Electrocardiogram interpretation and management in a pediatric emergency department.

2005 
Objectives: To determine the accuracy of electrocardiogram (ECG) interpretation by pediatric emergency physicians through comparison with a pediatric cardiologist and to determine the intrarater and interrater reliability for pediatric emergency physicians and cardiologists. Methods: This was a prospective cohort study in which pediatric emergency physicians ordering an ECG completed a standardized questionnaire. The same emergency physician, a second emergency physician, and a pediatric cardiologist also completed the questionnaire for all ECGs at a later time. A randomly selected subset of ECGs was also interpreted by the same cardiologist and a second pediatric cardiologist. Major outcome variables were (1) whether the ECG was normal or abnormal, and if abnormal, (2) whether the abnormality represented a minor or major concern, and (3) whether the ECG warranted referral to a pediatric cardiologist. Results: For pediatric emergency physicians, the intrarater and interrater K values were 0.56 and 0.24 for the presence of an abnormality, 0.49 and 0.36 for level of concern, and 0.63 and 0.25 for need of cardiology follow-up. For pediatric cardiologists, the intrarater and interrater K values were 0.82 and 0.92 for the presence of an abnormality, 0.71 and 1.00 for level of concern, and 0.82 and 0.91 for need of cardiology follow-up. A comparison of the initial emergency physician and cardiologist interpretations yielded K values of 0.42 for the presence of an abnormality, 0.16 for level of concern, and 0.31 for need of cardiology follow-up. Conclusions: When compared with interpretation by a pediatric cardiologist, ECG interpretation by pediatric emergency physicians was relatively inaccurate; intrarater and interrater agreement among emergency physicians was good and poor, respectively, and the intrarater and interrater agreement among pediatric cardiologists was excellent.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    0
    References
    5
    Citations
    NaN
    KQI
    []