Hardening determinants among adult daily smokers in nine African countries: Data from the Global Adult Tobacco Survey

Abstract

Little has been reported about hardening nor softening indicators in Africa where smoking prevalence is low. We aimed to examine the determinants of hardening in nine African countries. We conducted two separate analyses using data from the most recent Global Adult Tobacco Survey in Botswana, Cameroon, Egypt, Ethiopia, Kenya, Nigeria, Senegal, Tanzania, and Uganda (total sample of 72,813 respondents): 1) multilevel logistic regression analysis to assess individual and country-level factors associated with hardcore, high dependence, and light smoking.; 2) a Spearman-rank correlation analysis to describe the association between daily smoking and hardcore, high dependence, and light smoking at an ecological level. Age-standardized daily smoking prevalence varied from 37.3% (95 %CI: 34.4, 40.3) (Egypt) to 6.1% (95 %CI: 3.5, 6.3) (Nigeria) among men; and 2.3% (95 %CI: 0.7, 3.9) (Botswana) to 0.3% (95 %CI: 0.2, 0.7) (Senegal) among women. The proportion of hardcore and high-dependence smokers was higher among men whereas for light smokers the proportion was higher among women. At the individual level, higher age and lower education groups had higher odds of being hardcore smokers and having high dependence. Smoke-free home policies showed decreased odds of both being hardcore and highly dependent smokers daily smoking correlated weakly and negatively with hardcore smoking (r = -0.243, 95 %CI: -0.781, 0.502) among men and negatively with high dependence (r = -0.546, 95 %CI: -0.888, 0.185) and positively with light smokers (r = 0.252, 95 %CI: -0.495, 0.785) among women. Hardening determinants varied between the countries in the African region. Wide sex differentials and social inequalities in heavy smoking do exist and should be tackled.

Keywords: Africa; Ecological analysis; GATS; Hardening hypothesis; Multilevel analysis; Smokers; Softening.

Fig. 1

Odds ratio (OR) and 95% confidence interval (CI) of being a hardcore (panel A), a high dependent (panel B), a light smoker (panel C) or a light smoker by country.* Footnote: *OR and 95%CI derived from logistic regression models adjusted for age, occupation, education level, smoking rules at home, knowledge score about health effects of cigarette smoking.