Methods: Primary care clinicians at three geographically diverse VA medical centers were randomly assigned to receive the ATHENA Hypertension Advisory (intervention) or a simple hypertension pop-up reminder (control) in the Computerized Patient Record System during a 15 month trial. This DSS made individual patient tailored recommendations to initiate, change, or intensify BP therapy, based on JNC-6 guideline knowledge. Clinicians were randomized into clusters when they shared patient care (ATHENA: 47; Control: 44). We compared clinician-clusters' rates of intensification of antihypertensive therapy at 1st visit with BP above target using generalized estimating equation analyses that adjusted for the inter-correlation of observations within clinician. We analysed patient SBP over 15 months, using multi-level regression models. Analysis was by intention to treat.

Results: 34,427 visits of 11,473 patients met inclusion criteria for SBP analysis. The mean number of drugs at baseline was 1.5 (sd 0.16) for control group and 1.5 (sd 0.17) for ATHENA group. The mean number of visits in the 15 month trial was 3.25 (sd 0.48) for control group and 3.27 (sd 0.64) for ATHENA group. 9,492 patients had a BP above target during the trial and were included in the intensification of therapy analysis. Control clinician-clusters intensifed therapy at 26% visits when patients first presented with BP above target during trial. Clinician-clusters receiving the ATHENA Hypertension Advisory were more likely to intensify antihypertensive therapy than control clinician- clusters (OR 1.26; CI 1.01,1.57). The estimated average SBP for patients treated by ATHENA clinicians-clusters was 138mmHg and 139mmHg for patients treated by control clinicians (p = ns).

Conclusions: Randomization to ATHENA DSS led providers to intensify anti-hypertensive therapy when BP was above target. ATHENA did not lead to improved control of SBP, perhaps due to the small number of visits during the trial. Clinical decision support systems such as ATHENA DSS may have a role in improving management of chronic illnesses like hypertension, but such interventions may require additional components to encourage more frequent return visits to adjust therapy.