The relative value of predischarge clinical variables, bicycle ergometry, radionuclide ventriculography and 24 hour ambulatory electrocardiographic monitoring for predicting survival during the first year in 351 hospital survivors of acute myocardial infarction was assessed. Discriminant function analysis showed that in patients eligible for stress testing the extent of blood pressure increase during exercise slightly improved the predictive accuracy beyond that of simple clinical variables (history of previous myocardial infarction, persistent heart failure after the acute phase of infarction and use of digitalis at discharge), whereas radionuclide ventriculography and 24 hour electrocardiographic monitoring did not. The predictive value for mortality was 12% with clinical variables alone and 15% with the stress test added. Radionuclide ventriculography and 24 hour electrocardiographic monitoring were slightly additive to clinical information in the whole group of patients independent of the eligibility for stress testing (predictive value for mortality 24% with clinical variables alone and 26% with radionuclide ejection fraction and 24 hour electrocardiographic monitoring added). It is concluded that the appropriate use of simple clinical variables and stress testing is sufficient for risk stratification in postinfarction patients, whereas radionuclide ventriculography and 24 hour electrocardiographic monitoring should be limited to patients not eligible for stress testing.