Non-pharmaceutical interventions (NPIs) remain the only widely available tool for controlling the ongoing SARS-CoV-2 pandemic. We estimated weekly values of the effective basic reproductive number (R-eff) using a mechanistic metapopulation model and associated these with county-level characteristics and NPIs in the United States (US). Interventions that included school and leisure activities closure and nursing home visiting bans were all associated with a median R-eff below 1 when combined with either stay at home orders (median R-eff 0.97, 95% confidence interval (CI) 0.58-1.39) or face masks (median R-eff 0.97, 95% CI 0.58-1.39). While direct causal effects of interventions remain unclear, our results suggest that relaxation of some NPIs will need to be counterbalanced by continuation and/or implementation of others. Disentangling the impacts of non-pharmaceutical interventions on COVID-19 transmission is challenging as they have been used in different combinations across time and space. This study shows that, early in the epidemic, school/daycare closures and stopping nursing home visits were associated with the biggest reduction in transmission in the United States.