A Bayesian approach to aggregated chemical exposure assessment

Human exposure to a chemical commonly arises from multiple sources, yet traditional assessments often treat these sources in isolation, overlooking their aggregated impact.

Introducing a novel approach to aggregated chemical exposure assessment that explicitly accounts for these intertwined exposure pathways.

A conceptual framework was developed that describes this approach in multiple steps, namely: 1) the comprehensive integration of diverse datasets, such as consumption surveys, demographics, chemical measurements, and market presence data; 2) the modeling of singular exposure to distinct sources using Bayesian inference; 3) the estimation of aggregated exposure through a simulation-based strategy reflecting the full spectrum of individual exposure scenarios. The value of this approach is demonstrated using the case of titanium dioxide, a chemical found in foods, dietary supplements, medicines, and personal care products that has been banned as a food additive in the European Union since 2022. A comparison was hereby made between exposure before and after the ban, as well as between sources.

Through this case study, common data challenges in the context of aggregated exposure assessment, including missing values, limited sample sizes, uncertainties, and inconsistencies, were effectively showcased and addressed by leveraging the advantages of Bayesian inference. As such, robust estimates of aggregated exposure were derived between and across sources and populations, while incorporating relevant prior knowledge.

By capturing the complexity of real-world exposures, this comprehensive Bayesian approach provides decision-makers with more reliable probabilistic estimates to inform public health policies.