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

Table 9 The summary table of simulation scenarios and results comparison

Scenarios	AE #1	AE #2	AE #3	AE #4	AE #5	AE #6	AE #7
Scenario I (Safe: #1234567)	Safe	Safe	Safe	Safe	Safe	Safe	Safe
Scenario I (Safe: #1234567)	Baseline proportions
Scenario II (Unsafe: #1234567 same effect)	Unsafe	Unsafe	Unsafe	Unsafe	Unsafe	Unsafe	Unsafe
	Compare to Baseline proportions scenario I
	Proportions increase a lot (All effect scenario)
Scenario III (Safe: #4,5,6,7; Unsafe: #1,2,3 same effect)	Unsafe	Unsafe	Unsafe	Safe	Safe	Safe	Safe
	Compare to Baseline proportions scenario I
	Higher proportions			Smaller proportions
Scenario IV (Unsafe: #4,5,7 flat effect; Unsafe: #1,2,3,6 same effect)	Unsafe	Unsafe	Unsafe	Unsafe	Unsafe	Unsafe	Unsafe
	Comparing to scenario II
	Similar proportions			Proportion decreases	Proportion decreases	Similar proportion	Proportion decreases
Scenario V (Unsafe: #1234567 flat effect; unsafe for control arms)	Unsafe	Unsafe	Unsafe	Unsafe	Unsafe	Unsafe	Unsafe
	The model flags potential signals at the blinded stage but not at the unblinded stage with fewer proportions comparing to other scenarios.

ISSN: 1471-2288