Author(s): TIMMERMAN, W., JASPERS, A., VEREECKEN, S., HOUTMEYERS, K., ACKIN, L., HOMBROUX, W., BOGAERTS, S., STAES, F., Institution: KU LEUVEN, Country: BELGIUM, Abstract-ID: 2047

INTRODUCTION:
Accurate alignment between coach and athlete perceptions of training intensity and recovery is important for training planning and injury prevention. Most studies reported no difference between coach and athlete self-reported training intensity (1). However, less is known about coach-athlete agreement based on coaches’ training observations in youth sports, especially for recovery. Therefore, the aim of this study was to examine whether coach observations during training can accurately estimate athlete-perceived intensity and recovery in elite youth volleyball.
METHODS:
Two coaches and thirty-four elite youth volleyball players (13.4±0.5y) participated in this prospective observational study for 4 weeks, consisting of 6 sessions per week. Athletes reported Total Quality of Recovery (TQR, 6-20 scale) before and Rating of Perceived Exertion (RPE, 6-20 Borg scale) after sessions. Coaches reported observed recovery (TQ-OR) after warm-up and observed exertion (ROE) after sessions for each athlete. This resulted in 750 paired RPE-ROE and 639 paired TQR-TQ-OR observations.
Coach-athlete agreement was assessed per coach on pooled athlete data using ICC (2,1) with 95% confidence intervals, coach-athlete correlation with Pearson's r, and Bland-Altman analysis using athlete-self report as the criterion reporting systematic bias (±SD) with Cohen’s d effect size, and proportional bias via linear regression. Statistical significance was set at p<0.05.
RESULTS:
RPE agreement was poor (Coach 1: ICC=0.28 [0.19-0.36]; Coach 2: ICC=0.24 [0.13-0.34]). Linear association was weak but significant (Coach 1: r=0.28; Coach 2: r=0.25, both p<0.001). Bland-Altman analysis revealed opposite systematic biases (Coach 1: 0.33±1.95, d=0.17 p<0.001; Coach 2: -0.61±2.22, d=-0.28, p<0.001) and significant proportional bias (Coach 1: slope=-0.70, zero-bias point=14.5; Coach 2: slope=-0.72, zero-bias point=13.2; both p<0.001).
TQR agreement was poor to negligible (Coach 1: ICC=0.22 [0.12-0.31]; Coach 2: ICC=0.05 [-0.07-0.16]). Linear association was weak for Coach 1 (r=0.23, p<0.001) and negligible for Coach 2 (r=0.05, p=0.445). Systematic bias was small and non-significant for Coach 2 (Coach 1: 0.35±2.07, d=0.17, p=0.001; Coach 2: 0.02±2.16, d=0.01, p=0.866), yet proportional bias was significant (Coach 1: slope=-0.84, zero-bias point=14.4; Coach 2: slope=-0.95, zero-bias point=14.1, both p<0.001).
CONCLUSION:
Coach observations cannot substitute for athlete self-reported intensity and recovery as results showed poor agreement. Coaches overestimated low and underestimated high athlete self-reported intensity and recovery. Therefore, we recommend directly collecting athlete self-report for training intensity and recovery monitoring as inaccurate coach estimates may lead to inappropriate training decisions, especially during intensive training.

1. Inoue, A., et al. (2022). Internal Training Load Perceived by Athletes and Planned by Coaches: A Systematic Review and Meta-Analysis. Sports Medicine - Open, 8(1), 35.