NOSI to Support Enhancement of Software Tools for Multilevel Mediation Analysis for Investigating Effects of Environmental and Individual Risk Factors on Respiratory Diseases

环境与个体调查效应的多层次中介分析

Multilevel Mediation Analysis for Investigating Effects of Environmental and Individual Risk Factors on Respiratory Diseases The link between particulate matter (PM) exposure and poor respiratory health is well established. Through Aim 3 of Project 1, the LSU Superfund Research Center postulates that environmentally persistent free radicals (EPFRs) from hazardous waste sites, incinerators/chemical fires, and other sources is the missing mechanistic link between PM exposure and poor respiratory health. We will investigate the mediation of this association using hierarchical mediation analysis to decompose the air pollutant adverse respiratory effects into direct and indirect (EPFR-mediated) effects. We developed a multilevel mediation analysis method that allows for both longitudinal assessments of residential environments and individual risk factors to be jointly utilized in determining the mechanistic link between exposure to PM and poor respiratory health. By including measures of individual behavioral factors our methods are capable of explaining existing disparities in health outcomes. We have implemented our method as an R package (mlma) to provide the research community with open access to software for performing multilevel hierarchical mediation analysis. However, use of our R package does require knowledge of R, which limits its broader use in research. To address this limitation, we propose to collaborate with a software engineer at LSU to create an API for an interactive web implementation of our mediation software. We expect that this would greatly expand access to our methods, as no programming will be required. The proposed app, with its intuitive interactive visual interface, will allow users to easily read in datasets and build a conceptual mediation model framework using drag-and-drop functionality. We will also provide graphical tools such as directed acyclic graphs to allow users to visually interpret the analysis results. In addition, we will transform part of the computing to an efficient low-level computer language to improve computational speed and plan to use cloud computing resources so that R and associated software packages do not have to be installed on individual computers to perform the analysis.