![Sample our Health and Social Care journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5938/TOC-Subject-Sample-2021-Health-Social-Care.jpg"})
Abstract
Aim: To identify correlates of bias in self-reported anthropometry among reproductive-aged Colombian women and to correct overweight/obesity and obesity prevalence based on self-reported data using two calibration techniques.
Subjects and methods: Self-reported and objectively measured anthropometry were obtained from 597 women aged 21–55 years from Bogotá, Colombia. This study identified correlates of reporting bias (self-reported minus measured anthropometry) by examining its distribution across categories of sociodemographic characteristics, objectively measured anthropometry and body shape perception using linear regression. Next, weight status misclassification was assessed using self-reported anthropometry. Finally, multivariable linear regression and ROC curves were used to calibrate weight status misclassification from self-reported data; these techniques were applied in half of the study population and validated in the remaining half.
Results: Women under-estimated weight by 2.0 ± 5.0 kg and over-estimated height by 0.6 ± 4.0 cm. Correlates of bias included objectively measured anthropometry and marital status. Self-reported BMI yielded spuriously low prevalences of overweight/obesity and obesity. The ROC approach effectively corrected overweight/obesity prevalence, while the regression method provided a more accurate estimate of obesity prevalence.
Conclusions: Bias in self-reported anthropometry varied with respect to objectively measured anthropometry and sociodemographic characteristics. BMI from self-reported anthropometry under-estimates overweight/obesity and obesity prevalence; calibration methods can effectively correct reporting bias.