Comparison of Validity and Accuracy of Circumference, SKF, and BIA Method in Predicting % Body Fat

The purpose of this study was to compare the difference in the percent body fat(% body fat) values estimated by the circumference method, the skinfold method(SKF), the bioelectrical impedance analysis(BIA), and the un-derwater weighing(UWW) and to examine the accuracy of three different field methods by comparing with ref-erence method. Twenty-nine adults(women = 13 and men =16) volunteered to participate(age = 25 to 49 years) in this study. Body weight and height were measured, and body mass index(BMI) was calculated. Body circum-ferences, SKF thicknesses, BIA, and UWW were measured from five technicians(beginner level) under the guid-ance of a trained technician. Several validated equations appropriated to race, gender, and age were used to estimate % body fat from SKF method. The prediction equations for estimating % body fat using circumference method were developed considering only gender and age. The % body fat from BIA was calculated with the value of free fatty acids(FFA) estimated from the preset equation for the device. The data were analyzed by using one-way ANOVA to examine the difference in the % body fat results measured from four different methods, and Pearson correlation coefficients and regression analysis were measured to assess the validity and accuracy for three body composition measurements. Statistical significant was set at P < 0.05. The results showed that the % body fat values measured by four different measurements were statistically similar. However, there were differ-ences in the accuracy for estimating % body fat among different measurements, and it was found that the SKF method was the most accurate method, and the circumference method was the lowest method for predicting % body fat. In conclusion, these results suggested that the estimation of body composition from the circumference and SKF method, and BIA can be estimated relatively accurately, even if measured by an untrained technician when following prescribed measurement methods and procedures. Our data also suggested that the population-specific equations appropriate to subject should be used to estimate body composition because they will probably predict more accurate estimates.