Survival Model for Foot and Leg High Rate Axial Impact Injury Data

Objectives: Understanding how lower extremity injuries from automotive intrusion and underbody blast (UBB) differ is of key importance when determining whether automotive injury criteria can be applied to blast rate scenarios. This article provides a review of existing injury risk analyses and outlines an approach to improve injury prediction for an expanded range of loading rates. This analysis will address issues with existing injury risk functions including inaccuracies due to inertial and potential viscous resistance at higher loading rates.Methods: This survival analysis attempts to minimize these errors by considering injury location statistics and a predictor variable selection process dependent upon failure mechanisms of bone. Distribution of foot/ankle/leg injuries induced by axial impact loading at rates characteristic of UBB as well as automotive intrusion was studied and calcaneus injuries were found to be the most common injury; thus, footplate force was chosen as the main predictor variable because of its proximity to injury location to prevent inaccuracies associated with inertial differences due to loading rate. A survival analysis was then performed with age, sex, dorsiflexion angle, and mass as covariates. This statistical analysis uses data from previous axial postmortem human surrogate (PMHS) component leg tests to provide perspectives on how proximal boundary conditions and loading rate affect injury probability in the foot/ankle/leg (n = 82).Results: Tibia force-at-fracture proved to be up to 20% inaccurate in previous analyses because of viscous resistance and inertial effects within the data set used, suggesting that previous injury criteria are accurate only for specific rates of loading and boundary conditions. The statistical model presented in this article predicts 50% probability of injury for a plantar force of 10.2 kN for a 50th percentile male with a neutral ankle position. Force rate was found to be an insignificant covariate because of the limited range of loading rate differences within the data set; however, compensation for inertial effects caused by measuring the force-at-fracture in a location closer to expected injury location improved the model's predictive capabilities for the entire data set.Conclusions: This study provides better injury prediction capabilities for both automotive and blast rates because of reduced sensitivity to inertial effects and tibia-fibula load sharing. Further, a framework is provided for future injury criteria generation for high rate loading scenarios. This analysis also suggests key improvements to be made to existing anthropomorphic test device (ATD) lower extremities to provide accurate injury prediction for high rate applications such as UBB.