Background:
The body mass index (BMI) is often assessed within national surveys from self-reported height and weight rather than measured values for the monitoring population health and the effectiveness of public health strategies. The discrepancies between measured and self-reported values were assessed using data collected within a health examination survey (HES) on the general adult population, and correction models were estimated and implemented on national interview survey data.

Methods:
The Italian National Institute of Health conducted the HES 2018-2019 within the CUORE Project measuring weight and height and 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 in women; weight -1 kg and -1 kg, respectively) corresponding to an underestimation of BMI (-1 kg/m2 and -1 kg/m2, respectively). Stable differences were found across age groups and educational levels, except for height, which discrepancy was greatest in women aged 65 years and over. Self-reported vs measured prevalence were: normal weight 40%-33% in men and 55%-45% in women, overweight 46%-46% and 26%-29%, obesity 14%-20% and 16%-24%. Linear regression models were assessed for height and weight adjusting by sex and age classes (R2>=0.92); they were 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:
CORRECTED Self-reported values ​, adjusted using correction models developed on the basis of the relationship between self-reported and measured height and weight values, could be considered to provide more accurate prevalence of normal weight, overweight and obesity.