To assess the effects of cognitive behavioural group training (CBGT) on glycaemic control, diabetes self-efficacy and well-being in Type 1 diabetes patients in persistent poor glycaemic control. In a randomized controlled trial, a total of 107 patients with Type 1 diabetes in poor glycaemic control (HbA(1c) > or = 8%) were assigned to a 6-week CBGT or blood glucose awareness training (BGAT) as control condition. The intervention was preceded by a 3-month run-in period. Glycaemic control (HbA(1c)), diabetes-specific self-efficacy (CIDS), diabetes-related distress (PAID) and depressive symptoms (CES-D), were assessed at baseline (T1), directly before (T2) and 3 months after (T3) the intervention. No significant changes in HbA(1c) were found after CBGT, whilst diabetes self-efficacy increased (mean CIDS score 71.6 +/- 14.0 to 74.3 +/- 12.2) and diabetes-related distress (mean PAID score 47.0 +/- 21.6 to 42.6 +/- 20.8) and depressive symptoms decreased (mean CES-D score 16.9 +/- 12.8 to 13.5 +/- 12.6). Changes in psychological outcomes were similar for both treatment groups. Diabetes self-care behaviours improved equally. Drop-out rate, which was higher among CBGT participants, was relatively low overall (total n = 15, 17.05%), and both interventions were well-appreciated by the participants. CBGT was successful in improving self-efficacy, diabetes-related distress and mood at 3 months' follow-up, but not in improving glycaemic control.