Aim: The purpose of this study was to test utility of different fitness field tests to predict self-reported low-back pain (LBP) in youth.
Methods: A sample of 376 school aged (4th through 10th grade students) first completed a survey instrument designed to assess past events (and degree) of LBP. On subsequent days, participants completed a battery of different muscular fitness tests including the current FITNESSGRAM trunk extension test. Receiver operator characteristic curves (ROC) were used to test the utility of single and combined scores from various tests to predict self-reported LBP.
Results: The results show that the trunk extension test used in FITNESSGRAM has poor utility for predicting self-reported LBP. In girls, the sit-and-reach, the static curl-up test, and a composite score using static and dynamic curl-up scores had the highest Area Under the Curve (AUC=0.80, 0.71, and 0.79, respectively). Sensitivity (Se) values ranged from 75.0-85.7 while Specificity (Sp) ranged from 59.1-81.4. In boys, the dynamic curl-up test alone was the best predictor of LBP in high-school boys (AUC=0.75, Sensitivity=75.0 and Specificity=74.1).
Conclusion: The reasonable Sensitivity values from the ROC analyses indicate that individual and aggregate indicators of musculoskeletal fitness can potentially identify girls that had LBP in the past. However, the low values for Specificity indicates that girls with "No LBP" can be misclassified as being at risk for LBP. The current FITNESSGRAM trunk extension test was not able to discriminate between students with and without LBP therefore, the utility of this test for fitness assessment in youth should be revisited.