Pattern of Five-Year Weight Change by Age and Birth Cohorts in a Swedish Adult Population

Abstract

The driving factors behind the global epidemic of obesity have not yet been fully elucidated. Investigating the effect of age and cohort on weight change might enable us to develop effective intervention strategies. Aims To analyse the pattern of BMI change by age and birth cohort over five-year follow-up in adult Swedish sample. Methods The data was drawn from the Scania Public Health Cohort. The baseline survey was conducted between 1999 and 2000, and the follow-up in 2005. A total of 10,373 individuals responded to both surveys; all were included in the present study. The data was stratified into three-year intervals according to a respondent's birth year and age at baseline. This was done in order to allow for age and cohort effect analysis by means of fixed effects linear mixed models. Results We observed a significant increase in BMI between baseline and follow-up for males and females (p < 0.001), as well as significant age and cohort effects resulted from ageperiod and cohortperiod models, respectively, for both gen- ders. The age effect on BMI varied between � 1.94 to +1.10 BMI units among males and � 3.15 to +0.54 among females. The cohort effect on BMI varied from � 1.67 to +1.61 BMI units in males, and from � 2.28 to +1.18 in females. Conclusions Our analyses suggest that cohort effects play a contributory role in obesity. Our results also show that it is difficult to distinguish between age-period and cohort-period effects in our study, due to on the one hand the linear relationship between age, period and cohort, and on the other hand to a short follow-up.