Area occupation and body mass index: Composition or context?

A36 Introduction : The prevalence of obesity has increased substantially in the United States (US) and obesity has been associated with the risk of several chronic diseases, including some forms of cancer. The causes of obesity primarily result from an energy imbalance and prevention will require targeting multiple levels of its influence, including both the individual and physical environment. The purpose of this study was to determine the association of area level occupational status and body mass index (BMI) in a sample of black and white men and women in Maryland. Methods : This cross-sectional study used a random sample of non-institutionalized adults, aged 18-64, residing in six Maryland jurisdictions: Baltimore City and Caroline, Dorchester, Kent, Prince George’s, and Somerset counties. Telephone interviews using Computer Assisted Telephone Interviewing (CATI) were administered to 906 participants between April 2006 and February 2007. These data were linked to zip code occupational measures from the 2000 US Census. BMI (height in kilograms (kg)/weight in meters (m) 2 ) was calculated based on self-report and ordinal categories were created based on World Health Organization classifications. We restricted the analysis to 712 (78.6%) black and white men and women who had valid BMI and zip code data. Gender-specific proportional odds logistic regression models were used to estimate odds ratios (OR) and 95 percent confidence intervals (CI) for area-level occupation and BMI after adjustment for confounding. All models incorporated generalized estimating equations (GEE) to adjust for the correlation between individuals residing in the same zip code. Results : BMI varied slightly between areas (Intraclass Correlation Coefficient = 2.0%), suggesting that most of the variation was between individuals. No association was observed between area percentage unemployment or area percentage in production, transportation or moving, sales and office, or service occupations. For black women, the odds of higher BMI among women living in areas composed of a moderate to high area percentage in management, professional and related occupations was 2.7 times the odds among those living in areas composed of a low percentage (OR=2.66; 95% CI= 1.47-4.94). This result was not observed for black men nor white men and women, and was not a result of confounding due to age, marital status, education, current cigarette use, physical inactivity or history of chronic diseases. Additional adjustment for area socioeconomic and racial composition did not attenuate this effect. Conclusions : Area management, professional and related occupation was strongly associated with BMI among black women. Future work should examine the sociocultural and physical features of areas with high management, professional, and related occupations so that appropriate policies and interventions targeted at both the individual and communities in which they reside may be developed.