Gender- and ethnic-specific associations with obesity: Individual and neighborhood-level factors

Objectives: Gender- and ethnicity-specific groups face different risks for obesity, but little is understood about the factors that predict groupspecific risks. We evaluated individual and neighborhood factors in relation to obesity. Design: Cross-sectional surveys of adults (ages 18-100 years) from southeastern Pennsylvania were analyzed. Individual-and neighborhoodlevel factors were included in fully-adjusted regression models to estimate relationships with obesity for specific gender-ethnic groups. The study included 679 Asian women, 655 Asian men, 4190 African-American women, 1568 African-American men, 1248 Hispanic women, 586 Hispanic men, 11791 European American women, and 6547 European American men. Results: There were significant differences in the predictors of obesity by gender-ethnic groups. Obesity was differentially associated with age (p<0.001, positively associated with middle age in African-American men and in all women except Asian; positively associated with older age in European American women but inversely in African-American men and European American men), employment (p<0.01, positively associated in African-American men and European American men) and poverty (p<0.001, positively associated in Asian men, African-American women, and European American women). Reporting good/excellent health was differentially associated with less obesity (p<0.01, no association for African-American men and Asians). Interestingly, neighborhood-level effects, however, did not differ significantly by gender-ethnic group. Inverse neighborhood effects on obesity prevalence were observed in most groups for higher neighborhood education and family income. Direct associations with obesity were observed for neighborhood poverty and neighborhood smoking. Conclusions: We observed that individual- and neighborhood-level variables are associated with obesity. Several individual-level effects differ by gender-ethnic group.