Diagnostic performance of a computer-based ECG rhythm algorithm
Kimble Poon, MD, Peter M. Okin, MD, Paul Kligfield, MD
T
Division of Cardiology, Department of Medicine, Weill Medical College of Cornell University and the Cornell Center of The New York – Presbyterian Hospital,
New York, NY 10021, USA
Received 14 October 2004; revised 11 January 2005; accepted 24 January 2005
Abstract
We examined the accuracy of computer-based rhythm interpretation from one electrocardiograph
manufacturer (GE Healthcare Technologies MUSE software 005C) in 4297 consecutive recordings
in a university hospital setting. Overreading was performed by either of 2 experienced cardiologists,
and all disagreements with the initial computer rhythm statement were reviewed by the second
cardiologist to achieve physician consensus used as the bgold standardQ for rhythm diagnosis.
Overall, 13.2% (565/4297) of computer-based rhythm statements required revision, but excluding
tracings with pacemakers, the revision rate was 7.8% (307/3954), including 3.8% involving the
primary rhythm diagnosis and 3.9% involving definition of ectopic complexes. The false-negative
rate for sinus rhythm was only 1.3%, but a computer diagnosis of sinus rhythm was incorrect in 9.9%
of other rhythms. The false-negative rate for atrial fibrillation was 9.2%, whereas a computer
diagnosis of atrial fibrillation was incorrect in 1.1% of other rhythms, including sinus. Computer
diagnosis of paced rhythms remains problematic, and physician overreading to correct computer-
based electrocardiogram rhythm diagnoses remains mandatory.
D 2005 Elsevier Inc. All rights reserved.
Keywords: Algorithm; Computer; Cardiac rhythm; Sensitivity; Specificity
1. Introduction
Computerized interpretation of the electrocardiogram
(ECG) has become widely applied during the past several
decades [1-7], and most digital ECG recorders are capable of
providing automated diagnostic statements that can assist or
at times mislead the electrocardiographer [8-15]. Interpretive
statements that depend on precise measurement of ECG
amplitudes and durations can approach experienced readers
in sensitivity, specificity, and reproducibility [16-18]. How-
ever, statements that depend on waveform configuration,
such as repolarization, and relationships between waveforms,
such as irregular P waves, may be less accurate [19,20].Asa
consequence, evolving algorithms for the interpretation of
cardiac rhythm remain imperfect and problematic [20-25].
Accordingly, the clinical importance of physician over-
reading of computer-based rhythm statements requires
periodic reevaluation.
2. Methods
We compared the initial unconfirmed computer-based
rhythm interpretation with the physician-confirmed
rhythm diagnosis in 4297 consecutive inpatient and out-
patient ECGs recorded in a university teaching hos-
pital during a 3-week period. All tracings were obtained
with GE Healthcare Technology (Milwaukee, Wis) simulta-
neous 12-lead digital recorders using a currently available
standard software algorithm for computer-based rhythm
diagnosis (005C, version 19). All unconfirmed tracings were
overread and corrected by 1 of 2 experienced electro-
cardiographers (P.M.O. and P.K.), and ECGs with rhythm
changes by an overreading physician were separated for
further review and analysis. All changes in rhythm interpre-
tation were reviewed by the second cardiologist to provide a
consensus bgold standardQ for each ECG rhythm for the
purpose of this study. In the relatively rare event of dis-
agreement between cardiologists, differences were resolved
by consensus, or the computer statement was accepted as
correct in further analysis.
0022-0736/$ – see front matter D 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.jelectrocard.2005.01.008
T Corresponding author. Tel.: +1 212 746 4686; fax: +1 212 746 8561.
Journal of Electrocardiology 38 (2005) 235– 238
www.elsevier.com/locate/jelectrocard