Field data studies have identified various factors influencing injury risk in frontal crashes. However, understanding these complex interactions solely from real-world data is challenging due to limited availability. This study investigated whether detailed human body models (HBMs) can replicate trends observed in field data through stochastic simulations. Specifically, it compared injury risk factors across four body regions between HBM simulations and field data, and between two HBM line-ups, THUMS and VIVA+. A simulation matrix was created based on parametric distributions from literature and crash data analysis. With all four HBMs, the same 200 simulations were performed. Results showed underprediction of lower extremity injuries and overprediction of brain injuries by all models except the small female THUMS. For rib fractures, the VIVA+ line-up overpredicted, while the THUMS line-up underpredicted. A linear correlation indicated that the main cause of risk differences was the strain response and its interpretation. Notably, VIVA+ HBMs exhibited age- and delta-v dependent differences between average female and male, with females having a higher rib fracture risk that increases with age. While this represents a significant finding, the study highlights the need for further work to harmonise injury risk assessments across different HBMs.