Modeling annual elevated blood lead levels among children in Maryland in relation to neighborhood deprivation.

Estimating environmental lead exposure using ecologic risk models is an inexpensive strategy to inform public health departments and to develop location-based intervention strategies such as targeted screening and mitigation. Importantly, studies in this area have not assessed temporal and spatio-temporal lead exposure risk trends. Due to lead abatement efforts and targeted screening efforts, it is anticipated that lead exposure risk has decreased over time. However, it is unknown if decreases have occurred, and if the decreases are evenly distributed across neighborhoods. Thus, the purpose of this study was to examine the association between neighborhood deprivation and risk of elevated blood lead levels (EBLLs) in both temporal and spatio-temporal contexts within the US state of Maryland in 2005-2015. To consider the temporal dimension of lead risk, we used a novel extension of Bayesian index models to estimate time-varying neighborhood deprivation indices along with time-varying index effects. The results showed that overall EBLL proportion decreased over time, from a high of 0.11 in 2006 to a low of 0.02 in 2015. The association between neighborhood deprivation and EBLL risk was positive and significant annually, but generally diminished over time. The most important variables in the neighborhood deprivation index were percent of houses built before 1940 and median household income. In summary, using Bayesian index models that can account for both temporal and spatio-temporal contexts is a promising approach to inform public health efforts to remediate lead and focus testing efforts and may be useful in studies in other geographic areas and times.