Geographical Analysis in Vaccine Efficacy Trials: Spatial Confounders, Effect Modifiers, and Herd Immunity Measurement

Portland State University, Portland OR

Investigators

Abstract

This research project will develop and test spatial analytical methods for vaccine trials because existing methods limit the extrapolation of results and are sometimes of questionable validity. It will use spatial methods to control for heterogeneous disease exposures (spatial effect modifiers) and spatial bias in disease outcomes (spatial confounders). In 1985, a community-based, individually-randomized oral cholera vaccine trial was conducted in Matlab, Bangladesh. The double-blind trial measured the efficacy of two vaccines, the B subunit-killed whole cell (BS-WC) and the killed whole cell only (WC) vaccine. Females aged 15 years and older, and children aged 2 to 15 were the target group in the trial. To identify the cholera cases from the study area, the investigators conducted a passive surveillance at one hospital and two community-based treatment centers. During 5 years of follow-up, the protective efficacy for the BS-WC group was 49% and for the WC group it was 47%. Passive surveillance can introduce bias because access to treatment centers, which is usually a function of distance, influences health-seeking behavior. Also, efficacy might differ in different parts of the study area because socio-environmental circumstances and therefore disease exposure levels vary in space. This study will use a geographic information system (GIS) to determine: (1) how cholera vaccine efficacy varies spatially in the study area; (2) what ecological socio-environmental variables are related to cholera vaccine efficacy (i.e., which variables are effect modifiers); (3) how protective efficacy varies with access to treatment facilities (i.e., whether access is a spatial confounder; and (4) whether cholera incidence in the placebo group is related to vaccine coverage rates (i.e., is herd immunity important). We will use three data sets, already collected by the investigators, including: (1) a large cholera vaccine trial database, (2) a comprehensive longitudinal demographic database of the rural population of approximately 200,000 from which the vaccine trial participants were selected, and (3) an accurate household-level spatial database of the same study area population. This spatial database, in conjunction with the demographic and vaccine datasets, will facilitate adding an integrated, comprehensive, and accurate spatial component to all three datasets. The broader impacts of this study are both theoretical and methodological. It will build on the medical geographic theory of disease ecology by showing that the effect of a health intervention (e.g., a vaccine) is modified by spatially heterogeneously distributed complexes of variables (i.e., effect modifiers). It will show that a spatial perspective is necessary because of spatial bias in existing methodologies. It will also show how a spatial perspective can be used to develop new methodologies for something that has until now been impossible to measure (i.e., herd immunity). Understanding human-environment interaction from a spatial perspective is one of the core concepts of geography and this project will expand this theoretical area. It will specifically build on existing medical geographic disease ecology theory and place it within modern scientific frameworks such as biocomplexity. It will also serve as a methodological case study that describes how geographical methods can be used in vaccine trials.