Skip to main content

Table 3 Operating characteristics for Bayesian designs for the Collaborative Ankle Support Trial

From: Bayesian adaptive designs for multi-arm trials: an orthopaedic case study

ScenarioProportion stopping early for efficacyProportion stopping early for futilityMSEMean proportion allocated to controlMean proportion allocated to bootMean proportion allocated to braceMean proportion allocated to below-knee cast
Null (50, 50, 50, 50) 
 Design 1NANANA0.250.250.250.25
 Design 20.00630.013NA0.250.250.250.25
 Design 30.00250.0248NA0.360.210.210.21
 Design 40.00220.0125NA0.330.220.220.22
 Design 50.00150.0134NA0.370.210.210.21
 Design 60.01170NA0.250.250.250.25
One arm works, 10 more (50, 50, 50, 60)
 Design 1NANA2.770.250.250.250.25
 Design 20.73205.030.250.250.250.25
 Design 30.69190.00223.680.400.110.110.39
 Design 40.79603.560.390.110.110.39
 Design 50.790903.290.360.100.100.44
 Design 60.997202.340.130.130.130.61
One arm works, 5 more (50, 50, 50, 55)
 Design 1NANANA0.250.250.250.25
 Design 20.10910.0015NA0.250.250.250.25
 Design 30.06240.0052NA0.390.130.130.35
 Design 40.07330.0008NA0.370.140.140.35
 Design 50.06770.001NA0.370.130.130.37
 Design 60.56540NA0.150.150.150.54
Better best (50, 55, 60, 65)
 Design 1NANA3.290.250.250.250.25
 Design 20.795305.110.250.250.250.25
 Design 30.68430.00014.160.370.100.190.34
 Design 40.817704.050.360.110.190.35
 Design 50.806903.860.360.100.180.37
 Design 60.898201.950.070.100.220.61
One worse, others work (50, 45, 55, 60) 
 Design 1NANA2.960.250.250.250.25
 Design 20.634105.100.250.250.250.25
 Design 30.61230.00053.860.380.070.180.36
 Design 40.687203.670.380.070.180.37
 Design 50.685603.430.360.070.170.40
 Design 60.897201.950.100.070.220.61
All work, two similar (50, 55, 60, 60)
 Design 1NANA3.390.250.250.250.25
 Design 20.270105.240.250.250.250.25
 Design 30.26920.00043.870.360.110.270.26
 Design 40.27440.27443.730.350.120.270.27
 Design 50.274403.540.370.100.260.27
 Design 60.549302.930.060.120.410.41
  1. MSE mean square error, NA not applicable