The diagnostic and therapeutic approach to the problem of mural thrombi after acute myocardial infarction is uncertain. It is clear that the main therapeutic goal is the prevention of embolic strokes. Although it is known that the incidence of thrombi is greatest after anterior wall infarctions, there is uncertainty concerning (1) the probability of a mural thrombus; (2) the sensitivity and specificity of echocardiography in making the diagnosis; (3) the probability that a thrombus will embolize and result in a cerebrovascular accident (CVA); (4) the efficacy of warfarin in preventing embolization; and (5) the probability of bleeding with and without warfarin. To study this problem in patients who have had an anterior wall myocardial infarction, a model was created in which reasonable estimates for the unknown parameters were determined from published medical studies. The model was designed to consider patients if they were or were not treated during the initial hospitalization with heparin. The probability of thrombus was estimated at 0.30, sensitivity and specificity of echocardiography at 0.85 and 0.85, probability that a thrombus will embolize at 0.15, efficacy of anticoagulation of 0.75, probability of bleeding with warfarin at 0.03 and probability of bleeding without warfarin at 0.005. Probabilities of a CVA and of bleeding with and without warfarin were determined if all patients were anticoagulated, if patients with positive echocardiographic results were treated, if patients with negative echocardiographic results were treated and if echocardiographically guided therapy was instituted in which patients with positive echocardiographic results are treated and patients with negative results are not treated.(ABSTRACT TRUNCATED AT 250 WORDS)