Type 1 diabetes (T1D) is a disease characterized by the loss of glycemic control following the immune-mediated destruction of the insulin-secreting β-cells, which reside in pancreatic cell clusters known as islets. Presently, insulin administration is the primary method of treatment for T1D. However, secondary complications and the requirement for constant attention render insulin therapy imperfect. The intraportal infusion of allogeneic islets, or more recently stem cell-derived β-cell clusters, have demonstrated the potential for more robust glycemic control. Nevertheless, intraportal islet therapy is constrained in applicability because it requires the co-administration of immunosuppressive drugs. Encapsulating insulin-secreting cells in a hydrogel-based system which prevents immune interference but allows free transport of nutrients and therapeutics promises to overcome these limitations and would constitute a bioartificial pancreas. Unfortunately, encapsulation demands that the cells are physically separated from the host circulation, introducing mass transport challenges. This thesis aims first to quantify those constraints, especially that of oxygen---what is likely the limiting nutrient---and then evaluates three strategies proposed to overcome them. First, a computational platform is developed to evaluate bioartificial pancreas devices, accounting for their intrinsic stochastic properties. We find that device potency is strikingly sensitive to variance in the size distributions of the encapsulated cell clusters. We also determine that large cell payloads and devices are required to provide clinical benefit. A novel accounting of the therapeutic cell mass, adjusted for oxygen-limited potency, is then proposed. Subsequently, an approach which enables the maintenance of a continuous gas phase within hydrogels is described, which increases the oxygen permeability in such systems significantly. The second approach proposes a system by which the cellular waste product carbon dioxide is recycled into oxygen in situ through an "inverse-breathing" chemical reaction in a physically separated encapsulation system. The final chapter explores the benefit of vascularizing the host site before device implantation. Characterizations using electron paramagnetic resonance oxygen imaging confirm model predictions that all three systems elevate graft oxygenation. Studies in mice demonstrate that all three strategies improve diabetes correction in comparison to suitable controls. Cumulatively, this work contributes to overcoming mass transport limitations in encapsulated cell therapy for T1D.