Sociodemographic and Health Factors Are Associated with Antimicrobial Resistance across Eight States in the United States

Recent studies suggest that country-level socioeconomic factors may explain antimicrobial resistance (AMR) patterns better than antimicrobial usage (AMU), but it remains unclear whether this holds for sociodemographic and health variation within countries. We used metagenomic analysis of untreated sewage to cross-sectionally characterize the bacterial resistome as a proxy for AMR at 44 wastewater treatment plants across eight USA states between 2019 and 2020. We examined associations between AMR with site-specific sociodemographic and health indicators and AMU. Spatial autocorrelation analyses were used to identify clusters of AMR. Gradient-boosted multivariate regression trees were applied to evaluate individual and joint predictor effects on AMR. Outpatient AMU explained negligible variation in AMR, whereas predictors related to economy, income, preventive health care, access to health care, social welfare, housing, and racial/ethnic composition showed the strongest associations. These relationships were observed across individual resistance classes and their combinations and predicted AMR nonlinearly, with thresholds where AMR shows sharp increases (risk factors) or decreases (protective factors). Significant interannual differences in resistome and bacteriome composition were observed between 2019 and 2020. Although causal inference is limited, the findings suggest that local-level indicators of health, economic conditions, well-being, and development may play an important role in shaping AMR within countries.