Application of social epidemiological models for evaluating the interaction of chemical and nonchemical stressors in cumulative risk assessment as a policy tool for protecting vulnerable populations

The National Research Council's Committee on Improving Risk Analysis recently called for incorporating nonchemical stressors into cumulative risk assessment to improve risk science and better protect vulnerable populations. Nonchemical stressors, such as neighborhood poverty and socioeconomic deprivation, are critical to consider in risk assessments because many chemical exposures occur in the context of multiple social disadvantages, which frequently impair human health and well being. Such nonchemical stressors and buffers are thought to modify the toxic effects of chemical stressors, however, the analytical tools and methods to evaluate these processes have been underdeveloped in the scientific literature. In estimating health risks associated with environmental exposures, researchers need to consider multiple levels of factors and to take into account potential confounding and effect modification within and across levels. This paper discusses how social epidemiology can be employed in evaluating the health impacts of the interaction of nonchemical and chemical stressors in the cumulative risk assessment of chronic exposure to air toxics. The approach proposed in this paper is the development of “dose-response” model for the cumulative effects of nonchemical stressors relating to population “vulnerability”. Upon validating a model of vulnerability using selected census tract demographic characteristics, the paper demonstrates how this vulnerability construct can be used to test for effect modification of toxic exposure in accounting for health endpoints such as incident respiratory hospitalization rates. Incorporating nonchemical stressors into cumulative risk assessment presents a useful policy tool for addressing environmental justice issues in toxic exposure and protecting overburdened communities and vulnerable populations.