OBJECTIVES: To study the association between rates of sulfonylurea (SU) use at a physician’s practice and average health care costs and complication rates among the physician’s patients with type 2 diabetes mellitus (T2DM). METHODS: We performed a retrospective group-level analysis on a sample of 7,905 patients (214,230 patient-months) insured by Humana, aged 18-64 with an incident T2DM diagnosis between 2007 and 2011. We regressed physician-level monthly complication rates (cardiovascular, lower extremity, ophthalmic, renal, neuropathy, and hypoglycemia) and average health care costs on 3-month-lagged rates of SU use controlling for use of 10 other T2DM therapy classes in each physician’s practice, and patient and practice characteristics. Costs were estimated using a generalized linear model with log link to account for common zero costs. RESULTS: SU use was associated with increased rates of lower extremity, ophthalmic, and renal complications relative to no drug use (p<0.05).  For each complication class, we identified the best performing therapy and estimated the effect of a hypothetical switch in prescribing patterns from 100% use of this class to 100% use of SU, as percentages of the respective average complication rates.  Our estimates were: 93% increase in cardiovascular when switching from meglitinides, 133% increase in lower extremity when switching from alpha glucosidase inhibitors (AGI), 433% increase in ophthalmic when switching from amylinomimetics, 47% increase in renal when switching from biguanides, 147% increase in neuropathy when switching from AGIs and 252% increase in hypoglycemic complications when switching from amylinomimetics.  Despite SU’s association with higher complication rates, the association between average costs and rate of SU use was not significant. Other therapy classes were associated with increased costs with the exception of premixed insulin, meglitinides and amylinomimetics (no significant association) and biguanides and alpha glucosidase inhibitors (negative association). CONCLUSIONS: Use of SU could potentially increase complications in type 2 diabetes.