BACKGROUND: Diabetes self-management education has an important role in diabetes management. The efficacy of education has been proven in several randomized trials. However, the status of diabetes education programs in real Korean clinical practice has not yet been evaluated in terms of patient compliance with the education prescription. METHODS: We retrospectively analyzed clinical and laboratory data from all patients who were ordered to undergo diabetes education during 2009 at Samsung Medical Center, Seoul, Korea (n=2,291). After excluding ineligible subjects, 588 patients were included in the analysis. RESULTS: Among the 588 patients, 433 received education. The overall compliance rate was 73.6%, which was significantly higher in the subjects with a short duration or living in a rural area compared to those with a long duration (85.0% vs. 65.1%, respectively; P<0.001) or living in an urban area (78.2% vs. 70.4%, respectively; P=0.037). The hemoglobin A1c decreased greater in the compliant group (from 7.84+/-1.54 at baseline to 6.79+/-1.06 at 3 months and 6.97+/-1.20 at 12 months after prescription in the compliant group vs. from 7.74+/-1.25 to 7.14+/-1.02 and 7.24+/-1.24 in the non-compliant group; P=0.001). The decrease in hemoglobin A1c was greater in the subjects with a short duration (P=0.032). CONCLUSION: In our study a large percent of patients refuse to get education despite having a prescription from their physician. This refusal rate was higher in the patients with long-standing diabetes or in urban residence. Furthermore, education was more effective in patients with a short duration of diabetes in clinical practice.