Independent blinded reference standard, applied in all patients from a consecutive appropriate spectrum.
Reference standard: Diagnostic test: Physician read ST-segment and T-wave changes of the ECGs. Physicians assigned a diagnosis of acute myocardial infarction, angina pectoris or not acute ischaemic heart disease to each patient based on clinical presentation. Trainee physicians (in post-graduate years one, two and three, and those physicians identifying themselves as trainees during post-graduate years four and five) were analysed separately.

The evidence


Cardiac ischaemia: physician read ECGs were not very accurate.

Clinical bottom line (level 1b)

  1. A third of patients with suspected abnormal ECGs had them.
  2. In patients with suspected normal ECGs, a normal ST segment according to physicians made it more likely (LR+3.34) , and an abnormal one made it less likely (LR-0.38) .
  3. A normal T wave according to physicians made it more likely (LR+3.02) , and an abnormal one made it less likely (LR-0.34) .
  4. Trainee physicians were slightly less accurate than other physicians.
Jayes et al: Journal of General Internal Medicine 1992; 7: 387-392
Expires March 2003

The study

Setting: four medical school teaching hospitals and two nonteaching hospitals, USA

2320 patients (aged ?, ?% male) ECGs of patients presenting to emergency departments with complaints of chest, jaw or left arm pain, shortness of breath, abdominal pain or nausea and fainting or lightheadedness.

Excluded if
  • paced rhythms
  • left ventricular hypertrophy
  • right or left bundle branch block
  • men <30 years old
  • women <40 years old



    • Independent blinded reference standard, applied in all patients from a consecutive appropriate spectrum.
    Reference standard:
    • ECGs read by electrocardiographers. The ECG readings were compared with: ST-segment depression or elevation of 1 mm or more, ST-segment straightening or depression of less than 1 mm, T-wave depression of 1 mm or more, T-wave flattening and T-wave elevation.
    Diagnostic test: Physician read ST-segment and T-wave changes of the ECGs. Physicians assigned a diagnosis of acute myocardial infarction, angina pectoris or not acute ischaemic heart disease to each patient based on clinical presentation. Trainee physicians (in post-graduate years one, two and three, and those physicians identifying themselves as trainees during post-graduate years four and five) were analysed separately.

    The evidence


    diagnostic test normal ECG abnormal ECG LR+
    (95% CI)    		 post-test probability LR-
    (95% CI)    		 post-test probability
    physician reading of ST segments (all physicians) 855 142 3.34
    (2.87 to 3.88)     		86.0% 0.38
    (0.35 to 0.42)     		41.0%
    trainee physician reading of ST-segments 226 64 2.57
    (2.05 to 3.20)     		78.0% 0.43
    (0.36 to 0.51)     		37.0%
    other physicians reading of ST segments 577 69 4.04
    (3.24 to 5.03)     		89.0% 0.36
    (0.32 to 0.40)     		43.0%
    physician reading of T waves (all patients) 878 178 3.02
    (2.65 to 3.45)     		83.0% 0.34
    (0.31 to 0.38)     		36.0%
    trainee physicians readings of T waves 225 83 2.67
    (2.21 to 3.23)     		73.0% 0.30
    (0.24 to 0.38)     		23.0%
    other physicians readings of T waves 602 83 3.39
    (2.79 to 4.12)     		88.0% 0.31
    (0.31 to 0.39)     		43.0%
    normal triage with correct or incorrect reading of abnormal ST segments 328 40 1.36
    (1.10 to 1.69)     		89.0% 0.57
    (0.43 to 0.77)     		77.0%
    normal triage with abnormal T waves 270 45 1.10
    (0.91 to 1.32)     		86.0% 0.82
    (0.57 to 1.17)     		82.0%
    total 997 910

    Comments

    1. Agreement bewteen observers was good (>90%). No assessment of computerised readings was made.

    Citation

    1. Jayes RL, Larsen GC, Beshansky JR, et al: Physician electrocardiogram reading in the emergency department- Accuracy and effect on triage decisions: findings from a multicenter study. Journal of General Internal Medicine 1992; 7: 387-392
    Contributor: Clare Wotton and Musab Hayatli, February 2000
    Reviewer: William Rhoton

    Clinical Question.
    Patient possible acute cardiac ischaemia
    Intervention or Exposure physician reading of ECG
    Comparison electrocardiographer reading
    Outcome diagnosis
diagnostic test normal ECG abnormal ECG LR+
(95% CI)		 post-test probability LR-
(95% CI)		 post-test probability
physician reading of ST segments (all physicians) 855 142 3.34
(2.87 to 3.88) 		86.0% 0.38
(0.35 to 0.42) 		41.0%
trainee physician reading of ST-segments 226 64 2.57
(2.05 to 3.20) 		78.0% 0.43
(0.36 to 0.51) 		37.0%
other physicians reading of ST segments 577 69 4.04
(3.24 to 5.03) 		89.0% 0.36
(0.32 to 0.40) 		43.0%
physician reading of T waves (all patients) 878 178 3.02
(2.65 to 3.45) 		83.0% 0.34
(0.31 to 0.38) 		36.0%
trainee physicians readings of T waves 225 83 2.67
(2.21 to 3.23) 		73.0% 0.30
(0.24 to 0.38) 		23.0%
other physicians readings of T waves 602 83 3.39
(2.79 to 4.12) 		88.0% 0.31
(0.31 to 0.39) 		43.0%
normal triage with correct or incorrect reading of abnormal ST segments 328 40 1.36
(1.10 to 1.69) 		89.0% 0.57
(0.43 to 0.77) 		77.0%
normal triage with abnormal T waves 270 45 1.10
(0.91 to 1.32) 		86.0% 0.82
(0.57 to 1.17) 		82.0%
total 997 910

Comments

  1. Agreement bewteen observers was good (>90%). No assessment of computerised readings was made.

Citation

  1. Jayes RL, Larsen GC, Beshansky JR, et al: Physician electrocardiogram reading in the emergency department- Accuracy and effect on triage decisions: findings from a multicenter study. Journal of General Internal Medicine 1992; 7: 387-392
Contributor: Clare Wotton and Musab Hayatli, February 2000
Reviewer: William Rhoton

Clinical Question.
Patient possible acute cardiac ischaemia
Intervention or Exposure physician reading of ECG
Comparison electrocardiographer reading
Outcome diagnosis