Ecological level analysis of the relationship between smoking and residential-fire mortality.

OBJECTIVES To examine the association between tobacco smoking and residential-fire mortality and to investigate whether this association is explained by the confounding effects of selected socioeconomic factors (ie, educational attainment and median household income). DESIGN An ecological analysis relating state-level residential-fire mortality to state-level percentages of adults who smoke was conducted. Negative binomial rate regression was used to model this relationship, simultaneously controlling for the selected socioeconomic factors. RESULTS After educational attainment and median household income had been controlled for, smoking percentages among adults correlated significantly with state-level, population-based residential-fire mortality (estimated relative rate for a 1% decrease in smoking = 0.93; 95% CI 0.89 to 0.97). CONCLUSIONS Mortality from residential fires is high in states with high smoking rates. This relationship cannot be explained solely by the socioeconomic factors examined in this study.