Age, period and birth cohort effects on prevalence of overweight and obesity in Australian adults from 1990 to 2000

Objective: To assess the individual contributions of age, period and birth cohort to prevalence of overweight and obesity in the Australian population during 1990 to 2000. Design: Age-period-cohort Poisson regression modelling of data from National Health Surveys conducted in Australia in 1990, 1995 and 2000. Subjects: Adults aged 20 years and over. Weightings were applied to account for differences in sampling and participation rates so that the sample is representative of the entire Australian adult population. Methods: Twelve age groups, based on 5-year intervals from 20 to 24 years to greater than 75 years, three survey periods and 14 cohorts, also based on 5-year intervals from pre-1915 up to 1976-1980, were used in the analysis. The data were age-standardized to the 2000 population and body mass index (BMI) was calculated. Log-linear models, for the prevalence rates of overall overweight (BMI >= 25) and of obesity (BMI >= 30) were fitted to the data. Results: Age (P<0.001), period (P<0.001) and cohort (P<0.002) all showed significant independent effects on prevalence of overall overweight in the Australian population such that prevalence rises with increasing age, recency of period and cohorts born since 1960. Age (P<0.001) and period (P<0.001) demonstrated strong effects on prevalence of obesity but birth cohort (P=0.07) was not significant. The effects were similar for men and women except that the overall effect of birth cohort on overall overweight was significant in women (P<0.05) but not men (P=0.09). Conclusion: The prevalence of overweight and obesity in Australian adults continued to rise during the 1990s. The obesogenic environment seems to have worsened and more recently born cohorts may be at increased risk of overweight.