Your neighborhood matters: an ecological social determinant study of the relationship between residential racial segregation and the risk of firearm fatalities

The study addresses how structural social determinants of health—particularly residential racial segregation, income inequality, and community resilience—relate to firearm fatalities in the United States, a country with uniquely high levels of firearm-related deaths. Prior work links firearm outcomes to gun ownership, homicide and suicide rates, inequality, trust in institutions, welfare spending, and police violence. Residential racial segregation restricts access to resources and opportunities for underserved groups and is implicated in adverse health outcomes. However, gaps remain regarding how these SDOH jointly influence firearm fatalities beyond specific populations or city-level analyses. The research question is whether residential racial segregation is associated with increased firearm fatality risk at the county level in Wisconsin, controlling for income inequality, community resilience, and urbanicity. The purpose is to inform scalable, community-led interventions by contextualizing firearm fatalities within structural and socioeconomic determinants, given the pressing need to reduce firearm-related mortality and its health system burden.

The literature indicates associations between firearm outcomes and multiple systemic factors: gun ownership, homicide/suicide rates, inequality, institutional trust, welfare spending, and police violence. Residential racial segregation has been linked to health disparities and violent outcomes. Studies (e.g., Knopov et al.) examined Black-White dissimilarity and firearm homicide disparities over time; others (e.g., Wong et al.) used city-level segregation measures but often did not incorporate broader SDOH such as income inequality, community resilience, or non-Black minority populations. Research also shows income inequality correlates with firearm homicide, and community social capital relates to neighborhood deprivation and health. Despite these insights, few studies jointly assess residential racial segregation, income inequality, and community resilience as concurrent determinants of firearm fatalities across all ages and geographies within a state, motivating the present ecological analysis.

Design and setting: County-level ecological study of Wisconsin, 2019. Data sources: AHRQ Social Determinants of Health (SDOH) Database (2015–2019 ACS 5-year estimates) linked with County Health Rankings (CHR) for firearm fatality counts; GIS analyses via ArcMap. The SDOH database aggregates ACS and CHR variables at county level; zip code and tract data were not uniformly available. Firearm fatality counts were obtained from CHR because AHRQ SDOH provided only rates. The CDC suppresses firearm fatality data for counties with fewer than 10 deaths; thus, 13 counties (Buffalo, Crawford, Florence, Forest, Green Lake, Iron, Kewaunee, Lafayette, Menominee, Pepin, Richland, Rusk, Taylor) had missing counts, leaving 59 counties for count-based analyses (though descriptive SDOH measures covered 72 counties). Outcome: Number of firearm fatalities per county in 2019 (counts), defined by ICD-10 codes W32–W34, X72–X74, X93–X95, Y22–Y24, Y35.0; deaths assigned to county of residence of decedent. Exposure: Residential racial segregation measured by the Dissimilarity Index (DI; non-White vs White distribution; 0=complete integration, 100=complete segregation), categorized into tertiles (low, moderate, high). Covariates: Income inequality via Gini index (0–1) categorized into tertiles (low, moderate, high); Community resilience measured as number of individual-level risk factors within the county (income-to-poverty ratio, single/zero caregiver household, crowding, communication barrier, no full-time year-round employment, disability, no health insurance, age 65+, no vehicle, no broadband), categorized as low risk (0 factors), moderate risk (1–2), high risk (3+); Rural-urban classification using CDC NCHS 2013 scheme recoded to urban (metropolitan) vs rural (nonmetropolitan). Statistical analysis: Descriptive statistics; unadjusted negative binomial regressions of firearm fatalities on each SDOH variable; adjusted negative binomial regression estimating incidence rate ratios (IRRs) for DI categories controlling for income inequality, community resilience, rural-urban classification, and county population weights. Significance threshold p≤0.05. Software: STATA/MP v17.0. Spatial analysis: ArcMap used to map DI and firearm fatalities; Moran’s I assessed spatial autocorrelation for fatalities (0.084) and firearms per 100k (0.378) and continuous segregation (0.156). Spatial models were not estimated due to missingness and non-contiguity from suppressed counties.

- Total firearm fatalities in Wisconsin in 2019: 802. Descriptive (N=72 counties): residential racial segregation predominantly low (33.3%), moderate (31.9%), high (31.9%); income inequality low (36.1%), moderate (33.3%), high (30.6%); community resilience risk moderate (38.9%), high (31.9%), low (29.2%); rural counties 63.9%, urban 36.1%. - Unadjusted associations (negative binomial; coefficients reported): compared to low segregation, high segregation associated with higher likelihood of fatalities (Coef 1.28, 95% CI 0.72–1.85), and moderate segregation with Coef 0.86 (95% CI 0.28–1.45). High income inequality vs low: Coef 1.17 (95% CI 0.67–1.67). High community resilience risk vs low risk: Coef 1.86 (95% CI 1.37–2.36). Rural vs urban: Coef −1.37 (95% CI −1.77 to −0.97), indicating fewer fatalities in rural areas in unadjusted analysis. - Adjusted model (controlling for income inequality, community resilience, rural-urban, population weights): • Residential racial segregation: High vs low IRR 1.26 (95% CI 1.04–1.52), Moderate vs low IRR 1.09 (95% CI 0.89–1.33). • Income inequality: High vs low IRR 1.18 (95% CI 1.00–1.40), Moderate vs low IRR 1.12 (95% CI 0.93–1.34). • Community resilience: Moderate risk vs low risk IRR 0.61 (95% CI 0.48–0.78); High risk vs low risk IRR 0.53 (95% CI 0.41–0.68). • Rural vs urban IRR 1.13 (95% CI 0.95–1.34). - GIS mapping: Areas with higher residential racial segregation overlapped with higher firearm fatality rates.

Findings indicate that counties with higher residential racial segregation experience higher firearm fatality risk independent of income inequality, community resilience, and urbanicity, supporting the hypothesis that structural segregation contributes to firearm mortality disparities. High income inequality also correlates with increased firearm fatalities, aligning with prior research linking inequality and homicide. The study further incorporates community resilience as a structural SDOH; while adjusted IRRs for moderate/high risk categories were below 1, the narrative interpretation emphasizes that low community resilience (i.e., more risk factors) is associated with greater firearm fatality risk. The results underscore that firearm fatalities are not isolated events but are embedded within broader structural and community contexts. Integrating SDOH data with health outcomes enables a more nuanced understanding of social and environmental mechanisms underlying firearm violence, informing public health approaches that prioritize place-based and equity-focused interventions.

Living in highly racially segregated counties is associated with increased firearm fatality risk. The risk is further elevated in contexts of higher income inequality and low community resilience. These findings highlight the importance of multi-level, community-engaged, and policy-driven strategies—including equitable implementation of effective gun safety laws, investment to reduce extreme wealth disparities, efforts to lessen residential racial segregation, and initiatives to strengthen community resilience. Future research should assess these relationships in other states, examine ethnic segregation, include nonfatal injuries and intent, and apply multilevel and spatial models to capture neighborhood-level variation and human-environment interactions.

- Outcome assignment by county of residence rather than incident location may misclassify exposure context, serving as a proxy for place-based risk. - Suppression of firearm fatality counts for 13 counties (<10 deaths) led to missing data and analyses based on 59 counties for count outcomes, limiting representativeness and generalizability within Wisconsin. - Wisconsin focus limits generalizability to other states; Wisconsin is among the most racially segregated states. - Nonfatal firearm injuries and intent (homicide vs suicide vs unintentional) were not analyzed due to data constraints. - Segregation measure captured racial (non-White vs White) residential segregation; ethnic segregation (e.g., Hispanic/Latino, American Indian/Alaska Native) was not assessed. - High missingness precluded multilevel analyses; spatial autocorrelation noted but spatial models not estimated due to missingness and county contiguity issues. - Did not incorporate broader human-environmental developmental trajectories (psychological, behavioral, socioeconomic, political factors) over the life course.