Sensitivity, reliability and construct validity of GPS and accelerometers for quantifying peak periods of rugby competition

Abstract

Training prescription and monitoring of team-sport athletes rely on accurate quantification of player movement. Our aim was to determine the sensitivity, reliability and construct validity of measures derived from a wearable device incorporating Global Positioning System (GPS) and accelerometer technology to quantify the peak periods of rugby competition. Match movement data were collected from 30 elite and 30 sub-elite rugby union players across respective competitive seasons. Accelerometer and GPS measures were analysed using a rolling average to identify peak movement for epochs ranging from 5 to 600 seconds. General linear mixed modelling was used to quantify the effects of playing position and match-half on the peak movement and variabilities within and between players represented reliability of each measure. Mean positional differences and match-half changes were assessed via standardisation and magnitude-based decisions. Sensitivity of measures was quantified via evaluation of ("signal") and typical error of measurement ("noise"). GPS and accelerometer measures had poor sensitivity for quantifying peak movement across all epochs and both levels of rugby union competition (noise 4× to 5× the signal). All measures displayed correspondingly low reliability across most epochs and both levels of competition (ICC<0.50). Construct validity was evident in mean differences between playing positions and match halves that were consistent with expected activity profiles in rugby union. However, it was clear from the pattern of differences across epoch durations and levels of competition that GPS and accelerometer measures provided different information about player movement. The poor sensitivity and low reliability of GPS and accelerometer measures of peak movement imply that rugby union players need to be monitored across many matches to obtain adequate precision for assessing individuals. Although all measures displayed construct validity, accelerometers provided meaningful information additional to that of GPS. We recommend using accelerometers alongside GPS to monitor and prescribe match respresentative training.

Peak Locomotor Intensity in Elite Handball Players: A First Insight Into Player Position Differences and Training Practices

“…Previous studies have investigated match demands (e.g., total distance, high-intensity events) through time-motion analyses, over either the entire match or shorter arbitrary time periods (e.g., block of 15 minutes) (27,33,34,39,42). These demands were described both in absolute and relative to minutes played (27,33,34,39,42). Although those segmental approaches provide meaningful primary information for practitioners, it is likely that such methods underestimate the peak locomotor intensity fluctuations that can occurs over short periods (e.g., over 1–5 minutes) (9,14,25,27,30,41,44).…”

Wearable Applications in Rugby for Performance Quantification and Player Health Assessment: A Brief Review

“…However, there remains limitations to the accuracy and reliability of wearable sensors used in such cases (Howe et al, 2020). Most experiments are in-doors or inside a lab-controlled setup in addition to lack of various important parameters and variables that are needed to build concrete conclusions.…”

Sensitivity of Countermovement Jump Variables in Professional Rugby Union Players Within a Playing Season

“…Similarly, individual week-to-week changes during a competitive season (i.e., "signal") can be used to characterize variability during the season (also commonly reported as a CV %). Signal can then be assessed with reference to the noise by combining these measures in a signal-to-noise ratio (SNR) (17,26), providing a means to quantify group-level sensitivity. This approach has previously been applied in team-sport settings (15,33,35) to identify which monitoring variables are responsive to changes related to daily training and competition (11).…”