Skip to main content

Table 7 Simulation results evaluating Rule 3 when the threshold probability is set at 0.80

From: Evaluation of a multi-arm multi-stage Bayesian design for phase II drug selection trials – an example in hemato-oncology

Sample size True benefit Posterior mean estimate biases Enrolled sample sizes % Early stopping
  dB dC pA pB pC A B C B C
40 −0.15 0.00 0.0083 0.0320 0.0562 39.5748 38.6963 33.8148 3.64 % 18.64 %
  −0.05 0.00 0.0063 0.0495 0.0568 38.7270 35.8578 33.7310 12.09 % 18.95 %
  0.00 0.00 0.0049 0.0555 0.0581 38.1052 33.9250 33.5867 18.38 % 19.20 %
  0.05 0.00 0.0016 0.0632 0.0563 37.5141 31.3987 33.8662 26.92 % 18.51 %
  0.10 0.00 −0.0008 0.0691 0.0570 36.8763 28.0621 33.8537 38.65 % 18.55 %
  0.15 0.00 0.0001 0.0720 0.0578 35.9576 24.3662 33.6137 51.87 % 19.22 %
  0.20 0.00 −0.0002 0.0673 0.0561 35.5246 20.7723 33.8744 64.82 % 18.40 %
  0.25 0.00 −0.0017 0.0575 0.0568 34.9601 16.9213 33.7372 77.56 % 18.91 %
  0.30 0.00 −0.0043 0.0409 0.0575 34.3862 13.5320 33.6114 87.33 % 19.19 %
  0.35 0.00 −0.0016 0.0196 0.0560 34.5481 10.7458 34.0309 94.31 % 17.94 %
  0.40 0.00 −0.0033 −0.0069 0.0547 34.2171 8.5161 33.9205 97.80 % 18.43 %
  0.45 0.00 −0.0023 −0.0372 0.0542 34.1941 6.8534 33.9756 99.23 % 18.33 %
100 −0.05 0.00 0.0032 0.0229 0.0559 98.8583 96.4597 82.6701 3.69 % 18.83 %
  −0.05 0.00 −0.0006 0.0433 0.0563 96.1819 89.1834 82.5699 11.51 % 18.92 %
  0.00 0.00 −0.0038 0.0577 0.0571 94.3563 82.2688 82.2694 19.21 % 19.14 %
  0.05 0.00 −0.0060 0.0689 0.0570 92.3215 73.7509 82.5095 29.51 % 19.02 %
  0.10 0.00 −0.0075 0.0778 0.0567 89.5994 63.0008 82.5545 43.27 % 18.97 %
  0.15 0.00 −0.0088 0.0822 0.0574 87.6736 49.5431 82.5278 61.69 % 18.88 %
  0.20 0.00 −0.0102 0.0771 0.0530 86.2653 37.2856 83.2202 77.99 % 18.15 %
  0.25 0.00 −0.0113 0.0636 0.0558 84.2998 25.2895 82.5265 91.74 % 19.02 %
  0.30 0.00 −0.0112 0.0448 0.0576 83.4348 17.4124 82.3130 97.39 % 19.14 %
  0.35 0.00 −0.0114 0.0204 0.0558 83.0775 11.9089 82.5449 99.58 % 18.89 %
  0.40 0.00 −0.0102 −0.0073 0.0552 83.3131 9.0623 82.9624 99.95 % 18.47 %
  0.45 0.00 −0.0111 −0.0375 0.0585 82.5830 6.9659 82.3697 100.00 % 19.09 %
  1. pA, pB and pC refer to the posterior means of response probability in arms A, B and C, respectively; δ*=0.15. Bold data refer to the null hypothesis of absence of any treatment difference (dB=dC=0), or to the alternative hypothesis of an expected true 0.15 benefit of treatment B (dB=0.15)