Evaluation of a multi-arm multi-stage Bayesian design for phase II drug selection trials – an example in hemato-oncology

Table 6 Simulation results evaluating Rule 2 when the threshold probability is set at 0.10

Sample size	True benefit		Posterior mean estimate biases			Enrolled sample sizes			% Early stopping
	dB	dC	p_A	p_B	p_C	A	B	C	B	C
40	−0.20	0.00	0.0625	0.0585	−0.0123	30.6959	13.8404	29.9078	92.23 %	34.59 %
	−0.15	0.00	0.0601	0.0311	−0.0120	31.2312	18.3044	29.5173	79.84 %	35.56 %
	−0.10	0.00	0.0567	0.0107	−0.0119	32.3427	22.4957	29.6238	63.89 %	35.43 %
	−0.05	0.00	0.0506	−0.0027	−0.0126	33.5772	26.5306	29.7194	48.34 %	35.68 %
	0.00	0.00	0.0480	−0.0122	−0.0123	34.5352	29.3818	29.5852	35.89 %	35.74 %
	0.05	0.00	0.0386	−0.0166	−0.0109	35.7946	32.2020	29.7191	25.14 %	35.03 %
	0.10	0.00	0.0337	−0.0178	−0.0107	36.7805	34.2756	29.8086	17.75 %	34.52 %
	0.15	0.00	0.0300	−0.0176	−0.0122	37.6091	35.9146	29.5542	12.22 %	35.64 %
	0.20	0.00	0.0268	−0.0180	−0.0111	38.1120	36.8978	29.8112	8.92 %	34.91 %
100	−0.20	0.00	0.0704	0.0581	−0.0229	66.7855	15.4632	65.8818	99.60 %	42.83 %
	−0.15	0.00	0.0673	0.0252	−0.0230	68.2876	23.4374	66.0635	96.28 %	42.43 %
	−0.10	0.00	0.0601	0.0007	−0.0229	71.7312	36.8674	66.5112	84.36 %	42.43 %
	−0.05	0.00	0.0571	−0.0160	−0.0231	75.3970	51.5977	65.8379	64.18 %	43.02 %
	0.00	0.00	0.0467	−0.0237	−0.0240	81.2572	66.4872	65.8151	42.25 %	42.95 %
	0.05	0.00	0.0379	−0.0255	−0.0233	86.5612	76.7592	66.9099	27.22 %	41.95 %
	0.10	0.00	0.0338	−0.0241	−0.0242	89.8261	83.4433	64.7309	18.46 %	43.92 %
	0.15	0.00	0.0279	−0.0229	−0.0232	92.3335	88.1690	66.3140	12.66 %	42.48 %
	0.20	0.00	0.0234	−0.0208	−0.0245	94.1323	91.5214	65.5106	8.98 %	43.34 %

p_A, p_B and p_C refer to the posterior means of response probability in arms A, B and C, respectively; Δ=0. Bold data refer to the null hypothesis of absence of any treatment difference (dB=dC=0), or to the alternative hypothesis of an expected true 0.15 benefit of treatment B (dB=0.15)

ISSN: 1471-2288