We used descriptive statistics to evaluate trends in the utilization of each therapy of interest. We focused on eight medication classes depicted in Table 1: sulfonylureas (glucotrol XL), biguanides (metformin HCl), glitazones (pioglitazone), DPP-4 inhibitors (sitagliptin), incretins (liraglutide), meglitinides (repaglinide), α-glucosidase inhibitors (acarbose), insulins (insulin glargine), and amylin analogs (pramlintide). We classified drugs within these therapeutic groups based on their chemical composition, using the IMS Health Universal System of Classification (USC) codes. We counted fixed-dose combination products as contributing to each of their constituent classes when computing total compounds; for example, a combination product such as Janumet (sitagliptin and metformin) was counted as contributing once to biguanides and once to DPP-4 inhibitors. Thus, our analysis of trends in DPP-4 use includes visits where they were used as fixed-dose combination products containing DPP-4 inhibitors as well as where they were used as an individual product. We also calculated therapeutic intensity, which we assessed by dividing total number of compounds used by the total number of unique treatment visits in a given year. For these calculations, a visit with a fixed-dose combination would be considered as similarly intense as a treatment where two separate products were used, since both would similarly reflect the use of two compounds during a single visit. For estimates from the NDTI, we calculated 95% CIs using tables of relative standard errors that are derived accounting for the survey’s complex sampling design. Because the NPA is based on such a large sample of pharmacies, the uncertainty surrounding estimates of national prescription expenditures is small.