Accuracy of ECG Interpretation in the Pediatric Emergency Department
Received 1 September 2004; received in revised form 5 February 2005 and 8 March 2005; accepted 17 March 2005. published online 16 June 2005.
Study objective
We assess accuracy of ECG interpretation and indications for obtaining ECGs and develop a clinical classification system of ECG abnormalities.
Methods
Prospectively acquired ECG data on patients 0 to 21 years of age and presenting to our pediatric emergency department (ED) were obtained. Clinical indications were documented. The initial ECG interpretation (pediatric ED attending physician) was compared with the criterion standard (pediatric cardiologist). A blinded cardiology panel reviewed discrepancies, and a final concordance rate was determined. An ECG abnormality classification system was developed and used to categorize these abnormal ECGs.
Results
One thousand six hundred fifty-three ECGs from 1,501 patients, aged 2 days to 21 years (median 10.0 years), were obtained during 3.5 years. Fifty-one percent were male patients. ECG indications included chest pain (21%), seizure or syncope (18%), arrhythmias (17%), apparent life-threatening event or respiratory symptoms (16%), ingestions (10%), cardiac abnormality (10%), and miscellaneous (8%). From 1,631 ECGs, 1,160 (71%) were normal (class 0), 259 (16%) were minimally abnormal (class I), 174 (11%) were moderately abnormal (class II), and 38 (2%) were severely abnormal (class III). Kendall's τ-b test showed concordance of 0.73 (95% confidence interval 0.70 to 0.77) between pediatric ED and cardiology interpretation. The sensitivity of pediatric ED interpretation was 75%, and the specificity was 98.5%. The positive predictive value of pediatric ED interpretation was 88.3%, and the negative predictive value was 96.3%.
Conclusion
We conclude that, overall, a high rate of concordance exists between the pediatric emergency physician's and the cardiologist's ECG interpretation. The majority of discordant ECGs are not clinically significant. However, among the clinically significant ECGs, there is a higher rate of discordance. These data suggest that review of pediatric ECGs by pediatric cardiologists may significantly reduce underdetection of clinically important ECG findings in children.
From the Department of Pediatrics, Division of Emergency Medicine (Wathen, Rewers), and Division of Pediatric Cardiology (Yetman, Schaffer), University of Colorado at Denver and Health Sciences Center/The Children's Hospital, Denver, CO
Address for correspondence: Joe E. Wathen, MD, The Children's Hospital, Emergency Department B251, 1056 E. 19th Avenue, Denver, CO 80218; 303-837-5464, fax 303-764-8694
Supervising editor: David M. Jaffe, MD
Author contributions: JEW, ABR, and MSS conceived the study and designed the trial. JEW and ABR supervised data collection. ATY and MSS reviewed ECGs and assigned blinded interpretations. ABR provided statistical analyses of the data. JEW drafted the manuscript, and all authors contributed substantially to its revision. JEW takes responsibility for the paper as a whole.
Funding and support: The authors report this study did not receive any outside funding or support.
Presented at the Pediatric Academic Societies' meeting, May 2003, Seattle, WA.
ABSTRACT