Towards High-Accuracy Athletic Injury Predictions Using a First-Principles Modelling Approach: Theory to (Future) Practice.

Abstract

Athletic injury has previously been defined as tissue damage or other derangement of normal physical function due to participation in sports, resulting from rapid or repetitive transfer of kinetic energy. Alternatively, athletic injury has also been described as occurring when the stresses and strains experienced by a tissue result in damage severe enough to be considered an athletic injury. In mechanical models quantifying damage and the accumulation of damage over time, damage is commonly represented using a damage variable (D) ranging between 0 and 1, where D = 0 corresponds to an undamaged state and D = 1 corresponds to complete mechanical failure. Adopting this approach, a mathematically deterministic definition for athletic injury can be established, allowing precise predictions of athletic injury occurrence in mathematical models. Specifically, an athletic injury can be mathematically defined as occurring when the damage (D) experienced by a tissue exceeds a critical damage threshold (D_c), that is, D > D_c, with complete tissue failure occurring at a damage variable of 1. Alternatively, athletic injury can also be mathematically defined as an applied mechanical load (L) exceeding a critical tissue strength threshold (S_c), that is, L > S_c, with complete tissue failure occurring when L > failure strength (S_f). While determinism is valuable for establishing a precise definition of athletic injury for mathematical models, in practice, probabilistic models are needed to account for the inherent variability and uncertainty in estimating the primary variables determining athletic injury outcomes. By examining the overlap of probability density functions of tissue load and strength estimations, or the probability that D > D_c, the probability of an athletic injury occurring can be appropriately quantified. However, to offer practical utility for the prevention of athletic injuries in real-world settings, athletic injury predictions must provide sufficient windows for intervention. For injuries involving sudden unanticipated loads that exceed the possible physiological ranges of tissue strength, this method of athletic injury risk assessment may be of little value, as there is no window of opportunity for intervention. In scenarios where future loads experienced by an athlete can be estimated, tissue damage accumulates over time, or tissue load and strength values converge over time, actualising this approach by obtaining accurate estimations of these variables is likely crucial for achieving high-accuracy athletic injury predictions that offer specific data-driven windows for intervention.