Table 1.

Approaches to Group Sequential Monitoring of HIV Vaccine Efficacy in Past Efficacy Trials.

Efficacy Trial	Monitoring Type	Number and Timing of Interim Analyses	Null and Alternative Hypotheses	Alpha level	Boundary Type
Vax004 Phase 3 1998-2003	Efficacy	1; when 50% infections expected	H0: VE ≤ 30% vs. H1: VE > 30%	0.025	O'Brien-Fleming
Vax003 Phase 3 1999-2003	Efficacy	1; when 50% infections expected	H0: VE ≤ 30% vs. H1: VE > 30%	0.025	O'Brien-Fleming
Step/HVTN 502 Phase 2b 2004-2007	Non-efficacy	1; 30 PP infections1	H0: VE ≤ 0% vs. H1: VE > 60%	0.05	Conditional Power < 20%
	Efficacy	1; 30 PP infections	H0: VE ≤ 0% vs. H1: VE > 0%	0.05	Custom
Phambili/HVTN 503 Phase 2b 2005-2007	Non-efficacy	1; 60 PP infections	H0: VE ≤ 50% vs. H1: VE > 50%	N/A	Conditional Power < 20%
	Efficacy	1; 60 PP infections	H0: VE ≤ 0% vs. H1: VE > 0%	0.05	Custom
RV144 Phase 2b 2004-2009	Harm	Monthly	H0: VE ≥ 0% vs. H1: VE < 0%	0.05	Pocock-type2
	Non-efficacy	8; every 6 to 12 months	H0: VE ≤ 0% vs. H1: VE > 50%	N/A	Conditional Power < 10%
	Efficacy	1; 2/3 of follow-up information	H0: VE ≤ 30% vs. H1: VE > 30%	0.025	O'Brien-Fleming

¹Per-protocol (PP) infections are those diagnosed after the Week 12 visit in volunteers HIV negative at baseline and who received the first two doses of either vaccine or placebo, excluding those who were either diagnosed with HIV infection before or at the Week 12 visit or who violated the protocol on the basis of pre-defined criteria (Buchbinder et al., 2008). The interim analysis was triggered by the 30^th PP infection in the primary analysis group of subjects with Adenovirus-5 neutralization titers ≤ 200.

²Continuous stopping boundary of Betensky (1998).