We conducted a randomized controlled trial using mobile health technology in an ethnically diverse sample of 137 patients with complicated diabetes. Patients in the intervention group (n = 72) were trained to measure their blood glucose with a sensor which transmitted the readings to a mobile phone via a Bluetooth wireless link. Clinicians were then able to examine and respond to the readings which were viewed with a web-based application. Patients in the control arm of the study (n = 65) did not transmit their readings and received care with their usual doctor in the outpatient and/or primary care setting. The mean follow-up period was 9 months in each group. The default rate was higher in the patients in the intervention arm due to technical problems. In an intention-to-treat analysis there were no differences in HbA(1c) between the intervention and control groups. In a sub-group analysis of the patients who completed the study, the telemonitoring group had a lower HbA(1c) than those in the control group: 7.76% and 8.40%, respectively (P = 0.06).