Abstract Background In monitoring population health and the effectiveness of public health strategies, the body mass index (BMI) is often assessed within national surveys from self-reported height and weight rather than measured values. Using data collected within a recent health examination survey (HES), the discrepancies between self-reported and measured values were assessed, and correction models were estimated and implemented on national interview survey data. Methods Within the CUORE Project, the Italian National Institute of Health conducted the HES 2018-2019 measuring height and weight as well as collecting data on self-reported values in random samples of general population aged 35-74 years residing in ten (of 20) Italian regions distributed in the North, Centre and South: 1033 men and 1061 women. Results Self-reported and measured data comparison showed greater differences in mean values of height than weight and in women than in men (height +2 cm in men and +3.2 in women; weight -0.7 kg and -1.4 kg, respectively) and a corresponding underestimation of BMI (-0.7 kg/m2 and -1.4 kg/m2, respectively). Differences were stable across age groups and educational levels, except for height discrepancy, which was greatest in women aged 65-74 years. Self-reported vs measured prevalence were: normal weight 39.7%-33.3% in men and 54.8%-44.7% in women, overweight 45.8%-46.1% and 26.0%-29.2%, obesity 13.8%-20.1% and 15.7%-23.9%. Linear regression models adjusted by sex and age classes were assessed for height and weight (R2 > =0.92) and implemented to estimate adjusted BMI and normal weight/overweight/obesity prevalence on the national multi-purpose interview survey data collected by the Italian National Institute of Statistics. Conclusions To provide more accurate prevalence of normal weight, overweight and obesity, self-reported values could be adjusted using correction models developed on the basis of the relationship between self-reported and measured height and weight values. Key messages • Discrepancies between self-reported and measured values of height and weoght were found. • Self-reported values could be adjusted using correction models developed on the basis of the relationship between self-reported and measured height and weight values.
Read full abstract