Seamless phase 2/3 design for trials with multiple co-primary endpoints using Bayesian predictive power

Table 2 The performance of BPP in comparison with CP approach when the null hypothesis is true

n₁	ρ	Type 1 Error (%)		Stop Percentage(%)		Sample size
n₁	ρ	BPP	CP	BPP	CP	BPP	CP
50	0	< 0.01	< 0.01	78.40	85.54	374.00	266.90
	0.3	< 0.01	< 0.01	68.22	75.65	526.46	413.34
	0.6	0.04	0.03	56.46	64.10	701.74	575.49
100	0	< 0.01	< 0.01	87.38	87.45	289.30	288.20
	0.3	< 0.01	< 0.01	78.17	78.32	427.35	423.77
	0.6	0.01	0.03	67.71	68.01	581.81	571.25
150	0	< 0.01	< 0.01	89.33	86.60	310.05	351.00
	0.3	< 0.01	< 0.01	80.30	77.08	445.50	493.46
	0.6	0.01	0.03	69.91	66.28	600.34	651.15
200	0	< 0.01	< 0.01	93.89	90.26	291.65	346.10
	0.3	< 0.01	< 0.01	85.93	81.64	410.93	475.06
	0.6	0.01	0.02	77.37	71.69	537.13	619.83

The percentage signs were omitted for type 1 error and stop percentage in the table; Sample size represents the sample size used for each group (eg. selected dose group) during the whole phase 2/3 trial

ISSN: 1471-2288