Ethnic variation in validity of classification of overweight and obesity using self-reported weight and height in American women and men: the Third National Health and Nutrition Examination Survey

Background: Few data have been published on the validity of classification of overweight and obesity based on self-reported weight in representative samples of Hispanic as compared to other American populations despite the wide use of such data. Objective: To test the null hypothesis that ethnicity is unrelated to bias of mean body mass index (BMI) and to sensitivity of overweight or obesity (BMI >= 25 kg/m(2)) derived from self-reported (SR) versus measured weight and height using measured BMI as the gold standard. Design: Cross-sectional survey of a large national sample, the Third National Health and Nutrition Examination Survey (NHANES III) conducted in 1988-1994. Participants: American men and women aged 20 years and over (n = 15,025). Measurements: SR height, weight, cigarette smoking, health status, and socio-demographic variables from home interview and measured weight and height. Results: In women and Mexican American (MA) men SR BMI underestimated true prevalence rates of overweight or obesity. For other men, no consistent difference was seen. Sensitivity of SR was similar in non-Hispanic European Americans (EA) and non-Hispanic African Americans (AA) but much lower in MA. Prevalence of obesity (BMI >= 30 kg/m(2)) is consistently underestimated by self-report, the gap being greater for MA than for other women, but similar for MA and other men. The mean difference between self-reported and measured BMI was greater in MA (men -0.37, women -0.76 kg/m(2)) than in non-Hispanic EA (men -0.22, women -0.62 kg/m(2)). In a regression model with the difference between self-reported and measured BMI as the dependent variable, MA ethnicity was a significant (p < 0.01) predictor of the difference in men and in women. The effect of MA ethnicity could not be explained by socio-demographic variables, smoking or health status. Conclusion: Under-estimation of the prevalence of overweight or obesity based on height and weight self-reported at interview varied significantly among ethnic groups independent of other variables.