Objectives: Many studies have been proposed to evaluate the effectiveness of the diabetes mellitus medical benefit improvement project in Taiwan. Most of these evaluations, however, lack a control group and pay little attention to rural areas. To evaluate the long-term effectiveness of the Diabetes Shared Care Network on related metabolic indices for diabetic patients in a rural area, we compared repeat measurements from the intervention group and a non-intervention group. Methods: Subjects were identified retrospectively from a database of outpatients with diabetes mellitus in a regional hospital in Mid-Taiwan between December 2002 and July 2007. Analysis of covariance (ANCOVA) and mixed models were used to demonstrate differences and trends in the related metabolic indices of the two groups. Results: The difference in fasting blood sugar levels between the two groups was significant after 3 months of network intervention. After 9 months, hemoglobulin A1C (HbA1C), total cholesterol, and triglycerides also showed significant differences. Changes in high density lipoprotein cholesterol and low density lipoprotein cholesterol did not. Conclusions: The effectiveness of the Diabetes Shared Care Network in improving HbA1C, fasting blood sugar, total cholesterol, and triglyceride levels was significant. For most patients; however, the related metabolic indices did not achieve the target suggested by the American Diabetes Association. In conclusion, the Diabetes Shared Care Network was somewhat effective in improving metabolic indices and can still serve as a model for the longterm education and care of diabetic patients.