Smoking and inequalities.

摘要

Prabhat Jha and colleagues (July 29, p 367) show that 51-65% of social inequalities in total mortality among men were due to smoking. As Michael Marmot2 indicates, this estimate is substantially larger than previously reported. Estimates from cohort studies have shown that smoking explains at the most 24% of mortality inequalities. This inconsistency stems from a profound discrepancy between assessment of relative and assessment of absolute inequalities.To illustrate this point, I have used data from the Gtobe study. Among men, those with only primary education had a higher mortality than men with university education (age-adjusted rate ratio 1.67, 95% CI 1.38-2.02; figure). After adjustment for smoking, the corresponding rate ratio was 1.54 (1.27-1.86). Thus, smoking explained 19% of relative inequalities in mortality, consistent with previous reports.Lesser educated men had an age-adjusted excess mortality of 5-5 per 100 person-years compared with men with a university education (figure).I modelled the reduction that would occur in absolute mortality inequalities if smoking could be reduced to a theoretical minimum-ie, zero smoking prevalence. Under this counterfactual scenario, absolute mortality inequalities would be reduced to 2-6 per 100 person-years, a reduction of 53%. This is consistent with estimates reported by Jha and colleagues.1