Interpretation errors in focused cardiac ultrasound by novice pediatric emergency medicine fellow sonologists.

Abstract

Background: Focused cardiac ultrasound (FOCUS) is a core competency for pediatric emergency medicine (PEM) fellows. The objectives of this study were (1) to evaluate test characteristics of PEM-fellow-performed FOCUS for pericardial effusion and diminished cardiac function and (2) to assess image interpretation independent of image acquisition.

Methods: PEM fellows performed and interpreted FOCUS on patients who also received cardiology service echocardiograms, the reference standard. Subsequently, eight different PEM fellows remotely interpreted a subset of the PEM-acquired and cardiology-acquired echocardiograms.

Results: Eight PEM fellows performed 54 FOCUS exams, of which two had pericardial effusion and four had diminished function. PEM fellow FOCUS had a sensitivity of 50.0% (95% CI 9.19-90.8) and specificity of 100.0% (95% CI 91.1-100.0) for detecting diminished function, and sensitivity of 50.0% (95% CI 2.67-97.33) and specificity of 98.1% (95% CI 88.42-99.9) for detecting pericardial effusions. When PEM fellows remotely interpreted 15 echocardiograms, the sensitivity was 81.3% (95% CI 70.7-88.8) and specificity 75% (95% CI 67.0-81.0) for detecting diminished function, and sensitivity of 76.3% (95% CI 65.0-85.0) and specificity 94.4% (95% CI 89.0-97.0) for detecting pericardial effusion. There were no differences in sensitivity and specificity of PEM fellows' interpretation of FOCUS studies compared to their interpretation of cardiology echocardiograms. Interrater reliability for interpretation of remote images (kappa) was 0.66 (95% CI 0.59-0.73) for effusion and 0.31 (95% CI 0.24-0.38) for function among the fellows.

Conclusion: Novice PEM fellow sonologists (a physician who performs and interprets ultrasound) in the majority of instances were able to acquire and remotely interpret FOCUS images with limited training. However, they made real-time interpretation errors and likely need further training to incorporate real-time image acquisition and interpretation into their practice.