- Email: [email protected]
Screening for Lead Poisoning: A Geospatial Approach to Determine Testing of Children in At-Risk Neighborhoods
Screening for Lead Poisoning: A Geospatial Approach to Determine Testing of Children in At-Risk Neighborhoods AMBARISH VAIDYANATHAN, MS, FORREST STALEY, MUP, JEFFREY SHIRE, MS, SUBRAHMANYAM MUTHUKUMAR, MCRP, CHINARO KENNEDY, DRPH, MPH, PAMELA A. MEYER, PHD, AND MARY JEAN BROWN, SCD, RN
Objective
To develop a spatial strategy to assess neighborhood risk for lead exposure and neighborhood-level blood lead testing of young children living in the city of Atlanta, Georgia. Study design This ecologic study used existing blood lead results of children aged <36 months tested and living in one of Atlanta’s 236 neighborhoods in 2005. Geographic information systems used Census, land parcel, and neighborhood spatial data to create a neighborhood priority testing index on the basis of proxies for poverty (Special Supplemental Nutrition Program for Women, Infants and Children [WIC] enrollment) and lead in house paint (year housing built). Results In 2005, only 11.9% of Atlanta’s 18 627 children aged <36 months living in the city had blood lead tests, despite a high prevalence of risk factors: 75 286 (89.6%) residential properties were built before 1978, and 44% of children were enrolled in WIC. Linear regression analysis indicated testing was significantly associated with WIC status (P < .001) but not with old housing. Conclusions This neighborhood spatial approach provided smaller geographic areas to assign risk and assess testing in a city that has a high prevalence of risk factors for lead exposure. Testing may be improved by collaboration between pediatricians and public health practitioners. (J Pediatr 2009;154:409-14)
levated blood lead levels (BLLs) in young children have been associated with acute and long-term adverse health impacts.1,2 Very elevated BLLs (⬎70 g/dL) can result in encephalopathy, seizures, or death2 but are not common in the United States today.3 Elevated BLLs (ⱖ10 g/dL) in young children have been linked to learning disabilities and behavioral disorders; in addition, increasing evidence suggests that cognitive impairment occurs at BLLs ⬍ 10 g/dL.4-10 Importantly, no safe BLL has been established for children. Childhood lead poisoning prevention and control efforts in the United States have focused on testing children.11-14 Since 1997 the Centers for Disease Control and Prevention (CDC) has recommended testing children at highest risk for lead poisoning.14 Children at high risk for lead poisoning are those whose families are poor and live in old From the Centers for Disease Control and Prevention (A.V., J.S., P.M., M.B.), the Georhousing.15-17 Old housing often contains lead in paint, especially properties built before gia Childhood Lead Poisoning Prevention 1950. Beginning in the 1950s the paint industry began to reduce the amount of lead added Program (F.S.), the Center for GIS, Georgia to residential paint. Paint containing more than 0.06% (600 ppm) lead was banned for Institute of Technology (S.M.), and the Georgia, Division of Public Health, Maternal residential use in the United States in 1978.18 and Child Health Epidemiology (C.K.), AtBecause the risk for lead poisoning is not equally distributed across populations, lanta, GA. CDC recommends that states develop targeted screening plans and assess screening levels The authors declare no conflicts of interest, real or perceived. The findings and concluamong the groups targeted. Researchers and public health practitioners have developed sions in this paper are those of the authors targeted testing strategies using elevated blood lead levels and other risk factors for specific and do not necessarily represent the views 19-23 22,24 25 of the Centers for Disease Control and geographic areas, such as ZIP code, census tract, block group, or tax parPrevention. cel19,22,23,26-27 and neighborhood.28 Some have assessed testing of children enrolled in Submitted for publication Feb 8, 2008; last 19-20,29 Medicaid, a proxy for poverty and a group that should be tested at ages 12 and 24 revision received Aug 7, 2008; accepted 19,22,23 Sep 12, 2008. months. Others have focused on areas with old housing.
E
BLL CDC CMS GACLPPP
Blood lead level Centers for Disease Control and Prevention Centers for Medicare and Medicaid Services Georgia Childhood Lead Poisoning Prevention Program
GIS WIC
Geographic information system Special Supplemental Nutrition Program for Women, Infants, and Children
Reprint requests: Forrest Staley, MUP, MPHc, Georgia Lead Poisoning Prevention Programs, 2 Peachtree Street, Ste 14-472, Atlanta, GA 30303. E-mail: [email protected]. ga.us. 0022-3476/$ - see front matter Copyright © 2009 Mosby Inc. All rights reserved. 10.1016/j.jpeds.2008.09.027
409
Defining high-risk areas by ZIP code has the advantage that residents know their ZIP code and if providers are told which ZIP codes are at high risk, they can test children who reside in those areas. However, ZIP codes can cover large areas that can be heterogeneous in levels of risk, and they can change over time. Defining high-risk areas by census tract, census block, or block group or tax parcel is not typically recognizable to parents. However, this strategy can be used to create software to tag patient addresses in high-risk areas or to provide outreach to those areas. A strategy to identify risk at the neighborhood level would have two advantages. First, neighborhoods are small areas and probably more homogenous. Second, children’s parents and guardians can easily identify them. This project’s goals were to develop a spatial strategy for categorizing risk for lead exposure by neighborhood and assess lead testing in neighborhoods with several categories of risk in the city of Atlanta, Georgia.
METHODS This study focused on the city of Atlanta. To define neighborhood level categories of risk of lead exposure and to evaluate the extent of childhood blood lead testing in neighborhoods with varying levels of risk, the study integrated the following data: (1) childhood blood lead testing data, (2) residential land parcel data, (3) 2000 U.S. Census data within a geographic information system (GIS), and (4) neighborhood spatial data from the Georgia Department of Community Affairs.
Blood Lead Test Data The Georgia Childhood Lead Poisoning Prevention Program (GACLPPP) recommends routine testing of all children (1) enrolled in Medicaid—a proxy measure for poverty—at ages 12 months and 24 months and (2) who are ⬍6 years old and live in or visit properties built before 1978.30 This ecological study did not fit the definition of human subjects research. CDC analyzed existing blood lead data that was aggregated, ie, not individually identifiable. GACLPPP provided aggregated neighborhood level blood lead testing information on all children aged ⱕ36 months who were tested in 2005 and had a residential address of Atlanta, Georgia. Blood lead tests for children ⬍12 months and 24 to 36 months were included because many children are tested when they visit providers, which may not be at their first or second birthdays. With GIS, each child’s street address obtained from the blood lead data was located as a point, which was grouped by the neighborhood in which they were located. The GACLPPP created a neighborhood dataset with derived fields, including the number of children (1) tested for lead poisoning, (2) with elevated BLLs (ⱖ10 g/dL), and (3) enrolled in Georgia’s Special Supplemental Nutrition Program for Women, Infants, and Children (WIC). Percentage of children tested was calculated using number of children aged ⱕ36 months tested for lead divided by the population of children aged ⱕ36 months. 410
Vaidyanathan et al
Residential Land Parcel Data The study obtained parcel-based data for Atlanta for 1999 from the Center for GIS at the Georgia Institute of Technology in Atlanta. The land parcel dataset had graphic and descriptive components. Each graphic parcel had an associated unique identifier label. The descriptive component contained 1 record for each parcel and the same unique identifier label as the graphic component. With GIS, the descriptive attributes of each parcel were related to the parcel’s corresponding graphic feature. This process resulted in a spatial dataset of parcels with property type, such as residential, commercial, or industrial; date of structure construction; and appraised value of the structure. GIS allowed identification of the geographic center of all residential parcels and the corresponding neighborhood. Census Data The study used U.S. Census block group-level data because this level is the lowest resolution at which the Census releases disaggregated population data. However, the boundaries of the neighborhoods and the block groups in the study did not coincide. The demographic age data of children were transferred from the block groups to the neighborhoods in 2 steps with GIS. Geometric intersections between the neighborhoods and the block groups were created to enable boundary coincidence. The demographic age data of children aged ⱕ36 months were weighted by the area of the block group within each neighborhood and estimated the total number in each neighborhood. The demographic and land parcel data were then integrated with the child blood lead dataset created earlier at the neighborhood level. Spatial Strategy An index for risk of lead exposure was created for each neighborhood. The index was based on 2 surrogates for known risk factors: old housing and poverty. Housing units were categorized for 2 levels of risk: built before 1950 (pre1950), and built before 1978 (pre-1978). Housing units built before 1950 (pre-1950) indicate the highest risk for containing lead paint and for having higher concentrations of lead in the paint. Housing units built before 1978 (pre-1978), the year more than 0.06% lead in residential paint was banned, indicate a risk for containing lead paint. The percentage of pre-1950 housing was calculated with number of residential land parcels with pre-1950 housing units divided by the total number of residential land parcels. Similarly, the percentage of pre-1978 housing was calculated with the number of residential land parcels with pre-1978 housing units divided by total number of residential land parcels. The percentage of children enrolled in WIC was calculated by use of the number of children aged ⱕ36 months enrolled in WIC divided by the population of children aged ⱕ36 months. The Medicaid enrollment data could not be obtained. Instead, WIC enrollment data were used as a proxy measure for poverty. In Georgia, it is estimated that nearly 100% of children eligible for WIC are enrolled before age The Journal of Pediatrics • March 2009
Figure 1. Percentage of pre-1978 housing and percentage of WIC children and percentage of children ⱕ36 months old tested by neighborhood.
1 year, after which enrollment drops. Medicaid has stricter income requirements than WIC. An estimated 70% to 80% of the WIC program participants are enrolled in Medicaid. A score from 1 to 4 was assigned for each neighborhood-level risk factor based on the quartile breaks for (1) pre-1978 housing, (2) pre-1950 housing, and (3) children enrolled in WIC (1 ⫽ 0 to 25th quartile; 2 ⫽ 26th to 50th quartile; 3 ⫽ 51st to 75th quartile; and 4 ⫽ 76th to 100th quartile). Two priority testing indexes were created by summing (1) the neighborhood pre-1978 housing score and the WIC score and (2) the neighborhood pre-1950 housing score and the WIC score. Summing these scores created a cumulative risk for each neighborhood.31 For example, a neighborhood with a high percentage of pre1950 housing, but a low percentage of children enrolled in WIC would have a lower risk than a neighborhood with the same percentage of old housing but a high percentage of children enrolled in WIC. This resulted in 2 simple hazard scores of 2 to 8 for each neighborhood, which were used to assign neighborhood priority for testing of “low,” “low-medium,” “medium,” “high-medium,” or “high.” The neighborhood level risk factors for housing and WIC in Atlanta were not well correlated. To test the statistical significance of the association between a neighborhood risk factor and the percentage of children tested, the Kruskal-Wallis test for association and the Jonckheere-Terpstra test for trend32 were used. Linear regression was used to evaluate simultaneously the association of housing age and WIC status with the percentage of children tested in neighborhoods. Finally, maps were created in GIS (ArcGIS 9, ESRI Redlands, California) to evaluate visually the percentages of older housing units in the neighborhoods, children enrolled in WIC, and children tested for lead poisoning in Atlanta.
RESULTS Neighborhood Risk The city of Atlanta has 236 neighborhoods (median area: 0.29 square miles; range: 0.02-4.36 square miles) containing 87 791 properties, of which 95.7% were residential. Of the residential properties, 75 286 (89.6%) were pre-1978 and 47 142 (56.1%) were pre-1950. The median number of housing units built before 1950 per neighborhood is 47 (range, 0-2124). The median number of housing units built before 1978 per neighborhood is 163 (range, 0-2480). There were 18 627 children aged ⱕ36 months living in Atlanta with an estimated median of 39 children (range, 0-406) per neighborhood. There were 8229 (44.2%) children aged ⱕ 36 months enrolled in WIC living in the city of Atlanta. The median number of children aged ⱕ 36 months enrolled in WIC per neighborhood was 11 (range, 0-254). South Atlanta neighborhoods tended to have lower median incomes leading to higher percentages of children enrolled in WIC (Figure 1). Neighborhoods in east central Atlanta had a higher percentage of pre-1950 housing and neighborhoods in west and south Atlanta had a lower percentage of pre-1950 housing than the overall median percentage of pre-1950 housing for Atlanta (Figure 2). Blood Lead Testing In 2005, 2231 (11.9%) children, were tested for lead poisoning; of these, 22 (1.0%) had elevated BLLs. The median number of children tested for lead poisoning by neighborhood was 4 (range: 0-135) and the median percentage tested for lead by neighborhoods was 8.3%. Testing did not significantly increase with increasing percentages of either pre-1978 housing or pre-1950 housing (Kruskal-Wallis, P ⫽ 0.991 and 0.520, respectively). However, testing significantly
Screening for Lead Poisoning: A Geospatial Approach to Determine Testing of Children in At-Risk Neighborhoods
411
Figure 2. Percentage of pre-1950 housing and percentage of children tested by neighborhood.
increased in neighborhoods with higher percentages of children enrolled in WIC (Kruskal-Wallis, P ⬍ .001). The Table shows the median percentage of children tested by priority testing indexes. Testing was high in neighborhoods with both a high percentage of old housing and children enrolled in WIC, whereas it was low in neighborhoods with a high percentage of old housing and a low percentage of children enrolled in WIC. The median percentage of children tested increased from low- to high-priority neighborhoods and demonstrated a significant monotonic trend (Jonckheere-Terpstra, P ⬍ .001). Linear regression showed that the percentage of children being tested in neighborhoods was significantly associated with WIC status (P ⬍ .001) but was not associated with the percentage of either pre-1950 or pre-1978 housing.
DISCUSSION This neighborhood spatial approach provided smaller geographic areas to assign risk and assess testing in a city that has a high prevalence of risk factors for lead exposure. Categorizing risk for lead poisoning is important, both for clinicians to determine which child to test and for public health practitioners to assess the extent to which children at high risk 412
Vaidyanathan et al
are being tested—and being tested in compliance with testing guidelines and laws—and to target lead poisoning prevention resources. This study explored the feasibility of using a small geographic resolution that can be easily recognized by residents: the neighborhood. Usually, neighborhoods are named after major landmarks and their boundaries align with road segments. Neighborhood geography is often used by planning departments. Neighborhoods often have 1 or more community organizations that address key issues of interest to the community, including health. These organizations can be used to address community health issues and are convenient for outreach and education. It is increasingly recognized that variations in neighborhood conditions are critical to health outcomes and program options—in almost all urban areas, serious health problems are highly concentrated in a fairly small number of distressed neighborhoods.40 The need for local data to address local problems initiated the National Neighborhood Indicators Project. There are currently 30 cities that are developing neighborhood level indicators for use in local policymaking and community building. In some areas of the United States, neighborhood geography may not be available and in those areas, state and local health departments may find other geographic scales that work better depending on their local resources. However, this simple summing approach followed by categorizing the risk scores to create the priority testing index was easy to implement in a GIS, independent of geographic scale and could be replicated on a state-wide basis. The neighborhood risk assessment focused on 2 welldocumented risk factors: old housing and poverty. The lack of association between testing and old housing is a particular concern in Atlanta, a city in which nearly 90% of residential units were built before 1978. Strategically testing children who live in old housing is a good prevention strategy because lead-based residential paint and lead-contaminated house dust and soil are the most common high-dose sources of lead exposure for children in the United States.33-35 Targeting children living in properties built before 1978 is a conservative approach because it assumes that children living in all pre1978 housing may be exposed to lead paint and dust. However, if large numbers of children live in pre-1978 housing and they cannot all be tested, focusing on those children at highest risk because they live in pre-1950 housing may be necessary and advantageous. Testing strategies consider local resources and the local sources of lead, which vary across the United States. CDC and the state childhood lead poisoning prevention programs recommend that providers assess young children’s risk for lead poisoning. A commonly asked question is: Does the child live in or visit properties built before 1978 or 1950? However, one study found that parents and guardians may not know when their homes were built.36 Furthermore, children who live in old housing, regardless of income level, can be exposed to lead, especially if the home is renovated or remodeled.37,38 Another group of children that should be tested for lead poisoning are those enrolled in Medicaid. The recommendaThe Journal of Pediatrics • March 2009
Table. Median percentage of children tested for lead poisoning in neighborhoods based on housing characteristics and children enrolled in WIC by priority testing category Neighborhoods calculated using pre-1978 housing and children enrolled in WIC Priority testing index (housing ⴙ WIC scores) 2 3 4 5 6 7 8 Total
Priority testing category Low Low-medium Medium High-medium High
Neighborhoods calculated using pre-1950 housing and children enrolled in WIC
Number (and percentage) of neighborhoods
Median percentage of children tested
Number (and percentage) of neighborhoods
Median percentage of children tested
18 (7.6%) 33 (14.0%) 41 (17.4%) 57 (24.2%) 36 (15.3%) 28 (11.9%) 23 (9.7%) 236 (100.0%)
0.00 5.20 4.20 10.80 8.30 12.80 18.10 P ⬍ .001*
16 (6.8%) 25 (10.6%) 38 (16.1%) 70 (29.7%) 51 (21.6%) 23 (9.7%) 13 (5.5%) 236 (100.0%)
0.00 1.80 4.80 11.10 8.80 12.10 25.00 P ⬍ .001*
*Jonckheere-Terpstra test for trend in association between priority type and lead testing, P ⬍ .05 statistically significant.
tion to test this group is based on a national survey that found that 83% of children with BLLs ⱖ 20 g/dL were Medicaid recipients.39 Although GACLPPP did not have access to Medicaid data when this analysis was initiated, GACLPPP did have access to data on children enrolled in WIC, another proxy for poverty. Both Medicaid and WIC have income eligibility requirements, but Medicaid has stricter income eligibility requirements than WIC, and consequently fewer children are eligible for Medicaid. The findings indicate that testing increased in neighborhoods with higher percentages of children who were enrolled in the WIC program. This finding is consistent with a South Carolina study.19 WIC clinics may provide a viable source for testing children at risk of lead poisoning and may help to improve testing rates. Although testing children in Medicaid is legally mandated, not all WIC children are enrolled in Medicaid. Thus a successful partnership with WIC may help to target and test most children at the highest risk for lead poisoning. This study had several limitations. First, analyzing neighborhood-level data is an ecologic design that lacks the precision of using individual child-level data. This approach was chosen because it could be replicated relatively easily on a statewide basis. Second, the datasets used in these analyses covered different time periods. The census and the land parcels data represented a snapshot of the neighborhood layout in 1999, and the WIC enrollment and the blood lead testing data represented information collected in 2005. This difference may cause potentially inaccurate calculations for risk and percent tested. Furthermore, children migrating within Atlanta could have caused some miscalculation in the number of children enrolled in the WIC program. Third, the land parcels dataset did not include the total number of housing units in multifamily residential parcels, and each multifamily parcel was treated as having 1 unit. This treatment resulted in somewhat conservative estimates of housing risk; however, even with an accurate number of units, high testing would not be
expected in neighborhoods with high housing risk, given that a conservative estimate also showed low testing rates. Fourth, although the land parcels dataset included dates of construction, it did not contain any details about whether or not a structure had been remediated or made safe from lead. Fifth, area-weighted analyses using census data by block groups to derive demographic data by neighborhoods assumes uniform population density. The error in calculating percent tested and percent of WIC children diminishes if the area under consideration is highly populated. In Atlanta, block groups are small and uniformly populated; whether or not this characteristic is a source of bias in this analysis is unclear. Finally, this analysis does not consider other sources of lead poisoning, such as imported food items or cooking utensils that are used in immigrant populations. The low percentage of children tested for lead exposure, especially in neighborhoods with old homes prompted the GACLPPP to apply a risk index approach across the state. In an effort to increase testing of high risk children, the GACLPPP is working with the state immunization program to include the lead risk index information into the registry. When a child presents to a provider’s office, the provider can check the immunization registry and know the child’s risk for lead exposure. Blood lead testing is the only simple way to know if a child has recently been exposed to lead. Unlike other diseases in which a child’s symptoms motivate parents to seek health care, elevated BLLs may cause no specific symptoms to suggest exposure to lead. When children are found to have elevated BLLs, state and local lead poisoning prevention programs typically conduct investigations to identify potential sources of lead in the child’s environment. Removing or controlling lead sources is the best way to manage elevated blood BLLs.41 However, children have already been harmed when an elevated BLL is found. Blood lead testing is secondary prevention and will remain an important prevention strategy. However, it is im-
Screening for Lead Poisoning: A Geospatial Approach to Determine Testing of Children in At-Risk Neighborhoods
413
portant to identify and remove sources before children are ever exposed, that is, primary prevention. One approach is to screen housing.34 The risk index can help not only identify children who should be tested, but also identify neighborhoods that could benefit from remediation. Understanding the areas of high risk, especially at smaller geographic resolution and the extent of testing among high-risk groups will allow clinicians and public health practitioners to make informed decisions about providing interventions. Improving testing of children at greatest risk and making homes safe from lead must be priorities to achieve the national goal of eliminating elevated BLLs among young children in the United States by 2010.42
REFERENCES 1. Agency for Toxic Substances and Disease Registry. US Department of Health and Human Services, Public Health Service. Draft toxicological profile for lead. Atlanta: Agency for Toxic Substances and Disease Registry; 2005 [cited 2006 Oct 20]. Available from: http://www.atsdr.cdc.gov/toxprofiles/tp13.html. 2. Lidsky TI, Schneider JS. Lead neurotoxicity in children: basic mechanisms and clinical correlates. Brain 2003; 126:5-19. 3. CDC. Surveillance for elevated blood lead levels among children—United States, 1997—2001. MMWR 2003 [cited 2006 Oct 20]; 52(SS10); 1-21. Available from: http://www.cdc.gov/mmwr/preview/mmwrhtml/ss5210a1.htm. 4. Schwartz J. Low-level lead exposure and children’s IQ: a meta-analysis and search for a threshold. Environ Res 1994; 65:42-55. 5. Lanphear BP, Dietrich K, Auinger P, Cox C. Cognitive deficits associated with blood lead concentrations ⬍10 g/dL in US children and adolescents. Public Health Rep 2000; 115:521-9. 6. Canfield RL, Henderson CR, Cory-Slechta DA, Cox C, Jusko TA, Lanphear BP. Intellectual impairment in children with blood lead concentrations below 10 g per deciliter. N Engl J Med 2003; 348:1517-26. 7. Bellinger DC, Needleman HL. Intellectual impairment and blood lead levels. N Engl J Med 2003; 349:500-502. 8. Lanphear BP, Hornung R, Khoury J, Yolton K, Baghurst P, Bellinger DC et al., Low-level environmental lead exposure and children’s intellectual function: an international pooled analysis. Environ Health Perspect 2005;113:894-899. 9. Tellez-Rojo MM, Bellinger DC, Arroyo-Quiroz C, Lamadrid-Figueroa H, Mercado-Garcia A, Schnaas-Arrieta L. et al., Longitudinal associations between blood lead concentrations lower than 10 microg/dL and neurobehavioral development in environmentally exposed children in Mexico City. Pediatrics 2006;118:e323-e330. 10. CDC. Preventing lead poisoning in young children: a statement from the Centers for Disease Control and Prevention. Atlanta: CDC; 2005 [cited 2006 Oct 20]. Available from: http://www.cdc.gov/nceh/lead/publications/PrevLeadPoisoning.pdf. 11. CDC. Increased lead absorption and lead poisoning in young children: a statement by the Center for Disease Control. Atlanta: US Department of Health, Education, and Welfare, CDC; 1975 [cited 2006 Oct 20]. Available from: http://www.cdc.gov/ nceh/lead/Publications/books/plpyc/plpyc_history/increaseLead_1975.pdf. 12. CDC. Preventing lead poisoning in young children: a statement by the Centers for Disease Control, DHHS publication no. (CDC) 99-2230. Atlanta: US Department of Health and Human Services, CDC; 1985 [cited 2006 Oct 20]. Available from: http://www.cdc.gov/nceh/lead/Publications/books/plpyc/plpyc_history/CLP_1985.pdf. 13. CDC. Preventing lead poisoning in young children: a statement by the Centers for Disease Control - October 1991. Atlanta: US Department of Health and Human Services, Public Health Service, CDC; 1991 [cited 2006 Oct 20]. Available from: http://www.cdc.gov/nceh/lead/Publications/books/plpyc/contents.htm. 14. CDC. Screening young children for lead poisoning: guidance for state and local public health officials. Atlanta: US Department of Health and Human Services, Public Health Service, CDC; 1997 [cited 2006 Oct 20]. Available from: http://www.cdc. gov/nceh/lead/guide/guide97.htm. 15. Pirkle JL, Kaufmann RB, Brody DJ, Hickman T, Gunter EW, Paschal DC. Exposure of the U.S. population to lead, 1991-1994. Environ Health Perspect 1998;106:745-50. 16. Kaufmann RB, Clouse TL, Olson DR, Matte TD. Elevated blood lead levels and blood lead screening among US children aged one to five years: 1988--1994. Pediatrics 2000; 106:E79. 17. CDC. Blood lead levels—United States, 1999 —2002. MMWR 2005 [cited 2006 Oct
414
Vaidyanathan et al
20]; 54(20); 513-516. Available from: http://www.cdc.gov/mmwr/preview/mmwrhtml/ mm5420a5.htm. 18. US Consumer Product Safety Commission. 1977. Ban of Lead-Containing Paint and Certain Consumer Products Bearing Lead-Containing Paint. 16 CFR 1303. Fed Reg 42:44199. 19. Roberts JR, Reigart JR. Screening for lead poisoning in South Carolina. J SC Med Assoc 2001;87:459-464. 20. Kemper AR, Cohn LM, Fant KE, Dombkowski KJ. Arch Pediatr Adolesc Med 2005;159:646-650. 21. Haley VB, Talbot TO. Geographic analysis of blood lead levels in New York State Children Born 1994-1997. Environ Health Perspect 2004;112:1577-1582. 22. Reissman DB, Staley F, Curtis GB, Kaufmann RB. Use of geographic information system technology to aid health department decision making about childhood lead poisoning prevention activities. Environ Health Perspect 2001;109:89-94. 23. Roberts JR, Hulsey TC, Curtis GB, Reigart JR. Using geographic information systems to assess risk for elevated blood lead levels in children. Pub Health Reports 2003;118:221-229. 24. Sargent JD, Bailey A, Simon P, Blake M, Dalton MA. Census tract analysis of lead exposure in Rhode Island children. Environ Res 1997;74:159-68. 25. Lanphear BP, Byrd RS, Auinger P, Schaffer SJ. Community characteristics associated with elevated blood lead levels in children. Pediatrics 1998;101:264-271. 26. Reyes NL, Wong L-Y, MacRoy PM, Curtis G, Meyer PA, Evens A, Brown MJ. Identifying housing that poisons: a critical step in eliminating childhood lead poisoning. J Public Health Management Practice 2006;12:564-569. 27. Miranda ML, Dolinoy DC, Overstreet MA. Mapping for prevention: GIS Models for directing childhood lead poisoning prevention programs. Environ Health Perspect 2002;110:947-953. 28. Oyana TJ, Margai FM. Geographic analysis of health risks of pediatric lead exposure: A golden opportunity to promote healthy neighborhoods. Arch Environ Occup Health 2007/62:93-104. 29. Vivier PM, Hogan JW, Simon P, Leddy T, Dansereau LM, Alario AJ. A statewide assessment of lead screening histories of pre-school children enrolled in a Medicaid managed care program. Pediatrics 2001;108 (2). 30. Georgia Department of Human Resources, Division of Public Health. Georgia Childhood Lead Poisoning Prevention Program (GCLPPP). Screening Guidelines: Children. Available from: http://health.state.ga.us/programs/lead/screeningtypes.asp 31. Wartenberg D. Screening for lead exposure using a geographic information system. Environ Res 1992;59:310-317. 32. Jonckheere AR. A distribution free K sample test against ordered alternatives. Biometrika 1954; 41:133-145. 33. Lanphear BP, Hornung R, Ho M, Howard CR, Eberle S, Knauf K. Environmental lead exposure during early childhood. J Pediatr 2002;140:40-7. 34. Lanphear BP, Hornung R, Ho M. Screening housing to prevent lead toxicity in children Public Health Reports 2005;120:305-310. 35. CDC. Preventing lead exposure in young children: a housing-based approach to primary prevention of lead poisoning. Atlanta: CDC; 2004 [cited 2006 Oct 20]. Available from: http://www.cdc.gov/nceh/lead/publications/Primary%20Prevention%20 Document.pdf. 36. Schwab LT, Roberts JR, Reigart JR. Inaccuracy in parental reporting of the age of their home for lead screening purposes. Arch Pediatrics and Adolesc Medicine 2003;157:584-586. 37. Reissman DB, Matte TD, Gurnitz KL, Kaufmann RB, Leighton J. Is home renovation or repair a risk factor for exposure to lead among children residing in New York City? J of Urban Health 2002;79:502-11. 38. CDC. Children with elevated blood lead levels attributed to home renovation and remodeling activities - New York, -1994. 1993MMWR 1997;45:1120-23. 39. US General Accounting Office. Medicaid: Elevated blood lead levels in children. Washington, DC: GAO; 1998. 40. CDC. Managing elevated blood lead levels among young children: recommendations from the Advisory Committee on Childhood Lead Poisoning Prevention. Atlanta: US Department of Health and Human Services, Public Health Service, CDC; 2002 [cited 2006 Oct 20]. Available from: http://www.cdc.gov/nceh/lead/ CaseManagement/caseManage_main.htm. 41. Pettit KLS, Kingsley GT, Coulton CJ, Cigna J. Neighborhoods and health: building evidence for local policy. Submitted to The Office of the Assistant Secretary for Planning and Evaluation, US Department of Health and Human Services. Submitted by The Urban Institute. May 2003. Available from: http://aspe.hhs.gov/ hsp/neighborhoods-health03 42. US Department of Health and Human Services. Healthy people 2010 (conference ed, in 2 vols). Washington, DC: US Department of Health and Human Services; 2000 [cited 2006 Oct 20]. Available from: http://www.healthypeople.gov.
The Journal of Pediatrics • March 2009
Objective
To develop a spatial strategy to assess neighborhood risk for lead exposure and neighborhood-level blood lead testing of young children living in the city of Atlanta, Georgia. Study design This ecologic study used existing blood lead results of children aged <36 months tested and living in one of Atlanta’s 236 neighborhoods in 2005. Geographic information systems used Census, land parcel, and neighborhood spatial data to create a neighborhood priority testing index on the basis of proxies for poverty (Special Supplemental Nutrition Program for Women, Infants and Children [WIC] enrollment) and lead in house paint (year housing built). Results In 2005, only 11.9% of Atlanta’s 18 627 children aged <36 months living in the city had blood lead tests, despite a high prevalence of risk factors: 75 286 (89.6%) residential properties were built before 1978, and 44% of children were enrolled in WIC. Linear regression analysis indicated testing was significantly associated with WIC status (P < .001) but not with old housing. Conclusions This neighborhood spatial approach provided smaller geographic areas to assign risk and assess testing in a city that has a high prevalence of risk factors for lead exposure. Testing may be improved by collaboration between pediatricians and public health practitioners. (J Pediatr 2009;154:409-14)
levated blood lead levels (BLLs) in young children have been associated with acute and long-term adverse health impacts.1,2 Very elevated BLLs (⬎70 g/dL) can result in encephalopathy, seizures, or death2 but are not common in the United States today.3 Elevated BLLs (ⱖ10 g/dL) in young children have been linked to learning disabilities and behavioral disorders; in addition, increasing evidence suggests that cognitive impairment occurs at BLLs ⬍ 10 g/dL.4-10 Importantly, no safe BLL has been established for children. Childhood lead poisoning prevention and control efforts in the United States have focused on testing children.11-14 Since 1997 the Centers for Disease Control and Prevention (CDC) has recommended testing children at highest risk for lead poisoning.14 Children at high risk for lead poisoning are those whose families are poor and live in old From the Centers for Disease Control and Prevention (A.V., J.S., P.M., M.B.), the Georhousing.15-17 Old housing often contains lead in paint, especially properties built before gia Childhood Lead Poisoning Prevention 1950. Beginning in the 1950s the paint industry began to reduce the amount of lead added Program (F.S.), the Center for GIS, Georgia to residential paint. Paint containing more than 0.06% (600 ppm) lead was banned for Institute of Technology (S.M.), and the Georgia, Division of Public Health, Maternal residential use in the United States in 1978.18 and Child Health Epidemiology (C.K.), AtBecause the risk for lead poisoning is not equally distributed across populations, lanta, GA. CDC recommends that states develop targeted screening plans and assess screening levels The authors declare no conflicts of interest, real or perceived. The findings and concluamong the groups targeted. Researchers and public health practitioners have developed sions in this paper are those of the authors targeted testing strategies using elevated blood lead levels and other risk factors for specific and do not necessarily represent the views 19-23 22,24 25 of the Centers for Disease Control and geographic areas, such as ZIP code, census tract, block group, or tax parPrevention. cel19,22,23,26-27 and neighborhood.28 Some have assessed testing of children enrolled in Submitted for publication Feb 8, 2008; last 19-20,29 Medicaid, a proxy for poverty and a group that should be tested at ages 12 and 24 revision received Aug 7, 2008; accepted 19,22,23 Sep 12, 2008. months. Others have focused on areas with old housing.
E
BLL CDC CMS GACLPPP
Blood lead level Centers for Disease Control and Prevention Centers for Medicare and Medicaid Services Georgia Childhood Lead Poisoning Prevention Program
GIS WIC
Geographic information system Special Supplemental Nutrition Program for Women, Infants, and Children
Reprint requests: Forrest Staley, MUP, MPHc, Georgia Lead Poisoning Prevention Programs, 2 Peachtree Street, Ste 14-472, Atlanta, GA 30303. E-mail: [email protected]. ga.us. 0022-3476/$ - see front matter Copyright © 2009 Mosby Inc. All rights reserved. 10.1016/j.jpeds.2008.09.027
409
Defining high-risk areas by ZIP code has the advantage that residents know their ZIP code and if providers are told which ZIP codes are at high risk, they can test children who reside in those areas. However, ZIP codes can cover large areas that can be heterogeneous in levels of risk, and they can change over time. Defining high-risk areas by census tract, census block, or block group or tax parcel is not typically recognizable to parents. However, this strategy can be used to create software to tag patient addresses in high-risk areas or to provide outreach to those areas. A strategy to identify risk at the neighborhood level would have two advantages. First, neighborhoods are small areas and probably more homogenous. Second, children’s parents and guardians can easily identify them. This project’s goals were to develop a spatial strategy for categorizing risk for lead exposure by neighborhood and assess lead testing in neighborhoods with several categories of risk in the city of Atlanta, Georgia.
METHODS This study focused on the city of Atlanta. To define neighborhood level categories of risk of lead exposure and to evaluate the extent of childhood blood lead testing in neighborhoods with varying levels of risk, the study integrated the following data: (1) childhood blood lead testing data, (2) residential land parcel data, (3) 2000 U.S. Census data within a geographic information system (GIS), and (4) neighborhood spatial data from the Georgia Department of Community Affairs.
Blood Lead Test Data The Georgia Childhood Lead Poisoning Prevention Program (GACLPPP) recommends routine testing of all children (1) enrolled in Medicaid—a proxy measure for poverty—at ages 12 months and 24 months and (2) who are ⬍6 years old and live in or visit properties built before 1978.30 This ecological study did not fit the definition of human subjects research. CDC analyzed existing blood lead data that was aggregated, ie, not individually identifiable. GACLPPP provided aggregated neighborhood level blood lead testing information on all children aged ⱕ36 months who were tested in 2005 and had a residential address of Atlanta, Georgia. Blood lead tests for children ⬍12 months and 24 to 36 months were included because many children are tested when they visit providers, which may not be at their first or second birthdays. With GIS, each child’s street address obtained from the blood lead data was located as a point, which was grouped by the neighborhood in which they were located. The GACLPPP created a neighborhood dataset with derived fields, including the number of children (1) tested for lead poisoning, (2) with elevated BLLs (ⱖ10 g/dL), and (3) enrolled in Georgia’s Special Supplemental Nutrition Program for Women, Infants, and Children (WIC). Percentage of children tested was calculated using number of children aged ⱕ36 months tested for lead divided by the population of children aged ⱕ36 months. 410
Vaidyanathan et al
Residential Land Parcel Data The study obtained parcel-based data for Atlanta for 1999 from the Center for GIS at the Georgia Institute of Technology in Atlanta. The land parcel dataset had graphic and descriptive components. Each graphic parcel had an associated unique identifier label. The descriptive component contained 1 record for each parcel and the same unique identifier label as the graphic component. With GIS, the descriptive attributes of each parcel were related to the parcel’s corresponding graphic feature. This process resulted in a spatial dataset of parcels with property type, such as residential, commercial, or industrial; date of structure construction; and appraised value of the structure. GIS allowed identification of the geographic center of all residential parcels and the corresponding neighborhood. Census Data The study used U.S. Census block group-level data because this level is the lowest resolution at which the Census releases disaggregated population data. However, the boundaries of the neighborhoods and the block groups in the study did not coincide. The demographic age data of children were transferred from the block groups to the neighborhoods in 2 steps with GIS. Geometric intersections between the neighborhoods and the block groups were created to enable boundary coincidence. The demographic age data of children aged ⱕ36 months were weighted by the area of the block group within each neighborhood and estimated the total number in each neighborhood. The demographic and land parcel data were then integrated with the child blood lead dataset created earlier at the neighborhood level. Spatial Strategy An index for risk of lead exposure was created for each neighborhood. The index was based on 2 surrogates for known risk factors: old housing and poverty. Housing units were categorized for 2 levels of risk: built before 1950 (pre1950), and built before 1978 (pre-1978). Housing units built before 1950 (pre-1950) indicate the highest risk for containing lead paint and for having higher concentrations of lead in the paint. Housing units built before 1978 (pre-1978), the year more than 0.06% lead in residential paint was banned, indicate a risk for containing lead paint. The percentage of pre-1950 housing was calculated with number of residential land parcels with pre-1950 housing units divided by the total number of residential land parcels. Similarly, the percentage of pre-1978 housing was calculated with the number of residential land parcels with pre-1978 housing units divided by total number of residential land parcels. The percentage of children enrolled in WIC was calculated by use of the number of children aged ⱕ36 months enrolled in WIC divided by the population of children aged ⱕ36 months. The Medicaid enrollment data could not be obtained. Instead, WIC enrollment data were used as a proxy measure for poverty. In Georgia, it is estimated that nearly 100% of children eligible for WIC are enrolled before age The Journal of Pediatrics • March 2009
Figure 1. Percentage of pre-1978 housing and percentage of WIC children and percentage of children ⱕ36 months old tested by neighborhood.
1 year, after which enrollment drops. Medicaid has stricter income requirements than WIC. An estimated 70% to 80% of the WIC program participants are enrolled in Medicaid. A score from 1 to 4 was assigned for each neighborhood-level risk factor based on the quartile breaks for (1) pre-1978 housing, (2) pre-1950 housing, and (3) children enrolled in WIC (1 ⫽ 0 to 25th quartile; 2 ⫽ 26th to 50th quartile; 3 ⫽ 51st to 75th quartile; and 4 ⫽ 76th to 100th quartile). Two priority testing indexes were created by summing (1) the neighborhood pre-1978 housing score and the WIC score and (2) the neighborhood pre-1950 housing score and the WIC score. Summing these scores created a cumulative risk for each neighborhood.31 For example, a neighborhood with a high percentage of pre1950 housing, but a low percentage of children enrolled in WIC would have a lower risk than a neighborhood with the same percentage of old housing but a high percentage of children enrolled in WIC. This resulted in 2 simple hazard scores of 2 to 8 for each neighborhood, which were used to assign neighborhood priority for testing of “low,” “low-medium,” “medium,” “high-medium,” or “high.” The neighborhood level risk factors for housing and WIC in Atlanta were not well correlated. To test the statistical significance of the association between a neighborhood risk factor and the percentage of children tested, the Kruskal-Wallis test for association and the Jonckheere-Terpstra test for trend32 were used. Linear regression was used to evaluate simultaneously the association of housing age and WIC status with the percentage of children tested in neighborhoods. Finally, maps were created in GIS (ArcGIS 9, ESRI Redlands, California) to evaluate visually the percentages of older housing units in the neighborhoods, children enrolled in WIC, and children tested for lead poisoning in Atlanta.
RESULTS Neighborhood Risk The city of Atlanta has 236 neighborhoods (median area: 0.29 square miles; range: 0.02-4.36 square miles) containing 87 791 properties, of which 95.7% were residential. Of the residential properties, 75 286 (89.6%) were pre-1978 and 47 142 (56.1%) were pre-1950. The median number of housing units built before 1950 per neighborhood is 47 (range, 0-2124). The median number of housing units built before 1978 per neighborhood is 163 (range, 0-2480). There were 18 627 children aged ⱕ36 months living in Atlanta with an estimated median of 39 children (range, 0-406) per neighborhood. There were 8229 (44.2%) children aged ⱕ 36 months enrolled in WIC living in the city of Atlanta. The median number of children aged ⱕ 36 months enrolled in WIC per neighborhood was 11 (range, 0-254). South Atlanta neighborhoods tended to have lower median incomes leading to higher percentages of children enrolled in WIC (Figure 1). Neighborhoods in east central Atlanta had a higher percentage of pre-1950 housing and neighborhoods in west and south Atlanta had a lower percentage of pre-1950 housing than the overall median percentage of pre-1950 housing for Atlanta (Figure 2). Blood Lead Testing In 2005, 2231 (11.9%) children, were tested for lead poisoning; of these, 22 (1.0%) had elevated BLLs. The median number of children tested for lead poisoning by neighborhood was 4 (range: 0-135) and the median percentage tested for lead by neighborhoods was 8.3%. Testing did not significantly increase with increasing percentages of either pre-1978 housing or pre-1950 housing (Kruskal-Wallis, P ⫽ 0.991 and 0.520, respectively). However, testing significantly
Screening for Lead Poisoning: A Geospatial Approach to Determine Testing of Children in At-Risk Neighborhoods
411
Figure 2. Percentage of pre-1950 housing and percentage of children tested by neighborhood.
increased in neighborhoods with higher percentages of children enrolled in WIC (Kruskal-Wallis, P ⬍ .001). The Table shows the median percentage of children tested by priority testing indexes. Testing was high in neighborhoods with both a high percentage of old housing and children enrolled in WIC, whereas it was low in neighborhoods with a high percentage of old housing and a low percentage of children enrolled in WIC. The median percentage of children tested increased from low- to high-priority neighborhoods and demonstrated a significant monotonic trend (Jonckheere-Terpstra, P ⬍ .001). Linear regression showed that the percentage of children being tested in neighborhoods was significantly associated with WIC status (P ⬍ .001) but was not associated with the percentage of either pre-1950 or pre-1978 housing.
DISCUSSION This neighborhood spatial approach provided smaller geographic areas to assign risk and assess testing in a city that has a high prevalence of risk factors for lead exposure. Categorizing risk for lead poisoning is important, both for clinicians to determine which child to test and for public health practitioners to assess the extent to which children at high risk 412
Vaidyanathan et al
are being tested—and being tested in compliance with testing guidelines and laws—and to target lead poisoning prevention resources. This study explored the feasibility of using a small geographic resolution that can be easily recognized by residents: the neighborhood. Usually, neighborhoods are named after major landmarks and their boundaries align with road segments. Neighborhood geography is often used by planning departments. Neighborhoods often have 1 or more community organizations that address key issues of interest to the community, including health. These organizations can be used to address community health issues and are convenient for outreach and education. It is increasingly recognized that variations in neighborhood conditions are critical to health outcomes and program options—in almost all urban areas, serious health problems are highly concentrated in a fairly small number of distressed neighborhoods.40 The need for local data to address local problems initiated the National Neighborhood Indicators Project. There are currently 30 cities that are developing neighborhood level indicators for use in local policymaking and community building. In some areas of the United States, neighborhood geography may not be available and in those areas, state and local health departments may find other geographic scales that work better depending on their local resources. However, this simple summing approach followed by categorizing the risk scores to create the priority testing index was easy to implement in a GIS, independent of geographic scale and could be replicated on a state-wide basis. The neighborhood risk assessment focused on 2 welldocumented risk factors: old housing and poverty. The lack of association between testing and old housing is a particular concern in Atlanta, a city in which nearly 90% of residential units were built before 1978. Strategically testing children who live in old housing is a good prevention strategy because lead-based residential paint and lead-contaminated house dust and soil are the most common high-dose sources of lead exposure for children in the United States.33-35 Targeting children living in properties built before 1978 is a conservative approach because it assumes that children living in all pre1978 housing may be exposed to lead paint and dust. However, if large numbers of children live in pre-1978 housing and they cannot all be tested, focusing on those children at highest risk because they live in pre-1950 housing may be necessary and advantageous. Testing strategies consider local resources and the local sources of lead, which vary across the United States. CDC and the state childhood lead poisoning prevention programs recommend that providers assess young children’s risk for lead poisoning. A commonly asked question is: Does the child live in or visit properties built before 1978 or 1950? However, one study found that parents and guardians may not know when their homes were built.36 Furthermore, children who live in old housing, regardless of income level, can be exposed to lead, especially if the home is renovated or remodeled.37,38 Another group of children that should be tested for lead poisoning are those enrolled in Medicaid. The recommendaThe Journal of Pediatrics • March 2009
Table. Median percentage of children tested for lead poisoning in neighborhoods based on housing characteristics and children enrolled in WIC by priority testing category Neighborhoods calculated using pre-1978 housing and children enrolled in WIC Priority testing index (housing ⴙ WIC scores) 2 3 4 5 6 7 8 Total
Priority testing category Low Low-medium Medium High-medium High
Neighborhoods calculated using pre-1950 housing and children enrolled in WIC
Number (and percentage) of neighborhoods
Median percentage of children tested
Number (and percentage) of neighborhoods
Median percentage of children tested
18 (7.6%) 33 (14.0%) 41 (17.4%) 57 (24.2%) 36 (15.3%) 28 (11.9%) 23 (9.7%) 236 (100.0%)
0.00 5.20 4.20 10.80 8.30 12.80 18.10 P ⬍ .001*
16 (6.8%) 25 (10.6%) 38 (16.1%) 70 (29.7%) 51 (21.6%) 23 (9.7%) 13 (5.5%) 236 (100.0%)
0.00 1.80 4.80 11.10 8.80 12.10 25.00 P ⬍ .001*
*Jonckheere-Terpstra test for trend in association between priority type and lead testing, P ⬍ .05 statistically significant.
tion to test this group is based on a national survey that found that 83% of children with BLLs ⱖ 20 g/dL were Medicaid recipients.39 Although GACLPPP did not have access to Medicaid data when this analysis was initiated, GACLPPP did have access to data on children enrolled in WIC, another proxy for poverty. Both Medicaid and WIC have income eligibility requirements, but Medicaid has stricter income eligibility requirements than WIC, and consequently fewer children are eligible for Medicaid. The findings indicate that testing increased in neighborhoods with higher percentages of children who were enrolled in the WIC program. This finding is consistent with a South Carolina study.19 WIC clinics may provide a viable source for testing children at risk of lead poisoning and may help to improve testing rates. Although testing children in Medicaid is legally mandated, not all WIC children are enrolled in Medicaid. Thus a successful partnership with WIC may help to target and test most children at the highest risk for lead poisoning. This study had several limitations. First, analyzing neighborhood-level data is an ecologic design that lacks the precision of using individual child-level data. This approach was chosen because it could be replicated relatively easily on a statewide basis. Second, the datasets used in these analyses covered different time periods. The census and the land parcels data represented a snapshot of the neighborhood layout in 1999, and the WIC enrollment and the blood lead testing data represented information collected in 2005. This difference may cause potentially inaccurate calculations for risk and percent tested. Furthermore, children migrating within Atlanta could have caused some miscalculation in the number of children enrolled in the WIC program. Third, the land parcels dataset did not include the total number of housing units in multifamily residential parcels, and each multifamily parcel was treated as having 1 unit. This treatment resulted in somewhat conservative estimates of housing risk; however, even with an accurate number of units, high testing would not be
expected in neighborhoods with high housing risk, given that a conservative estimate also showed low testing rates. Fourth, although the land parcels dataset included dates of construction, it did not contain any details about whether or not a structure had been remediated or made safe from lead. Fifth, area-weighted analyses using census data by block groups to derive demographic data by neighborhoods assumes uniform population density. The error in calculating percent tested and percent of WIC children diminishes if the area under consideration is highly populated. In Atlanta, block groups are small and uniformly populated; whether or not this characteristic is a source of bias in this analysis is unclear. Finally, this analysis does not consider other sources of lead poisoning, such as imported food items or cooking utensils that are used in immigrant populations. The low percentage of children tested for lead exposure, especially in neighborhoods with old homes prompted the GACLPPP to apply a risk index approach across the state. In an effort to increase testing of high risk children, the GACLPPP is working with the state immunization program to include the lead risk index information into the registry. When a child presents to a provider’s office, the provider can check the immunization registry and know the child’s risk for lead exposure. Blood lead testing is the only simple way to know if a child has recently been exposed to lead. Unlike other diseases in which a child’s symptoms motivate parents to seek health care, elevated BLLs may cause no specific symptoms to suggest exposure to lead. When children are found to have elevated BLLs, state and local lead poisoning prevention programs typically conduct investigations to identify potential sources of lead in the child’s environment. Removing or controlling lead sources is the best way to manage elevated blood BLLs.41 However, children have already been harmed when an elevated BLL is found. Blood lead testing is secondary prevention and will remain an important prevention strategy. However, it is im-
Screening for Lead Poisoning: A Geospatial Approach to Determine Testing of Children in At-Risk Neighborhoods
413
portant to identify and remove sources before children are ever exposed, that is, primary prevention. One approach is to screen housing.34 The risk index can help not only identify children who should be tested, but also identify neighborhoods that could benefit from remediation. Understanding the areas of high risk, especially at smaller geographic resolution and the extent of testing among high-risk groups will allow clinicians and public health practitioners to make informed decisions about providing interventions. Improving testing of children at greatest risk and making homes safe from lead must be priorities to achieve the national goal of eliminating elevated BLLs among young children in the United States by 2010.42
REFERENCES 1. Agency for Toxic Substances and Disease Registry. US Department of Health and Human Services, Public Health Service. Draft toxicological profile for lead. Atlanta: Agency for Toxic Substances and Disease Registry; 2005 [cited 2006 Oct 20]. Available from: http://www.atsdr.cdc.gov/toxprofiles/tp13.html. 2. Lidsky TI, Schneider JS. Lead neurotoxicity in children: basic mechanisms and clinical correlates. Brain 2003; 126:5-19. 3. CDC. Surveillance for elevated blood lead levels among children—United States, 1997—2001. MMWR 2003 [cited 2006 Oct 20]; 52(SS10); 1-21. Available from: http://www.cdc.gov/mmwr/preview/mmwrhtml/ss5210a1.htm. 4. Schwartz J. Low-level lead exposure and children’s IQ: a meta-analysis and search for a threshold. Environ Res 1994; 65:42-55. 5. Lanphear BP, Dietrich K, Auinger P, Cox C. Cognitive deficits associated with blood lead concentrations ⬍10 g/dL in US children and adolescents. Public Health Rep 2000; 115:521-9. 6. Canfield RL, Henderson CR, Cory-Slechta DA, Cox C, Jusko TA, Lanphear BP. Intellectual impairment in children with blood lead concentrations below 10 g per deciliter. N Engl J Med 2003; 348:1517-26. 7. Bellinger DC, Needleman HL. Intellectual impairment and blood lead levels. N Engl J Med 2003; 349:500-502. 8. Lanphear BP, Hornung R, Khoury J, Yolton K, Baghurst P, Bellinger DC et al., Low-level environmental lead exposure and children’s intellectual function: an international pooled analysis. Environ Health Perspect 2005;113:894-899. 9. Tellez-Rojo MM, Bellinger DC, Arroyo-Quiroz C, Lamadrid-Figueroa H, Mercado-Garcia A, Schnaas-Arrieta L. et al., Longitudinal associations between blood lead concentrations lower than 10 microg/dL and neurobehavioral development in environmentally exposed children in Mexico City. Pediatrics 2006;118:e323-e330. 10. CDC. Preventing lead poisoning in young children: a statement from the Centers for Disease Control and Prevention. Atlanta: CDC; 2005 [cited 2006 Oct 20]. Available from: http://www.cdc.gov/nceh/lead/publications/PrevLeadPoisoning.pdf. 11. CDC. Increased lead absorption and lead poisoning in young children: a statement by the Center for Disease Control. Atlanta: US Department of Health, Education, and Welfare, CDC; 1975 [cited 2006 Oct 20]. Available from: http://www.cdc.gov/ nceh/lead/Publications/books/plpyc/plpyc_history/increaseLead_1975.pdf. 12. CDC. Preventing lead poisoning in young children: a statement by the Centers for Disease Control, DHHS publication no. (CDC) 99-2230. Atlanta: US Department of Health and Human Services, CDC; 1985 [cited 2006 Oct 20]. Available from: http://www.cdc.gov/nceh/lead/Publications/books/plpyc/plpyc_history/CLP_1985.pdf. 13. CDC. Preventing lead poisoning in young children: a statement by the Centers for Disease Control - October 1991. Atlanta: US Department of Health and Human Services, Public Health Service, CDC; 1991 [cited 2006 Oct 20]. Available from: http://www.cdc.gov/nceh/lead/Publications/books/plpyc/contents.htm. 14. CDC. Screening young children for lead poisoning: guidance for state and local public health officials. Atlanta: US Department of Health and Human Services, Public Health Service, CDC; 1997 [cited 2006 Oct 20]. Available from: http://www.cdc. gov/nceh/lead/guide/guide97.htm. 15. Pirkle JL, Kaufmann RB, Brody DJ, Hickman T, Gunter EW, Paschal DC. Exposure of the U.S. population to lead, 1991-1994. Environ Health Perspect 1998;106:745-50. 16. Kaufmann RB, Clouse TL, Olson DR, Matte TD. Elevated blood lead levels and blood lead screening among US children aged one to five years: 1988--1994. Pediatrics 2000; 106:E79. 17. CDC. Blood lead levels—United States, 1999 —2002. MMWR 2005 [cited 2006 Oct
414
Vaidyanathan et al
20]; 54(20); 513-516. Available from: http://www.cdc.gov/mmwr/preview/mmwrhtml/ mm5420a5.htm. 18. US Consumer Product Safety Commission. 1977. Ban of Lead-Containing Paint and Certain Consumer Products Bearing Lead-Containing Paint. 16 CFR 1303. Fed Reg 42:44199. 19. Roberts JR, Reigart JR. Screening for lead poisoning in South Carolina. J SC Med Assoc 2001;87:459-464. 20. Kemper AR, Cohn LM, Fant KE, Dombkowski KJ. Arch Pediatr Adolesc Med 2005;159:646-650. 21. Haley VB, Talbot TO. Geographic analysis of blood lead levels in New York State Children Born 1994-1997. Environ Health Perspect 2004;112:1577-1582. 22. Reissman DB, Staley F, Curtis GB, Kaufmann RB. Use of geographic information system technology to aid health department decision making about childhood lead poisoning prevention activities. Environ Health Perspect 2001;109:89-94. 23. Roberts JR, Hulsey TC, Curtis GB, Reigart JR. Using geographic information systems to assess risk for elevated blood lead levels in children. Pub Health Reports 2003;118:221-229. 24. Sargent JD, Bailey A, Simon P, Blake M, Dalton MA. Census tract analysis of lead exposure in Rhode Island children. Environ Res 1997;74:159-68. 25. Lanphear BP, Byrd RS, Auinger P, Schaffer SJ. Community characteristics associated with elevated blood lead levels in children. Pediatrics 1998;101:264-271. 26. Reyes NL, Wong L-Y, MacRoy PM, Curtis G, Meyer PA, Evens A, Brown MJ. Identifying housing that poisons: a critical step in eliminating childhood lead poisoning. J Public Health Management Practice 2006;12:564-569. 27. Miranda ML, Dolinoy DC, Overstreet MA. Mapping for prevention: GIS Models for directing childhood lead poisoning prevention programs. Environ Health Perspect 2002;110:947-953. 28. Oyana TJ, Margai FM. Geographic analysis of health risks of pediatric lead exposure: A golden opportunity to promote healthy neighborhoods. Arch Environ Occup Health 2007/62:93-104. 29. Vivier PM, Hogan JW, Simon P, Leddy T, Dansereau LM, Alario AJ. A statewide assessment of lead screening histories of pre-school children enrolled in a Medicaid managed care program. Pediatrics 2001;108 (2). 30. Georgia Department of Human Resources, Division of Public Health. Georgia Childhood Lead Poisoning Prevention Program (GCLPPP). Screening Guidelines: Children. Available from: http://health.state.ga.us/programs/lead/screeningtypes.asp 31. Wartenberg D. Screening for lead exposure using a geographic information system. Environ Res 1992;59:310-317. 32. Jonckheere AR. A distribution free K sample test against ordered alternatives. Biometrika 1954; 41:133-145. 33. Lanphear BP, Hornung R, Ho M, Howard CR, Eberle S, Knauf K. Environmental lead exposure during early childhood. J Pediatr 2002;140:40-7. 34. Lanphear BP, Hornung R, Ho M. Screening housing to prevent lead toxicity in children Public Health Reports 2005;120:305-310. 35. CDC. Preventing lead exposure in young children: a housing-based approach to primary prevention of lead poisoning. Atlanta: CDC; 2004 [cited 2006 Oct 20]. Available from: http://www.cdc.gov/nceh/lead/publications/Primary%20Prevention%20 Document.pdf. 36. Schwab LT, Roberts JR, Reigart JR. Inaccuracy in parental reporting of the age of their home for lead screening purposes. Arch Pediatrics and Adolesc Medicine 2003;157:584-586. 37. Reissman DB, Matte TD, Gurnitz KL, Kaufmann RB, Leighton J. Is home renovation or repair a risk factor for exposure to lead among children residing in New York City? J of Urban Health 2002;79:502-11. 38. CDC. Children with elevated blood lead levels attributed to home renovation and remodeling activities - New York, -1994. 1993MMWR 1997;45:1120-23. 39. US General Accounting Office. Medicaid: Elevated blood lead levels in children. Washington, DC: GAO; 1998. 40. CDC. Managing elevated blood lead levels among young children: recommendations from the Advisory Committee on Childhood Lead Poisoning Prevention. Atlanta: US Department of Health and Human Services, Public Health Service, CDC; 2002 [cited 2006 Oct 20]. Available from: http://www.cdc.gov/nceh/lead/ CaseManagement/caseManage_main.htm. 41. Pettit KLS, Kingsley GT, Coulton CJ, Cigna J. Neighborhoods and health: building evidence for local policy. Submitted to The Office of the Assistant Secretary for Planning and Evaluation, US Department of Health and Human Services. Submitted by The Urban Institute. May 2003. Available from: http://aspe.hhs.gov/ hsp/neighborhoods-health03 42. US Department of Health and Human Services. Healthy people 2010 (conference ed, in 2 vols). Washington, DC: US Department of Health and Human Services; 2000 [cited 2006 Oct 20]. Available from: http://www.healthypeople.gov.
The Journal of Pediatrics • March 2009