A thrombolytic decision tree.

We constructed a decision analysis model based on data in the medical literature to estimate the possible outcomes of thrombolytic therapy in patients 50 to 80 years old with possible myocardial infarction. We used the model to test the most likely effects of treatment (determined by averaging the values in reports of large studies) and the worst effects reported so far. The program begins by asking the patient's age, the hours from the onset of pain, and the probability of acute myocardial infarction. It then provides an opportunity to perform sensitivity analyses by changing the values for these variables and for the probability of death in the absence of thrombolytic therapy, as well as for the probability of major stroke and hemorrhage. The counterintuitive findings observed with this program are that the benefits of thrombolytic therapy increase with age and that young patients derive surprisingly little benefit from it.