Monitoring Internal Load with a Mobile App in Professional Soccer Players after the COVID-19 Lockdown (Readiness Soccer): Observational Study (Preprint)

Abstract

BACKGROUND Heavy physical and mental loads are typical for professional soccer players during the competitive season. COVID-19 lockdowns had recently forced competitions to be interrupted and later disputed in a shrunken calendar. Wearable sensors and mobile phones could be potentially useful in monitoring players’ training load in such highly demanding environments. OBJECTIVE The aim of this study was to explore whether remote heart rate variability (HRV) monitoring and self-reported wellness of professional soccer players could be useful to monitor players’ internal training load and to estimate their performance during the continuation of the 2020 season after the COVID-19 lockdown in Spain. METHODS A total of 21 professional soccer players participated in a 6-week study. Participants used an Android or iOS-based smartphone and a Polar H10 wearable ECG monitor for the duration of the study. Every morning they recorded their HRV and answered a questionnaire about their perceived recovery, muscle soreness, stress and sleep satisfaction. Smallest worthwhile change (SWC) and coefficient of variation (CV) were calculated for the logarithm of the root mean square of the successives differences (LnRMSSD) of the HRV. Players’ in-game performance was evaluated subjectively by independent observers and classified as high, normal and low. In order to find which variables could be potentially linked to performance, we studied their correlation and tested for significant differences among distributions. We also trained random forest models with cross-validation and bootstrapping to find the wellness and HRV features with best predictive ability for performance. RESULTS We found the usability of Readiness Soccer in a real scenario to be very good, with 81.36 points in the System Usability Scale. A total of 241 measurements of HRV and self-reported wellness were recorded. For a entire training microcycle (ie, time in between matches), self-reported high recovery (Mann-Whitney U, P=.003), low muscle soreness (P=.002), high sleep satisfaction (P=.02), low stress (Anderson-Darling, P=.03), and not needing more than 30 minutes to sleep since going to bed (Chi-Squared, P=.02), were found significant to differentiate high from normal match performance. Performance estimation models achieved the highest accuracy (73.4%) when combining self-reported wellness and HRV features. CONCLUSIONS HRV and self-reported wellness data were useful to monitor the evolution of professional soccer players’ internal load and to predict match performance levels out of measures in a training microcycle. Despite the limitations, these findings highlight opportunities for long-term monitoring of soccer players during the competitive season as well as real-time interventions aimed at early management of overtraining and boosting individual performance.