PURPOSE: The purpose of this study was to determine the validity and reliability of the LeanScreen (LS) mobile application that estimates percent body fat (%BF) using estimates of circumferences from photographs. METHODS: The %BF of 148 weight-stable adults was estimated once using dual-energy x-ray absorptiometry (DXA). Each of two administrators assessed the %BF of each subject twice using the LS app and manually measured circumferences. A mixed-model ANOVA and Bland-Altman analyses were used to compare the estimates of %BF obtained from each method. Interrater and intrarater reliabilities values were determined using multiple measurements taken by each of the two administrators. RESULTS: The LS app and manually measured circumferences significantly underestimated (P < 0.05) the %BF determined using DXA by an average of -3.26 and -4.82 %BF, respectively. The LS app (6.99 %BF) and manually measured circumferences (6.76 %BF) had large limits of agreement. All interrater and intrarater reliability coefficients of estimates of %BF using the LS app and manually measured circumferences exceeded 0.99. CONCLUSIONS: The estimates of %BF from manually measured circumferences and the LS app were highly reliable. However, these field measures are not currently recommended for the assessment of body composition because of significant bias and large limits of agreements.
PURPOSE: The purpose of this study was to determine the validity and reliability of the LeanScreen (LS) mobile application that estimates percent body fat (%BF) using estimates of circumferences from photographs. METHODS: The %BF of 148 weight-stable adults was estimated once using dual-energy x-ray absorptiometry (DXA). Each of two administrators assessed the %BF of each subject twice using the LS app and manually measured circumferences. A mixed-model ANOVA and Bland-Altman analyses were used to compare the estimates of %BF obtained from each method. Interrater and intrarater reliabilities values were determined using multiple measurements taken by each of the two administrators. RESULTS: The LS app and manually measured circumferences significantly underestimated (P < 0.05) the %BF determined using DXA by an average of -3.26 and -4.82 %BF, respectively. The LS app (6.99 %BF) and manually measured circumferences (6.76 %BF) had large limits of agreement. All interrater and intrarater reliability coefficients of estimates of %BF using the LS app and manually measured circumferences exceeded 0.99. CONCLUSIONS: The estimates of %BF from manually measured circumferences and the LS app were highly reliable. However, these field measures are not currently recommended for the assessment of body composition because of significant bias and large limits of agreements.