Comparing lead poisoning risk assessment methods: census block group characteristics vs. zip codes as predictors.

Area of Science:

• Environmental health
• Pediatric public health
• Biostatistics

Background:

• Universal blood lead level (BLL) screening faces financial and practical barriers.
• Existing screening strategies may not optimally identify at-risk children.
• Centers for Disease Control and Prevention (CDC) guidelines permit tailored BLL screening approaches.

Purpose of the Study:

• To develop a more accurate prediction equation for identifying children who require blood lead level (BLL) testing.
• To improve upon current pediatric screening practices that rely on zip code risk and Medicaid status.
• To provide a tool for health departments to optimize BLL screening efforts.

Main Methods:

• Utilized Michigan BLL test data (1998-2005) including address, Medicaid, and race information.
• Linked residential addresses to U.S. Census 2000 block group data for sociodemographic and housing characteristics.
• Employed Hierarchical Linear Modeling to create a continuous variable prediction equation for BLL.

Main Results:

• Neighborhood characteristics (housing age, SES, race/ethnicity) at the census block group level explained over 41% of BLL variance.
• This model significantly outperformed traditional zip code risk and Medicaid status, which explained only 15% of variance.
• The derived prediction equation accurately forecasted BLL for subsequent years (2002-2005).

Conclusions:

• The developed equation demonstrates superior sensitivity and specificity for identifying high-risk children compared to zip code and Medicaid-based methods.
• This approach offers significant potential for cost savings in public health lead screening programs.
• An accessible, internet-based tool can be developed for healthcare providers to utilize this prediction equation.