Evaluating Protective Effects of Vaccination

Halloran, M. Elizabeth; Longini, Ira M.; Struchiner, Claudio J.

Evaluating the direct protective effects of vaccines in the individuals who were vaccinated has been the focus of vaccine studies over the past century. Generally, interest has been in the ability of vaccination to prevent or to ameliorate disease rather than to prevent infection (Clements-Mann 1998). Ascertainment of cases is often done by finding suspected cases in the population under study in people who exhibit a set of symptoms. The suspected cases are then tested for biological confirmation of the infectious agent of interest. Alternatively, surveillance can ascertain cases reported in central registries. However they are ascertained, with most vaccines, clinical disease is the primary outcome of interest. When ascertainment is on clinical cases, most asymptomatic infections may go undetected. A different situation arises when infection is the primary outcome. To ascertain infections in asymptomatic people, an active follow-up method of testing asymptomatic people is needed. In this chapter we consider estimation and inference for direct protective effects of vaccination, VE(S) and VE(SP), in studies that do not condition on exposure to infection.We consider aspects of the design of such studies. Several examples of randomized, double-blind (double-masked) controlled vaccine trials illustrate the standard approach to design and analysis of such studies. Our choice of studies to present was motivated largely because of their use as illustrations in other sections of the book. Most randomized and pivotal studies of vaccines have been based on VE(SP) or VE(S). Much has been written on studies to meet the approval of the regulatory agencies, and the design of clinical trials generally. Our goal here is to consider how VE(S) and VE(SP) relate to other measures of vaccine effects within the dependent happening context, and consider a few design considerations. Because VE(S) and VE(SP) do not condition on exposure to infection, assumptions about the relative exposure opportunity in the vaccine and control groups are important.