Identifying Geographic Areas Where Children May Be At-risk of Lead Poisoning and Assessing the Need for Lead Abatement in Urban Areas: A Case Study in Hamilton, Ontario

Abstract

Thirty years since the removal of lead from gasoline, lead still poses a health risk. Children are most at-risk for adverse health outcomes caused by lead toxicity due to both behavioural (e.g., hand-to-mouth behaviour) and physiological differences (i.e., increased intake of lead by body weight, higher uptake rate and a higher vulnerability to the effects of lead) compared to adults (Yeoh et al., 2009). As a result, governments must identify children that may be at-risk of lead poisoning and develop practical methods to mitigate lead exposure. Before a government can develop a policy to help mitigate exposure of lead for children, we need to understand the spatial distribution of lead within the city. A popular spatial model used within air pollution research may allow more accurate, and more localized predictions than the most common interpolation method, kriging. Land use regression (LUR) is a technique leveraging multiple predictor variables to help estimate the spatial distribution of the dependent variable. By using historical sources of lead, LUR can be used to model soil lead levels (SLL) with localized variation. Unfortunately, spurious relationships can be the basis of a LUR model, which may lead to an overfitted spatial model resulting in a model with little generalizability and questionable ability to estimate the dependent variable at unobserved locations. Ultimately, Empirical Bayesian Kriging may be the best option for soil contamination research due to its ability to provide a smoothed prediction surface and its dependence on the spatial structure of the data to provide estimations. The benefit to society and the return on investment (ROI) is often the justification for lead remediation. Gould (2009) estimates a $17 to $221 ROI for every dollar spent on lead hazard control. One of the main components of this estimate of ROI comes from the decrease in intelligence quotient (IQ) that a child may experience as a consequence of lead toxicity. There are three main ways that a decrease in IQ can negatively impact the economy, (i) lower potential lifetime earnings, (ii) reduced tax revenues, and (iii) higher spending on special education (Gould, 2009). Since IQ has such a significant role in the ROI estimates, chapter 3 seeks to achieve a greater understanding of the relationship between blood lead levels (BLLs) and IQ. The loss of IQ points for an increase in blood lead concentration proposed by Lanphear et al. (2005) and referenced by Gould (2009) is significantly higher than what we found in our meta-analysis. Thus, the projected ROI proposed by Gould (2009) may be much lower than previously calculated. In the final chapter, the cost associated with permanent lead abatement is investigated based on ROI projections as a case study in Hamilton, Ontario. We show that, in most cases, permanent lead remediation is far too expensive for a municipal government. Furthermore, the capital initially invested may not be distributed back into the local economy, as the ROI suggests. We suggest that municipal governments make decisions based on need, rather than basing remediation decisions on ROI projections. Furthermore, we recommend the use of hazard quotient maps to help justify lead remediation as a more accurate representation of potential lead toxicity, instead of only looking at SLL exceedances across the city.