Ethnic variation in validity of the estimated obesity prevalence using self-reported weight and height measurements

BACKGROUND We examined ethnic differences between levels of body mass index (BMI) based on self-reported and measured body height and weight and the validity of self-reports used to estimate the prevalence of obesity (BMI≥30 kg/m2) in Turkish, Moroccan, and Dutch people in the Netherlands. Furthermore, we investigated whether BMI levels and the prevalence of obesity in Turkish and Moroccan people with incomplete self-reports (missing height or weight) differ from those with complete self-reports. METHODS Data on self-reported and measured height and weight were collected in a population-based survey among 441 Dutch, 414 Turks and 344 Moroccans aged 18 to 69 years in Amsterdam, the Netherlands in 2004. BMI and obesity were calculated from self-reported and measured height and weight. RESULTS The difference between measured and estimated BMI was larger in Turkish and Moroccan women than in Dutch women, which was explained by the higher BMI of the Turkish and Moroccan women. In men we found no ethnic differences between measured and estimated BMI. Sensitivity to detect obesity was low and specificity was high. In participants with available self-reported and measured height and weight, self-reports produced a similar underestimation of the obesity prevalence in all ethnic groups. However, many obese Turkish and Moroccan women had incomplete self-reports, missing height or weight, resulting in an additional underestimation of the prevalence of obesity. Among men (all ethnicities) and Dutch women, the availability of height or weight by self-report did not differ between obese and non obese participants. CONCLUSIONS BMI based on self-reports is underestimated more by Turkish and Moroccan women than Dutch women, which is explained by the higher BMI of Turkish and Moroccan women. Further, in women, ethnic differences in the estimation of obesity prevalence based on self-reports do exist and are due to incomplete self-reports in obese Turkish and Moroccan women. In men, ethnicity is not associated with discrepancies between levels of BMI and obesity prevalence based on measurements and self-reports. Hence, our results indicate that using measurements to accurately determine levels of BMI and obesity prevalence in public health research seems even more important in Turkish and Moroccan migrant women than in other populations.