(212) Personalized Prescribing Portraits and Evidence Summaries on First-line Antihypertensive Prescribing: A Factorial Randomized Trial of Audit-and-Feedback

Publication Number: 1204

Background: In British Columbia (BC), Canada, family physicians were mailed prescribing Portraits that provided a snapshot of their recent individualized prescribing patterns compared to other family physicians, plus evidence-based prescribing recommendations. The Portrait “How do you prescribe angiotensin-converting enzyme inhibitors and angiotensin II receptor blockers for hypertension?” was accompanied by a Therapeutics Letter summarizing evidence favouring thiazides as first-line antihypertensive (AH) therapy.

Objectives: To estimate the impact of this Portrait and Letter on physicians’ preferences for thiazide as initial AH therapy.

Methods: BC family physicians (n = 3,863) were randomized to three arms. In March 2021, we mailed the Portrait and Letter to 1,287 family physicians (Arm A) and the Letter alone to an additional 1,288 family physicians (Arm C). The 1,288 family physicians in Arm B served as the control group. We excluded ten physicians who opted out from our primary (intention-to-treat) analysis and an additional 42 who did not receive the mail from the per-protocol analysis. The primary endpoint was the odds of thiazide monotherapy among patients who started AH therapy. The odds were expected to increase in the 9-month follow-up among patients who did not have heart or renal failure or cardiovascular disease and were not pregnant. We assessed the impact of the Portrait and Letter (Arm A vs. B), the impact of the Letter alone (Arm C vs. B), and the added impact of the Portrait on the Letter (Arm A vs. C). The odds of thiazide monotherapy after the mailings versus before were compared between Arms using ratios of odds ratios (OR) and 95% confidence intervals (CI).

Results: There were 37,766 AH initiations in the 9 months before (8.0% thiazide monotherapy) and 37,967 in the 9-month follow-up period. In the primary analysis, the Portrait and Letter combination had an OR ratio of 1.62 (CI: 1.42-1.84). For the Letter alone, it was 1.27 (CI: 1.12-1.44). Compared to the Letter alone, the combination was 1.27 (CI: 1.11-1.45.) Results were consistent in a per-protocol analysis.

Conclusions: The personalized prescribing Portrait and the Letter were each associated with significant increases in preference for thiazides as first-line AH therapy in BC.