Do socio-environmental variables mediate the relationship between socioeconomic status and obesity among African Americans?

The association of excess body weight with several chronic diseases makes the obesity epidemic a major public health concern. African American women have the highest obesity rate, 51%, compared to all other groups; and, 31% of African American men are obese. It has been well documented in the literature that among United States adults, obesity is more prevalent in lower socioeconomic status (SES) groups. However, recent evidence suggests this relationship may vary with socio-environmental characteristics, including perceived neighborhood safety, the presence of sidewalks, and number of grocery stores within walking distance. Understanding how socio-environmental variables influence the association between SES and obesity may help identify strategies to prevent and manage obesity.

We are conducting cross-sectional analyses using data from an ongoing intervention study to examine whether the relationship between SES and obesity is mediated by socio-environmental variables. Participants were 235 adults in the baseline examination of the Nutrition and Fitness for Life (NFL) Study, a community-based, randomized control intervention for cancer prevention among African Americans based at the University of California San Francisco. SES is defined by annual household income and education. Obesity is measured using Body Mass Index (BMI) of 30 or higher. Socio-environmental characteristics are measured using the International Physical Activity Questionnaire Environmental Module.

Since the NFL study targeted people who were already overweight, 84%of participants were obese. The mean age was 45 years and 87% of participants were female. Income was broadly distributed with 31% reporting incomes of $50,000 or more. Thirty-seven percent of participants reported having some college training after high school, and 33% percent of participants were college graduates. We will present results from structured logistic regression models, with obesity modeled as a function of SES and socio-environmental variables, controlling for gender, age, race/ethnicity and other demographic variables. Given that most study participants were obese by the conventional definition (BMI>30), we will explore different BMI cut-points for the dependent variable. Investigating this relationship in the African American community will help identify strategies to reduce the burden of obesity and related diseases in this specific population.