Introduction: The assessment of obesity in health surveys includes the study of the physical characteristics of human body. Anthropometrical dates can be obtained by measurement or by interview. According the method that we were used we can estimate different values for frequencies of this problem. Objectives: Measure the discrepancies of the frequency of obesity and overweight estimated with data of height and weight obtained by interview respect to those measured in a population survey. Methods: Cross-sectional survey of a sample of 1,153 adult subjects, where height and weight data reported and measured were compared by Pearson correlation. It was estimated the prevalence of obesity and overweight by two methods and we calculated the Kappa’s index, the sensitivity and specifi city for to define the predictive capacity of reported data. Results: The values of weight, height and BMI by declared and measured data presented a correlation of 0.96, 0.93 and 0.91, respectively. The prevalence of obesity was 16.4% with declared data and 23.3% for measured data (kappa = 0.68), sensitivity 63.2% and specifi city 97.9%. The prevalence of excess of weight was 51.4% with declared data and 61.1% for measured data (kappa = 0.73), sensitivity 81.3% and specificity 95.3%. These differences were statistically signifi cant. Conclusions: There is an underestimation of the prevalence of obesity and excess of weight in populations when anthropometric data were used but in good agreement and specificity was found when measured data are used.
Introduction: The assessment of obesity in health surveys includes the study of the physical characteristics of human body. Anthropometrical dates can be obtained by measurement or by interview. According the method that we were used we can estimate different values for frequencies of this problem. Objectives: Measure the discrepancies of the frequency of obesity and overweight estimated with data of height and weight obtained by interview respect to those measured in a population survey. Methods: Cross-sectional survey of a sample of 1,153 adult subjects, where height and weight data reported and measured were compared by Pearson correlation. It was estimated the prevalence of obesity and overweight by two methods and we calculated the Kappa’s index, the sensitivity and specifi city for to define the predictive capacity of reported data. Results: The values of weight, height and BMI by declared and measured data presented a correlation of 0.96, 0.93 and 0.91, respectively. The prevalence of obesity was 16.4% with declared data and 23.3% for measured data (kappa = 0.68), sensitivity 63.2% and specifi city 97.9%. The prevalence of excess of weight was 51.4% with declared data and 61.1% for measured data (kappa = 0.73), sensitivity 81.3% and specificity 95.3%. These differences were statistically signifi cant. Conclusions: There is an underestimation of the prevalence of obesity and excess of weight in populations when anthropometric data were used but in good agreement and specificity was found when measured data are used.