Session 4: Evidence-based Toxicology

Introduction to Evidence-based Toxicology
Dr. Martin L. Stephens, Johns Hopkins University
PRESENTING AUTHOR: 

Dr. Martin L. Stephens

INSTITUTION / COMPANY : 

Johns Hopkins University

ABSTRACT CONTENT / DETAILS: 

Evidence-based approaches are structured, objective tools for assessing the existing evidence on a given topic in a manner that limits bias and enhances transparency. These approaches were pioneered in medicine and are beginning to be applied in toxicology, giving rise to the new discipline of Evidence-Based Toxicology or EBT (the parallel to Evidence-Based Medicine or EBM).

The quintessential evidence-based tool is the systematic review, which assesses the evidence relevant to a specific question. Systematic reviews involve a number of steps, such as assessing the “internal validity” of included studies, i.e., the degree to which the outcomes reflect the experimental variability and not sources of bias.

EBM provides guidance on framing questions of interest, searching the literature, selecting relevant studies, extracting data, assessing data quality, and analyzing data. The latter step constitutes data integration, and is accomplished via a meta-analysis, where appropriate.

Whereas data integration in EBM focuses on the outcomes of clinical trials and other human studies, the challenge in toxicology is how to integrate data across a variety of sources, including in vivo, in vitro, or epidemiological studies. EBM offers little specific guidance on combining evidence across evidence streams. However, even in this situation, one can honor the principles of EBM by developing structured approaches for data integration that address bias and ensure transparency. Another challenge in toxicology is the use of model systems (e.g., in vivo studies), which raise the question of the “external validity” of relevant studies, i.e., how well the model reflects the real-world situation in humans. Ideally, this question should be addressed in any proposed data integration scheme.