The 130th Annual Meeting of APHA

Abstract #46688

Stan Kaplowitz, PhD¹, Harry Perlstadt, PhD, MPH², Yasmina Bouraoui³, Sean Frost¹, Warren Brown, PhD⁴, Lori Ann Post, PhD⁵, and Marguerite Barratt, PhD⁵. (1) Dept of Sociology, Michigan State University, 316 Berkey Hall, East Lansing, MI 48824-1111, 517 355-6640, Stan.Kaplowitz@ssc.msu.edu, (2) Department of Sociology, Michigan State University, 316 Berkey Hall, East Lansing, MI 48824-1111, (3) Community Living, Children and Families Administration, Michigan Department of Community Health, 3423 N. Martin L. King Jr. Blvd, PO Box 30195, Lansing, MI 48909, (4) Institute for Social and Economic Research, Cornell University, 201 Caldwell Hall, Ithaca, NY 14853-2602, (5) Institute for Children, Youth, and Families, Michigan State University, Suite 27 Kellogg Center, East Lansing, MI 48824

Many surveys designed to predict lead risk use questions about ZIP code, age of house, presence of paint chips and Medicaid participation, exposure through hobbies occupations and cosmetics. But these surveys have had very limited predictive value. Our research attempts to create a better screening instrument, by first gathering larger amounts of information and then statistically evaluating which pieces of information have high predictive value. We are collecting both survey data (residence address and survey responses) and the blood lead level test results on 4000 children. Sampling is targeted to attain 3000 cases from urban and 1000 cases from rural areas and at least 1000 African-American children. We attempt to improve on existing prediction methods in the following ways: a) Self-report questions are easier to understand, and are more specific than in the typical screening survey (e.g. questions about previous housing and specific hobbies, occupations, or cosmetics. b) Current residence will be associated with a census block group, a smaller, more homogeneous area, that should be more predictive than ZIP code. The census block groups permits further analysis by income, housing characteristics, and race/ethnicity. c) Distance between residence and Environmental Protection Agency identified lead contamination sites using Geographic Information Systems.

Learning Objectives: At the conclusion of the session, attendees will be able to

• identify key predictors of risk for elevated blood lead levels
• describe a procedure for collecting and combining survey and census data
• assess different methods of analyzing the predictiveness of indicators

Keywords: Lead, Child Health

Presenting author's disclosure statement:
I do not have any significant financial interest/arrangement or affiliation with any organization/institution whose products or services are being discussed in this session.