Currently 1 in 9 Arizonan has T2D. African-American, Hispanics, American Indians and Asian-Americans, that account for 40% of Arizona residents, are about twice as likely to have T2D as are Whites. In 2008, 9,883 hospitalizations in Arizona were due to diabetes that make diabetes one of most costly diseases. Treatment regime for T2D is rather complex. After failure of diet and lifestyle efforts, step-wise addition of glucose-lowering medications is the usual course of T2D therapy. The decision of prescribing subsequent medications in the best sequence after initiation of the generally agreed upon initial oral medication (metformin) is strikingly challenging due to the unclear advantages of 2, 3, 4 and 5 drug regimens and the increased potential for adverse effects. Right now, most T2D treatments guidance are designed for the average patient. But one size doesn't fit all, and treatments that are very successful for some patients don't work for others. In addition, there are also new uncertainties regarding the benefits of intensive glycemic control on macrovascular complications and the ideal target goals for therapy. Comparative effectiveness studies are the traditional tools to preform comparisons. However, it is impossible due to its complexity, cost and length of the study. In this proposal, we focus on developing a data-driven paradigm to understand the medication treatment heterogeneity effects of T2D and to provide an evidence-based treatment guidance that is tailored to subgroups of patients sharing similar characteristics (precision medicine). Our data-driven approach will be based on the study of Veteran Healthcare Database, using the VA Informatics and Computing Infrastructure (VINCI). The clinical data includes the longitudinal data profiles starting from the year 2000.