Background The need for evidence to inform interventions to prevent mass shootings (MS) in the USA has never been greater.

Methods Data were abstracted from the Gun Violence Archive, an independent online database of US gun violence incidents. Descriptive analyses consisted of individual-level epidemiology of victims, suspected shooters and weapons involved, trends and county-level choropleths of population-level incident and fatality rates. Counties with and without state-level assault weapons bans (AWB) were compared, and we conducted a multivariable negative binomial model controlling for county-level social fragmentation, median age and number of gun-related homicides for the association of state-level AWB with aggregate county MS fatalities.

Results 73.3% (95% CI 72.1 to 74.5) of victims and 97.2% (95% CI 96.3 to 98.3) of shooters were males. When compared with incidents involving weapons labelled ‘handguns’, those involving a weapon labelled AR-15 or AK-47 were six times more likely to be associated with case-fatality rates greater than the median (OR=6.1, 95% CI 2.3 to 15.8, p<0.00001). MS incidents were significantly more likely to occur on weekends and during summer months. US counties in states without AWB had consistently higher MS rates throughout the study period (p<0.0001), and the slope for increase over time was significantly lower in counties with AWB (beta=−0.11, p=0.01). In a multivariable negative binomial model, counties in states with AWB were associated with a 41% lower incidence of MS fatalities (OR=0.58, 95% CI 0.37 to 0.97, p=0.02).

Conclusions Counties located in states with AWB were associated with fewer MS fatalities between 2014 and 2022.