A probabilistic approach to exposure risk assessment

摘要

The introduction of hazardous substances into the environment has long been recognized as being a cause of several diseases in humans, wildlife, and plants. The damaging character of suspected contaminants is usually assessed via a "reject/retain" design with no explicit link between levels of exposure and intensities of the potential adverse health effects even though this connection may be important for the development of public health regulations that limit exposure to hazardous substances. Here, we propose a probabilistic approach to exposure risk assessment as a way around this typical flaw. We develop a Bayesian model using proximity to the source of an alleged contaminant as a surrogate for exposure. Subsequently, we carry out an experimental study based on simulated data to illustrate the model implementation with real world data. We also discuss a possible way of extending the model to accommodate potential heterogeneity in the spatial distribution of the focal disease.