BACKGROUND AND OBJECTIVES: Indications for insulin pump therapy (IPT) in children with type 1 diabetes (T1D) are relatively non-specific and therefore subject to provider discretion. Health professionals' perceptions of which people will have difficulty with IPT, for example, those with higher hemoglobin A1c (HbA1c ), may not be correct. This study examined the effect of IPT on HbA1c , and the role of pre-pump HbA1c on this effect. METHODS: All children with T1D started on IPT at British Columbia Children's Hospital from January 2011 through June 2016 were included if they had HbA1c values available both before and after IPT (n = 125). Generalized estimating equations was used to estimate the effects of IPT on HbA1c , stratified by pre-pump HbA1c levels (good: <7.5% [<58 mmol/mol], moderate: 7.5%-9.0% [58-75 mmol/mol], poor: >9.0% [>75 mmol/mol]). RESULTS: After adjusting for potential confounders, mean HbA1c decreased by 0.48% [5.2 mmol/mol] (95% confidence interval: -0.64, -0.33% [-7.0, -3.6 mmol/mol]; P < 0.0001) after IPT initiation. The adjusted mean HbA1c decreased by 0.14% [1.5 mmol/mol] (-0.35, 0.07% [-3.8, 0.8 mmol/mol]; P = 0.188), 0.54% [5.9 mmol/mol] (-0.74, -0.34% [-8.1, -3.7 mmol/mol]; P < 0.0001), and 1.08% [11.8 mmol/mol] (-1.69, -0.46% [-18.5, -5.0 mmol/mol]; P = 0.0006) after pump initiation in the good, moderate, and poor pre-pump metabolic control groups, respectively. CONCLUSIONS: Pre-pump HbA1c appears to play a significant role in the effects of IPT on HbA1c , with the largest decrease in HbA1c seen in the poor pre-pump HbA1c group. Eligibility and consideration for IPT should be expanded to routinely include these children.