Impact of a comprehensive guideline dissemination strategy on diabetes diagnostic test rates: an interrupted time series

Background — Diabetes Canada launched a comprehensive Dissemination and Implementation (D&I) strategy to optimize uptake of their 2013 Clinical Practice Guidelines; the strategy involved continuing professional development courses, webinars, an interactive website, applications for mobile devices, point-of-care decision support tools, and media awareness campaigns. It included a focus on promoting HbA1c as the recommended diagnostic test for diabetes.

Objective — To determine the impact of Diabetes Canada’s 2013 D&I strategy on physician test-ordering behavior, specifically HbA1c testing, for the diagnosis of diabetes, using provincial healthcare administrative data.

Design — Population-based interrupted time series.

Setting — Ontario, Canada.

Participants — Ontario residents aged 40–79 not previously diagnosed with diabetes.

Measurements — For each quarter between January 2005 and December 2014, we conducted an interrupted time series analysis on the first-order difference of the proportion of patients receiving HbA1c tests per quarter with an autoregressive integrated moving average model with the intervention step occurring in quarter 2 of 2013. Subgroup analyses by rurality, physician graduation year, and practice size were also conducted.

Results — There were 32 quarters pre-intervention and 6 post-intervention; average sample size per quarter was 5,298,686 individuals. Pre-intervention, the quarter-to-quarter growth was 1.51 HbA1c tests per quarter per 1000 people. Post-intervention, the quarter-to-quarter growth increased by 8.45 tests per 1000 people (p < 0.005). Growth of HbA1c ordering differed significantly by region, years since physician graduation, and practice size.

Limitations — Incomplete data collection, inadequate stratification, and other unidentified confounders.

Conclusion — The D&I strategy resulted in a significant increase in the growth of HbA1c tests. The successful uptake of this recommendation may be due to its simplicity; guideline developers should consider this when drafting recommendations. Furthermore, differential uptake by user groups suggests that future strategies should include targeted barrier analysis and interventions to these groups.