OBJECTIVES: To determine the factors that influence HbA1c improvement amongst newly diagnosed patients with poor glycaemic control (HbA1c >8%). METHODS: The sample includes all newly diagnosed diabetic patients, with poor glycaemic control, seen between January and December 2006 at three restructured hospitals. Patients were followed up for a 12-month period. Rate of HbA1c improvement was plotted using the Kaplan-Meier approach, where an “improvement” was defined as achieving at least 0.5% absolute reduction from baseline. Cox proportional hazards regression model was used to assess differences in HbA1c improvement by age, gender, baseline HbA1c and serum creatinine levels. RESULTS: There were 505 newly diagnosed diabetic patients. Baseline HbA1c ranged from 8.1% to 17.1% with a median of 9.4%. An “improvement” in HbA1c control was achieved by 76% of the cohort within 12 months. The median time to achieve the HbA1c “improvement” was 126 days (95%CI: 112-138) with the lower and upper quartiles being 82 days (95%CI: 77-84) and 227 days (95%CI: 210-252) respectively. Age, gender and baseline serum creatinine levels were not significantly associated with HbA1c improvement. Patients with baseline HbA1c 9.1% to 12.0% and those greater than 12.0% were 1.4 and 1.7 times more likely than those with baseline HbA1c 8.1% to 9.0%, to improve their HbA1c, respectively (p <0.01). CONCLUSIONS:The glycaemic control of three quarters of newly diagnosed diabetic patients improved within 12 months, and those with poorer baseline glycaemic control were more likely to achieve an “improvement”. Physicians treating patients with poorer glycaemic control can aim for higher improvement goals within the same time frame.