Abstract 14418: A Practical Approach to Quality Assurance Implementation in Clinical Echocardiographic Laboratories
BACKGROUND: Echocardiography is subject to interpretative variability that can limit quantitative value. Reproducibility is key for clinical management, however a practical framework for quality assurance implementation in the clinical setting is lacking.
METHODS: We devised a method to assess inter-reader reliability (IIR) for key echocardiographic parameters without a gold standard, or complex statistical analysis, by using pairwise comparisons among all readers. Reproducibility is deemed acceptable if 80% or more of all possible pairwise comparisons fall within a pre-defined acceptable difference (≥0.80 coverage probability). Readers who do not meet this agreement standard undergo retraining guided by descriptive results that indicate magnitude and direction of disagreement. The process is repeated until all readers achieve acceptable reproducibility for each parameter.
RESULTS: Five readers interpreted 10 echoes for the determination of IIR for left ventricular end-diastolic volume (LVEDV), ejection fraction (EF), and mitral (MR) and aortic (AR) regurgitation. The acceptable differences were pre-defined as 30mL for LVEDV, 10% for EF, and 1 grade on a 1-4 scale for AR and MR. All five readers demonstrated acceptable reproducibility for EF, AR and MR. One reader (A) failed to meet the minimum standard of agreement (coverage probability of 0.675) for LVEDV (Figure 1a). This reader underwent retraining and all readers reviewed their results. Upon repeating the process, there was improved inter-reader agreement and all readers demonstrated acceptable reproducibility for LVEDV (Figure 1b).
CONCLUSIONS: We devised a feasible strategy to assess and reduce interpretative variability in clinical echocardiographic laboratories. This practical approach yields actionable results to guide retraining, significantly improve reproducibility and assure quality. Further work is needed to assess the impact of widespread implementation of this method.
- © 2012 by American Heart Association, Inc.