Feasibility of Prophylactic HIV Vaccine Trials: Some Statistical Issues

摘要

In 1915, Greenwood and Yule (1915) articulated that a valid vaccine efficacystudy requires equal exposure to infection in the vaccinated and unvaccinatedgroups. One issue in planning prophylactic HIV vaccine trials is that peoplewho are randomized to a blinded, placebo trial may have themselves tested tosee if they have a positive HIV antibody response. If the test is positive, theymight believe that they have received the vaccine. Then, depending on theirbelief about the protective effects of the vaccine, they may increase theirsexual activity, thus increasing their exposure to infection. This increasedexposure to infection in the vaccinated population could bias the estimate ofthe biological efficacy of the vaccine to protect against infection if it is notincluded in the analysis.