Skip to main content

Table 3 Simulated operating characteristics for shortlisted designs

From: Improving clinical trials using Bayesian adaptive designs: a breast cancer example

Design Type I error (%) Power (%) Sample size, mean (SD)
Null effect Target effect
Fixed 2.50 80.8 235 (1.9) 235 (1.9)
Small sample size
  Posterior 2.46 81.3 184 (59.8) 211 (25.3)
  PPS 2.26 81.4 149 (59.2) 199 (50.4)
  CPS 2.39 81.8 154 (44.5) 207 (34.8)
  Goldilocks 2.45 81.7 149 (59.0) 194 (47.6)
  PPBS 2.42 81.5 153 (43.3) 205 (29.5)
High power
  Posterior 2.45 82.1 223 (30.5) 213 (20.7)
  PPS 2.44 85.9 198 (29.1) 225 (25.4)
  CPS 2.36 84.8 187 (28.2) 222 (19.2)
  Goldilocks 2.42 85.9 198 (29.0) 224 (17.4)
  PPBS 2.44 85.9 191 (29.4) 223 (18.6)
  1. PPS Predictive probability of success, CPS Conditional probability of success, PPBS Predictive probability of Bayesian success. The target effect represents the scenario of a true median survival of 15 months in the experimental arm and 10 months in the control arm