Introduction: To support the adoption of guideline concordant care by primary care practices, the New York Diabetes Coalition (NYDC) promoted use of an electronic diabetes registry and developed an interactive educational module on using the registry and improving patient communication. The NYDC hypothesized that use of a registry with immediate feedback would achieve measurable and clinically meaningful improvement in the proportion of patients at goal for diabetes health metrics. Research Design and Methods: In 2006-2007, the NYDC recruited 7 small to midsized primary care practices to implement the registry and to receive education and coaching on registry use, practice work flow, and patient engagement. The patient cohort included those with 2 or more visits with a diagnosis of diabetes within a 12-month period. Each patient's health measure status (at goal, above goal, not recorded) was assessed quarterly for hemoglobin A1C, low-density lipoprotein (LDL), and blood pressure (BP), and most recent A1C value was noted. A cohort analysis was performed using random effects regression models to assess the impact of the registry over time for each diabetes health metric. Results: After controlling for variability between sites, with each subsequent quarter during the registry period patients were 1.4 times more likely to have A1C less than/equal to 9, almost twice (OR = 1.8) as likely to have LDL less than 100, and 1.3 times more likely to have BP less than 140/90. These improvements in compliance were statistically significant. Average A1C also improved over time, though this did not reach statistical significance. Discussion: Utilizing a Web-based registry and interactive education, the project demonstrated improved patient outcomes, as well as the feasibility of collecting aggregate data from unrelated, independent practices. (Contains 3 tables and 3 figures.)