DEVELOPMENT OF A SCREENING TOOL FOR PREDICTION OF CHILDREN AT RISK FOR LEAD-EXPOSURE IN A MIDWESTERN CLINICAL SETTING

Objective. Universal screening for childhood lead poisoning is becomin g quite common, with many states having legislation requiring screenin g. We set out to determine whether a questionnaire could be used to id entify children at risk for exposure to lead to determine whether sele ctive screening of those at risk was possible. Methods. Parents of 370 children 12 to 36 months of age having well-child examinations comple ted a questionnaire and their children were screened by a fingerstick capillary blood lead test at two clinics. Results. Of patients from cl inic A, 5.4% had lead levels greater than or equal to 10 mu g/dL compa red with 16.8% of those from clinic B (P < .001). This difference betw een clinics could not be explained by the demographic characteristics of the patients or by differences in their potential exposures to lead . We evaluated the five questions suggested by Centers for Disease Con trol and Prevention for anticipatory guidance for their ability to ide ntify children with elevated blood lead levels. In clinic A, this inst rument had a sensitivity of 76.9% and a negative predictive value of 9 6.5%. In clinic B, it had a sensitivity of 63.6% and a negative predic tive value of 81.4%. Based on an assessment of significant items from a large questionnaire, we determined five questions that were the best predictors of risk. On the basis of this risk assessment, 100% of the children from clinic A with elevated lead levels and 90.9% of the chi ldren from clinic B with elevated lead levels were classified as being at ''high risk.'' Had this risk assessment been used as an initial sc reen in this sample, 40% of the patients from clinic A and 37% of the patients from clinic B would not have been screened with a blood lead test, because they were classified as being at ''low risk.'' Conclusio ns. Results of this study suggest that there is great variability in t he prevalence of elevated lead levels and potential risks between clin ics within a fairly homogeneous community; however, selective screenin g with a community-specific questionnaire may be feasible if the preva lence is low and the risks to the population are known.