A Question of Policy: When is a PBPK Model “Good Enough” for Chemical Risk Assessment?
On this page:
Modern chemical risk assessments frequently make use of physiologically based pharmacokinetic (PBPK) models, which provide mathematical descriptions of the absorption, distribution, metabolism, and excretion of substances by organisms. These models can be used to relate external doses (such as oral or inhaled doses associated with adverse health effects) to internal dose metrics (such as blood concentrations) that may be more directly related to toxicity. One important question at the interface of science and policy for chemical risk assessment is, “When is a PBPK model good enough to be used for a given application?” For the purposes of this roundtable, we assume that quality assurance (QA) protocols that ensure a given model is biologically plausible and that it has been implemented correctly have already been applied and focus on quantitative evaluation criteria for assessing accuracy and precision of model predictions. One commonly applied “closeness” criterion requires that PBPK models should produce predictions that are generally within a factor of two of experimental data before accepting them for use in chemical risk assessment. In this session, we will explore the basis for and adequacy of various evaluation criteria, including the “factor of two” criterion. A diverse group of panelists will share their experiences and perspectives with respect to evaluation of PBPK model fitness, and then the panel will address specific prepared questions as well as questions from the audience on this topic.