The standard approach to evaluating the statistical significance of randomized clinical trials is to determine whether the difference in measured outcomes between the treatment and control groups is meaningful, where “meaningful” is typically taken to mean a p-value of 5% or less. Why 5%? And why the same value regardless of whether the trial is for acne medication or a potentially life-saving therapy for pancreatic cancer? Bayesian decision analysis provides a framework for making decisions about therapeutic efficacy by weighing the impact of false positives against false negatives using patient values and burden of disease instead of arbitrary p-values. The practical uses of this framework will be illustrated with specific applications in oncology and Parkinson’s disease.