RESUMO
Although the effects of antimicrobial resistance (AMR) are most obvious at clinical treatment failure, AMR evolution, transmission, and dispersal happen largely in environmental settings, for example within farms, waterways, livestock, and wildlife. We argue that systems-thinking, One Health approaches are crucial for tackling AMR, by understanding and predicting how anthropogenic activities interact within environmental subsystems, to drive AMR emergence and transmission. Innovative computational methods integrating big data streams (eg, from clinical, agricultural, and environmental monitoring) will accelerate our understanding of AMR, supporting decision making. There are challenges to accessing, integrating, synthesising, and interpreting such complex, multidimensional, heterogeneous datasets, including the lack of specific metrics to quantify anthropogenic AMR. Moreover, data confidentiality, geopolitical and cultural variation, surveillance gaps, and science funding cause biases, uncertainty, and gaps in AMR data and metadata. Combining systems-thinking with modelling will allow exploration, scaling-up, and extrapolation of existing data. This combination will provide vital understanding of the dynamic movement and transmission of AMR within and among environmental subsystems, and its effects across the greater system. Consequently, strategies for slowing down AMR dissemination can be modelled and compared for efficacy and cost-effectiveness.