Measurement agreement of FITNESSGRAM aerobic capacity and body composition standards

Abstract

<p>Introduction: FITNESSGRAM is a fitness testing program developed by The Cooper Institute. The FITNESSGRAM program provides two different tests for assessing aerobic capacity (VO2 max), the PACER and the one-mile run. The FITNESSGRAM program also provides two major options for assessing body composition, body fat using skinfold measurements or bioelectrical impedance analyzers (BIA) and body mass index (BMI). Criterion-referenced standards are used in FITNESSGRAM to determine the child's level of fitness compared to the level of fitness needed for good health. New standards were recently developed that classify student's fitness level into 3 zones based on potential risk for developing metabolic syndrome (Healthy Fitness Zone (low risk), Needs Improvement - Low Risk (moderate risk), and Needs Improvement - High Risk (high risk)). The new standards were also designed to improve classification agreement between alternative fitness assessments. Preliminary results have supported the utility of the new standards but additional validation research is needed. Purpose: The present study examined the measurement agreement and classification agreement between alternative indicators of aerobic capacity and body composition assessments in FITNESSGRAM. Methods: Data were collected in partnership with a local school district that agreed to let researchers administer aerobic capacity and body composition tests as part of normal physical education testing. Data were collected at a high school (grades 9 - 12) and a junior high school (grades 7, 8). Data were combined across grades to enable measurement agreement to be examined at both the junior high school and high school level. Separate two-way (gender x school type) ANOVA analyses were used to evaluate grade and gender differences in fitness levels with both sets of indicators (4 total ANOVA analyses). Percent agreement was computed between the assessments and kappa statistics were used to understand the practical differences in agreement. Results: A total of 701 students had data for both measures of aerobic capacity (PACER and one-mile run) and 889 had all measures of body composition (BMI and body fat). The correlation between predicted aerobic capacity from the PACER and one-mile run tests was high (r = 0.94). The percent of students achieving the HFZ was similar for the one-mile run (69.6%) and the PACER (72.6%). The total percent classified similarly by the two tests was 92.7%. The kappa score was also very good (0.84). The overall correlation between predicted measure of percent of body fat using the BIA and skinfold was high (0.86). The percent of students achieving the HFZ was BIA (58.8%), skinfold (61.3%), and BMI (55.2%). The classification agreement averaged 70.6% for the various pairwise comparisons (skinfold/BIA: 76.2%, skinfold/BMI: 68%, BIA/BMI: 67.7%). The classification agreement based on kappa scores were fair to moderate (skinfold/BIA: 0.575, skinfold/BMI: 0.454, BIA/BMI: 0.462). Conclusion: The findings of this study suggest that there is strong agreement between measures of aerobic capacity, while there are moderate levels of agreement between measures of body composition.</p>