Computational models of flow in patient-derived arterial geometries have become a key paradigm of biomedical research. These fluid models are often challenging to visualize due to high spatial heterogeneity and visual complexity. Virtual immersive environments can offer advantageous visualization of spatially heterogeneous and complex systems. However, as different VR devices offer varying levels of immersion, there remains a crucial lack of understanding regarding what level of immersion is best suited for interactions with patient-specific flow models. We conducted a quantitative user evaluation with multiple VR devices testing an important use of hemodynamic simulations-analysis of surface parameters within complex patient-specific geometries. This task was compared for the semi-immersive zSpace 3D monitor and the fully immersive HTC Vive system. The semi-immersive device was more accurate than the fully immersive device. The two devices showed similar results for task duration and performance (accuracy/duration). The accuracy of the semi-immersive device was also higher for arterial geometries of greater complexity and branching. This assessment demonstrates that the level of immersion plays a significant role in the accuracy of assessing arterial flow models. We found that the semi-immersive VR device was a generally optimal choice for arterial visualization.