Heterogeneity of income and lifestyle determinants of body weight among adult women in Mexico, 2006.

In Mexico, the combined prevalence of overweight and obesity among Mexican women increased from 64% in 2000 to 72% in 2006. In this paper, we report our findings on the relation of women's body mass index (BMI) with income and lifestyles choices using data from a cross-sectional survey conducted in 2006. The two following approaches were executed. First, we estimated a two-stage least-squares regression to control for the potential endogeneity of income stratified by urban or rural residency. The second approach was aimed at exploring whether the determinants of weight varied among different weight levels using latent class models. Our findings from the two-stage least-squares regression show a positive non-significant association between income and BMI in the overall and urban samples but a significant positive relationship among rural women. Our results suggest that one unit increase in income is associated with 4.1% increase in body weight in rural areas. Estimates from the latent class model (LCM) show a positive but marginally significant association between income and BMI in the overall sample only in the class where there is a greater likelihood that women have normal weight or overweight compared to the class with a higher probability of being obese, but we also found a large association in rural areas for both classes. Lifestyle choices were associated with BMI. Results from the two-stage least-squares regressions reveal that more hours sitting per day and a higher percentage of expenditures in sugary beverages were associated with higher BMI levels. In the LCM, for women who eventually belong to the higher body weight class, lifestyles seem to matter more. Findings from this research suggest that policies to tackle the obesity epidemic among adult women should be different for women living in urban and rural areas and women with different weight levels.