Educational disparities of daily smoking in Estonia: an age-period-cohort approach

Objective This study analyses the educational disparities in daily smoking (DS) by age, period, and cohort during 1996–2022 in Estonia. Methods Individual-level data on adults aged 25–64 years (n=27262) from nationally representative biennial cross-sectional surveys conducted between 1996 and 2022 (response rates varied between 49%–77%) were used. Hierarchical age-period-cohort (HAPC) analysis using cross-classified random effects modelling was performed to study the independent effects of age, period, and cohort on predicted probability of DS by educational level and sex. Results The observed prevalence of DS varied between 60.7% among males with low education in 2000/2002 to 6.5% among females with high education in 2020/2022. The HAPC models by sex and education demonstrated a distinct age pattern with the highest predicted probability for smoking found at ages below 40 years, followed by a gradual decline. Period effects indicate a significant decreasing trend in the predicted probability of smoking after 2004/2006 except for females with low education, for whom it continued to increase until 2012/2014. Across all educational groups, smoking was highest in the 1960s cohort. Relative educational disparities in DS between low and high education groups have increased over the study period for males (p=0.016 for trend) and females (p=0.003), whereas respective changes were not found for cohort effects. Conclusion Significant differences in the prevalence of DS were found across educational levels, age, periods, and birth cohorts. The findings highlight considerable social inequalities in smoking in Estonia, with individuals having lower educational levels being more exposed to smoking-related harm. Implications Despite the general decline of daily smoking prevalence in Estonia over the last three decades, significant educational inequalities persist. Individuals with lower educational levels consistently show a higher risk of smoking across all ages, time periods, and birth cohorts. Additionally, the declining trend does not apply to females with low education, for whom rates remain stable or are increasing, and the relative disparities in smoking between educational groups are widening for both males and females. Smoking prevention should focus more on lower-educated groups.