A randomized controlled trial was conducted to determine whether an education program specifically designed for patients with non-insulin-dependent diabetes and limited literacy could improve and sustain glucose and weight control. From a referral clinic, 120 obese (greater than 130 per cent of ideal body weight) diabetic patients who were not taking insulin were recruited. Of these, 55 per cent were female and 49 per cent were black; the mean age was 53 years. Mean glycosylated hemoglobin (HbA1%) was 10.2 per cent. Each subject was assigned to one of three groups: monthly group sessions with videotapes for diabetic persons with low literacy skills; monthly group sessions without videotapes; or no monthly sessions. After seven months, there had been 16 dropouts (13 per cent). Differences in weight changes between groups were significant (p less than 0.05); group 1 lost a median of 1 kg of weight (p less than 0.05) compared with a 0.1-kg loss and no change in groups 2 and 3, respectively. This weight loss was not sustained at 11 months. There was no significant change in HbA1%. Age, education, and compliance beliefs did not predict outcome. The authors conclude that the patient education programs did not result in sustained glucose or weight control.