A história de tabagismo no Brasil segundo coortes de nascimento, sexo e escolaridade e seus efeitos prováveis sobre a mortalidade adulta futura

Abstract

The general objective of the thesis was to study the historical pattern of smoking in Brazil and to evaluate the extent to which changes in smoking prevalence will affect the country's mortality by sex until 2030. To articulate the history of smoking with the levels of past and of the future, the thesis proposed three specific objectives: 1) the reconstruction of the history of smoking in Brazil between 1948 and 2008, by age, period and cohort; 2) the estimation of mortality attributable to smoking between 1980 and 2015; 3) projections of the effects of smoking on Brazilian mortality by 2030. The main sources of data used in the thesis were the Special Smoking Survey (2008), Demographic Censuses of 1980, 1991, 2000 and 2010, mortality Information on Mortality (SIM) and the United Nations Population Division (2017). Corrections of excess mortality for smokers followed the same methodological approach suggested by Vechhia et al. (1986) and Brenner (1993). The indirect method of Peto et al. (1992) for the estimation of smoking attributable mortality between 1980 and 2015. In addition, the method suggested by Janssen et al. (2013) and Palloni et al. (2013), developed from Lee-Carter's projection method (1992) to project total mortality and excluding smoking by 2030. The pattern of smoking history between 1948 and 2008 is different for men and women. The prevalence of smoking among men was considerably higher than that of women, regardless of age, period and birth cohort. There was a delay of at least 15 years in the spread of the epidemic among women compared to men. The educational level plays an important role as a predictor of tobacco consumption. In the results presented for the men, a higher prevalence of smoking was observed for individuals with lower level of schooling in relation to individuals with higher level of schooling during the whole period and among all cohorts. However, among women, this relationship was weaker, especially among birth cohorts between 1934-1938, 1939-1943, 1944-1948, and 1954-58. Smoking has left an important mark on the levels and trends of Brazilian past mortality. Between 1980 and 2015, smoking was responsible for 4,716,986 deaths among men and 1,802,057 deaths among women, with a total of 6,519,043 deaths. Estimates of life expectancy including smoking-attributable mortality (all causes) and excluding attributable smoking mortality show the impact of smoking in the last decades. The adult life expectancy excluding smoking attributable mortality is higher in relation to the life expectancy of the total population during the analyzed period. The projections were performed over the 15-year period between 2015 and 2030. Life expectancy gains among men during this period was 3.9 years for all causes and 1.1 years excluding attributable mortality to smoking. Life expectancy gains among women were 1.9 years for all causes and 1.6 years excluding smoking attributable mortality. This result shows the contribution of the fall in the prevalence of smoking in the last decades and, consequently, in the reduction of mortality attributed to smoking in the future gains of life expectancy.