Frequentist statistics typically examines the null hypothesis that no difference exists between the competing strategies. Bayesian statistics permits the calculation of the probability that a treatment is superior based on observed data and prior information. Output of Bayesian analysis allows the observation of how prior information affects the output, especially as new information builds on the prior information. OBJECTIVES: To further analyze using Bayesian methods clinical trial results for the treatment of asymptomatic carotid artery stenosis (surgery vs. medical management) whose results were previously reported using frequentist methods. METHODS: The outcome of interest was the mean difference in the probability of stroke or perioperative death between carotid endarterectomy (CEA) and aggressive medical management (AMM). The prior distribution came from the results of the Asymptomatic Carotid Atherosclerosis Study (ACAS). The likelihood distribution was sourced from the Asymptomatic Carotid Surgery Trial (ACST). The Metropolis-Hastings Markov chain Monte Carlo sampling algorithm was used to approximate the posterior distribution. A 4% mean difference was chosen as the threshold for clinical-economic significance. In sensitivity analysis, the prior distribution was replaced with results of The European Carotid Surgery Trialists Collaborative Group (ECST) study. RESULTS: The likelihood distribution (ASCT results) had an 86.5% probability that the outcome of interest exceeded the threshold. The posterior distribution had a 93.0% probability. When data from the ECST study formed the prior distribution, the posterior distribution had only a 57.5% probability of exceeding the threshold. This reflects a revised prior which included a lower (relative to ACAS) expectation of stroke in the AMM population as informed by ECST data. CONCLUSIONS: Bayesian analysis allows the incorporation of differing prior information, whether representing clinical opinion or clinical trials. This permits the observation of how differing priors affect the posterior distribution and, hence, interpretations of predicted clinical outcomes between treatments.