Trends in smoking in Sweden from 1968 to 2002: Age, period, and cohort patterns

Objective. Smoking is related to many later life health outcomes. We examined age, period, and cohort patterns in smoking between 1968 and 2002. Methods. A nationally representative panel study allowed repeated cross-sectional comparisons of ages 18-75 (5 waves n approximate to 5000), and ages 77+ at later waves (2 waves n approximate to 500). Cross-sectional 10-year age group differences in 5 waves, time-lag differences between waves for age groups, and within-cohort differences between waves for 10-year birth cohorts were evaluated using graphs and ordered logistic regressions. Results. Age-period-cohort models suggested that period and age effects dominated smoking patterns, showing decreases over time and age. The 1935-44 and 1945-54 cohorts, however, showed lesser period decline. Moreover, men showed a period reduction of smoking rates but no age related decrease, while women showed an age related decrease but no period effect. The genders' cohort patterns were similar, with higher smoking rates in the last waves for some cohorts, for men the 1945-54 cohort and women the 1935-44 cohort. Conclusions. Cross-sectional studies of cohorts must be aware of age effects. Due to the coming of age of the 1940s' cohorts smoking may increase among women in the oldest age groups. (C) 2008 Elsevier Inc. All rights reserved.