We reviewed the literature on clinical trials of beta-adrenergic blockade for treatment of heart failure, seeking evidence of reductions in hospital admissions. To analyze the economic implications of six clinical trials, we developed a stochastic cost model to generate estimates of total medical costs resulting from heart failure and related causes. The model includes inpatient, outpatient, and professional cost estimates based on Medicare claims data, and it is driven by traditional endpoint statistics reported in the clinical trial literature. It provides a common framework for comparing cost effectiveness across clinical trials in the absence of detailed cost information collected in the trial. The incremental expected cost per year of life saved is $3,300 for bisoprolol, $2,500 for metoprolol, and $6,700 for carvedilol. The cost per year of life saved for each compound is well below accepted standards for cost effectiveness. These results are sensitive to the cost of drug therapy and the relative mortality rate for the experimental group. For example, if the relative mortality rate of the experimental group were to increase from the reported 40% to 82%, and if the annual cost of the drug were to decrease from $2,000 to $500, then we estimate that carvedilol would break even and the cost per year of life saved would drop to zero. Whether beta-blocker therapy, as assumed, sustains its differential effectiveness in terms of relative mortality risk beyond the study duration has not been demonstrated.