Using Machine Learning to Predict Lower-Extremity Injury in US Special Forces.

INTRODUCTION: Musculoskeletal injury rates in military personnel remain unacceptably high. Application of machine learning algorithms could be useful in multivariate models to predict injury in this population. The purpose of this study was to investigate if interaction between individual predictors, using a decision tree model, could be used to develop a population-specific algorithm of lower-extremity injury (LEI) risk. METHODS: One hundred forty Air Force Special Forces Operators (27.4 ± 5.0 yr, 177.6 ± 5.8 cm, 83.8 ± 8.4 kg) volunteered for this prospective cohort study. Baseline testing included body composition, isokinetic strength, flexibility, aerobic/anaerobic capacity, anaerobic power, and landing biomechanics. To evaluate unilateral landing patterns, subjects jumped off two-feet from a distance (40% of their height) over a hurdle and landing single-legged on a force plate. Medical chart reviews were conducted 365 d postbaseline. χ automatic interaction detection (CHAID) was used, which compares predictor variables to LEI and assigns a population-specific "cut-point" for the most relevant predictors. RESULTS: Twenty-seven percent of operators (n = 38) suffered LEI. A maximum knee flexion angle difference of 25.1% had the highest association with injury in this population (P = 0.006). Operators with >25.1% differences in max knee flexion angle (n = 13) suffered LEI at a 69.2% rate. Seven of the 13 Operators with >25.1% difference in max knee flexion angle weighed >81.8 kg, and 100% of those operators suffered LEI (P = 0.047; n = 7). Only 33% of operators with >25.1% difference in max knee flexion angle that weighed <81.8 kg suffered LEI. CONCLUSIONS: This study demonstrated increased risk of LEI over a 365-d period in Operators with greater differences in single-leg landing strategies and higher body mass. The CHAID approach can be a powerful tool to analyze population-specific risk factors for injury, along with how those factors may interact to enhance risk.

Shared Neuromuscular Performance Traits in Military Personnel with Prior Concussion.

Concussions are common in military personnel and may result in increased risk of musculoskeletal injury. One plausible explanation for this risk could be that neuromotor deficiencies enhance injury risk after a concussion through altered muscular activation/contraction timing. PURPOSE: To compare military personnel with at least one concussion during the past 1 month to 2 yr (CONCUSSED) to military branch-matched, age-matched, and Special Operations Forces group-matched controls (CONTROL) on physiological, musculoskeletal, and biomechanical performance. METHODS: A total of 48 (24 CONCUSSED, 24 CONTROL) male Air Force and Naval Special Warfare Operators age 19 to 34 yr participated in the study. Participants self-reported demographics/injury history and completed the following assessments: 1) physiological-body composition, anaerobic power and capacity, aerobic capacity and lactate threshold; 2) musculoskeletal-lower extremity isokinetic strength testing, including time to peak torque; and 3) biomechanical-single-leg jump and landing task, including landing kinematics of the hip, knee and ankle. A machine learning decision tree algorithm (C5.0) and one-way ANOVA were used to compare the two groups on these outcomes. RESULTS: Despite nonsignificant differences using ANOVA, the C5.0 algorithm revealed CONCUSSED demonstrated quicker time to peak knee flexion angle during the single-leg landing task (≤0.170 s; CONCUSSED: n = 22 vs CONTROL: n = 14), longer time to peak torque in knee extension isokinetic strength testing (>500 ms; CONCUSSED: n = 18 vs CONTROL: n = 4) and larger knee flexion angle at initial contact (>7.7°; CONCUSSED: n = 18 vs CONTROL: n = 2). CONCLUSION: The findings supported the hypothesis that CONCUSSED military personnel would demonstrate altered neuromuscular control in landing strategies and muscular activation. Future research should assess prospectively neuromuscular changes after a concussion and determine if these changes increase risk of subsequent musculoskeletal injuries.