Can prioritized clinical decision support in primary care reduce cardiovascular risk? [poster] Conference Poster uri icon
Overview
abstract
  • Background: The objective of this project was to develop and implement sophisticated point-of-care EHR-based clinical decision support (CDS) that (a) identifies and (b) prioritizes all available evidence-based treatment options to reduce a given patient’s cardiovascular risk (CVR). Methods: We randomized 19 primary care clinics with 102 primary care providers (PCPs) and 39,025 adults with diabetes, cardiovascular disease (CVD), or 10-year ACC/AHA reversible CV risk >=10% into one of two experimental conditions: Group 1 included 10 clinics that received the CV Wizard. Group 2 included 9 usual care clinics. The study formally tested the hypothesis that after control for baseline CVR, post-intervention American College of Cardiology/ American Heart Association (ACC/AHA) 10-year CVR (risk of fatal or non-fatal heart attack or stroke) will be significantly better in Group 1 than Group 2 in the post-intervention period Results: The CV Wizard system was integrated successfully into the workflow of primary care visits, and use rates at targeted visits in intervention clinics ranged from 44% to 77% and improved over time. In the high reversible CV risk sample (N=7,595), 10-year ACC/AHA CVR declined by -.030% per visit in the control group (p=.001) and by -0.46% per visit in the intervention group (p=0.28), this difference in annual rate of change in CVR was statistically significant and favored the intervention group (p<0.001). In the diabetes sample (N=5,510), the observed change in 10-year ACC/AHA CVR was +0.06% per visit in the control group (p=.56) and by -0.23% per visit in the intervention group (p<.03); this difference in rate of change in CVR was statistically significant and favored the intervention group (p=.049). The predicted annual change in CVR was +0.91% in the control group (p=.16) and +0.38% in the intervention group (p=0.55), this difference in annual rate of change in CVR was not statistically significant (p=.56). In the CVD sample (N 2,078), 10-year ACC/AHA CVR change over visits (p=.42), and over time (p=.92) was not significant when comparing intervention and usual care clinics. Conclusion: The overall pattern of change in CVR, whether measured by visit or time, was consistent with the assertion that CVR decreased at a faster rate (or increase at a slower rate) in the Wizard intervention clinics relative to control clinics for those with diabetes or high reversible CV risk, but not for those with known CVD. The difference in trajectories reached statistical significance over the course of visits among diabetes patients, and over time among patients with high reversible CV risk. Use rates and PCP satisfaction with the CV Wizard were very high, and economic analysis suggests improved care is cost-effective. Based on these and other research results, the CV Wizard clinical decision support system is currently being used at three large health care delivery systems in 4 states that provide care to 1,500,000 patients.

  • publication date
  • 2017
  • Research
    keywords
  • Cardiovascular Diseases
  • Clinical Decision Support Systems
  • Diabetes
  • Economics
  • Medical Records Systems, Computerized
  • Primary Health Care
  • Quality of Health Care
  • Randomized Controlled Trials
  • Risk Reduction