The authors conducted a diabetes quality improvement project in 5 privately owned primary care practices serving at least 25% minority patients. Interventions included group-specific and practice-specific training on an electronic patient registry, cultural competency practices and tools, and selected quality improvement strategies. The authors conducted a comprehensive evaluation involving quantitative and qualitative data to assess project impact. Although overall clinical performance did not improve over the 14- to 20-month project time frame, other practice structural characteristics and processes did show improvement: successful implementation of the registry and clinician reminders in all practices, institution of team care and patient reminders in 4 practices, and collection of patient race/ethnicity data in 4 practices. These results highlight the difficulty of bringing about clinical improvement in this subset of practices and also the importance of conducting comprehensive evaluations to fully understand and interpret multicomponent quality improvement projects.