Data Challenge: Uncovering drivers and clusters of antibiotic resistance

Antibiotic resistance represents one of the largest global health threats of the 21st century. As bacteria evolve to resist the drugs we rely on, even simple infections can prove deadly. This research leverages Pfizer’s ATLAS database to transform how resistance patterns and their clinical drivers are understood on a global level.
Employing advanced regression modelling, allied to latent class analysis, we aim to uncover hidden resistance phenotypes that are often overlooked by conventional resistance methods that primarily focus on specific bug-drug combinations or on a multidrug resistance status. Our approach identifies nuanced resistance patterns and describes patient characteristics, clinical settings, and institutional factors that are associated with these phenotypes.
The methodology addresses critical gaps in current antibiotic stewardship by primarily, (1) revealing diverse resistance patterns beyond the conventionally explore categories (e.g. ESBL or carbapenem resistance); and (2) identifying patient populations at a particularly high risk for specific cluster membership/resistance phenotype; and secondarily, (3) assessing the transportability of resistance patterns between countries and socioeconomic contexts. This understanding is essential for developing more precise, targeted treatment strategies.
Integrating global surveillance data with sophisticated statistical methods has to potential to deliver actionable insights for antibiotic stewardship programmes worldwide. It can provide tools to healthcare systems that help at improving patient risk stratification, optimising empirical antibiotic selection, and identifying institutional practices that may inadvertently lead to specific resistance patterns.
These findings aim to preserve antibiotic effectiveness while improving patient outcomes by providing a framework through which more personalised, risk-stratified treatment approaches - accounting for the complex, multidimensional natural of bacterial resistance - can be prioritised in clinical practice.