Table 2.

Operational characteristics of the proposed adaptive method based on the simulation results.

Simulation condition		Interim outcome	Probability (per cent)	Proposed adaptive method
				Average Z value	Power∗1 (per cent)	ASN∗2
Diff. of mean	Com. SD
0	8.1	Futility	66.83	–	0.00	19.0
		Unfavorable	20.00	0.474	0.57	47.0
		Promising	9.64	0.783	0.97	70.4
		Favorable	3.53	1.358	0.86	47.0
		Total	100.00	0.658	2.40	30.5
5.5	8.1	Futility	5.04	–	0.00	19.0
		Unfavorable	12.32	2.470	9.18	47.0
		Promising	24.24	3.667	23.59	67.3
		Favorable	58.40	3.749	57.16	47.0
		Total	100.00	3.562	89.93	50.5
		Total	100.00	4.491	98.31	49.4
0	10	Futility	66.55	–	0.00	19.0
		Unfavorable	20.65	0.495	0.70	47.0
		Promising	9.41	0.721	0.72	70.3
		Favorable	3.39	1.462	0.91	47.0
		Total	100.00	0.657	2.33	30.6
5.5	10	Futility	10.87	–	0.00	19.0
		Unfavorable	19.23	2.132	11.24	47.0
		Promising	27.33	3.119	24.91	68.0
		Favorable	42.57	3.301	39.93	47.0
		Total	100.00	2.993	76.08	49.7
0	12	Futility	66.71	–	0.00	19.0
		Unfavorable	20.36	0.487	0.70	47.0
		Promising	9.47	0.741	0.69	70.2
		Favorable	3.46	1.385	0.88	47.0
		Total	100.00	0.653	2.27	30.5
5.5	12	Futility	16.82	–	0.00	19.0
		Unfavorable	22.65	1.878	10.51	47.0
		Promising	27.24	2.716	21.94	68.6
		Favorable	33.29	2.939	28.92	47.0
		Total	100.00	2.577	61.37	48.2

*1: Power unconditional on interim data.

*2: Average sample size.

Futility boundary = 0.05; Unfavorable boundary = 0.42; alpha = 0.05 (2-sided); Original N per group = 47; N per group included in IA = 19; Maximum N per group = 75; number of simulations for each simulation condition = 10000.