Cause of Device-Related Incident
User errors

Clinical Specialty or Hospital Department
Cardiology / Cardiac Catheterization; Clinical/Biomedical Engineering

Device Factors
*Not stated

Document Type
User Experience Network (UEN) reports

External Factors
*Not stated

Mechanism of Injury or Death
Misdiagnosis

Support System Failures
*Not stated

Tampering and/or Sabotage
*Not stated

User Errors
Inappropriate reliance on an automated feature; Incorrect control settings

UMDNS
Electrocardiographs, Multichannel [11-411]; Electrocardiographs, Multichannel, Interpretive [16-231]; Electrocardiographs, Multichannel, Interpretive, Signal-Averaging [18-330]; Electrocardiographs, Single-Channel [11-413]

ST-Segment Distortion in Manual Report Mode of Electrocardiographs



User Experience Network™ [Health Devices Aug-Sep 1995;24(8-9):362-3]

Hospital

Using a 12-lead electrocardiograph with a 0.5 Hz baseline wander filter (low-frequency filter) engaged, we generated both an auto report and a manual report from the same patient. When comparing the two reports, we noticed that the manual report contained distortion of the ST segments; that is, certain leads had elevated ST segments compared with those on the auto report. We are concerned that ST-segment distortion of this type could be misinterpreted as representing a cardiac abnormality (e.g., myocardial infarction, myocardial ischemia). Why does this distortion occur, and how can it be prevented?

ECRI

An electrocardiograph's manual report mode generates a printed report displaying continuous ECG data, whereas the auto report mode generates reports in a variety of formats displaying short segments of ECG data. Each of these modes has different filtering characteristics, as evidenced by the reported problem. Although no distortion was observed on the auto report when the 0.5 Hz low-frequency filter setting was used, the use of this filter setting, when the unit was in the manual report mode, resulted in the ST-segment distortion observed by the hospital.

An electrocardiograph's low-frequency filter is used to minimize drifting of the ECG baseline. Many units provide a range of settings (e.g., 0.01 to 0.5 Hz) for low-frequency filtering of the ECG signal. Higher frequency settings eliminate a greater amount of baseline wander from the ECG signal than lower settings. However, as mentioned in the operator's manuals for several manufacturers' units, improperly selecting a higher frequency setting may cause distortion of the ST segment because some essential, low-frequency, ST-segment content may be filtered out. Thus, choosing the proper low-frequency filter setting is important.

In 1975, theAmerican Heart Association (AHA) addressed the problem of ST-segment distortion by recommending that frequencies as low as 0.05 Hz be preserved to prevent distortion.(1) This recommendation is valid for analog filter designs typically used at that time. Similar requirements will still exist on current units in the real-time manual report mode. Digital technology typically used in the auto report mode allows the lower frequency limit to be relaxed to 0.67 Hz with minimal risk of ST-segment distortion.(2) (This is possible because digital filter design allows phase distortion to be eliminated. For example, digital circuit design allows the ECG data to be stored in memory and passed through the filter both forward and backward. Passing the data through the filter in reverse cancels any distortion caused by the first pass.) Therefore, a 0.5 Hz low-frequency limit is adequate for the auto report mode, but if the unit is switched to the manual report mode and the 0.5 Hz limit is maintained, ST-segment distortion may occur.

Recommendations

  1. In the manual report mode, set the low-frequency filter no higher than 0.05 Hz to avoid distortion of the ST segment. Be especially careful when switching from the auto report mode to the manual report mode to ensure that an adequate low-frequency filter is used.
  2. If baseline wander occurs in the auto report mode, follow the manufacturer's recommendations concerning appropriate filter settings that can be used without introducing significant waveform distortion.
  3. In either report mode, make sure that proper procedures are observed for electrode placement and skin preparation and that patient movement is limited. Noise on the ECG recording is often caused by poor electrode contact. Also, respiration may contribute to noise on an ECG; thus, minimizing the patient's deep breathing during ECG recording may reduce the interfering noise.

Notes

  1. Pipberger HV, Arzbaecher RC, Berson AS, et al. Recommendations for standardization of leads and specifications for instruments in electrocardiography and vectorcardiography--Report of the Committee of Electrocardiography, American Heart Association [AHA Committee Report]. Circulation 1975 Aug;52(2):11-31.
  2. Bailey JJ, Berson AS, Garson A Jr, et al. Recommendations for standardization and specifications in automated electrocardiography: Bandwidth and digital signal processing--A report for health professionals by an ad hoc writing group of the Committee on Electrocardiography and Cardiac Electrophysiology of the Council on Clinical Cardiology, American Heart Association [Special Report]. Circulation 1990 Feb;81(2):730-9.

UMDNS Terms

  • Electrocardiographs, Multichannel [11-411]
  • Electrocardiographs, Multichannel, Interpretive [16-231]
  • Electrocardiographs, Multichannel, Interpretive, Signal-Averaging [18-330]
  • Electrocardiographs, Single-Channel [11-413]

Cause of Device-Related Incident

User errors: Inappropriate reliance on an automated feature; Incorrect control settings

Mechanism of Injury or Death

Misdiagnosis


[Home]    [About]    [Help]    [Site Map]
Copyright © 2017 ECRI
All rights reserved
www.ecri.org