Simple computer-assisted diagnosis of acute myocardial infarction in patients with acute thoracic pain

Methods Inf Med. 1992 Nov;31(4):263-7.


In order to minimize the initial diagnostic uncertainty in patients suspected of having acute myocardial infarction, we prospectively extracted predictive variables from previous history, ECG, and clinical chemical parameters of 87 patients, who were admitted for acute thoracic pain. The variables thus extracted were: Thoracic pain in previous history, duration of pain, white blood cell count, blood glucose, creatine-kinase, and S-T elevation in the ECG. These parameters were used for formulating a mathematical model based upon univariate and multivariate statistical methods. The sensitivity of the model in the study population was 95% and the specificity 77%. Correct classification was achieved in 89% of cases. In a second phase, the prognostic index was prospectively evaluated in a second set of 122 consecutive patients. In this test population, the sensitivity was 89% and the specificity 86%. 87% of patients were classified correctly.

MeSH terms

  • Adult
  • Aged
  • Chest Pain / etiology*
  • Diagnosis, Computer-Assisted*
  • Electrocardiography
  • Female
  • Humans
  • Male
  • Middle Aged
  • Myocardial Infarction / complications
  • Myocardial Infarction / diagnosis*
  • Prospective Studies
  • Sensitivity and Specificity