Skip to main content

Table 1 Estimated type I error rates, where pw,F is the cumulative probability of stopping for futility at look w or earlier, pE is the probability of stopping early for efficacy and p12m is the probability of stopping for efficacy at the end of the study; N=85, for (a) one look N1=60,N2=45,N3=25, (b) two looks N1=(55,70),N2=(40,55),N3=(20,35) and (c) three looks, N1=(50,65,75),N2=(35,50,60),N3=(15,30,40), ρ=ρ13=ρ23=ρ12 and \(\sigma ^{2}_{1} = \sigma ^{2}_{2} = \sigma ^{2}_{3} = 20\) (10,000 simulations)

From: An adaptive two-arm clinical trial using early endpoints to inform decision making: design for a study of sub-acromial spacers for repair of rotator cuff tendon tears

Futility bound (\(\alpha ^{*}_{L}\))

ρ

pE

p1,F

p2,F

p3,F

p12m

(a) One look; \(\alpha ^{*}_{U}=(0.001,0.025)\)

(0.0,0.975)

0.0

0.002

0.000

-

-

0.025

(0.5,0.975)

0.0

0.002

0.504

-

-

0.023

(0.0,0.975)

0.5

0.002

0.000

-

-

0.026

(0.5,0.975)

0.5

0.002

0.504

-

-

0.026

(b) Two looks; \(\alpha ^{*}_{U}=(0,0.001,0.025)\)

(0.0,0.0,0.975)

0.0

0.001

0.000

0.000

-

0.025

(0.2,0.5,0.975)

0.0

0.001

0.202

0.499

-

0.025

(0.0,0.0,0.975)

0.5

0.001

0.000

0.000

-

0.024

(0.2,0.5,0.975)

0.5

0.002

0.199

0.505

-

0.025

(c) Three looks; \(\alpha ^{*}_{U}=(0,0,0.001,0.025)\)

(0.0,0.0,0.0,0.975)

0.0

0.001

0.000

0.000

0.000

0.024

(0.1,0.3,0.5,0.975)

0.0

0.002

0.110

0.306

0.503

0.025

(0.0,0.0,0.0,0.975)

0.5

0.001

0.000

0.000

0.000

0.025

(0.1,0.3,0.5,0.975)

0.5

0.001

0.108

0.307

0.506

0.025