Data Challenge: Global co-resistance networks in Carbapenem Resistant Enterobacterales (CRE): A predictive framework for Low and Middle Income Countries (LMICs)

Abstract

Detection of Co-resistance patterns in Carbapenem Resistant Enterobacterales (CRE) is essential for preventing sequential treatment failures, particularly in low- and middle-income healthcare settings where treatment options and diagnostics remain limited. While conventional antimicrobial resistance surveillance focuses on phenotypes, emerging evidence suggests that phenotypic co-resistance networks may reveal hidden resistance associations that remain undetected by standard analyses. Resistance to one antimicrobial class may therefore indirectly indicate resistance to multiple unrelated classes through co-selection mechanisms driven by plasmid-mediated resistance dissemination. Our project aims to investigate global co-resistance networks in CRE using large-scale phenotypic surveillance datasets from Pfizer ATLAS and Venatorx GEARS. Using MIC-derived resistance profiles, we will apply association rule mining and network-based analyses to identify recurrent co-resistance relationships, characterize clinically significant resistance connectivity patterns, and detect antibiotics that may function as potential co-selection hubs accelerating multidrug resistance dissemination. Furthermore, we will explore how co-resistance networks vary across geographic regions and over time, with focus on clinically important CRE. This project seeks to move beyond conventional prevalence reporting toward a system-level understanding of Antimicrobial Resistance (AMR) evolution in resource-limited settings. Using this approach, we aim to build a practical framework that assists clinicians and antimicrobial stewardship programs in low- and middle-income countries to interpret co-resistance patterns and make more informed therapeutic decisions based solely on phenotypic surveillance data without requiring molecular diagnostics infrastructure.