Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
December 19, 1991New England Journal of MedicineOpen Access

The Diagnostic Performance of Computer Programs for the Interpretation of Electrocardiograms

View Full Paper
Ask AI
Bookmark
Share

Why the study?

Do electrocardiographic computer programs match the diagnostic accuracy of cardiologists in interpreting ECGs for various cardiac disorders?

Population

1220 clinically validated cases of various cardiac disorders, including control patients, left ventricular…

Comparison

Interpretation of ECGs by nine… vs Interpretation of ECGs by eight cardiologists

Design

Cross-sectional

Authors

JWJos L. WillemsUnited States Department of Veterans AffairsCACassiano Abreu‐LimaUniversidade do PortoPAP ArnaudHôpital Privé Jacques Cartier

Discussion

Loading...

Member takes

Implication

ECG programs warrant cardiologist oversight; leaves open whether optimized algorithms achieve equivalence in clinical use.

Key Points

  • To systematically assess the performance of various computer programs in interpreting electrocardiograms compared to cardiologists.
  • Comparative analysis of nine computer programs and eight cardiologists interpreting 1220 ECG cases.
  • Inclusion of ECGs from different cardiac disorders.
  • Validation against independent clinical diagnoses.
  • Computer programs correctly classified a median of 91.3% of ECGs, lower than cardiologists' 96.0% (P<0.01).
  • Median sensitivity for diagnosing left ventricular hypertrophy was 56.6% for programs vs. 63.9% for cardiologists (P<0.02).
  • The best programs' performance nearly matched that of the most accurate cardiologists.

Structured PICO

Do electrocardiographic computer programs match the diagnostic accuracy of cardiologists in interpreting ECGs for various cardiac disorders?

P
Population
1220 clinically validated cases of various cardiac disorders, including control patients (n=382), left ventricular hypertrophy (n=183), right ventricular hypertrophy (n=55), biventricular hypertrophy (n=53), anterior myocardial infarction (n=170), inferior myocardial infarction (n=273), combined myocardial infarction (n=73), and combined infarction and hypertrophy (n=31).
I
Intervention
Interpretation of ECGs by nine electrocardiographic computer programs
C
Comparator
Interpretation of ECGs by eight cardiologists
O
Outcome
Diagnostic performance (percentage of ECGs correctly classified, sensitivity, and total accuracy level) compared with clinical diagnoses made independently of the ECGs

While computer programs for ECG interpretation generally have lower overall diagnostic accuracy than cardiologists, the best performing programs can achieve comparable results.

Cite This Study

Willems et al. (1991) studied this question.

synapsesocial.com/papers/69fd28f230a474415f89dfd5https://doi.org/10.1056/nejm199112193252503

Topics

Artificial intelligence in cardiology
View Full Paper
Ask AI
Bookmark
Share

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

  1. 1Effect of combining elecrocardiographic interpretation results on diagnostic accuracy1988 · 32 citations
  2. 2A Method for Evaluating Computer Programs for Electrocardiographic Interpretation1974 · 48 citations
  3. 3Comparative evaluation of the IBM (12-lead) and Royal Infirmary (orthogonal three-lead) ECG computer programs.1981 · 17 citations
  4. 4Criteria for intraventricular conduction disturbances and pre-excitation1985 · 395 citations
  5. 5A reporting system on patients evaluated for coronary artery disease. Report of the Ad Hoc Committee for Grading of Coronary Artery Disease, Council on Cardiovascular Surgery, American Heart Association1975 · 3,260 citations