A novel method for environmental risk assessment: A case study of coarse particulate matter and infant birth weight

Abstract

In this study, we demonstrate how a Bayesian belief network (BN) model can be used to quantify hypothesis-based weight-of-evidence (HBWoE) evaluation for the particulate matter (PM) exposure network risk factors. Most studies of particulate matter exposure and adverse health outcomes have been performed for PM2.5(aerodynamic diameter ≤ 2.5 μm) and PM10(aerodynamic diameter ≤ 10 μm). However, the lack of direct measurement of coarse PM (PM2.5-10, aerodynamic diameter from 2.5 to 10 μm) causes inconsistent results. There are limited studies on exposure to PM2.5-10during gestation and low birth weight. It is important to investigate the impacts of PM2.5-10on birth weight by weighing all of the evidence that can help to understand and predict the potential risks. The results of the designed model indicate that maternal exposure to ambient coarse PM pollution (>16 μg/m3) provides high correlation between low birth weight and coarse PM concentration with false negative rate (FNR) of 15%. The results show an effective way of evaluating the cumulative results of studies to measure how much the evidence supports the hypothesis. The proposed framework can be implemented in several other environmental contaminant exposure health-risk problems to understand the value of information in each study and overall cumulative information. © ASCE.