Evaluation of risk assessment questions used to target blood lead screening in Illinois

Objective. Beginning in 1995, Illinois law permitted targeted-as opposed to universal-blood lead screening in low-risk areas, which were defined by ZI P code characteristics. State guidelines recommended specific lead risk ass essment questions to use when targeting screening. This study was designed to evaluate the sensitivity and specificity of Illinois lead risk assessmen t questions. Design. Parents bringing their 9- or 10- or 12-month and 24-month-old child ren for health supervision visits at 13 pediatric practices and parents of children (aged 6 through 25 months and who needed a blood lead test) receiv ing care at 5 local health departments completed a lead risk assessment que stionnaire concerning their child. Children had venous or capillary blood l ead testing. Venous confirmation results of children with a capillary level greater than or equal to 10 mu g/dL were used in analyses. Children. There were 460 children with both blood and questionnaire data re cruited at the pediatric practices (58% of eligible) and 285 children (51% of eligible) recruited at local health departments. Of the 745 children stu died, 738 provided a ZIP code that allowed their residence to be categorize d as in a low-risk (n = 456) or high-risk (n = 282) area. Results. Sixteen children (3.5%) living in low-risk areas versus 34 childre n (12.1%) living in high-risk areas had a venous blood lead level (BLL) gre ater than or equal to 10 mu g/dL; 1.8% and 5.3%, respectively, had a venous BLL greater than or equal to 15 mu g/dL. For children living in low-risk a reas, Illinois mandated risk assessment questions (concerning ever resided in home built before 1960, exposure to renovation, and exposure to adult wi th a job or hobby involving lead) had a combined sensitivity of .75 for lev els greater than or equal to 10 mu g/dL and .88 for levels greater than or equal to 15 mu g/dL; specificity was .39 and .39, respectively. The sensiti vity of these questions was similar among children from high-risk areas; sp ecificity decreased to .27 and .28, for BLLs greater than or equal to 10 mu g/dL and greater than or equal to 15 mu g/dL, respectively. The combinatio n of items requiring respondents to List house age (built before 1950 consi dered high risk) and indicate exposure to renovation had a sensitivity amon g children from low-risk areas of .62 for BLLs greater than or equal to 10 mu g/dL with specificity of .57; sensitivity and specificity among high-ris k area children were .82 and .36, respectively. For this strategy, similar sensitivities and specificities for low and high-risk areas were found for BLLs greater than or equal to 15 mu g/dL. Conclusions. The Illinois lead risk assessment questions identified most ch ildren with an elevated BLL. Using these questions, the majority of Illinoi s children in low-risk areas will continue to need a blood lead test. This first example of a statewide screening strategy using ZIP code risk designa tion and risk assessment questions will need further refinement to limit nu mbers of children tested. In the interim, this strategy is a logical next s tep after universal screening.