Seasonal and regional heterogeneities in antibiotic resistance (AbR) across meatpacking plants in the USA

Abstract A foundational question in public health is why certain infectious diseases emerge in some locales at certain moments, but not others. Here we leverage a multiyear dataset on nearly all meatpacking plants in the USA handling chicken, cattle, swine, and turkey products as a case study to present a novel approach to elucidating the spatiotemporal dynamics of a major human disease agent: antibiotic resistant (AbR) bacteria. In particular, we apply geographical and temporal neural-network weighted regression, a novel spatiotemporal analysis based on deep-learning, to characterize how infection risks at discrete facilities shift over space and time within the United States. Bracketing the dataset from 2021-2024 by seasons and processing plants across the US, our analyses find seasonal and regional differences in AbR bacteria detection by plant size and commodity type. These factors differentially affected when and where AbR bacteria were found — from distinct outbreaks specific in time and place to broader and persistent trends. While highlighting environmental and occupational safety hazards posed by AbR bacteria, our analyses have the potential to help local communities within plant environs and beyond to operationalize interventions into epidemiological risk profiles posed by dynamic commodity production regimes across spatial scales.