Comparison of Body Mass Index and Percent Body Fat Measurements used to Assess Obesity for Participants in Health Fairs

Introduction: Bioelectric impedance analysis (BIA) using footpad style electrodes and body mass index (BMI) are convenient methods to estimate weight status in large populations. In particular, BMI has been used to assess obesity prevalence and trends nationally. However, BMI is an imprecise method to assess true adiposity. This study evaluated the concordance between the two methods and evaluated optimal BMI cut-points for better agreement with BIA estimated adiposity.

Methods: Physical and demographic data on 249 adults were collected during several health fairs in Oklahoma in 2005 and 2006. Percent body fat was measured by BIA. The weighted kappa statistic was used to test for agreement in categorizing body composition between BMI and by BIA. Receiver operating characteristic curves were used to evaluate the optimal BMI cut-point value for obesity based on percent body fat estimates from BIA.

Results: Prevalence rates of overweight and obesity were similar to the National Health and Nutrition Examination Survey (66.7% vs. 66.3%) based on BMI. The kappa statistic was higher for women than for men (0.84 vs. 0.69). The BMI cut-points with minimum misclassifications of obesity for men, women, and overall was 29 (sensitivity 0.84, specificity 0.85), 28 (sensitivity 0.93, specificity 0.88), and 29 (sensitivity 0.86, specificity 0.89) respectively. The standard BMI for obesity nationally is 30.

Conclusions: There is substantial agreement between BMI and BIA for the categorization of obesity. However, lower BMI cut-points provide better agreement with BIA estimates of obesity.