Universal screening for childhood lead poisoning is widely debated. Our purpose was to compare screening results at three pediatric clinics within Houston and to evaluate the effectiveness of screening according to published criteria. The clinics were chosen for their geographic and socioeconomic diversity. Children between 6 months and 6 years of age were tested, and the results were classified according to current guidelines. We screened 864 children. Results between sites were significantly different, P = 0.002. No children with blood lead levels greater than 0.45 mumol/L (9 micrograms/dL) were identified at Clinic C compared to 76 (8.8%) from Clinics A and B, but no site had children with levels greater than or equal to 2.20 mumol/L (45 micrograms/dL). The prevalence of childhood lead poisoning can vary even within the city. If regional screening is to replace universal screening, statewide as well as citywide data are needed to identify high-risk areas. This could be done by clinic site, zip code, or census track data with a minimum of 3000 children.