Tobacco retail landscapes, structural disadvantage, and adult smoking in Tennessee: A spatially adjusted mediation analysis.

INTRODUCTION: Tennessee remains a high-smoking state with a permissive retail policy environment. We used spatially adjusted Bayesian mediation models to quantify how census-tract socioeconomic disadvantage and tobacco retailer density jointly shape adult smoking prevalence. METHODS: Using a Bayesian joint path model with spatial adjustment, we estimated direct and indirect associations of poverty, education, racial/ethnic composition, rurality, Appalachian status, and historical redlining through Home Owners' Loan Corporation (HOLC) maps. RESULTS: Greater density was associated with higher smoking: +2.59% higher odds per +1 SD (OR = 1.03). Direct associations with smoking were strongest for lower education: +16.88% per SD (OR = 1.17), followed by % non-Hispanic Black: +12.48% per SD (OR = 1.12), poverty: +8.84% per SD (OR = 1.09), non-Appalachian versus Appalachian tracts: +4.33% (OR = 1.04). Relative to ungraded tracts, HOLC D tracts showed slightly higher smoking odds. Mediation via retailer density was modest at a 20% baseline prevalence: rural vs urban: +0.15 percentage points (pp), HOLC D vs ungraded: +0.12 pp, poverty: +0.07 pp, and %<HS: +0.04 pp; HOLC A vs ungraded was negative (-0.17 pp), while non-Appalachia versus Appalachia showed a small negative mediated effect (-0.03 pp). CONCLUSION: In Tennessee, tobacco retailing is concentrated in socially and historically marginalized places and is linked to higher local smoking. Reducing prevalence and inequities will require paired strategies: retail-focused controls and broader statewide measures, higher excise taxes (cigarettes and e-cigarettes), sustained mass-media campaigns, and strengthened tobacco control infrastructure for retailer licensing systems, enforcement capacity, cessation access and coverage.