Related Experiment Videos
Accuracy of computer-interpreted electrocardiography in selecting patients for thrombolytic therapy. MITI Project
P J Kudenchuk1, M T Ho, W D Weaver
1Department of Medicine, University of Washington, Seattle.
Insights
Prehospital computer-interpreted electrocardiograms (ECGs) are feasible for identifying acute myocardial infarction. While slightly less sensitive than physicians, computer ECGs show high specificity and positive predictive value for thrombolytic therapy candidates.
Area of Science:
- Cardiology
- Medical Informatics
- Emergency Medicine
Background:
- Acute myocardial infarction (AMI) requires rapid diagnosis and treatment.
- Prehospital electrocardiograms (ECGs) can aid in early diagnosis.
- Computer interpretation of ECGs offers potential for standardized, rapid analysis.
Purpose of the Study:
- To evaluate the feasibility and accuracy of prehospital computer-interpreted ECGs.
- To assess the diagnostic performance of computer-interpreted ECGs compared to physician interpretation in patients with chest pain.
- To determine the utility of computer-interpreted ECGs for screening candidates for prehospital thrombolytic therapy.
Main Methods:
- A prospective trial involving 1,189 patients with chest pain suggestive of cardiac origin.
- Paramedics performed computer-interpreted ECGs 1.5 ± 1.2 hours after symptom onset.
- Comparison of computer interpretation with expert electrocardiographer diagnosis for ST-segment elevation (acute injury) and exclusion of infarction.
Main Results:
- Computer ECG identified 52% of AMI patients with ST-segment elevation versus 66% by physicians (p<0.001).
- Computer ECG appropriately excluded non-infarction patients with 98% accuracy versus 95% for physicians (p<0.001).
- Positive predictive value was 94% for computer ECG and 86% for physician ECG; negative predictive values were 81% and 85%, respectively.
Conclusions:
- Prehospital computer-interpreted ECGs are feasible and highly specific for screening AMI patients.
- Computer interpretation demonstrates a high positive predictive value for identifying patients eligible for thrombolytic therapy.
- Further enhancements to computer ECG algorithms could improve sensitivity and expedite care for AMI patients.
Abstract:
A prehospital computer-interpreted electrocardiogram (ECG) was obtained in 1,189 patients with chest pain of suspected cardiac origin during an ongoing trial of prehospital thrombolytic therapy in acute myocardial infarction. Electrocardiograms were performed by paramedics 1.5 +/- 1.2 h after the onset of symptoms. Of 391 patients with evidence of acute myocardial infarction, 202 (52%) were identified as having ST segment elevation (acute injury) by the computer-interpreted ECG compared with 259 (66%) by an electrocardiographer (p less than 0.001). Of 798 patients with chest pain but no infarction, 785 (98%) were appropriately excluded by computer compared with 757 (95%) by an electrocardiographer (p less than 0.001). The positive predictive value of the computer- and physician-interpreted ECG was, respectively, 94% and 86% and the negative predictive value was 81% and 85%. Prehospital screening of possible candidates for thrombolytic therapy with the aid of a computerized ECG is feasible, highly specific and with further enhancement can speed the care of all patients with acute myocardial infarction.