Following the discovery of high levels of lead in dust samples taken from a recreation center in the Eure department (27), a modeling study conducted by Cire Normandie estimated high levels of lead exposure (exceeding the threshold set by health authorities to define a case of lead poisoning) among children who attended the center under various exposure scenarios. To check for cases of childhood lead poisoning in the target population, a screening campaign for childhood lead poisoning was implemented. Among the children screened, no cases of childhood lead poisoning were diagnosed. This discrepancy between observation and modeling can be explained by (i) the application of an unfavorable scenario, (ii) the use of environmental data affected by construction activities, and (iii) the time lag between the screening and the children’s last exposure. Furthermore, this study reinforced the observation that blood lead levels in children depend not only on environmental lead levels but also on individual behaviors that modeling is unable to account for in detail. (R.A.)