On-Field Assessment of Joint Load in Football Using Machine Learning (Part II)

Area of Science:

• Sports Medicine
• Biomechanics
• Injury Prevention

Background:

• Anterior cruciate ligament (ACL) injuries are prevalent in female youth football.
• Previous research on knee joint loading is limited to controlled lab settings, hindering understanding of in-game injury risk.
• Field-based kinetic analysis is needed for realistic injury risk assessment.

Purpose of the Study:

• To characterize knee joint loading during football-specific movements using field-based wearable sensors and machine learning.
• To identify potentially high-risk knee abduction moments (KAMs) during agility tasks in female youth football players.
• To correlate movement strategies and performance with high KAMs.

Main Methods:

• Fifty-two female youth football players performed agility tasks (structured exercises and game-based play).
• Wearable inertial measurement units captured full-body kinematics.
• A validated machine learning model classified trials based on estimated knee abduction moments (KAMs).

Main Results:

• Approximately 9-12% of change-in-direction trials were classified as high KAM in both exercise and game conditions.
• High KAM movements were associated with reduced knee and pelvis flexion, increased hip flexion, and greater pelvis rotation.
• No significant differences in peak velocity were observed, but cut angles differed between exercises and game play.

Conclusions:

• Field-based kinetic screening for ACL injury risk is feasible using wearable technology and machine learning.
• Potentially high-risk knee loading occurs during routine actions in female youth football.
• Movement strategies like upright, stiff postures contribute to elevated knee joint loading, supporting a complex-systems approach to ACL injury prevention.