www.medscape.com
COMMENTARY
How Computer-Interpreted ECGs May Lead to Errors
Amal Mattu, MD
November 28, 2017
Computer-Interpreted Electrocardiograms: Benefits and Limitations
Schläpfer J, Wellens HJ
J Am Coll Cardiol. 2017;70:1183-1192
Background
ECG was invented more than 100 years ago.[1] Despite many technical advances in the field of emergency cardiology, the basic 12-lead ECG still remains a cornerstone test in the acute diagnosis of many types of cardiac conditions, both acute and chronic.
Automated ECG analysis was initiated in the 1950s in an attempt to assist providers who had less training in ECG interpretation.[1] Unfortunately, despite improvements in the automated interpretations over the past 60 years, computer ECG interpretations remain far from perfect. Despite the common teaching that we should not trust the computer interpretations, many healthcare providers still rely heavily on these interpretations in their clinical practice.
The following review provides some excellent information regarding the benefits and limitations of the computer interpretation programs.
Study Summary
The authors reviewed technical specifications about the various computer programs and also numerous articles regarding accuracy data on the various algorithms. The following is a list of key points they made.
This article sheds some much-needed light on the true accuracy of CIE programs. Although these programs may help providers who are inexperienced in ECG interpretation, they appear to decrease the accuracy of experienced providers by frequently providing a false sense of security with benign interpretations and also by encouraging less scrutiny of the ECG. I truly believe that if the computer interpretations were removed, providers would pay closer attention to the ECG, resulting in more accurate interpretations and fewer errors. I also believe that our trainees would feel compelled to work harder at their ECG interpretation skills if the computer interpretations were removed.
During the past 15 years, I've seen at least a half-dozen malpractice cases in which the computer interpretation simply indicated "nonspecific" findings on ECGs that demonstrated fairly obvious ischemia. In each of these cases, the treating physician was misled by the computer, and only in retrospect, too late, did the physician recognize the ischemic findings. These cases are typically nearly impossible to defend.
ECG interpretation is a life-saving skill. Proficiency is a must for us all. We must exercise extreme caution regarding the utility of these CIE programs and gain enough skill and confidence to avoid using them entirely.
Abstract
References
Any views expressed above are the author's own and do not necessarily reflect the views of WebMD or Medscape.
Cite this article: How Computer-Interpreted ECGs May Lead to Errors - Medscape - Nov 28, 2017.
COMMENTARY
How Computer-Interpreted ECGs May Lead to Errors
Amal Mattu, MD
November 28, 2017
Computer-Interpreted Electrocardiograms: Benefits and Limitations
Schläpfer J, Wellens HJ
J Am Coll Cardiol. 2017;70:1183-1192
Background
ECG was invented more than 100 years ago.[1] Despite many technical advances in the field of emergency cardiology, the basic 12-lead ECG still remains a cornerstone test in the acute diagnosis of many types of cardiac conditions, both acute and chronic.
Automated ECG analysis was initiated in the 1950s in an attempt to assist providers who had less training in ECG interpretation.[1] Unfortunately, despite improvements in the automated interpretations over the past 60 years, computer ECG interpretations remain far from perfect. Despite the common teaching that we should not trust the computer interpretations, many healthcare providers still rely heavily on these interpretations in their clinical practice.
The following review provides some excellent information regarding the benefits and limitations of the computer interpretation programs.
Study Summary
The authors reviewed technical specifications about the various computer programs and also numerous articles regarding accuracy data on the various algorithms. The following is a list of key points they made.
- There is no international accepted standard for computer interpretations. As a result, significant variability exists among interpretations from different manufacturers' algorithms.
- Direct comparative evaluations of the various commercially available computer-interpreted ECG (CIE) programs has never been performed.
- A 1991 study of nine CIE programs versus eight cardiologists demonstrated consistently lower accuracy among the CIEs compared with the cardiologists, and also (surprisingly) significant variation in accuracy among the various programs.[2]
- CIE programs have a frequent tendency to overcall atrial fibrillation, especially in elderly persons, potentially leading to inappropriate administration of harmful medications.
- CIE programs have a tendency to double-count the rate due to large T-waves (eg, in the setting of hyperkalemia).
- CIE programs are particularly inaccurate in diagnosing pacemaker rhythms.
- CIE programs demonstrate wide variations in the false-positive (0%-42%) and false-negative (22%-42%) rates of diagnosis of ST-segment myocardial infarction; therefore, it is not recommended that CIEs be used as a sole means of activation of the cardiac catheterization lab.
- CIE programs frequently tend to underestimate the QT interval, especially in the presence of artifact or improperly placed leads.
- The authors cite a 2008 study indicating that "It has been roughly estimated that [CIE] misdiagnoses may account for up to 10,000 adverse effects or avoidable deaths worldwide annually."[3]
- The authors summarize, "Computer-based analysis of the ECG may lead to erroneous diagnosis with useless, inappropriate, or even dangerous care of the patient."
This article sheds some much-needed light on the true accuracy of CIE programs. Although these programs may help providers who are inexperienced in ECG interpretation, they appear to decrease the accuracy of experienced providers by frequently providing a false sense of security with benign interpretations and also by encouraging less scrutiny of the ECG. I truly believe that if the computer interpretations were removed, providers would pay closer attention to the ECG, resulting in more accurate interpretations and fewer errors. I also believe that our trainees would feel compelled to work harder at their ECG interpretation skills if the computer interpretations were removed.
During the past 15 years, I've seen at least a half-dozen malpractice cases in which the computer interpretation simply indicated "nonspecific" findings on ECGs that demonstrated fairly obvious ischemia. In each of these cases, the treating physician was misled by the computer, and only in retrospect, too late, did the physician recognize the ischemic findings. These cases are typically nearly impossible to defend.
ECG interpretation is a life-saving skill. Proficiency is a must for us all. We must exercise extreme caution regarding the utility of these CIE programs and gain enough skill and confidence to avoid using them entirely.
Abstract
References
- Al Ghatrif M, Lindsay J. A brief review: history to understand fundamentals of electrocardiography. J Community Hosp Intern Med Perspect. 2012;2.
- Willems JL, Abreu-Lima C, Arnaud P, et al. The diagnostic performance of computer programs for the interpretation of electrocardiograms. N Engl J Med. 1991;325:1767-1773. Abstract
- Mele PF. The ECG dilemma: guidelines on improving interpretation. J Healthc Risk Manag. 2008;28:27-29.
Any views expressed above are the author's own and do not necessarily reflect the views of WebMD or Medscape.
Cite this article: How Computer-Interpreted ECGs May Lead to Errors - Medscape - Nov 28, 2017.