Does the association between smoking and mortality differ by educational level?

Some researchers suggest that the effect of smoking on health depends on socioeconomic status; while others purport that the effect of smoking on health is similar across all social groups. This question of the interaction between smoking and socioeconomic status is important to an improved understanding of the role of smoking in the social gradient in mortality and morbidity. For this purpose, we examined whether educational level modifies the association between smoking and mortality. Information on smoking by age, gender and educational level was extracted from the Belgian Health Interview Surveys of 1997 and 2001. The mortality follow up of the survey respondents was reported until December 2010. A Poisson regression was used to estimate the hazard ratio of mortality for heavy smokers, light smokers, and former smokers compared with never smokers by educational level controlling for age and other confounders. Among men, we found lower hazard ratios in the lowest educational category compared with the intermediate and high-educated categories. For instance, for heavy smokers, the hazard ratios were 2.59 (1.18-5.70) for those with low levels of education, 4.03 (2.59-6.26) for those with intermediate levels of education and 3.78 (1.52-9.43) for the highly educated. However, the interaction between smoking and education was not statistically significant. For women, the hazard ratios were not significant for any educational category except for heavy smokers with intermediate levels of education. Also here the interaction was not statistically significant. Our results support the hypothesis that educational attainment does not substantially influence the association between smoking and mortality.