Back

State-level trends in urban-rural differences in cigarette smoking in the United States

Ulm, C.; Golden, S. D.; Hill, F.; Wiesen, C. A.; Mills, S. D.

2026-08-10 public and global health
10.64898/2026.08.06.26359850 medRxiv
Show abstract

Introduction Smoking prevalence remains higher in rural than in urban populations in the United States. To examine recent trends, we assessed state-level differences in cigarette smoking between urban and rural areas from 2018 to 2024. Methods Using repeated cross-sectional data from the Behavioral Risk Factor Surveillance System, we estimated state-specific logistic regression models to examine the relationship between urban-rural county residence and cigarette smoking. Unadjusted models (model 1) included urban-rural county status and year. Subsequent models (model 2) added age, sex, and race/ethnicity. A final model (model 3) included education and an interaction term between urban-rural county status and year to examine whether gaps in urban-rural smoking changed over time. In states with significant interactions, simple effects tests compared trends for urban-rural groups separately. Results Compared to urban adults, rural adults had higher unadjusted odds of cigarette smoking (odds ratio [OR] range:1.07-1.88) in 88.4% (38/43) of states. Adjusting for demographic covariates (model 2) increased the proportion of states with significant marginal effects of rurality to 90.7% (ORs:1.09-1.87). A final model that also controlled for education (model 3) decreased the proportion of states with significant marginal effects of rurality to 60.5% (ORs:1.10-1.54). Among the 14 states with significant interaction terms, the odds of smoking declined faster among urban than rural residents. Conclusion Urban-rural differences in smoking persist across most states. No state showed a reduction in urban-rural disparities over time, and the urban-rural gap widened in 14 states. Demographic variation accounted for some, but not the majority, of observed urban-rural differences.

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.