Objective — To compare glycaemic control, as reflected in the A1c level, of diabetic patients with primary care vs. with specialist care.
Methods — The study used administrative data from eastern Ontario, Canada, and a database containing the results of all A1c tests from this region between 1 September 1999 and 1 September 2000. To avoid referral bias, diabetic patients with an index specialist visit were selected and separated into those with exclusively primary care previously (n = 974) and those with prior specialist care (n = 3533). We compared A1c levels measured within 30 days of the index visit and hence attributable to the prior care. To control for confounding between the groups, both multiple linear regression and propensity score-based matching were used.
Results — After controlling for confounders, patients with prior specialist care had significantly lower A1c levels (P < 0.0001). Other predictors of lower A1c included older age, shorter diabetes duration, rural residence and higher neighbourhood income. In propensity score-matched cohorts, the A1c level was 8.3 ± 2.0% with prior primary care vs. 7.9 ± 1.6% with prior specialist care (P < 0.0001).
Conclusions — Specialist care prior to the index visit was associated with a lower A1c level than prior primary care. This difference would result in reductions in diabetes complications for patients with ongoing specialist care.
Health care evaluation
Primary care/clinical practice