A Bayesian approach to pilot-pivotal trials for bioequivalence assessment

Table 1 Probability of making a Go decision (%) to implement a pivotal trial, based on different methods to establish average bioequivalence

Comparator	\(\eta\)	(w, 1 - w)	Sc. 1	Sc. 2	Sc. 3	Sc. 4	Sc. 5	Sc. 6	Sc. 7	Sc. 8	Sc. 9
Comparator	\(\eta\)	(w, 1 - w)	GMR = 0.5	GMR = 0.7	GMR = 0.8	GMR = 0.9	GMR = 1.0	GMR = 1.1	GMR = 1.25	GMR = 1.43	GMR = 2.0
CI method (\(\alpha\) = 5%)
C\(_\text {max}\)			0	0.24	4.23	20.35	31.19	21.74	4.31	0.19	0
AUC			0	0.01	4.45	51.68	88.20	56.90	5.20	0	0
CI method (\(\alpha\) = 10%)
C\(_\text {max}\)			0	0.58	9.11	38.61	56.28	41.00	9.81	0.55	0
AUC			0	0.03	9.33	67.84	95.52	72.91	9.94	0.05	0
CI method (\(\alpha\) = 20%)
C\(_\text {max}\)			0	1.71	19.56	59.70	79.67	62.38	20.35	1.67	0
AUC			0	0.14	19.37	82.84	98.87	86.42	20.03	0.15	0
PAR method
C\(_\text {max}\)			0	1.22	16.69	55.44	76.41	58.89	17.90	1.29	0
AUC			0	0.10	16.53	80.18	98.56	84.10	17.28	0.11	0
Proposed method
C\(_\text {max}\)	0.85	(0.9, 0.1)	0	1.99	21.60	62.75	81.90	64.61	21.22	1.84	0
AUC	0.85	(0.9, 0.1)	0	0.15	20.32	83.99	99.02	86.95	20.72	0.19	0
C\(_\text {max}\)	0.85	(1.0, 0)	0	5.33	37.07	78.48	92.88	80.91	36.83	4.83	0
AUC	0.85	(1.0, 0)	0	0.75	36.33	92.73	99.85	94.36	36.14	0.70	0
C\(_\text {max}\)	0.95	(1.0, 0)	0	1.53	18.52	58.14	78.13	59.66	18.41	1.32	0
AUC	0.95	(1.0, 0)	0	0.13	17.89	81.36	98.67	84.64	17.89	0.13	0

CI method: Confidence Interval Method; \(\alpha\), significance level
PAR method: Pilot Acceptance Range method
Proposed method: the proposed robust Bayesian hierarchical model with a scaling factor \(\gamma\)

ISSN: 1471-2288