Table 2.

Posterior probabilities for stopping criteria by agent, action, and outcome: modified intention-to-treat population.

Outcome	Action	Criterion	Arm	Posterior probabilitya
Recovery (HR > 1 is better)	Graduation	Pr(HRr > 1.0 | data) ≥ 0.975	Apremilast	0.101
			Cenicriviroc	0.225
			Icatibant	0.163
			IC14	0.017
			Celecoxib/famotidine	0.011
			Dornase	0.125
	Futility	Pr(HRr < 1.5 | data) ≥ 0.900	Apremilast	0.999
			Cenicriviroc	0.998
			Icatibant	1.000
			IC14	1.000
			Celecoxib/famotidine	1.000
			Dornase	0.998
Death (HR < 1 is better)	Graduation	Pr(HRm < 1.0 | data) ≥ 0.900	Apremilast	0.428
			Cenicriviroc	0.138
			Icatibant	0.389
			IC14	0.716
			Celecoxib/famotidine	0.089
			Dornase	0.383
	Futility	Pr(HRm < 1.0 | data) ≤ 0.500	Apremilast	0.428
			Cenicriviroc	0.138
			Icatibant	0.389
			IC14	0.716
			Celecoxib/famotidine	0.089
			Dornase	0.383

Bold posterior probability indicates criterion was met.

Abbreviations: Hazard ratios (HR) as follows: HRr (recovery, in which values >1 indicate better outcomes, i.e., higher instantaneous recovery rate), HRm (mortality, in which values <1 indicate better outcomes, i.e., longer time to death), Pr (probability).

^aThe posterior probability column shows the probabilities Pr(HRr > 1.0 | data), Pr(HRr < 1.5 | data), and Pr(HRm < 1.0 | data) (i.e., the corresponding probabilities after having observed the data in the trial and given the priors and the model, as specified in the SAP). Since the same posterior probability, Pr(HRm < 1.0 | data), is used in both the graduation and the futility criteria for death, its numerical value is the same for all agents in the rows corresponding to the graduation and the futility criteria (e.g., 0.716 for IC14). This probability is however compared to different thresholds in the graduation and futility criteria (≥0.9 for graduation and ≤0.5 for futility).