Data Challenge : Therashift

Antimicrobial resistance poses a growing threat in low- and middle-income countries (LMICs), where treatment options are already limited and surveillance infrastructure remains underdeveloped. Standard resistance surveillance relies on clinical breakpoints to classify pathogens as susceptible or resistant. This binary system identifies resistance only after it is clinically established, often too late to prevent treatment failure or guide timely stewardship decisions. Team Therashift proposes an early-warning surveillance framework that uses Epidemiological Cutoff Values (ECOFFs) to detect subtle MIC drift before clinical resistance thresholds are crossed. Rather than waiting for breakpoint reclassifications, ECOFF analysis identifies biological shifts in bacterial wild-type populations that precede resistance emergence. This pre-clinical signal creates an earlier window for intervention and supports the extended use of existing antibiotics through evidence-based stewardship. We are requesting isolate-level, longitudinal MIC data for Escherichia coli and Staphylococcus aureus from the ATLAS (Pfizer) and GEARS (Venatorx) datasets, with geographic emphasis on Nigeria, Ghana, and Kenya. These pathogens were selected for their significance at the human-animal interface and their high clinical relevance in LMIC infection settings. External open-access veterinary resistance data from NARMS and WOAH will be integrated to build the One Health component of the analysis. Using joinpoint regression and temporal trend analysis, we will compare ECOFF-based MIC distribution shifts against traditional breakpoint-based alarm systems to test whether ECOFF analysis detects resistance emergence at an earlier stage. This directly addresses the fragmented nature of current One Health surveillance by harmonizing human clinical and animal resistance data under a shared interpretive standard. The expected output is an open-source, lightweight decision-support tool that accepts standard MIC inputs and returns both breakpoint and ECOFF interpretations alongside an automated early-warning alert. This tool is designed to operate in low-resource clinical and public health settings, enabling clinicians, veterinarians, and policymakers to transition from descriptive resistance reporting to predictive, proactive antimicrobial stewardship.