E-Consult Protocoling to Improve the Quality of Cardiac Stress Tests

Rarely appropriate cardiac stress tests remain prevalent in the range of 10% to 20% (1) and unnecessarily prolong wait times. To address this ongoing problem, we designed the EPIQ-Stress workflow, which includes a structured electronic consult (“e-consult”) with all outpatient stress test orders. The e-consult was designed to provide real-time decision support guiding appropriate test selection, individualized feedback to primary care clinicians (PCCs) regarding ordered test appropriateness, and the patient-specific rationale for any changes or cancellations of test orders. In this study, we assessed whether EPIQ-Stress implementation was associated with a reduction in rarely appropriate testing and in order-to-report wait times.

The pre-implementation cohort included consecutive veteran patients aged 40 to 75 years without a history of coronary artery disease referred by PCCs for exercise treadmill testing, stress single-photon emission computed tomography myocardial perfusion imaging, or stress echocardiography between March 1, 2017 and August 30, 2017. In these veterans, we determined the Appropriate Use Criteria (AUC) indication under which stress test referral occurred (2) and calculated “order-to-report” wait time as the number of days between order entry and test report generation. The EPIQ-Stress workflow (Figure 1A) was implemented at the Providence Veterans Affairs (VA) Medical Center on April 1, 2018. The e-consult was a checkbox-based hierarchical questionnaire structured to efficiently define the most appropriate AUC indication and then guide PCC ordering of an appropriate test (or no test). E-consult responses were required within 48 h and were templated (Figure 1B). The post-intervention cohort consisted of consecutive veterans meeting the pre-intervention cohort criteria who were referred for stress testing between April 1, 2018 and March 31, 2019. Between-cohort differences in the proportion of rarely appropriate tests were assessed using a chi-square test. Differences in order-to-report wait times were assessed using a 2-sample Wilcoxon rank-sum test. The Providence VA Medical Center granted institutional review board approval.

FIGURE 1. EPIQ-Stress.

(A) EPIQ-Stress Workflow. Swimlane diagram showing key differences between EPIQ-Stress and traditional workflows. (B) Example E-Consult Responses. CAD = coronary artery disease; ECG = electrocardiogram; PCP = primary care provider.

The pre- and post-intervention cohorts included 90 and 255 veterans, respectively, and both were predominantly composed of middle-aged non-Hispanic white males. AUC indications 1 to 6, representing tests requested for detection of coronary artery disease in symptomatic patients (2), accounted for ~85% of PCC referrals in both cohorts. Comparing the pre- and post-intervention cohorts, there was a nonsignificant reduction in the prevalence of rarely appropriate tests ordered (8% to 5%; chi-square test = 0.87; p = 0.35), but there was a significant reduction in the prevalence of rarely appropriate tests performed (8% to 2%; v = 8.31; p = 0.004). The median order-to-report wait time decreased from 26 days (interquartile interval: 18 to 32 days) in the pre-intervention cohort (n = 89) to 21 days (interquartile interval: 17 to 27 days) in the post-intervention cohort (n = 195), a statistically significant reduction (2-sample Wilcoxon rank-sum test p = 0.0433).

This study showed that EPIQ-Stress implementation was associated with a significant reduction in the prevalence of rarely appropriate cardiac stress tests performed and in order-to-report wait times. Our results suggest real-time clinical decision support may help PCCs order more appropriate tests but must be paired with structured auditing (including test changes or cancellations as appropriate) and feedback after order entry to achieve significant reduction in the prevalence of rarely appropriate tests. Moreover, full capture of feedback within the health record is a unique feature of EPIQ-Stress compared with prior clinical decision support tools (3) and simplifies retrospective audits of stress test appropriateness. We believe the improvement in order-to-report wait time was largely driven by the novel EPIQ-Stress mandate to answer e-consults (i.e., complete pre-test protocoling) within 48 h. Lastly, we believe EPIQ-Stress represents a viable policy alternative to the Centers for Medicare and Medicaid AUC Program, which shifts the burden of pre-test protocoling to nonspecialist referring providers. In contrast, EPIQ-Stress encourages appropriate testing through ground-level collaboration between referring providers and specialists.

Given the encouraging results at a single clinical site, there may be value in expanding EPIQ-Stress to other VA and non-VA clinical sites to determine if similar gains in quality care delivery can be achieved.