Based on the thrombolytic predictive instrument (TPI), we sought to create electrocardiographically based, real-time decision support to immediate identification of patients with ST-segment elevation myocardial infarction (STEMI) likely to benefit from primary percutaneous coronary intervention (PCI) compared with thrombolysis. Using data from the Atlantic Cardiovascular Patient Outcomes Research Team (C-PORT) Trial, we tested a mathematical model predicting mortality in patients with STEMI if treated with PCI and if treated with thrombolytic therapy. We adapted the model for incorporation into computerized electrocardiograms as a PCI-TPI. For patients with STEMI in the C-PORT Trial, the model yielded unbiased mortality predictions: for those receiving thrombolysis, it predicted 6.3% mortality and actual mortality was 6.0% (95% confidence interval 3.0 to 10.6); for those receiving PCI, it predicted 4.5% mortality and actual mortality was 3.9% (95% confidence interval 1.4 to 8.2). Excellent discrimination was reflected by its receiver operating characteristic curve area of 0.86. According to the model, and validated by actual trial outcomes, 1/3 of subjects accounted for all the mortality benefit from PCI. In conclusion, for STEMI, the PCI-TPI accurately predicts mortality for treatment with PCI and with thrombolytic therapy. Incorporated into electrocardiogram, it may assist targeting PCI to those who benefit most and identifying patients before hospitalization for whom a receiving hospital should prepare for PCI.