Purpose: Compare the effectiveness of two educational teaching methods for diabetic patients. Design: Quasi-experimental study comparing two interventions using a pretest/post-test design. Setting: Three clinics within a western U.S. regional health system. Subjects: 818 adult diabetic participants (60.5 mean age, 52% female) attended one to four sessions between 2013-2017, and had A1c tests within 180 days of first attended session and 30 to 365 days after last attended session. Intervention: A group-based, highly interactive learning experience (n = 561) and a traditional, lecture-style class (n = 257). Measures: Pre and post measures of A1c. Analysis: Paired t-tests measured change within each group pre-post intervention. Two-sample t-tests measured mean change pre-post intervention between the two groups. Multivariable linear regression measured mean change in A1c between groups, adjusted for pre-test scores and controlling for demographic variables. Results: Both interactive and traditional teaching interventions were effective at significantly reducing patient A1c levels by 1.3 (p < 0.001) and 1.0 (p < 0.001) points respectively. The between groups difference in A1c was not significant, t(512) = 1.66, p = 0.0985, but when controlling for age, pre-A1c and days post-A1c, the interactive intervention was significantly (p < 0.05) more effective reducing patient A1c levels by 0.19 points than the traditional intervention. Conclusion: Group-based, interactive diabetes self-management education programs may be an effective model for reducing patient A1c levels.