Sieve Analysis in HIV Vaccine Efficacy Trials with Multivariate and Missing Marks

• Activity Number: 616
• Type: Contributed
• Date/Time: Thursday, August 4, 2011 : 8:30 AM to 10:20 AM
• Sponsor: Biometrics Section
• Author(s): Michal Juraska*+ and Peter B. Gilbert
• Companies: University of Washington and University of Washington
• Keywords: vaccine efficacy ; competing risks ; mark variable ; density ratio model ; Cox model ; (augmented) inverse probability weighting

Abstract

In a randomized placebo-controlled HIV vaccine efficacy trial, an objective is to evaluate the relationship between vaccine efficacy to prevent infection and the genetic distances of the exposing HIV strains to the multiple HIV sequences included in the vaccine construct, where the set of genetic distances is considered as the continuous multivariate 'mark' observable in infected subjects only. We propose an inferential method in the framework of competing risks failure time analysis for the evaluation of mark-specific vaccine efficacy that improves efficiency of estimation, relative to current alternative approaches, by using the semiparametric method of maximum profile likelihood estimation in the vaccine-to-placebo mark density ratio model combined with a more efficient estimation method for the overall log hazard ratio in the Cox model. The marks of greatest scientific relevance are often missing. We propose two estimation approaches to accommodate a mark missing at random: (i) the inverse probability weighted (IPW) complete-case technique, and (ii) augmentation of the IPW estimating equation for improved efficiency by leveraging auxiliary information predictive of the mark.