Despite the importance of self-management support (SMS), few studies have compared SMS interventions, involved diverse populations, or entailed implementation in safety net settings. We examined the effects of two SMS strategies across outcomes corresponding to the Chronic Care Model. A total of 339 outpatients with poorly controlled diabetes from county-run clinics were enrolled in a three-arm trial. Participants, more than half of whom spoke limited English, were uninsured, and/or had less than a high school education, were randomly assigned to usual care, interactive weekly automated telephone self-management support with nurse follow-up (ATSM), or monthly group medical visits with physician and health educator facilitation (GMV). We measured 1-year changes in structure (Patient Assessment of Chronic Illness Care [PACIC]), communication processes (Interpersonal Processes of Care [IPC]), and outcomes (behavioral, functional, and metabolic). Compared with the usual care group, the ATSM and GMV groups showed improvements in PACIC, with effect sizes of 0.48 and 0.50, respectively (P < 0.01). Only the ATSM group showed improvements in IPC (effect sizes 0.40 vs. usual care and 0.25 vs. GMV, P < 0.05). Both SMS arms showed improvements in self-management behavior versus the usual care arm (P < 0.05), with gains being greater for the ATSM group than for the GMV group (effect size 0.27, P = 0.02). The ATSM group had fewer bed days per month than the usual care group (-1.7 days, P = 0.05) and the GMV group (-2.3 days, P < 0.01) and less interference with daily activities than the usual care group (odds ratio 0.37, P = 0.02). We observed no differences in A1C change. Patient-centered SMS improves certain aspects of diabetes care and positively influences self-management behavior. ATSM seems to be a more effective communication vehicle than GMV in improving behavior and quality of life.