Predicting and improving oxygen transport within bioartificial liver (BAL) devices continues to be an important engineering challenge since oxygen is one of the critical nutrients necessary for maintaining hepatocyte viability and function. Such a computational model would not only help predict outcomes but it would also allow system modifications to be analyzed prior to developing experimental protocols. This would help to facilitate future design improvements while reducing both experimental time and capital resource costs, and is the focus of the current study. Specifically, a computational model of O(2) transport through collagen and microporous collagen ECMs is analyzed for hollow fiber (HF), flat plate (FP), and spheroid BAL designs. By modifying the O(2) boundary conditions, hepatocyte O(2) consumption levels, O(2) permeability of the ECM, and ECM void fractions, O(2) transport predictions are determined for each system as a function of time and distance. Accuracy of the predictive model is confirmed by comparing computational vs. experimental results for the HF BAL system. The model's results indicate that O(2) transport within all three BAL designs can be improved significantly by incorporating the enhancement technique. This technique modifies a diffusion-dominant gel ECM into a porous matrix with diffusive and convective flows that mutually transport O(2) through the ECMs. Although tortuous pathways increase the porous ECM's overall effective length of O(2) travel, the decreased transport resistances of these pathways allow O(2) to permeate more effectively into the ECMs. Furthermore, because the HF design employs convective flow on both its inner and outer ECM surfaces, greater control of O(2) transport through its ECM is predicted, as compared with the single O(2) source inputs of the flat plate and spheroid systems. The importance of this control is evaluated by showing how modifying the O(2) concentration and/or transfer coefficients of the convective flows can affect O(2) transport.