Calibrating self-reported BMI in national surveillance: impact on obesity misclassification and socioeconomic inequalities in Portugal

Background

Self-reported height and weight are prone to misreporting, which can bias BMI estimates. This study identifies misreporting determinants, develops calibration equations and examines how measured, self-reported, and calibrated BMI affect estimates of obesity prevalence and socioeconomic inequalities.

Methods

We analysed survey-weighted, sex stratified data from 3,404 adults (18-64 years) in the Portuguese National Food, Nutrition and Physical Activity Survey (IAN-AF 2015–2016), including self-reported and measured anthropometry. Misreporting determinants were assessed using multinomial logistic regression. Calibration equations for height and weight were estimated using measured values, self-reports, age, region of residence and education level. Calibrated BMI was derived from predicted values. Obesity prevalence was estimated for each BMI assessment method (≥ 30 kg/m ² ). Education, income and employment inequalities in obesity were compared across BMI methods using prevalence difference and ratio, slope index and relative indexes of inequality.

Results

Height is systematically overreported and weight underreported, with misreporting increasing with age and BMI. Calibration eliminates underestimation of obesity prevalence from self-reported BMI, bringing calibrated estimates close to measured values. Regarding education-related inequalities in obesity, calibration widen disparities among women, whereas among men corrects the overestimation observed from self-reported BMI. Income and employment-inequality patterns are similar across BMI methods.

Conclusions

Among Portuguese adults, the systematic and socially patterned misreport of self-reported anthropometry affects obesity prevalence and inequality estimates. Calibration based on simple sociodemographic models improves validity and equity of obesity surveillance and could be routinely integrated into national surveys to strengthen monitoring of obesity and its socioeconomic distribution.

Key points

• The magnitude and determinants of misreporting in height and weight vary by sex and socioeconomic characteristics and are context-specific, indicating that misreporting is not uniform across population groups or settings.

• Simple sex-specific calibration equations using routinely collected variables (age, sex, region, education) substantially reduce obesity misclassification.

• Calibrated BMI almost fully aligns obesity prevalence with measured BMI, correcting underestimation from self-reported data in both sexes.

• Calibration changes education-related obesity inequalities, slightly widening them among women and attenuating their overestimation among men.

• Wider adoption of similar calibration approaches in European national surveys that rely on self-reported BMI could improve the validity of obesity surveillance and inequality monitoring.