Validity of self-reported measures of body weight and height in participants of the São Paulo Megacity Mental Health Survey

Abstract Background The use of valid indicators for the evaluation of the nutritional status of the population was very important. Objective The concordance between measured and self-reported data of body weight and height and BMI were evaluated and the factors that exert influence on the self-report. Method This is a cross-sectional population-based study of adults from the Metropolitan Region of SP (N=766). Calculation of the intraclass correlation coefficients and graphical analyzes of Bland & Altman were conducted, evaluating the agreement of these measures. Results In the total sample and for both genders, the self-reported body weight and the BMI were underestimated, while the self-reported height was overestimated. The imprecision in self-report was influenced by gender, age, school level,and nutritional status. The agreement between the measures was high in both genders and in the total sample (weight ICC 0.951/CI 0.938-0.961; height ICC 0.870/CI 0.597-0.939; BMI ICC 0.865/CI 0.677-0.928). The prevalence of overweight by self-reported measures was underestimated (13%) when compared to that calculated by the measured data. Conclusion Self-report can be influenced by several factors, producing imprecise measures. Its use in population surveys may lead to an important underestimation of the risk of illness and premature mortality for cardiovascular and metabolic diseases associated with overweight.