An upward trend in body mass index has been observed in the Australian adult population for over three decades. This trend may not continue, as recent evidence for high-income countries suggests decelerating rates of increase or even a plateau. The objective of this paper is to evaluate the predictive performance of an existing two-factor age-cohort regression model and estimate it with the addition of new data.
Population-based cross-sectional datasets from 1980 to 2012 are used in the analysis, including Risk Factor Prevalence Surveys from 1980, 1983, and 1989, National Nutrition Survey 1995, Australian Diabetes, Obesity and Lifestyle Study 2000, National Health Survey 2007-2008, and National Nutrition and Physical Activity Survey 2011-2012. Previous analysis included data up to 2000, i.e. two additional datasets have been included in this study. Body mass index (BMI) is calculated from measured weight and height, for Australians aged 18 years and older who were not pregnant at the time of evaluation.
Age-adjusted trend is projected to be 2.47% increase per decade for males and 3.18% increase per decade for females, compared to 2.74% and 3.91% from earlier predictions. The respective mean BMI for men and women in 2025 is predicted to be 28.51 kg/m2 (95% CI 27.90-29.13) and 28.13 kg/m2 (95% CI 27.36-28.92), when age-standardised to population level in 2012. It is found that the average female BMI will become surpass that of males starting from the year 2045, rather than 2021 as predicted before. Due to the increase in sample size, the uncertainty around the point estimates has been reduced by approximately 30%.
Validation results indicate that previous predictions are fairly accurate when compared to the observed values in the latest surveys. There is evidence for a slowing trend for both genders. Further research is required to explicitly model the slowing down of BMI increase.