Abstract 87: Comparison of Pretest Probability Clinical Assessment and 2011 ACCF/ASE/AHA/ASNC/HFSA/HRS/ SCAI/SCCM/SCCT/SCMR Appropriate Use Criteria for Transthoracic Echocardiography and Association with Clinical Impact

Abstract

Background: Although most transthoracic echocardiograms (TTEs) performed today meet Appropriate Use Criteria (AUC), the ability of AUC to predict whether these TTEs will impact clinical management remains unclear. Methods: The charts of 143 consecutive inpatient or outpatient TTEs at an academic medical center were retrospectively reviewed. Patients with a history of cardiac transplant or left ventricular assist device, or with inadequate pre-/ post- TTE clinical documentation were excluded; 123 studies were analyzed. Each TTE was assigned an AUC, pre-test, and clinical impact category. TTE indications were assigned by a cardiologist blinded to pre-study and post-study clinical data and categorized as: 1) inappropriate; 2) appropriate; or 3) uncertain based on 2011 AUC. Pre-test clinical course was analyzed by two independent cardiologists blinded to AUC and post-TTE clinical course and categorized as 1) TTE should not be performed, 2) TTE would likely alter clinical management, 3) TTE should be performed for reassurance/ continuity of care. Post-test clinical course was analyzed by two independent cardiologists blinded to AUC and pre-test assignment and categorized as 1) TTE did not alter clinical management, 2) TTE led to active change in management, 3) TTE led solely to reassurance/ continuity of current care. Results: By 2011 AUC, 85% of studies were appropriate, 11% were inappropriate, and 4% were uncertain. Active change resulted from 30% of studies, while 49% of studies provided reassurance/ continuity of care, and 21% had no clinical impact. These proportions were not statistically significant across AUC categories (p=0.54). Pretest clinical assessment suggested that 31% should not have been performed, 63% should have been performed for reassurance / continuity of care alone, and 7% were likely to alter clinical management. These pretest assessments were significantly associated with impact on clinical management (p=0.0067), with 75% (n=6/8) of those studies expected to alter clinical management actually doing so, 56% (n=43/77) of those predicted to primarily result in reassurance actually doing so, and 37% (n=14/38) of those predicted not to alter clinical management actually having no clinical impact. Conclusions: Classification by AUC is not associated with clinical impact. Although classification by pre-test probability is associated with clinical impact, a significant number of studies not predicted to have clinical impact actually had clinical impact. Further studies on methods to improve prediction of clinical utility for transthoracic echocardiography are warranted.