Antimicrobial resistance (AMR) in humans is inter-linked with AMR in other populations, especially farm animals, and in the wider environment. The relatively few bacterial species that cause disease in humans, and are the targets of antibiotic treatment, constitute a tiny subset of the overall diversity of bacteria that includes the gut microbiota and vast numbers in the soil. However, resistance can pass between these different populations; and homologous resistance genes have been found in pathogens, normal flora and soil bacteria. Farm animals are an important component of this complex system: they are exposed to enormous quantities of antibiotics (despite attempts at reduction) and act as another reservoir of resistance genes. Whole genome sequencing is revealing and beginning to quantify the two-way traffic of AMR bacteria between the farm and the clinic. Surveillance of bacterial disease, drug usage and resistance in livestock is still relatively poor, though improving, but achieving better antimicrobial stewardship on the farm is challenging: antibiotics are an integral part of industrial agriculture and there are very few alternatives. Human production and use of antibiotics either on the farm or in the clinic is but a recent addition to the natural and ancient process of antibiotic production and resistance evolution that occurs on a global scale in the soil. Viewed in this way, AMR is somewhat analogous to climate change, and that suggests that an intergovernmental panel, akin to the Intergovernmental Panel on Climate Change, could be an appropriate vehicle to actively address the problem.

Background: Antimicrobial resistance can be transmitted between animals and humans both directly or indirectly, through transmission via the environment (such as fomites or sewage). However, there is a lack of understanding of, and quantitative evidence about, the contribution of the environment to AMR epidemiology. In this study we incorporate the transmission of resistance via the environment into a mathematical model to study the potential importance of this form of transmission for human resistance levels and any effects of the impact of interventions to reduce antibiotic consumption in animals. Methods: We developed a compartmental model of human-animal AMR transmission with an additional environmental compartment. We compared the outcomes of this model under different human-animal-environment transmission scenarios, conducted a sensitivity analysis, and investigated the impact of curtailing antibiotic usage in animals on resistance levels in humans. Results: Our findings suggest that human resistance levels are most sensitive to both parameters associated with the human compartment (rate of loss of resistance from humans) and parameters associated with the environmental compartment (rate of loss of resistance from the environment and the transmission rate from the environment to humans). The impact of curtailing antibiotic consumption in animals on long term prevalence of AMR in humans was weaker when environmental transmission was assumed to be high. Conclusions: This study highlights that environment-human sharing of resistance can influence the epidemiology of resistant bacterial infections in humans and reduce the impact of interventions that curtail antibiotic consumption in animals. More data on the types and dynamics of resistance in the environment and frequency of human-environment transmission is crucial to understanding the population dynamics of antibiotic resistance.