The health impact of smoking in manual and non-manual social class men and women: a test of the Blaxter hypothesis

Blaxter has hypothesized that harmful behavioral habits like smoking have a greater impact on health in the non-manual than in the manual social classes, possibly because other adverse exposures have a more important role in the manual social classes. However, the outcome measure used was a composite measure of physiological indices of morbidity and the relevance of this to other health problems is uncertain. We have therefore investigated the effect of smoking on mortality, to test whether the risk of death associated with smoking differs between manual and non-manual social classes. Data on 6831 men and 7993 women, aged 45-64 when screened in the Renfrew and Paisley study, a large prospective observational study in the West of Scotland, have been analyzed. All cause mortality rate ratios for smokers compared with never smokers have been calculated within manual and non-manual social classes. Although the age adjusted rate ratios are slightly higher among the non-manual men and women (2.19 [1.83-2.61] versus 1.92 [1.71-2.17] for non-manual and manual men respectively, and 1.75 [1.54-1.99] versus 1.65 [1.50-1.82] for non-manual and manual women), this difference between social classes is not statistically significant (p-values for test of difference 0.26 and 0.47 for men and women respectively). When additionally adjusted for other risk factors, cardiorespiratory symptoms and deprivation, this picture remained the same (p-values for test of difference are 0.41 and 0.50 for men and women respectively). Similar results were found when the cohort was divided by deprivation categories rather than social classes or when smoking related mortality rather than mortality from all causes was used as the outcome measure. We therefore conclude that the health impact of smoking is similar in each socio-economic group. The relative health improvement consequent on smoking cessation is thus similar in different socio-economic groups.