Longitudinal Data Collection via Consumer-Grade Wearable Sensors in Adolescent Student-Athletes: Feasibility and Potential (Preprint)

Abstract

Recent advancements in sports medicine have been fueled by innovative technologies, particularly consumer-grade wearable devices like Fitbit, Apple Watch, and Garmin. These devices, offering real-time physiological and biomechanical data, have paved the way for integrating machine learning algorithms, holding promise for personalized and real-time assessments in athlete recovery. However, challenges such as data accuracy and ethical considerations underscore the importance of careful navigation in incorporating these technologies into healthcare research.

The primary objective of this study was to assess the feasibility of integrating consumer-grade wearable technology into the recovery monitoring of adolescent student-athletes.

The study enrolled 34 high school student-athletes aged 14-18 diagnosed with either concussion or acute orthopedic injuries within 10 days of injury. Participants were equipped with a Fitbit Sense, facilitating continuous monitoring of cardiovascular metrics, physical activity levels, and sleep patterns. The data collection phase extended 4-6 weeks beyond respective injury clearance, during which adherence rates were assessed at both hourly and daily intervals. Adherence was defined as the portion of each hour period that participants were compliant in wearing a functioning device. Additionally, daily adherence was determined based on the presence of recorded subject activity.

The study demonstrated high participant adherence to wearing the Fitbit Sense devices. The orthopedic injury cohort exhibited a median adherence rate of 95%, with individual rates ranging from 79% to 98%. Similarly, the concussion cohort demonstrated a median adherence rate of 85%, with adherence rates spanning from 39% to 98%. Notably, we encountered minimal issues related to device functionality, with only one participant necessitating a device replacement.

These findings demonstrate successful integration of wearable technology in data collection for adolescent student-athletes recovering from sports-related injuries. However, it is important to consider current limitations, including factors that may influence data accuracy and precision. This study serves as the first step of a project that will incorporate machine learning to predict recovery prognosis following pediatric concussion. In conclusion, this feasibility study demonstrates the practicality and acceptability of employing consumer-grade wearable technology for the collection of physiological and biomechanical parameters in adolescent student-athletes recovering from sports-related injuries. The high level of adherence underscores the potential applicability of wearable devices in this population. Study findings lay the foundation for future investigations with larger and more diverse cohorts, offering promise for optimizing assessment and management of injured athletes through wearable technology integration. As this field continues to advance, it holds the potential to guide healthcare providers, coaches, and athletes in optimizing injury recovery protocols.