The study introduces a novel AI-based framework for predicting injury risk in professional sports, combining survival analysis, improved statistical calibration, and decision theory to provide reliable injury probabilities and actionable recommendations on when players should rest. Using four seasons of data from FC Barcelona’s women’s football team, the researchers demonstrated that the new approach outperforms traditional machine learning models, identifying cumulative fatigue, particularly distance covered over the preceding 21 days, as the strongest predictor of injury risk. The framework has the potential to improve injury prevention and player availability in professional sports, with possible applications in amateur sports and physically demanding occupations.
The full press release is availabe here