Background: Studies explaining educational differences in health often employ current smoking as a mediator. To what extent does lifetime smoking mediate the association between education and self-reported health better than current smoking?
Methods: Analysis of cross-sectional data from a representative general population sample of Dutch men and women (n = 1,561) with complete retrospective smoking histories.
Results: Educational differences in smoking over the life course are more pronounced than educational differences in current smoking, especially among men. The association between education and self-reported health is reduced when controlling for smoking for men. Among women, smoking is not such a mediator. The odds ratio for men with primary education to report less than good health is reduced from 2.94 (95% CI: 1.20-6.30) to 2.62 (95% CI: 1.13-6.05) when current smoking is taken into account and to 2.14 (95% CI: 0.90-5.04) when lifetime smoking is controlled for. This reduction in the difference between the highest and lowest educated is approximately 30% and statistically significant. For women, reductions are smaller and non-significant.
Conclusions: Educational differences in smoking among men are underestimated if current instead of lifetime smoking is studied. Consequently, the contribution of smoking to bringing about social inequalities in health is underestimated if current smoking is measured.
Copyright 2004 The Institute for Cancer Prevention and Elsevier Inc.