Work

Accuracy of Self-reported Height, Weight and Body Mass Index in Community Health Survey in South Korea

Year: 2017

Type: article

Abstract: Objectives: The purpose of this study is to check accuracy of height, weight and body mass index (BMI) and to verify usefulness of the obesity rate based on the self-reported height and weight.Methods: Self-reported height and weight values of those who took part in the 2016 Community Health Survey in 3 areas (2,198 persons) were investigated and measured.The mean difference between the self-reported and the measured height, weight and BMI was estimated with paired t-test and the Pearson correlation coefficient.Percent agreement (%) and Kappa Statistics of BMI were calculated per group ( < 18.5, 18.5-24.9,25-29.9,≥ 30) and validity (sensitivity, specificity and predictive factor) of the self-reported obesity rate (BMI ≥ 25) was estimated. Results:The mean values between the self-reported and the measured values were compared, and it was found that height was over-reported, weight and BMI was under-reported (p < 0.05).The correlation coefficient between the height, weight and BMI was all 0.9 or higher (p < 0.05).When it comes to the mean difference in height and weight per group (4 groups), both sexes in height group 4 under-reported the height and male in weight group 1 over-reported the weight.BMI % agreement per BMI group was 88.3% in male and 85.4% in female, and Kappa Statistics was 0.779 (95% CI, 0.76-0.80) in male and 0.673 (95% CI, 0.65-0.70) in female.The sensitivity of self-reported BMI based on BMI 25 was 82.7% in male and 67.8% in female, showing a higher value in male.Obesity prevalence based on the self-reported values was 87.1% and 71.3% of that based on the measured values in male and female respectively, and it got lower among elderly people who are 60 years old or more.Conclusions: In community-based epidemiological studies, height, weight, and obesity rate should be compared and evaluated based on measurement data.

Source: Journal of Health Informatics and Statistics

DOI: 10.21032/jhis.2017.42.3.241

Authors: Jin-young Jeong, Dong-Hyun Kim, Keon-Yeop Kim, So Yeon Ryu, Soon-Young Lee, Yong Soon Park