Data-reduction method and surface effects on accelerometer-based estimates of cumulative damage

Acceleration measures may be collected for entire training sessions in real-world conditions using wearable sensors. The resulting data must be reduced to provide meaningful metrics. The aims of this study were to: (1) quantify the influence of binning methods on a cumulative damage metric based on tibial acceleration, and (2) determine the sensitivity of the damage metric to different running surface conditions. Acceleration peaks collected over a 2.4 km run over variable surfaces were binned and used to calculate a damage metric. A sensitivity analysis determined that a minimum of 7 bins were required to calculate a damage metric within 5% of the reference value. Peak acceleration and cumulative damage were highest when running on grass, and these counter-intuitive results may reflect limitations of using tibial mounted accelerometers as surrogate measures of tibial load.