CIAO DATE: 03/06

Efficacy Calculation in Randomized Vaccine Trials: Global or Local Measures?

Abstract

Conventional phase III vaccine trials have an underlying assumption that the effect of the vaccine is the same throughout the trial area. This study tests whether this assumption is true for an oral cholera vaccine that was administered in an individually-randomized trial in Bangladesh. Established non-spatial vaccine trial results do not report whether vaccines vary by neighborhood within a trial area. This study uses three data sets, including a (1) cholera vaccine trial database, (2) longitudinal demographic database of the rural population from which the vaccine trial participants were selected, and (3) household-level geographic information system database. Protective efficacy was calculated within spatial neighborhoods of various sizes. The results show that efficacy varies significantly in space regardless of the neighborhood size. Analytical z-score maps identify unusually high and low efficacy values in certain parts of the trial area. The main conclusion is that vaccine efficacy varies in space and that locational information is necessary to reveal this fact. This suggests that global efficacy measures that are calculated using non-spatial information may be misleading and that local efficacy measures will reveal more comprehensive information for public health practitioners to make decisions about whether or not to bring a vaccine into public health practice in different populations.

Main Text (PDF format, 16 pages, 618.1 KB)
References and Back Cover (PDF format, 9 pages, 1.61 MB)