Heterogeneous impact of mask mandates on U.S. masking behavior: an interrupted time series study

Despite widespread implementation of mask mandates for COVID-19 transmission control, studies examining their effectiveness have yielded mixed results ranging from strong benefits to no effect. These inconsistencies may arise from a variety of methodological and measurement challenges, including the implicit assumption that mandates truly modify masking behavior-the essential mechanism for transmission interruption. Here, we leverage self-reported mask adherence data from >34 000 individuals collected via a digital participatory surveillance platform between June 2, 2020, and January 1, 2021, to examine this assumption. Using an interrupted time series approach, we aggregate masking observations at the county level to analyze the effect of mandates on masking uptake across 555 diverse U.S. counties. We evaluate masking during the 14 days premandate and postmandate issuance, finding a modest 1-3 percentage point overall increases in masking. However, substantial heterogeneity was observed, with larger changes seen in counties initially exhibiting low mask adherence, the U.S. West, and on masking uptake in public compared to private settings. Our findings suggest that conflicting estimates of the effect of mandates on transmission reduction may reflect modification by heterogeneity in the mandates' alteration of masking behavior. Future interventions should tailor mandates to local context and baseline adherence for maximal behavioral change.