Background/Purpose Given the inherent limitations of BMI, low-cost alternatives for assessing body composition are needed to accurately and efficiently estimate adiposity in school-aged children. The purpose of this study was to compare the agreement in body fat (%BF) estimates between the handheld Omron HBF-301 BIA device (BIA) and the two-site skinfold thickness assessment (SKF) in middle-school students.
Method Body composition was assessed on 134 students (69 girls; 65 boys) enrolled in the 6th-8th grades utilizing SKF and BIA on the same testing day. SKF consisted of averaging three skinfolds measurements at two sites (tricep, calf) with use of the Slaughter et al. (1988) formula to estimate %BF. BIA required entering physical characteristics into the analyzer and, with shoes off, having the students hold the device with arms extended until a %BF reading was displayed.
Analysis/Results Bland-Altman plots yielded 95% Limits of Agreement of (-11.93, 13.26) for Grade6, (-10.03, 7.89) for Grade7, and (-10.82, 6.08) for Grade8. A correlation coefficient between %BF differences (BIA-SKF) and mean %BF was r= -.3381 (p<.05) for Grade8. Classifying the students into obese/non-obese categories using critical %BF gender-specific obesity cutoffs, a kappa coefficient of .39 (p<.001) and proportion of agreement of 0.88 was found between methods.
Conclusions The results suggest that although BIA and SKF classified youth into obese/non-obese categories similarly, there were large differences in %BF estimation between the two methods with the BIA method tending to have more pronounced underestimation of %BF compared to SKF in older children with higher body fat percentages.