Date of Completion


Embargo Period



workload; sleep; sports science; gps; analytics; injury prediction; machine learning; multi-level modeling; mediation; moderation

Major Advisor

Douglas Casa

Associate Advisor

Craig Denegar

Associate Advisor

Lindsey Lepley

Associate Advisor

Robert Huggins

Associate Advisor

John Wilson and Tania Huedo-Medina

Field of Study



Doctor of Philosophy

Open Access

Open Access


The purpose of this work was to 1) examine injury risk, rates and physical and psychological wellbeing; 2) identify risk factors for injury; 3) investigate mechanistic pathways for changes in perceived fatigue and 4) investigate the ability of supervised machine learning techniques to predict injury in women and men’s student-athletes competing in national collegiate athletics association (NCAA) division I soccer. Injuries, workload, psychological well-being, sleep characteristics and physical activity disablement was longitudinally assessed for 256 athletes from 12 separate NCAA division I teams. Absolute injury risk and injury rates were calculated. Multi-level models were used to 1) assess differences in sleep and wellness inventories 2) identify injury risk factors, and 3) investigate causal pathways (moderators and mediators) of perceived fatigue. Supervised learning techniques were used to predict subsequent injury and area under the receiver operator characteristics curve (AUC) was used to evaluate model performance. Women’s collegiate soccer players experienced 2.05 (95%CI 1.20-3.51, p