To examine the two-year impact of Stepping Up, a general practice based model of care intervention for insulin initiation and titration in Australia.266 participants from 74 general practices participated in the Stepping Up cluster randomised controlled trial between 2012-2014. Control practices received training in the model of care on completion of the 12-month trial. Patients were followed for 24 months. Participant baseline characteristics, insulin and non-insulin medication use were summarised for each study group. Linear mixed-effects models with random intercepts were used to estimate differences in mean outcome (HbA1c and weight) between the study groups using restricted maximum likelihood estimation.At baseline 61% of patients were male, mean (SD) age 62 (10) years, diabetes duration 9 (5, 13) years and mean (95% CI) HbA1c was 8.9 (8.8-9.1)% (74 (73-76)mmol/mol) for both groups. There was a significant between group difference at 6 months which was sustained at 24 months; Mean (95% CI) HbA1c at 24 months in the intervention group was 7.6 (7.5-7.8)% (60 (58-62)mmol/mol) and 8.0 (7.7-8.4)% (64 (61-68)mmol/mol) in the control group. At 24 months 97 (71.3%) of the intervention group and 26 (31.0%) of the control group were prescribed insulin; there was no significant difference in weight. Use of non-insulin anti-hyperglycaemic agents was similar in both groups with the exception of dipeptidyl peptidase-4 inhibitors which were prescribed more frequently in the control group (30(36%) vs 21(16%)).Stepping Up was associated with improved glycaemic control compared to usual care for 24 months, suggesting that the model facilitated more timely treatment intensification. Ongoing RN-CDE support may be needed to facilitate ongoing treatment intensification.