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

Table 1 Design parameters for two three-stage TAMS ^a trials

						I and D outcomes: OS^b				I outcomes: PFS^c; D outcomes: OS^b
Design	Stage	α _i ^d	ω _i ^e	r _i ^f	δ _i ^g	e _i ^h	t _i ⁱ	N _i ^j	H R ¹ ^k	e _i ^h	t _i ⁱ	N _i ^j	H R ¹ ^k
1	1	0.50	0.95	250	1.00	73	1.53	382	0.75	73	1.53	382	0.75
	2	0.25	0.95	250	0.92	140	2.62	566	0.75	140	2.62	566	0.75
	3	0.025	0.90	250	0.84	262	3.40	851	0.75	264	4.36	1091	0.75
2	1	0.2	0.95	250	0.91	159	2.45	612	0.75	159	2.45	612	0.75
	2	0.1	0.95	250	0.89	217	3.00	750	0.75	217	3.00	750	0.75
	3	0.025	0.90	250	0.84	262	3.40	851	0.75	264	4.36	1091	0.75

^aTwo-arm multi-stage;
^boverall survival;
^cprogression-free survival;
^dnominal significance level at stage i;
^enominal power at stage i;
^frate of patient accrual per unit time during stage i;
^gpredefined threshold;
^hcumulative number of control arm events required at end of stage i;
ⁱduration (in time units) up to the end of stage i;
^jcumulative number of patients accrued to control arm by end of stage i;
^ktarget hazard ratio under H ₁, the target hazard ratio for inefficacy HR ⁰is 1 in all scenarios.

ISSN: 1745-6215