Repeatability and Variability of the 3-Min All-Out Test at the Subject Level

Abstract

Purpose

The constant work-rate to exhaustion tests must be repeated several times at each work-rate to estimate subject-level trial-to-trial variance (intra-individual variability, IIV) of critical power (CP) and work capacity (W'). Alternatively, these parameters and their variance can be estimated by repeating the 3-min all-out test (3MT) fewer times. The purpose of this study was to propose a method to determine subject-level repeatability of the 3MT and demonstrate the need to repeat the test multiple times to estimate IIV.

Methods

Seven cyclists performed a ramp test and four 3MTs on a CompuTrainer. The parameters CP, W', peak power (Pp), and total work (TW) were compared across trials using repeated measures ANOVA, Bland–Altman analysis, Intraclass Correlation Coefficients (ICC), Typical Error (TE) of measurement, and Coefficient of Variation (CV).

Results

For the group, average CP and W' were 284 ± 58 W and 10.214 ± 3.143 kJ. The reliability statistics, CP (ICC = 0.97, TE = 8 W, CV = 2.94%) and W' (ICC = 0.88, TE = 1.11 kJ, CV = 10.87%), indicated strong agreement. Subject-level repeatability was determined by comparing time-to-peak power (T_Pp), absolute difference in Pp (δ_Pp), and TW (δ_TW) for pairs of 3MTs. The average IIVs estimated by the 95% confidence intervals were ± 15 W for CP and ± 1.68 kJ for W'.

Conclusions

Thresholds are proposed for T_Pp (7 s), δ_Pp (10%), and δ_TW (3%) to determine subject-level repeatability of the 3MT before computing the IIV of CP and W'. It is suggested that the 3MT is repeated at least three times to estimate the IIV, which aids in personalized measurement of training improvements and performance optimization.