Background: The use of decision support systems (DSS) at the point of care may enhance the appropriateness of clinical cardiology versus human physician standard care (STD) bringing evidence-based medicine at the point-of-care.
Purpose: To analyze DSS results vs standard care (STD) in the clinical workflow of patients (pts) at low, low-to-intermediate pre test likelihood (L-LI) of coronary artery disease (CAD).
Methods: 692 pts (403 males and 289 females, age 57±7 years) with L-LI of CAD were referred for stable chest pain evaluation over a 16 month period in three different hospitals. A browsing computerized automated DSS and a human cardiologist STD were applied during the same day visit. Pre-test likelihood of CAD was based on clinical score + coronary artery calcium score (CACS). Significant CAD (>50% coronary stenosis) criteria were applied in all pts by computerized tomography coronary angiography (CTCA).
Results: Distribution of population for DSS and STD is shown in the table. 498 (72%) pts were classified as “No further test (NFT)”, 110 (15.9%) “Exercise test (ET)/Functional Imaging (FI)”, 84 (12.1%) “CTA” and 0 “(ICA)” by DSS. Of note, 483 (97%) of DSS “NFT” showed no significant CAD vs 576 (99%) of STD “CTA” (p=0.3). 110 (15.9%) pts were assigned by DSS to “ET/FI” as the first approach vs 27 (3.9%) of STD (p=0.0001). The remaining 38 of STD “ET/FI” performed the test after CTCA. The diagnostic accuracy was 97.8% by DSS in the “NFT” group.
Conclusions: DSS is a sensitive tool for applying good clinical practice in pts with a L-LI pre-test likelihood of CAD. For the “NFT” group DSS was demonstrated to be highly accurate to exclude CAD. It could be a promising tool to substantially improve health care quality avoiding unnecessary tests and reducing costs.