What inference for two-stage phase II trials?

Table 1 Performance of the different methods when second stage sample size was different from planned: average over the different design scenarios and differences between the planned and attained second stage sample size

Property	Method	Π= Π ₀	Π= Π ₀ + δ
Bias	${\hat{Π}}_{m}$	−0.015	−0.005
	${\hat{Π}}_{g}$	−0.004	0.001
	${\hat{Π}}_{u}$	0.000	0.000
	${\hat{Π}}_{c}$	−0.029	−0.012
	${\hat{Π}}_{p}$	−0.028	−0.009
	${\hat{Π}}_{k}$	−0.009	−0.012
RMSE	${\hat{Π}}_{m}$	0.060	0.071
	${\hat{Π}}_{g}$	0.063	0.067
	${\hat{Π}}_{u}$	0.071	0.067
	${\hat{Π}}_{c}$	0.061	0.076
	${\hat{Π}}_{p}$	0.062	0.064
	${\hat{Π}}_{k}$	0.062	0.070
Rejection probability	p _n	0.033	0.882
	_{p m}	0.036	0.887
	_{p u}	0.036	0.887
	p _c	0.012	0.800
	_{p k}	0.035	0.885
Coverage probability	Naive exact	0.940	0.916
	Stage-wise	0.937	0.933
	Mid-p	0.916	0.895
	Conditional exact	0.952	0.906
	Conditional score	0.935	0.851
	Conditional mid-p	0.936	0.860
	Koyama–Chen	0.937	0.931

ISSN: 1471-2288