Skip to main content

Table 6 Bootstrap results for the stopped trials based on ICON3 and RE04 in a two-stage design

From: Impact of lack-of-benefit stopping rules on treatment effect estimates of two-arm multi-stage (TAMS) trials with time to event outcome

  

I and D outcomes are OSa

I outcome is PFSb, D outcome is OSa

 
   

%Stop

Estimated H R D for trials stopped at stage 1

 

%Stop

Estimated H R D for trials stopped at stage 1

   

at

At interim point

After follow-up

 

at

At interim point

After follow-up

α 1 c

δ 1 d

e 1 e

stage 1

Mean (centiles)

%Bias

Mean (centile)

%Bias

e 1 e

stage 1

Mean (centiles)

%Bias

Mean (centiles)

%Bias

a) ICON3 –

H R D (95% bootstrap CI): 0.97(0.87–1.09)

0.50

1.00

118

38

1.12(1.00,1.40)

15

0.96(0.84,1.10)

1

116

55

1.27(0.88,1.83)

31

0.93(0.79,1.10)

− 4

0.40

0.97

153

55

1.08(0.97,1.31)

11

0.98(0.86,1.11)

1

152

52

1.15(0.83,1.56)

19

0.95(0.81,1.10)

− 2

0.30

0.94

195

52

1.03(0.94,1.22)

6

0.98(0.87,1.10)

1

194

33

1.09(0.81,1.45)

12

0.97(0.85,1.11)

0

0.20

0.91

252

42

0.99(0.91,1.15)

2

0.99(0.89,1.10)

2

250

12

1.16(0.89,1.46)

20

1.00(0.89,1.12)

3

0.10

0.89

342

44

0.95(0.89,1.08)

− 2

0.99(0.90,1.10)

2

339

29

1.06(0.86,1.31)

9

1.00(0.90,1.12)

3

b) RE04 –

H R D (95% bootstrap CI): 1.05(0.90–1.21)

0.50

1.00

118

34

1.09(1.00,1.27)

4

1.10(0.95,1.26)

5

116

33

0.92(0.66,1.32)

− 12

1.07(0.89,1.30)

2

0.40

0.97

155

36

1.06(0.97,1.26)

1

1.09(0.96,1.24)

4

152

67

0.89(0.67,1.19)

− 15

1.02(0.86,1.23)

−3

0.30

0.95

199

71

1.05(0.95,1.23)

0

1.05(0.93,1.20)

0

196

79

0.95(0.74,1.23)

− 10

1.05(0.90,1.24)

0

0.20

0.93

259

84

1.05(0.95,1.23)

0

1.06(0.94,1.21)

1

255

73

0.96(0.78,1.19)

− 9

1.04(0.90,1.21)

− 1

0.10

0.91

355

79

1.05(0.92,1.26)

0

1.06(0.92,1.22)

1

351

97

1.02(0.86,1.22)

− 3

1.03(0.89,1.20)

− 2

  1. aOverall survival;
  2. bprogression-free survival;
  3. cone-sided significance level at stage 1;
  4. dpredefined threshold at stage 1;
  5. ecumulative number of control arm events required at end of stage 1.