BACKGROUND: To prospectively investigate if the grand mean HbA1c and the differences in mean HbA1c between centers in Sweden could be reduced, thereby improving care delivered by pediatric diabetes teams.

METHODS: We used an 18-month quality improvement collaborative (QIC) together with the Swedish pediatric diabetes quality registry (SWEDIABKIDS). The first program (IQ-1), started in April 2011 and the second (IQ-2) in April 2012; together they encompassed 70% of Swedish children and adolescents with diabetes.

RESULTS: The proportion of patients in IQ-1 with a mean HbA1c <7.4% (57 mmol/mol) increased from 26.4% before start to 35.9% at 36 months (P < .001), and from 30.2% to 37.2% (P < .001) for IQ-2. Mean HbA1c decreased in both participating and non-participating (NP) centers in Sweden, thereby indicating an improvement by a spatial spill over effect in NP centers. The grand mean HbA1c decreased by 0.45% (4.9 mmol/mol) during 36 months; at the end of 2014 it was 7.43% (57.7 mmol/mol) (P < .001). A linear regression model with the difference in HbA1c before start and second follow-up as dependent variable showed that QIC participation significantly decreased mean HbA1c both for IQ-1 and IQ-2. The proportion of patients with high HbA1c values (>8.7%, 72 mmol/mol) decreased significantly in both QICs, while it increased in the NP group.

CONCLUSIONS: The grand mean HbA1c has decreased significantly in Sweden from 2010 to 2014, and QICs have contributed significantly to this decrease. There seems to be a spatial spill-over effect in NP centers.