimicrobial agent, with outputs including dynamic heat maps, resistance forecasts, and site-specific risk scores to facilitate clinical and public health decision-making processes. The objective of this tool is to enhance patient outcomes by improving empiric therapy decisions based on the probabilities of regional resistance. Furthermore, it will reinforce stewardship by identifying patterns of overuse in broad-spectrum antibiotics, thus enabling data-driven adjustments to prescribing guidelines and formulary policies. From a public health perspective, the model intends to act as an early warning system, flagging the emergence of resistance before it becomes entrenched. It supports health system strengthening by aligning surveillance outputs with operational decision-making across hospitals, health departments, and policy bodies. Deliverables will include an interactive dashboard, risk classification models, and reproducible code that can be integrated into AMR surveillance workflows. The approach ensures scalability, transparency, and adaptability across various settings. This model transforms the ATLAS dataset from a retrospective archive into a forward-looking system that informs real-time clinical action, surveillance prioritisation, and stewardship strategy. The goal is to empower the AMR response with precision, speed, and contextual relevance, meeting the core objectives of the Vivli AMR Data Challenge.