Two-stage Bayesian hierarchical modeling for blinded and unblinded safety monitoring in randomized clinical trials

Table 6 The probability of flagged trials for the AEs under the same effect for three AEs with safety issue (No effect for the rest) for scenario III

Example AEs (True Proportion; Expected Event Rate)	Blinded: Proportion with signal for early termination Unblinded: Termination Proportion with confirmed by unblinded data	(P_crit1, P_crit2)
Example AEs (True Proportion; Expected Event Rate)		(0.9, 0.7)	(0.9, 0.8)	(0.9, 0.9)
Pneumothorax Induced by HBO therapy^a (π = 3%; π_M = 2%)	Blinded	0.10	0.10	0.11
	Unblinded	0.68	0.48	0.27
	Overall Rate	0.07	0.05	0.03
Signs of Pulmonary Dysfunction^a (π = 31%; π_M = 25%)	Blinded	0.36	0.36	0.36
	Unblinded	0.70	0.55	0.36
	Overall Rate	0.25	0.20	0.13
Pneumonia^a (π = 48%; π_M = 40%)	Blinded	0.41	0.40	0.41
	Unblinded	0.68	0.57	0.37
	Overall Rate	0.28	0.23	0.15
Critical decreased CPP (< 60 mmHg) (π = 75%; π_M = 75%)	Blinded	0.15	0.16	0.15
	Unblinded	0.49	0.34	0.17
	Overall Rate	0.07	0.05	0.02
Critical hypotension (MAP< 70 mmHg) (π = 75%; π_M = 75%)	Blinded	0.15	0.16	0.16
	Unblinded	0.48	0.31	0.16
	Overall Rate	0.07	0.05	0.03
Seizures during HBO treatment (π = 1%; π_M = 1%)	Blinded	0.05	0.05	0.05
	Unblinded	0.59	0.39	0.20
	Overall Rate	0.03	0.02	0.01
Hypercarbia during transportation (PaCO2 > 45 mmHg) (π = 10%; π_M = 10%)	Blinded	0.10	0.11	0.10
	Unblinded	0.47	0.36	0.15
	Overall Rate	0.05	0.04	0.02

ISSN: 1471-2288