CONSIDERATIONS AND APPROACHES FOR STUDY EVALUATION IN CHEMICAL HEALTH ASSESSMENTS -Presentation
On this page:
Systematic review approaches developed within the U.S. EPA’s Integrated Risk Information System (IRIS) Program to evaluate the potential human health effects of chemical exposures emphasize a continuous focusing on the most informative evidence at each step of the review. This includes the prioritization, evaluation, and integration of individual studies from diverse scientific disciplines, across databases of varied complexity, and for use within different decision contexts. Individual human and animal studies of potential health effects are evaluated using consistent criteria addressing reporting quality, risk of bias, and study sensitivity for each outcome of interest. Implicit in these evaluations are reviewer interpretations of the potential influence of the identified limitations on the study results. These decisions then inform qualitative conclusions drawn across the evaluated criteria to arrive at a level of confidence regarding the utility of the studies for assessing the health effect in question. Examples from the application of these approaches to both epidemiological and animal toxicology studies demonstrate that sensitivity-related issues (or issues that inform the ability of a study to reliably detect an effect) about exposure and outcome ascertainment are often critical to determining the relative influence of each study to drawing conclusions. Approaches for evaluating mechanistic information (i.e., data not intended to directly identify an adverse outcome) will also be discussed. IRIS approaches prioritize mechanistic data based on the potential impact that analyses might have on conclusions regarding the available human and animal evidence. To date, while consistency in the evaluation of individual toxicity studies has proven critical, efficiency and flexibility in the prioritization of studies for review and the effective integration of the results of these evaluations into subsequent decision steps have proven equally important.