Socioeconomic disparities in obesity are well established. The burden of obesity is on the rise, and the prevalence is highest among minorities and those from lower incomes and education. The purpose of this study was to examine the role of dietary factors in explaining socioeconomic disparities in obesity. The data was collected as part of the Seattle Obesity Study, a population based study of adult residents of King County, WA. A series of analyses were conducted to address several gaps in the existing literature on this topic and to improve our understanding of the relations among socioeconomic status, dietary factors and obesity within the same population.
The first part of the analysis examines the relations among socioeconomic status (SES), diet cost and diet quality using a sample of 1,270 adults (468 men and 802 women). The second part of the analysis focuses on the role of diet cost in explaining socioeconomic disparities in diet quality. Income and education, used as indicators of SES, were obtained from a telephone survey. Diet cost estimates were obtained from the cost instrument attached to a Food Frequency Questionnaire (FFQ). The estimates reflect the monetary value of the diet using lowest prices at which foods were available in key supermarkets in the area. Diet quality measures were derived from the FFQ. These include intake of 16 key nutrients and two summary measures of overall diet quality—energy density and Mean Adequacy Ratio (MAR). For analyses, individual nutrient intake and diet cost were energy adjusted using the residual method and converted into quintiles. Multivariate regression analyses were conducted and covariates included age, gender, race/ethnicity, household size and total energy intake. Results showed consistent associations among SES, measures of diet quality and diet cost. First, a socioeconomic gradient in diet quality was established in the present sample such that higher income and education were associated with diets lower in energy density and higher in nutritional quality. Second, an association between SES and diet cost was established such that quintiles of energy adjusted diet cost were associated with higher proportions of those with higher income and higher education. Third, higher diet costs were in turn associated with higher diet quality. Nutrient intakes associated with better health (i.e. vitamin A, C, D, E, B12, beta-carotene, magnesium, potassium, calcium, iron, folate and dietary fiber) were associated with higher diet costs. On the other hand, dietary components associated with poor health outcomes (i.e., fats and added sugar) were associated with lower diet cost. Consistent associations were obtained with overall measures of diet quality such that diets lower in energy density and higher in nutritional quality were associated with higher diet costs. This is a first study to establish associations among SES, diet cost and multiple measures of diet quality using a population based sample of US adults from King County, WA.
Further, formal mediation analyses were conducted to examine the role of diet cost in explaining socioeconomic disparities in diet quality. Results showed that the significant effect of income on diet quality was mediated by diet cost. These associations were moderated by education level such that income-diet cost play a stronger role in determining diet quality among those with lower education as compared to those with higher education. This is the first study to establish that diet cost significantly mediates the relation between SES and diet quality.
The third part of the analysis extends these associations to the health outcome, body mass index (BMI). In addition to income and education separately, a SES index combining both income and education was created as another measure of SES. A series of analyses were conducted to test the overall hypothesis that diet quality and diet cost partly explain socioeconomic disparities in BMI. First, multivariate regression confirmed a socioeconomic gradient in BMI both among men and women such that those with higher income and/ or education had significantly lower BMI, after taking demographics and lifestyle factors into account. This is one of the few studies to examine the combined effects of both income and education on BMI. Second, the observed inverse association between SES and BMI somewhat attenuated when either of the diet quality variables were added into the model. Third, higher diet quality was in turn strongly determined by higher diet cost, after taking SES and demographics into account.
Results from the present study support the hypothesis that diet cost plays a significant role in explaining socioeconomic disparities in diet quality, which in turn somewhat explains higher BMI among them. These findings have implications for future epidemiological studies and public health policy.