WHAT IS ALREADY KNOWN ON THIS TOPICWHAT THIS STUDY ADDS

This study reports significant disparities across US states in unintentional firearms-related deaths. Rates in some states are over 10 times those in other states. Poverty, rurality, non-white population and firearms ownership are all associated with increased unintentional firearms mortality rates.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

Efforts to reduce disparities in unintentional firearms mortality are necessary, but must be multifaceted to address the complex nature of the injury risk. Efforts to improve safe use, transport and storage of firearms, plus policy change to promote safer handling of firearms handling and engineering of safety devices for firearms and firearms storage locations are likely to be most effective.

Introduction

In 2021, 48 830 Americans died from firearms-related injuries. A large portion of those deaths were the result of intentional acts such as homicide and suicide, but 549 deaths were attributed to unintentional causes.1

Strategies to prevent unintentional firearms-related mortality and injuries differ greatly from prevention of intentional injuries, and it is therefore valuable to analyse epidemiological patterns independently. Following reports that unintentional firearms injury rates are higher in rural than urban areas2 and that there are significant disparities across US states in overall firearm mortality,3 the present study considers disparities in unintentional firearm-related mortality rates by US state. We pursued two goals: (a) describe unintentional firearm-related mortality across the USA and within individual states between 2001 and 2021 and (b) consider factors that might explain disparities across states.

Methods

Unintentional firearms mortality between 2001 and 2021 was included, as this provided the most recent and largest range of data available in the CDC WISQARS database. We used the full span of years to obtain stable estimates in less populated states.

The following five predictors were included, all of them estimated by state: rurality, non-white population, poverty, population and gun ownership. Rurality was derived from US Census data using 2010 census data. Racial distribution and poverty distribution were estimated using 2011 Census Bureau data, derived from the American Community Survey, concerning the per cent of families falling below the poverty line. State population was derived from US Census Bureau data, effective 1 July 2011. Gun ownership was derived from estimates developed by RAND. No authoritative data on gun ownership exist, so these estimates were based on a wide range of predictors from 2011, the middle year in our outcome data range (Schell et al, 2020).

Data analysis proceeded in three steps. First, we detailed unintentional firearms mortality by US state, both in tabular and graphic form. Second, we computed bivariate correlations between predictor and outcome variables. Third, we computed a linear regression model to predict crude rate with all predictors. SPSS (V.28.0.1.1; IBM Corporation) was used to conduct all analyses.

Results

Table 1 lists deaths, population size, crude rate and age-adjusted rate of firearms-related mortality between 2001 and 2021 in each US state, plus the District of Columbia. Figure 1 shows the data graphically. As shown, the highest rates clustered in Southeastern states, followed by states in the Northern Plains and Mountain West. The highest crude rates were in Mississippi (0.68), Louisiana (0.64) and Alabama (0.60). The lowest rates were in the Northeast, followed by scattered states in the West and Midwest such as Washington, Utah and Michigan. The lowest crude rates were in Massachusetts (0.03), New York (0.06), Maryland (0.06) and Connecticut (0.06). Rhode Island (0.05) also had a low rate, although the estimate was unstable due to the low number of deaths in the state.

Figure 1

Crude rate of unintentional firearms mortality, by US state, 2001–2021.

Table 1

Firearms-related deaths, population, crude rate, age-adjusted rate, years of potential life lost and state population, by US state and District of Columbia, 2001–2021

Table 2 presents a correlation matrix between the outcome variable of unintentional firearms mortality crude rate and all predictor variables. As shown, unintentional firearms mortality correlated positively and significantly with the per cent of families below the poverty level in each state (r=0.54, p<0.01), the per cent of population that is rural (r=0.59, p<0.01) and the per cent of the population that owns firearms (r=0.72, p<0.01). Several of the predictor variables intercorrelated with each other.

Table 2

Descriptive data and correlation matrix, all predictor variables and outcome of unintentional firearms mortality crude rate, by US state

Table 3 presents a linear regression model predicting the crude rate of unintentional firearms mortality. The overall model was significant (R2=0.77, p<0.01). Three factors were significant predictors: the per cent of population that is white (β=−0.22, p<0.05), the per cent of families living below the poverty level (β=0.43, p<0.01) and the estimated firearms ownership in the state (β=0.54, p<0.01).

Table 3

Linear regression predicting unintentional firearms mortality crude rate

Discussion

We report large disparities in the rate of unintentional firearms mortality across the 50 US states and the District of Columbia. Crude rates in the states with the highest rates, clustered mostly in the Southeast, were about 10 times the rates in the states with the lowest rates, located mostly in the Northeast. Across the country, over 12 000 lives were lost to unintentional firearms mortality between 2001 and 2021.

Previous reports have noted large disparities in overall firearms-related mortality across states,3 but this is the first published report to consider disparities in unintentional firearms mortality. Factors that create the disparities are likely to be multifaceted. Previous research points to rurality as associated with unintentional firearms-related injuries.2 Our analysis suggests rurality, poverty and firearms ownership are each associated with unintentional firearms mortality rates.

Attempts to alleviate the complex nature of poverty in America are challenging, complex and unlikely to happen quickly.4 Attempts to reduce firearms ownership are equally unlikely, and arguably inappropriate as firearms represent an integral part of culture and life for many Americans, including especially Americans living in rural areas. Solutions to reduce unintentional firearms mortality might turn instead to alternative strategies for behaviour change to promote safe use, transport and storage of firearms, which can come from education and training for safe storage, use and transport of firearms; engineering of safety devices for firearms and firearms storage locations and policy change to promote safer firearms handling.5 6

The findings in this report are subject to at least four limitations. First, because unintentional firearms mortalities are rare events, we used data from a 20-year span to gather meaningful estimates from less populated states. Even with the use of this strategy, data estimate for four states (Delaware, Hawaii, Rhode Island, Vermont) plus the District of Columbia were unstable. Second, no precise measurement of firearm ownership exists. We used reputable estimates from RAND,7 but those estimates must be interpreted cautiously, particularly since data on firearms-related injuries and deaths are incorporated into the latent estimate. Third, we used available data that we felt were relevant to predicting unintentional firearms mortality but recognise causal factors are likely to be complex and multifaceted, and will include factors omitted from this analysis. Finally, we conducted state-level analyses but acknowledge that there is substantial diversity in culture and practice across regions of each state. The patterns we detect at the state level may not generalise to particular regions or counties within a state.

Prevention of unintentional firearms mortality is essential nationwide, but should be targeted especially to vulnerable populations, including those living in states identified by this research to have crude rates that far exceed national averages and fall 10 times higher than crude rates in the least vulnerable states. Prevention must be multifaceted5 6 and should include efforts along at least three pathways: education and training,8 policy9 and engineering of safer firearms and firearms storage.10

Ethics statementsPatient consent for publication

Not applicable.

Ethics approval

This study involved human participants. The Institutional Review Board at University of Alabama at Birmingham reviewed and approved analyses of these de-identified and anonymised datasets, considering it ‘not human subjects research’ . Consent was not obtained from participants because anonymous population-level data were analysed.