Females are more likely to suffer severe to fatal injuries than males in automotive crashes, and injury risk for females is especially pronounced in the lower extremity;​ females sustain both a higher number and higher severity of lower extremity injuries than males for comparable crash conditions. In the automotive safety field, male injury risk is generally predicted with measured male data, while female injury risk is typically predicted by applying dimensional analysis scaling to measured male data. Scale factors to predict female injury are predominantly derived from gross anthropometry measurement ratios between males and females, and scaling is dependent on the assumption of geometric similitude across the sexes. However, this assumption has never been well-examined or validated. Therefore, the goal of this dissertation is to evaluate the effects of geometric differences between the sexes on injury tolerance in the lower extremity, and the relative contributions of gross and local geometry to injury.

Methods to accomplish this goal include identifying current knowledge of sex differences and how they affect injury risk through literature review, performing a field data epidemiology study to determine which injuries or injury mechanisms differ between the sexes in real-world automotive crashes, measuring differences in injury tolerance by generating female cadaveric injury data, quantifying local bony geometry sex differences in joint morphology, creating subject-specific male and female finite element (FE) models, and using developed finite element models to evaluate the effects of local and gross geometry on injury prediction metrics. Results demonstrate that sex differences have been established in geometry (bone shape, skeletal alignment, external body shape) throughout the body;​ however, little is known about sex differences in several lower extremity joints. Furthermore, ankle fractures were shown to be significantly more prevalent for females in real-world crash data independent of occupant age and crash severity. Therefore, ankle injury tests were performed with female cadaver lower extremities, and results demonstrated that scaling with gross or local geometry measures was unable to predict measured failure tolerance. Rather, individual ankle joint variation seemed to dictate the injury response. Population ankle joint morphology was classified across sex and age, and the geometries of the subtalar and talocrural joints were not found to differ significantly with age. Sex differences that did not correlate to gross size were not found in clinical ankle measurements, but statistically significant local geometry sex differences were found in principal components of both the subtalar and talocrural joints. Using the sexually dimorphic principal components identified in the ankle morphology results, three male and three female foot and ankle geometries were chosen to create FE models representative of male and female local ankle joint geometries. Sensitivity studies of geometry effects of these models on injury prediction metrics suggested that the effect of local sex-related geometric variables on injury prediction metrics was not greater than the effect of inter-individual local geometric variables (regardless of sex), and therefore, it is unlikely that sex-related local ankle geometry differences have a large contribution to the sex discrepancy in ankle injury risk due to inversion/eversion loading in automotive crashes. Gross geometry was also shown to have a lesser contribution to variation in injury prediction metrics in inversion/eversion loading of the ankle when compared to the effect of inter-individual local geometry differences.

Contributions of this dissertation include:​ a series of literature reviews which serve as a reference manual of how sex differences can affect injury risk in the automotive safety field, the creation of the largest component-level female cadaveric injury dataset (in terms of number of individual specimens tested) currently in existence for any region of the body, a quantitative description of ankle joint morphology with sex and age, and a framework for classifying biological group variation versus individual variation for geometric measures. Additionally, sex-specific local ankle geometry was shown to have an equivalent effect on injury prediction metrics as inter-individual local geometry, and thus it is unlikely that the sex differences in local ankle geometry greatly contribute to ankle injury risk discrepancy. Most significantly, this dissertation was able to demonstrate both experimentally and computationally that the predominant methodology used to quantify injury tolerance for half of the automotive population (females) is unable to predict measured female injury tolerance for some anatomical structures and loading modes. This contribution casts doubt on the applicability of using scaling across other body regions and loading modes, and demonstrates the need for more measured female injury data to quantify relationships between sex and injury risk, which are necessary to lessen sex discrepancies in automotive injury and fatality risk.