The effect of food prices and community environment on changes in Body Mass Index in China 1991-2009

Introduction: The worldwide prevalence of overweight and obesity has more than doubled since 1980, also in developing countries like China, due to rapid social and economic development. Changes have occurred in dietary patterns, physical activity patterns and the disease patterns among the Chinese population. This study explores what accounted for the rise in body mass index (BMI) in China over time between 1991 and 2009. Methods: This study uses longitudinal data from the China Health and Nutrition Survey (CHNS). Food prices and information about the community have been collected for each community from the CHNS. 3361 adults aged 18-65 were included in this study. Ordinary least squares (OLS) regressions are performed to estimate the effects of various food prices and community environment variables on BMI for 1991 and 2009 separately. In addition OLS regressions are performed with overweight and underweight as outcome. The last part of the analysis uses the traditional Blinder-Oaxaca decomposition technique in order to understand whether changes in the mean of the determinants or changes in the effect of the determinants have driven the increase in average BMI and the technique is also used to approximate the contribution of each determinant to the increase in BMI for the years 1991 and 2009. Results: Results from the OLS regressions show that different food prices play a significant role. In 1991, an increase in the price of sugar resulted in a decrease in BMI and an increase in the price of green vegetables resulted in an increase in BMI. While in 2009, an increase in the price of rapeseed oil gives a decrease in BMI. In addition, age has a significant increasing effect on the BMI and a high share of the community engaged in agriculture has a significant decreasing effect on the BMI. The decompositions make clear that the constant plays an important role in explaining the gap in mean outcome between 1991 and 2009. For BMI and overweight changes in the effects of the determinants accounted for a large part of the gap in mean outcomes and for underweight changes in the means of the determinants were more important in explaining the gap in mean underweight. Conclusion: The results make clear that the constant reduced the total contribution of the variables used in the decomposition and that the changes in the effect of the determinants accounted most for the gap in mean BMI between 1991 and 2009. This means it were not the changes in the means of the determinants that drove the rise in average BMI between 1991 and 2009. To effectively prevent overweight and obesity changes in community environment and pricing policies are needed to tackle the rise in overweight and obesity, but first more research is necessary about the effect of more food item prices on the BMI and the effect of more community environment variables on BMI.