Background: An electronic decision support tool called Diabetes Wizard (DM Wizard) was implemented in four HealthPartners Clinics from October 2006 through May 2007 as part of the NIDDK funded study entitled ‘Reducing Clinical Inertia in Diabetes Care.’ The DM Wizard was integrated into the electronic health record (EHR) as an EPIC smart form and accessed through the visit navigator screen at a patient’s encounter. It was also printed for provider review prior to entering the exam room. It gave clinically relevant data on the patient’s A1c, blood pressure, lipids, and relevant comorbidities. In addition, it was programmed using complex algorithms derived through expert consensus to list 3-4 specific recommendations for drug intensification along with safety alerts and reminders for each of the three clinical domains. Methods: Sixty-one providers were randomized at the clinic level to one of three study arms: (1) Control (2) SimCare Plus (a simulated case-based learning program for providers) or (3) DM Wizard. Six clinics received staff training and, for 6 months, received feedback, utilization reminders and paid incentives for DM Wizard use. Providers also completed a ‘visit action resolution’ form (a checklist of reasons for clinical inertia). A provider survey assessed satisfaction. The following were tracked electronically during the 6-month intervention and incentive period, and for 3 months after: (1) rate of opening the DM Wizard at eligible encounters (all visits of patients with diabetes) (2) rate of nurse printing the DM Wizard for the provider (3) rate of provider completion of the visit resolution form. Results: One hundred percent of providers agreed with listed treatment suggestions and would recommend use of the DM Wizard to other physicians. The form printed by the nurse and available prior to entering the exam room was seen as valuable by most providers. The DM Wizard was opened during eligible diabetes encounters 75% of the time, with printing and visit resolution completion rates of 88% and 87% of opened forms, respectively. Three months after the incentive period, rates of opening the DM Wizard dropped to 50%. Conclusions: The DM Wizard EHR tool was implemented successfully with high provider use rates and satisfaction. Use declined substantially after the study intervention period correlating to discontinuation of feedback, reminders, and incentives. The cost of maintaining high rates of DM Wizard utilization and keeping algorithms up to date will need to be weighed against the clinical benefits achieved.