Education is associated with lower levels of abdominal obesity in women with a non-agricultural occupation: an interaction study using China’s four provinces survey

BACKGROUND The prevalence of obesity is increasing rapidly in low- and middle-income countries (LMICs) as their populations become exposed to obesogenic environments. The transition from an agrarian to an industrial and service-based economy results in important lifestyle changes. Yet different socioeconomic groups may experience and respond to these changes differently. Investigating the socioeconomic distribution of obesity in LMICs is key to understanding the causes of obesity but the field is limited by the scarcity of data and a uni-dimensional approach to socioeconomic status (SES). This study splits socioeconomic status into two dimensions to investigate how educated women may have lower levels of obesity in a context where labour market opportunities have shifted away from agriculture to other forms of employment. METHODS The Four Provinces Study in China 2008/09 is a household-based community survey of 4,314 people aged ≥60  years (2,465 women). It was used to investigate an interaction between education (none/any) and occupation (agricultural/non-agricultural) on high-risk central obesity defined as a waist circumference ≥80 cm. An interaction term between education and occupation was incorporated in a multivariate logistic regression model, and the estimates adjusted for age, parity, urban/rural residence and health behaviours (smoking, alcohol, meat and fruit & vegetable consumption). Complete case analyses were undertaken and results confirmed using multiple imputation to impute missing data. RESULTS An interaction between occupation and education was present (P = 0.02). In the group with no education, the odds of central obesity in the sedentary occupation group were more than double those of the agricultural occupation group even after taking age group and parity into account (OR; 95%CI: 2.21; 1.52, 3.21), while in the group with any education there was no evidence of such a relationship (OR; 95%CI: 1.25; 0.92, 1.70). Health behaviours appeared to account for some of the association. CONCLUSION These findings suggest that education may have a protective role in women against the higher odds of obesity associated with occupational shifts in middle-income countries, and that investment in women's education may present an important long term investment in obesity prevention. Further research could elucidate the mechanisms behind this association.