Physical training outcome predictions with biomechanics, Part II: Overuse injury modeling

Abstract

Objectives: In Part II of a two-part series, we develop a phenomenological model of a negative outcome of U.S. Army Basic Combat Training that affects a large proportion of trainees. Previous models have been epidemiological in nature and have focused on trainee risk factors such as previous injury, gender, and initial fitness. This approach is limited due to difficulties extrapolating results to other cohorts. In addition, training regimen is often neglected, limiting accuracy when applied to novel scenarios. Methods: The prognostic Training Adaptation Injury Model (TAIM) developed accounts for both individual characteristics as well as regimen by integrating validated submodels of physiological and biomechanical principles known to be important for tibial stress fracture. Results: We find that when used to predict any type of overuse injury, the TAIM is most accurate when the effect of training activities on both overall fitness as well as muscle fatigue during activities is accounted for area under the receiver-operator curve of 0.65. This compares favorably with statistical-based models that do not account for training regimen (area under the receiver-operator curve ≈ 0.56. Conclusions: The TAIM has the potential to both identify trainees at overuse injury risk as well as make recommendations on regimen changes to reduce that risk.