Validity and reliability of a repeated multi-changes of direction agility test in senior soccer players

Background: This study aimed to evaluate the validity, reliability, and sensitivity of repeated multi-changes of direction agility test (rMCOD) compared to a soccer-specific field test of repeated sprint ability (S-RSA) and repeated sprint ability test (RSA). Methods: Thirty-five healthy male soccer players (age: 18.4 +/- 1.3 years) from Tunisan national soccer league (elite and sub-elite) took part in this study. They performed the tests in a randomized order over five sessions interspaced by at least 72 h. The construct, predictive and discriminant validity, relative and absolute reliability, and sensitivity of the tests were analyzed. The total and best time of the test (the sum for all trials and the trial with the lowest duration on nine, six, and seven attempts for rMCOD, RSA, and S-RSA, respectively), fatigue index, rating of perceived exertion (RPE), and lactate concentration were recorded. Results: rMCOD correlated with both, S-RSA and RSA in total time (r = 0.85 and r = 0.52, respectively) and fatigue index (r = 0.74 and r = 0.83, respectively). Receiver operator characteristics were not able to discriminate between group levels (elite and sub-elite). When comparing training levels, only the fatigue index in S-RSA showed a difference between groups. Fatigue index, total time, and the best time in rMCOD showed excellent reliability, as well as the minimal change detectable (MCD = 0.89, MCD = 0.63, and MCD = 0.11, respectively) was higher than the standard error of the mean (SEM = 0.32, SEM = 0.23, and 0.04, respectively). Conclusion: rMCOD showed large to very large predictive validity compared with the S-RSA and RSA, being a reliable test for the following parameters: the best time and total time to perform the test. Nevertheless, this study design cannot ensure whether or not this test is able to detect real changes in performance in response to training since it did not include a training intervention; besides, rMCOD could not distinguish between elite and sub-elite players, which is a limitation.