TABLE 3.

Results of the simulations that revisit the NCT01257230 trial using the ORD

Design powered to reject all hypotheses
${\theta }^{(1)}$	${\theta }^{(2)}$	Max. SS	Stages	Reject all	Reject $H_{01}$ not $H_{02}$	Reject at least one $H_{0k}$	ESS
0	0	474	1	0.005	0.020	0.025	474.00
		534	2	0.004	0.021	0.025	316.39
120	0	474	1	0.025	0.856	0.881	474.00
		534	2	0.025	0.854	0.879	371.83
120	120	474	1	0.803	0.078	0.881	474.00
		534	2	0.802	0.081	0.883	399.81

Design powered to reject at least one hypothesis
${\theta }^{(1)}$	${\theta }^{(2)}$	Max. SS	Stages	Reject all	Reject $H_{01}$ not $H_{02}$	Reject at least one $H_{0k}$	ESS
0	0	381	1	0.005	0.021	0.025	381.00
		426	2	0.004	0.021	0.025	252.43
120	0	381	1	0.025	0.779	0.803	381.00
		426	2	0.024	0.774	0.798	304.67
120	120	381	1	0.691	0.112	0.803	381.00
		426	2	0.684	0.117	0.802	331.89

Note: For the two‐stage design, the triangular bounds are used. Proportions refer to $10^6$ replications and values of interest are in bold.

Abbreviations: ESS, expected sample size; Max. SS, maximum sample size.