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 socioeconomic group. The relative health improvement consequent on smoking cessation is thus similar in different socio-economic groups.@en

The present study aimed to compare survival to age 75 between men and women, by social circumstances and smoking behaviour. A 20 year follow up was carried out of a large representative cohort of Scottish men and women in the Renfrew and Paisley study, and was combined with Scottish mortality statistics. 6831 men and 7993 women aged 45-64 y at time of examination (between 1972 and 1976) were considered. Combining the estimates from the Renfrew and Paisley study with those from Scottish mortality statistics for men and women younger than 45 y of age, it was calculated that 28% of the male smokers and 44% of the male never smokers in disadvantaged social circumstances will have survived to age 75 y, compared with 46% and 56% of the female smokers and never smokers respectively. In more privileged social circumstances, 41% of the male smokers and 62% of the male never smokers will have survived to age 75 y, compared with 56% and 70% of the female smokers and never smokers respectively. The difference between male smokers in low social classes and female never smokers in high social classes is 42% in absolute terms (28% vs 70%), which gives an indication of the combined influence of gender, social circumstances and smoking on survival. These results show that gender, social circumstances and smoking are important determinants of mortality which have led to substantial differences in survival. The influence of additional factors related to mortality could usefully be compared to these to put their effects into perspective.@en

Study objective - To assess the size of mortality differentials in men by social class in Scotland as compared with England and Wales, and to analyse the time trends in these differentials. Subjects - Men from England and Wales and Scotland around each census from 1951 to 1981. Methods - Poisson regression analysis was used to calculate relative indices of inequality for disease specific and all cause mortality as a measure of mortality differentials between social classes. This measure is not dependent on the size of the social class groups, so it can be used to compare the magnitude of differentials over time periods during which the relative sizes of social class groups change. Main results - While overall death rates were higher in Scotland than in England and Wales around the 1951, 1961, and 1971 censuses the relative indices of inequality indicated smaller mortality differences between social classes in Scotland. Inequality, as indexed by the relative index of inequality, increased over time in both Scotland and England and Wales, but to a greater degree in Scotland, resulting in greater social class mortality differentials for Scotland in 1981 (the relative index of inequality increased from 1.40 to 2.43 for England and Wales, and from 1.22 to 2.57 for Scotland between 1951 and 1981). This greater increase in the magnitude of inequalities in all cause mortality in Scotland seemed to result from increasing social class differentials in cardiovascular disease, accidents and external causes, and 'all other causes of death'. Examining the trends in overall death rates, it seems that the greater increase in social class differences in Scotland occurred because of the greater decrease in death rates among the privileged social groups, in combination with a smaller decrease (or a greater increase) in the death rates in the lower social class groups. Conclusions - This study has shown that trends in mortality and in inequalities in mortality differ within Great Britain. Although death rates were higher in Scotland than in England and Wales, smaller mortality differentials by social class were found in Scotland over the period 1951 to 1971. By 1981, however, social class mortality differentials were greater in Scotland than in England and Wales. The greater increase in the social class differentials over time in Scotland, may have contributed to the worsening overall mortality profile in Scotland as compared with England and Wales that occurred between 1971 and 1981.@en

Background. Prescott et al. found that the relative risks associated with smoking for respiratory and vascular deaths were higher for women who inhale than for inhaling men, and found no gender differences in relative risks of smoking-related cancers. The purpose of the present study was to assess whether these findings are reproducible, using data from the Renfrew and Paisley study. Methods. Age-standardized mortality rate differences and age-adjusted mortality rate ratios (using Cox's proportional hazard model) were calculated for male and female smokers by amount smoked compared with never smokers. These analyses were repeated after excluding non-inhalers. Results. The all-cause mortality rate ratio was higher for men than for women in all categories of amount smoked, although this difference was only statistically significant in the light smokers (1.83 [95% CI : 1.61-2.07] for men and 1.41 [95% CI : 1.28-1.56] for women, P = 0.001). The cause-specific mortality rate ratios tended to be higher for men than for women, and this difference was most substantial for neoplasms (2.57 [95% CI : 2.01-3.29] for male light smokers and 1.35 [95% CI : 1.14-1.61] for female light smokers, P < 0.001) and, in particular, for lung cancer (11.10 [95% CI : 5.89-20.92] for male light smokers and 4.73 [95% CI : 2.99-7.50] for female light smokers, P = 0.03). Furthermore, looking at the rate differences the effects of smoking were uniformly greater in men than in women. These results were virtually unchanged after excluding non-inhalers. Conclusion. We found similar results to Prescott et al. when all smokers were considered, but could not reproduce their findings when non-inhalers were excluded. Given the fact that we showed greater rate differences in men than in women, we think that it is too early to conclude that women may be more sensitive than men to some of the deleterious effects of smoking.@en