Anthropological, socioeconomic, and health factors contribute to antimicrobial resistance across the United States. US Sewage Project

Recent studies have suggested that country-level socio-economic factors are more important than antimicrobial usage (AMU) in explaining the prevalence of AMR. It is however, not known whether this is also the case for different groups within countries. Here we use metagenomic analysis of untreated sewage to cross-sectionally characterize the bacterial resistome from 44 sites across 8 states in the USA, collected in 2019 and 2020. We explored these data and associations between site specific anthropological, socioeconomic and health determinants, as well as state-specific AMU and health data. Spatial autocorrelation analysis was used to identify spatial clusters of AMR. Gradient boosted multivariate regression trees were used to identify predictors with individual and joint effects on AMR. Drivers with the highest global relative influence on AMR included those related to economy, income, preventative health care, access to healthcare, social welfare, housing, and ethnicity. These drivers showed widespread effects on AMR both individually and in combination. These variables influence AMR in a nonlinear trend consisting of thresholds beyond which there is either a clear-cut or stepwise leap (risk factors) or dip (protective factors) in AMR. We found no evidence for outpatient AMU being significantly or strongly associated with increased AMR. A significant difference between the bacteriome and resistome in the years 2019 and 2020 indicated a possible effect of Covid-19 lockdown. Our findings suggest that improving general health, economy, well-being and development at the very local level within a country or community could contribute drastically to reducing AMR.