Incorporating Patient Preferences via Bayesian Decision Analysis

Optimal size and power for two-armed balanced fixed sample clinical trial according to Bayesian decision analysis (BDA) in which the expected loss of false positives and negatives is jointly minimized. The left figure displays the expected cost to patients as a function of the false positive rate (the BDA optimal value minimizes this expected cost), and the right figure displays the true positive rate (or power) as a function of the false positive rate. For an interactive version of these graphs, see the Supplemental Material.