Self-reported weight and height are valid measures to determine weight status: results from the Brazilian National Health Survey (PNS 2013)

Self-reported measures have been used to obtain weight and height information in some epidemiological surveys. The validation of such information is necessary to guarantee data quality. This study assessed the validity of self-reported weight and height to determine weight status. Data were obtained in the Brazilian National Health Survey, a Brazilian household-based nationwide survey carried out in 2013. In this survey, 40,366 individuals (aged >= 18 years) provided self-reported and measured information about weight and height. Student's paired t-test was used to verify the differences between self-reported and measured data. The agreement between measurements was obtained using the intraclass correlation coefficient (ICC) and Bland-Altman method. To evaluate variations in weight status categorizations, the weighted kappa coefficient and exact agreement were used. Sensitivity and specificity were estimated for the self-reported information to classify overweight and obese individuals. There was high agreement between self-reported and measured weight, height, and body mass index (ICC > 0.88). The mean agreements estimated by the Bland-Altman method were 99.6% for weight and 100.6% for height. The weighted kappa coefficient showed substantial agreement among the weight status categories (> 0.66); the exact agreement was 77%. Sensibility and specificity for overweight (83% and 87.5%, respectively) and obesity (73.4% and 96.7%, respectively) were considered high for the sociodemographic characteristics evaluated. According to our results, self-reported measurements of weight and height can be used cautiously as valid alternatives to determine weight status.