Associations between BMI, social strata and the estimated energy content of foods

OBJECTIVE: Many studies have shown that the prevalence of obesity is greater in lower social classes. The reasons for this effect however are unclear. Since there is also a link between education and social class, and an association between education and prevalence of obesity, one hypothesis is that lack of education about energy contents of foods may contribute to the effects of social class on obesity. SUBJECTS: We tested the hypothesis that knowledge of food energy contents is associated with differences in body mass index (BMI) in a sample of 346 people of both genders, aged between 18 and 45 y, of variable body mass index and drawn from different social strata. RESULTS: Estimates of food energy contents were on average well correlated with the actual energy contents, but individual estimates were very poor in all subpopulations of this sample. We found that subjects of different BMIs did not differentially estimate the energy contents of foods high in carbohydrate, but low in fat and protein (fruit and bread). However, foods that contained high fat contents, independent of the other macronutrients present, were generally perceived to have significantly lower energy contents by obese people than nonobese subjects (although this was not observed for all high-fat foods). Overall, this difference interacted with social class, such that the difference between the BMI groups was exaggerated in the lower social stratum but abolished in the higher social class. Binary logistic regressions revealed that the probability of being obese (BMI > 30 kg/m2) in the lower social class group was significantly negatively associated with the estimated food energy content of most high-fat foods. Such an association was not found in the higher social class group. In the lower social class group, overall food knowledge appeared superior in the leaner subject group (BMI < 30 kg/m2), but obese subjects were actually better at estimating the energy contents of snacks and alcoholic beverages. The leaner group significantly overestimated the energy contents of these items. CONCLUSION: Differences between individuals in estimates of food energy contents may contribute to the development of obesity in lower social strata. Whether this is driven by a protective effect in lean subjects of overestimating the energy contents of certain foods (snacks and alcoholic beverages) or a susceptibility in the obese because they underestimate the energy contents of other foods is not certain. Knowing which of these alternatives is true is important and may help design public health education programmes directed at these people to help alleviate the obesity epidemic.