Socioeconomic inequalities in smoking-attributable mortality in Europe; understanding trends 2000-2020 29 2 Discussion Using mortality data from ten European countries we investigated trends in partial life expectancy and hypothetical life expectancy in the absence of smoking histories, and decomposed the changing difference in life expectancy between educational groups into smoking-attributable mortality and other causes. Between 2000 and 2020 the contribution of smoking to the educational inequalities in partial life expectancy increased in most countries for women, while it decreased among men. At the same time, smoking histories continue to have a larger impact on the life expectancies of men compared to women. Previous studies have shown that higher educated men were the first to experience the gains to life expectancy by reductions in smoking-attributable mortality.19, 31 If these gains worked their way down to the lower educated groups, who stopped smoking later, it is expected that mortality inequalities would narrow. Instead, we found that inequalities were still increasing or persisting in many countries, despite reductions in smoking-related mortality leading to the largest life expectancy gains for lower educated men. This is because in most countries, life expectancy gains from non-smoking-related causes were even larger among the higher educated. Women, who generally picked up smoking later than men, are experiencing increases in smoking-attributable mortality at large, but in Sweden and Denmark SAF starts to decrease, which is in line with historic smoking patterns; women in Denmark and Sweden were among the first to start smoking, hence they were also the first to start quitting.32, 33 As with men, the peak of smoking attributable mortality was reached first among women in these countries, which is in line with the described variations in the smoking epidemic between gender and socioeconomic status.4 Within Europe large differences in smoking prevalence and implementation of tobacco control policies are observed,34 and country-level tobacco control scores are associated with lower smoking prevalence.35 Current tobacco control is not the cause of current changes in mortality due to the long period between smoking and onset of diseases. However, past tobacco control measures have had an influence on mortality due to smoking, but have not decreased educational inequalities due to smoking yet. Hence, the need for tobacco control policies that target lower educational groups remains high. We used an indirect method to estimate smoking-attributable mortality. Both a direct method, using smoking prevalence and relative mortality risk, and the
RkJQdWJsaXNoZXIy MTk4NDMw