On-call transthoracic echocardiographic interpretation by first year cardiology fellows: comparison with attending cardiologists

Abstract

Background Transthoracic echocardiograms (TTE) performed and interpreted by cardiology fellows during off-duty hours are critical to patient care, however limited data exist on their interpretive accuracy. Our aims were to determine the discordance rate between TTEs performed and interpreted by cardiology fellows and National Board of Echocardiography certified attending cardiologists and to identify factors associated with discordance. Methods Consecutive on-call TTEs acquired and interpreted by 1st year cardiology fellows over 4.6 years at an academic center were prospectively evaluated by attending cardiologists. Fellow interpretations were classified as concordant or discordant with the attending interpretation. We assessed the association of patient, imaging and fellow characteristics with discordance. Results A total of 777 TTE interpretations (730 patients) were performed/interpreted by 40 first year fellows and overread by 13 attendings. The most common indications were assessment of left ventricular function (40.9%) and pericardial effusion (37.3%). There was a major or minor discordance in 4.1 and 17.4% of studies, respectively with 42.1% of disagreements occurring in assessment of left ventricular size and function. The indication to assess left ventricular function [OR 2.19, 95% CI (1.32, 3.62), P = 0.002 vs. pericardial effusion] and greater duration of echocardiographic image acquisition (OR 1.02, 95% CI 1.01, 1.03, P = 0.004) were independently associated with overall discordance. Conclusions In this large prospective study we found that attending cardiologists disagreed with 1 in 5 fellow TTE interpretations. Standardized tools for evaluation of echocardiograms performed by fellows are needed to ensure quality of training and patient safety. Electronic supplementary material The online version of this article (10.1186/s12909-019-1634-7) contains supplementary material, which is available to authorized users.

A machine learning algorithm supports ultrasound-naïve novices in the acquisition of diagnostic echocardiography loops and provides accurate estimation of LVEF

“…When performed by humans, echocardiography and its measurements are subjective and greatly influenced by the level of expertise [ 5 , 10 ]. Since its first description there have been concerns about the reproducibility of LVEF due to significant intra-observer, inter-observer, and inter-institutional variability of measurements [ 11 – 13 ].…”

“…At the same time, calculation of LVEF is strongly dependent both on image quality and on the experience of the reader [4]. Thus calculation of LVEF by non-cardiologists and beginners is a significant limitation for the widespread application of this parameter [5].…”

Tool to Improve Qualitative Assessment of Left Ventricular Systolic Function

“…Comprehensive exposure to the assessment of LV function is a difficult task that requires evaluation of a large number of patients. Indeed, even among cardiology trainees, the most frequent source of disagreement with attending overread is regarding LV function (19). This program took trainees less than 1 hour in total time and allowed trainees to be exposed to almost 100 exams of the LV in much less time than real-time examination of 100 patients by echocardiography.…”

<p>Cardiology Fellow Diagnostic Accuracy and Data Interpretation Outcomes: A Review of the Current Literature</p>

“…Similarly, a study completed at Beth Israel demonstrated that in transthoracic echocardiogram interpretation there was a nearly 20% discrepancy rate (4.1% major and 17.4% minor) between cardiologists and the first-year cardiology fellows. 16 …”