Self-reported weight and height - Implications for obesity research

Background: Self-reported weight and height are under- and over-reported, r espectively, in epidemiologic studies. This tendency, which may adversely a ffect study operations, has not been evaluated among subjects being enrolle d into a weight-loss program. Methods: Self-reported weight, height, and body mass index (BMI) were compa red to measured values in 91 overweight or obese (BMI>27.3) women being enr olled into a randomized, controlled trial of two behavioral interventions f or weight loss. The effects of demographic factors, baseline weight, baseli ne height, and baseline BMI on weight and height reporting were assessed. Results: There was a significant difference between measured and reported w eight (mean difference= -3.75 lb, p=0.0001) and height (mean difference= +0 .35 in., p=0.0007). The mean difference between measured and reported BMI w as -1.14 kg/m(2) (p=0.0001). Unemployed, retired? or disabled women were mo re likely to under-report their BMI than employed women (p=0.001). Six perc ent of subjects who were initially considered eligible for the study on the basis of the self-report were eventually excluded from the study because t hey did not meet the inclusion criterion for BMI. Conclusions: Obese women who seek weight-loss assistance tend to under-repo rt their weight and over-report their height, suggesting that self-reported data are likely to be inaccurate. Misreporting is apparently influenced by employment and disability and has the potential to complicate recruitment of subjects for research studies.