Computer assisted ECG interpretation

• computer interpretation has been around since the 1960s and ECG machines
  • standards were developed in the 1980s (the CSE database) and ECG interpretation programs can be tested against these standards, but they have often been ignored. Thus, not all programs are of a similar standard
  • as long as an ECG machine is 'electrically' safe, it can be marketed in Europe regardless of the quality of its interpretation program (1)
• computerised ECG interpretation
  • arrhythmias
    • computer-assisted ECG are relatively poor at this
      • a good program will diagnose atrial fibrillation correctly 8 times out of 10, a bad program 6 times. Therefore, additional visual analysis is always advisable, and for this it is strongly recommended that lead II (and ideally lead VI as well) is printed as a fourth channel below each of the 4 sets of 3 leads of the ECG.
    • myocardial ischaemic changes and conduction abnormalities:
      • computer-assisted interpretation will highlight abnormalities in QRS and ST-T complexes to reinforce diagnostic suspicions of ischaemia or infarction
      • computer-assisted ECGs can identify left bundle branch block and can suggest left ventricular hypertrophy to indicate end organ disease and a high risk patient
        • however the Sokolow-Lyon criterion for left ventricular hypertrophy, i.e. R wave in V5 or V6 + S wave V1 greater than 3.5 mV, is a gross simplification
          • this criteria is satisfactory in white females, but 3.9 mV is a more realistic upper limit of normal for white males and 4.9 mV for black males, while black females steadily increase their upper limit of normal from 3.4 mV to 4.9 mV as they get older (1)
    • validation of computer-assisted ECG
      • making the correct measurements is clearly crucial for any computer interpretation of the ECG
        • checking the PR interval and QRS duration visually and comparing them with the computer?s estimate is a quick validation method
          • if the computer-assisted measurement of the QRS duration is wrong then it is likely that the interpretation is also wrong. Estimating the QT interval, from the onset of the QRS to the end of the T wave, is now important in view of the concern that some drugs in common use may lengthen the QT interval and increase the possibility of serious rhythm disturbances
    • common errors when using/reading computer-assisted ECGs
      • missing small R waves resulting in a wrong diagnosis of myocardial infarction
      • under and over diagnosis of T wave abnormalities
      • errors in lead placement include:
        • placement of the chest electrodes too high can reduce the likelihood of diagnosing left ventricular hypertrophy and increase the likelihood of a non-specific T wave changes
        • swapping the wires of the patient cable is not a rare event: programs will usually recognise when the right and left arm wires have been swapped, but they frequently fail to recognise when wires going to chest leads are swapped or those going to the right arm and the right leg are swapped.
    • in conclusion, computer-assisted ECG interpretation can identify important cardiac abnormalities. However, errors are common and the interpretation should not be accepted without consideration of its clinical context and a visual inspection of the ECG, preferably by an expert (1)

Reference:

1. British Heart Foundation Factfile (12/05). Computer-Assisted ECG Interpretation.