THE PATH OF LEAST RESISTANCE: Modeling Antimicrobial Resistance Trends in United States Livestock with Spatial and Temporal Methods

Abstract

The rapid emergence of antimicrobial resistance (AMR) to all modern antibiotics in the past decade poses grave threats to human health and to the health of our meat production systems. Over 700,000 deaths worldwide are attributed to AMR and costs the US healthcare system over 2 billion USD annually. The continued rise of AMR also threatens to collapse the livestock industry, costing the economy trillions of dollars every year. Understanding that the continued, uncontrolled proliferation of AMR has dire consequences, policymakers around the world have implemented multiple programs to slow its rise. From antibiotic consumption caps to enhanced vaccination programs, countries and regional alliances have invested billions into solving this century’s new medical crisis. Yet all of these programs leave out one vital aspect: monitoring. Without robust surveillance programs, policymakers are unable to allocate their limited resources to areas with the fastest growing resistance and are unable to monitor the effectiveness of their multibillion dollar programs. While in low and middle-income countries, these programs are rudimentary or nonexistent, high income countries such as the US have more systematic, centralized, and robust monitoring programs. However, when evaluating the resistance monitoring program in the US, we found severe deficiencies in its scope and data availability. Because surveillance serves as the backbone to every mitigation program, this study attempted to create a more robust analysis of resistance trends by combining existing surveillance data from the National Antimicrobial Resistance Monitoring System (NARMS) with data from point prevalence surveys (PPS) in the US. This study found that overall, MDR and XDR is decreasing in the US, primarily driven by significant decreases in resistance in chickens. When analyzing NARMS and PPS data separately, this study found systematic differences in the geographical and temporal scope of each of the data sources: NARMS data systematically reported lower and decreasing resistance over time while PPS showed higher but stabilizing resistance over time. This study showed that PPS data can serve as an important data source that can supplement existing surveillance programs. Additionally, spatial modeling showed higher rates of resistance in the eastern United States with no hotspots. The relatively uniform pattern of resistance across the US indicates that broad, not targeted interventions like those recommended in LMICs, will be more effective in resisting resistance in the US.