Forty-eight families with children less than 13 years old attending a paediatric diabetic clinic volunteered for a 2-year randomized crossover trial to determine whether an informal education programme (diabetic club) could improve diabetic control. Group A attended the diabetic club for 10 afternoons of informal education in the first year, while Group B continued at the routine clinic (5 visits per year). For the second year Group A returned to the clinic, Group B attended the club. Glycosylated haemoglobin (HbA1) remained stable while attending the club but rose significantly (p less than 0.01) while attending the clinic in both groups (HbA1 at baseline, 1 year, and 2 years: Group A, 9.6 (SD 1.2), 9.6(1.4), 10.7(2.1)%; Group B 8.9(1.3), 10.4(1.4), 10.5(1.4)% (normal reference range 4.7-7.9%)). Other indices of control were unchanged. Diabetic problem-solving scores of parents improved (p less than 0.01) but their knowledge of diabetes did not correlate with their child's HbA1. Dietary intake showed a reduction in percentage of energy taken as fat (40% vs 37.7%, p less than 0.05) during club attendance. The percentage of parents reporting helpful social contact between families increased during their club year (Group A 50 to 78%, Group B 32 to 57%, p less than 0.001). Psychological measurements remained unchanged. An education programme for diabetic children may stabilize diabetic control in the short term but this effect is not sustained. The main benefit was the support provided by increased social contact with families of other diabetic children within the informal framework of the diabetic club.