Finite element (FE) computational human body models (HBMs) have gained popularity over the past several decades as human surrogates for use in blunt injury research. FE HBMs are critical for the analysis of local injury mechanisms. These metrics are challenging to measure experimentally and demonstrate an important advantage of HBMs. The objective of this study is to evaluate the injury risk predictive power of localized metrics to predict the risk of pelvic fracture in a FE HBM.
The Global Human Body Models Consortium (GHBMC) 50th percentile detailed male model (v4.3) was used for this study. Cross-sectional and cortical bone surface instrumentation was implemented in the GHBMC pelvis. Lateral impact FE simulations were performed using input data from tests performed on post mortem human subjects (PMHS). Predictive power of the FE force and strain outputs on localized fracture risk was evaluated using the receiver operator characteristic (ROC) curve analysis.
The ROC curve analysis showed moderate predictive power for the superior pubic ramus and sacrum. Additionally, cross-sectional force was compared to a range of percentile outputs of maximum principal, minimum principal, and effective cortical element strains. From this analysis it was determined that cross-sectional force was the best predictor of localized pelvic fracture.