Tobacco-attributable burden of cancer according to socioeconomic position in France

Abstract

Introduction Smoking is a major preventable cause of cancers and is increasingly concentrated among the most deprived individuals leading to increasing socioeconomic inequalities in the incidence of cancers linked to smoking. It has been shown that a higher level of tobacco smoking partly accounted for the higher incidence of tobacco related cancers in low socioeconomic groups. In addition to quantifying the relative contribution of risk factors to socioeconomic inequalities in cancer incidence, assessing the absolute number of cancer cases attributable to smoking by socioeconomic position (SEP) at a national level is important for policy makers to assess the population burden of smoking on social inequalities in cancer. We aimed to estimate the tobacco-attributable cancer burden according to socioeconomic position in France. Methods The analysis was restricted to cancer sites for which tobacco smoking was recognized as a risk factor. Cancer cases by sex, age group and European Deprivation Index (EDI) among people aged 30–74 between 2006 and 2009 were obtained from cancer registries covering approximately 20% of the French population. Socioeconomic position was assessed using the quintiles of the EDI. The tobacco-attributable burden of cancer according to EDI was estimated applying the population attributable fraction (PAF) computed with the Peto-Lopez method. We quantified the contribution of smoking to socioeconomic inequalities in cancer incidence at the national level by comparing the number of tobacco-related cancers that would be prevented if the four lower EDI quintiles smoked as the least deprived EDI quintile with the number of tobacco-related cancers that would be prevented if the four lower EDI quintiles had the same IRs as the least deprived EDI quintile Results The PAF increased from 56% in the least deprived EDI quintile to 70% in the most deprived EDI quintile among men and from 26% to 38% among women. Among men and women, the PAF was higher in the younger age groups i.e. 30–59 years. The age difference was pronounced in the more deprived EDI quintiles whereas the PAFs were almost identical in both age groups in the least deprived EDI quintile. In total, 28% of the excess cancer cases in the four most deprived EDI quintiles in men and 43% in women could be prevented if smoking in these four EDI quintiles was similar to that of the least deprived EDI quintile. The proportions did not differ by sex in the age group 60–74 (around 30%), but was much higher in women than in men in the age group 30–59 (48.1% versus 26.3%). Conclusion A substantial smoking-attributable burden of cancer by socioeconomic position was observed in France. The results highlight the need for policies reducing tobacco consumption. More comprehensive interventions integrating the various dimensions of health determinants and proportionate according to socioeconomic position may essentially contribute to the reduction of socioeconomic inequalities in cancer.