Identifying predictors and barriers related to clinician decisions to prescribe thiazide diuretics for patients with uncomplicated hypertension

Background. Despite evidence-based guidelines which favor thiazide diuretics, use of these drugs remains low. To date, decision support tools have demonstrated modest effects on improving prescribing practices.

Objective: To identify factors related to clinician decisions to prescribe thiazide diuretics when provided point-of-care decision support

Methods: This is a secondary cross-sectional analysis of data from a 15-month randomized evaluation of the impact of ATHENA-HTN, an automated decision support system for hypertension management, on guideline concordance and blood pressure control. Primary care clinicians at three Veterans Affairs Health Care Systems received either ATHENA advisories with patient-specific recommendations or hypertension reminders. Our analysis included patients with uncomplicated hypertension who were treated with one or more antihypertensive medications, were not on furosemide, had no strong contraindication to thiazides, and were not on a thiazide diuretic at baseline. Thiazides were preferred drugs for these patients under the JNC-6 guidelines in effect at the time; recommendations to prescribe thiazides are also included in the current JNC-7 guidelines. Generalized linear mixed models were used to evaluate the relationship between patient demographic, clinical, and health care utilization factors and the probability that a patient received a new prescription for thiazide during the study period.

Results: Eighteen percent of patients in our sample received a new prescription for thiazide during the study period. Patients were more likely to be prescribed a thiazide if their primary care clinician was randomized to receive ATHENA advisories versus hypertension reminders (OR: 1.61, 95% CI: 1.14, 2.27); or if they had a SBP≥160 or DBP≥100 (OR: 10.37, 95% CI: 6.88, 15.63) or a blood pressure above target (OR: 3.82, 95% CI: 2.56, 5.71) compared to patients with blood pressure in good control. Patients were less likely to be prescribed a thiazide if they had gout (OR: 0.68, 95% CI: 0.48, 0.96), benign prostatic hypertrophy (OR: 0.72, 95% CI: 0.58, 0.90), a relative indication for a non-thiazide drug (OR: 0.78, 95% CI: 0.62, 0.99), or a history of an adverse drug event related to thiazide (OR: 0.54, 95% CI: 0.30, 0.96). Gender, age, race/ethnicity, having a diagnosis of diabetes, and number of antihypertensive medications prescribed at baseline were not associated with receipt of a new prescription for thiazide.

Conclusions. Clinician decisions not to prescribe thiazides may reflect clinical inertia as well as clinically appropriate decisions. Tailored clinical decision support such as that provided by ATHENA-HTN offers a nuanced strategy for improving clinician prescribing behavior.