Electrocardiogram Results: ***READ ME***
A 63-year-old woman with labile hypertension presented to the emergency department (ED) with new onset chest discomfort and an initial blood pressure of 210/100 mm/Hg. Her electrocardiogram (ECG) was unchanged from previous studies and her symptoms resolved with treatment of the hypertension. She was admitted overnight with orders for a morning ECG and serial troponin (ST) levels, the first of which was normal. At 6:00 AM, a nursing assistant obtained the ordered ECG and placed it in the patient's bedside chart without notifying a nurse or physician.
When the team was rounding 2 hours later, they reviewed the ECG, which was notable for new ST elevations inferiorly and laterally. The computer readout of the ECG stated, "****ACUTE MI****" and cited the ST elevations. On questioning the patient, she did report intermittent chest pressure overnight that was less severe than it was when she presented. Her morning troponin level also returned elevated, which was consistent with an acute myocardial infarction (MI). After urgent evaluation, she underwent successful coronary angiography and placement of two stents. She was discharged home without complications and on appropriate medical therapy.
The case raised concerns about the review of ECGs routinely performed in the hospital setting and often by those without the skills to interpret them. In this case, the nursing assistant later reported that she wasn't sure that "MI" meant "heart attack," but she also said that she doesn't routinely look at the computer interpretations. It was unclear whether the delay of 2 hours influenced the patient's outcome, but it could have.
This case involved an error in detecting a seriously abnormal ECG reading due to failure of oversight of the computer-generated interpretation. Fortunately, the patient survived, although the size of her infarct (i.e., the volume of irreversibly necrotic myocardium) was undoubtedly larger than it would have been had the acute myocardial infarction been detected earlier. Early identification of acute coronary syndromes (ACS) as a cause for chest pain is critical because it triggers a cascade of urgent therapies, often including a trip to the catheterization lab. Delays in making this diagnosis can lead to significant adverse outcomes for patients.
What errors can we identify from this case and how could they have been prevented? We'll focus on two serious errors that could have been prevented with a better systems approach in this hospital's ACS protocol. The first error was that the person performing the ECG (i.e., the nursing assistant) did not understand the terminology or the implication of the computer-generated diagnostic statement. It is essential that the person performing the ECG be able to read and understand English as well as the implications of highlighted, important interpretations such as "ACUTE MI." Although nurse's aides don't have the expertise to be able to read and correctly interpret an ECG tracing, they should be expected to act on seeing a statement in bold capital letters at the top of a tracing and realize it deserves immediate attention from the bedside nurse or attending physician. Other important diagnostic statements that nurse's aides should immediately bring to the attention of the care team include ventricular tachycardia, atrial fibrillation, marked bradycardia, and any form of heart block beyond first-degree atrioventricular block. It would have been appropriate for the hospital where this error occurred to instruct all nurses aides concerning these important terms and to give out cards with the diagnoses listed that demand immediate attention. Nurse's aides need not know how to recognize these entities on the ECG, but they must recognize the importance of such diagnostic terms in order to call attention to them.
The second error in this case was failure to inform and educate the patient concerning the importance of reporting the development or recurrence of key symptoms, such as chest discomfort. I assume that the chest discomfort present on admission had resolved by the time the patient was transferred to the inpatient ward. Resolution of chest discomfort likely resulted from appropriate control of elevated blood pressure and analgesic administration. All patients admitted on an ACS protocol should be repeatedly reminded that chest discomfort is an important symptom and that they must inform the staff if and when it develops or recurs. Had the patient told her nurse about the recurrence of chest discomfort, the nurse almost certainly would have immediately ordered a repeat ECG while informing the attending physician about the change in symptom status. Obtaining another ECG would have led to the diagnosis of acute myocardial infarction being discovered earlier. The patient then would have gone emergently to the catheterization laboratory for reperfusion therapy.(1) The latter would almost certainly have resulted in a smaller myocardial infarct than eventually developed.
What reasonable measures should hospitals take to prevent this type of error from occurring? First, all computer ECG readings should at least be reviewed by the nurse responsible for the patient. Any disturbing diagnosis should then immediately be called to the attention of the attending physician. In addition, an education program should be instituted for nurse's aides and others who obtain ECGs. Such a program should deeply imbed in the minds of staff which diagnostic statements must be communicated immediately to the care team. Printed reminder cards should be distributed and posted prominently on all floors caring for potential ACS patients. Second, the standing orders for all ACS patients should include directions for the nurses to educate patients concerning the importance of reporting immediately any chest discomfort that might develop.
Before leaving this case, it is worth a final note regarding the accuracy of computer-generated ECG reports. Computers are capable of performing a variety of calculations with remarkable speed and accuracy. However, the human brain is uniquely qualified to perform pattern recognition analysis. Hence, expert human ECG readers are far more accurate than computer-generated ECG analysis.(2) Computer ECG interpretation is inaccurate approximately 20% of the time when performing wave form analysis and much more often when attempting to analyze arrhythmias. Therefore, all computer-generated ECG readings must be over-read by an expert, usually a cardiologist. Furthermore, a number of ECG patterns may resemble tracings with ST-elevation or non–ST-elevation myocardial infarction. This can lead to an erroneous diagnosis of acute myocardial infarction (Table) or other clinically important entities.
- Staff and patient education is essential in all areas of patient care and clinical medicine. Education is particularly important in illnesses that can rapidly evolve and that require urgent and immediate attention.
- All ECGs done should be shown to the nurse responsible for the patient. Any disturbing statement in the computer reading of the ECG should be referred to the attending physician for any action required.
- Computer ECG interpretation is quite fallible and all ECGs should be over-read by an expert.
- Whoever obtains an ECG should be aware of key diagnostic terms (such as acute MI and ventricular tachycardia) that should trigger an immediate clinician review.
- Hospitals and physicians should develop systems approaches to patient care that minimize the chance for errors. In the case of ECGs obtained in patients with potential acute coronary syndromes, such approaches should include a clear protocol for clinical review of ECGs.
Joseph S. Alpert, MD Professor of Medicine University of Arizona College of Medicine Editor in Chief, The American Journal of Medicine
1. Wright RS, Anderson JL, Adams CD, et al; American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines. 2011 ACCF/AHA focused update incorporated into the ACC/AHA 2007 Guidelines for the Management of Patients with Unstable Angina/Non-ST-Elevation Myocardial Infarction: a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines developed in collaboration with the American Academy of Family Physicians, Society for Cardiovascular Angiography and Interventions, and the Society of Thoracic Surgeons. J Am Coll Cardiol. 2011;57:e215-e367. [go to PubMed]
2. Alpert JS. Can you trust a computer to read your electrocardiogram? Am J Med. 2012;125:525-526. [go to PubMed]
b>Table. Common ECG pitfalls in diagnosing myocardial infarction
• Early repolarization
• Left bundle branch block (LBBB)
• J point elevation syndromes (e.g., Brugada syndrome)
• Pulmonary embolism
• Subarachnoid hemorrhage
• Metabolic disturbances, such as hyperkalemia
• Lead transposition
• Persistent juvenile pattern
• Malposition of precordial ECG electrodes
• Tricyclic antidepressants or phenothiazines
• Prior myocardial infarction with Q-waves and/or persistent ST elevation
• Right ventricular pacing