The "true" prevalence of obesity. A comparison of objective weight and height measures versus self-reported and calibrated data.

A regression model that describes the relationship between body mass index (BMI, kg/m2) based on self-reported weight and height, and body mass index based on objective measurements, makes it possible to calibrate the BMI based on self-reports. Calibrated data on prevalence of obesity are much higher than those calculated from self-reports. The aim of the present study was to evaluate which estimates of prevalence of obesity agreed better with objective values, those based on self-reported data or those based on calibrated BMIs. Results from two different representative samples of the same Swedish rural county were compared. Objective values were based on measurements of 2190 men and 1511 women (25-65 years of age) in 1977. Self-reports from 148 men and 147 women (25-64 years of age) were derived from the ULF-study, 1980-81 (Survey on Living Conditions) and then calibrated. The comparison demonstrated that calibrated data, compared with self-reports, are much closer to objective values. There was a striking agreement between objective (14.0%) and calibrated (13.6%) estimates of prevalence of women with BMI greater than 30.0 kg/m2, in contrast to the under-estimation by self-reports (8.1%). The calibration method was reliable for the evaluation of the "true" prevalence of obesity in population studies.