Rewriting paracelsus: linking chemical assessment to public health.

Abstract

Human health is continuously challenged by exposure to complex mixtures of chemicals present in food, products, and the environment. Traditional risk assessment (TRA), rooted in mid-20th-century paradigms, relies on deterministic thresholds and animal testing to define safe exposure levels. This approach, while useful for regulatory simplicity, fails to represent real-world variability, mixture effects, and lifelong events, limiting its relevance for public health protection. Here we show that a probabilistic Next-Generation Risk Assessment (NGRA) framework integrating quantitative modelling, mechanistic biology, and New Approach Methodologies (NAMs), can transform toxicological evidence into probability distributions, estimating the likelihood of adverse outcomes under realistic exposure scenarios. Central to this approach are Health Impact Pathways (HIPs), which connect molecular perturbations to population-level health indicators, bridging toxicology and epidemiology. The framework explicitly distinguishes susceptible groups, such as individuals in physiologically sensitive stages like pregnancy, from vulnerable groups, including those with pre-existing diseases or higher exposure burdens, poorly covered in TRA. By covering these groups, probabilistic NGRA enables more inclusive, relevant, and informative evaluations that align with FAIR (Findable, Accessible, Interoperable, Reusable) data principles and minimize animal testing, offering a scientifically grounded framework for improving chemical safety evaluations and public health protection.