Transplantation of encapsulated insulin-secreting cell clusters represents a potential cure for type I diabetes, but development efforts so far have yet to live up to its promise. Recent studies have elucidated the importance of cluster size on the insulin response in a manner that affects viability and efficacy of transplanted clusters. Using microfabrication techniques, a method to fabricate uniformly-sized insulin-secreting cell clusters was developed and thoroughly characterized using water contact angle, x-ray photoelectron spectroscopy, Fourier transform infrared spectroscopy, an d fluorescent microscopy. This technique enabled the formation of both monolayered and multilayered cell clusters of a predetermined size and shape. Subsequent evaluation of the impact that cluster size has on insulin expression, content and secretion using RT-PCR, ELISA, and confocal quantitative immunocytochemistry suggested that two cluster-size dependent behavioral changes relevant to transplant efficacy exist: First, glucose stimulation causes increased insulin production for clusters exceeding 40 μm in size. Second, cluster sizes greater than 60 μm secrete insulin more efficiently after production than smaller sized clusters. These results suggest that an optimal cluster size exists between 100-120μm. Lastly, human embryonic stem cells were differentiated in patterned 120 μm clusters along the pancreatic lineage, an effort that could produce optimally sized insulin-secreting cell clusters from a renewable cell supply. The studies presented here may help overcome two remaining challenges preventing encapsulated cell transplantation therapy from truly providing a cure for type I diabetes.