Monitoring Gait Complexity as an Indicator for Running-Related Injury Risk in Collegiate Cross-Country Runners: A Proof-of-Concept Study

Allison H Gruber; James McDonnell; John J Davis 4th; Jacob E Vollmar; Jaroslaw Harezlak; Max R Paquette

doi:10.3389/fspor.2021.630975

Monitoring Gait Complexity as an Indicator for Running-Related Injury Risk in Collegiate Cross-Country Runners: A Proof-of-Concept Study

Front Sports Act Living. 2021 May 21:3:630975. doi: 10.3389/fspor.2021.630975. eCollection 2021.

Authors

Allison H Gruber¹, James McDonnell¹, John J Davis 4th¹, Jacob E Vollmar^{1

2}, Jaroslaw Harezlak³, Max R Paquette⁴

Affiliations

¹ Biomechanics Laboratory, Department of Kinesiology, School of Public Health - Bloomington, Indiana University, Bloomington, IN, United States.
² Performance Engineering Laboratory, Reebok International, Boston, MA, United States.
³ Department of Epidemiology and Biostatistics, School of Public Health - Bloomington, Indiana University, Bloomington, IN, United States.
⁴ Musculoskeletal Analysis Laboratory, College of Health Sciences, University of Memphis, Memphis, TN, United States.

Abstract

Dynamical systems theory suggests that studying the complexity of biological signals could lead to a single gait metric that reliably predicts risk of running-related injury (RRI). The purposes of this pilot study were to examine center of mass (COM) acceleration complexity at baseline, prior to RRI, and the change between timepoints between collegiate runners who developed RRI during a competitive season and those who remained uninjured, and to determine if complexity at these timepoints was associated with increased odds of RRI. Twenty-two collegiate runners from the same cross-country team wore a waist-mounted triaxial accelerometer (100 Hz) during easy-intensity runs throughout the competitive season. RRIs requiring medical attention were reported via an online survey. Control entropy was used to estimate the complexity of the resultant COM acceleration recorded during each run. Associations between complexity and RRI were assessed using a frequency-matching strategy where uninjured participants were paired with injured participants using complexity from the most time-proximal run prior to RRI. Seven runners sustained an RRI. No significant differences were observed between injured and uninjured groups for baseline complexity (p = 0.364, d = 0.405), pre-injury complexity (p = 0.258, d = 0.581), or change from baseline to pre-injury (p = 0.101, d = 0.963). There were no statistically significant associations found between complexity and RRI risk. Although no significant associations were found, the median effect from the models indicated that an increase in baseline complexity, pre-injury complexity, and change in complexity from baseline each corresponded to an increased odds of sustaining an RRI [baseline: odds ratio (OR) = 1.560, 95% CI = 0.587-4.143, p = 0.372; pre-injury: OR = 1.926, 95% CI: 0.689-5.382, p = 0.211; change from baseline: OR = 1.119; 95% CI: 0.839-1.491, p = 0.445). Despite non-significance and wide confidence intervals that included both positive and negative associations, the point estimates for >98% of the 10,000 frequency-case-control-matched model fits indicated that matching strategy did not influence the directionality of the association estimates between complexity and RRI risk (i.e., odds ratio >1.0). This pilot study demonstrates initial feasibility that additional research may support COM acceleration complexity as a useful single-metric monitoring system for RRI risk during real-world training. Follow-up work should assess longitudinal associations between gait complexity and running-related injury in larger cohorts.

Keywords: accelerometry; collegiate running; control entropy; gait complexity; running injury; wearable technology in sports.