AIMS: To test whether concomitant use of an automated bolus calculator for people with Type 1 diabetes carrying out advanced carbohydrate counting would induce further improvements in metabolic control.

METHODS: We conducted a 12-month, randomized, parallel-group, open-label, single-centre, investigator-initiated clinical study. We enrolled advanced carbohydrate counting-naïve adults with Type 1 diabetes and HbA1c levels 64-100 mmol/mol (8.0-11.3%), who were receiving multiple daily insulin injection therapy. In a 1:1-ratio, participants were randomized to receive training in either advanced carbohydrate counting using mental calculations (MC group) or advanced carbohydrate counting using an automated bolus calculator (ABC group) during a 3.5-h group training course. For 12 months after training, participants attended a specialized diabetes centre quarterly. The primary outcome was change in HbA1c from baseline to 12 months.

RESULTS: Between August 2012 and September 2013, 168 participants (96 men and 72 women) were recruited and randomly assigned to the MC group (n = 84) and the ABC group (n = 84). Drop-out rates were 23.8 and 21.4%, respectively (P = 0.712); 130 participants completed the study. The baseline HbA1c was 75 ± 9 mmol/mol (9.0 ± 0.8%) in the MC group and 74 ± 8 mmol/mol (8.9 ± 0.7%) in the ABC group. At 12 months, change in HbA1c was significant within both groups: MC group: -2 mmol/mol (95% CI -4 to -1) or -0.2% (95% CI -0.4 to -0.1; P = 0.017) and ABC group: -5 mmol/mol (95% CI -6 to -3) or -0.5% (95% CI -0.6 to -0.3; P < 0.0001), but HbA1c reductions were significantly greater in the ABC group (P = 0.033). No episodes of severe hypoglycaemia were reported.

CONCLUSIONS: People with Type 1 diabetes initiating advanced carbohydrate counting obtained significantly greater HbA1c reductions when guided by an automated bolus calculator (NCT02084498).