figure 1 - Lead Risk Inquiry
Childhood exposure to lead remains a critical health issue in the U.S. population. The widespread distribution of lead throughout the environment and the profound physiological and cognitive affects it has on children, even at low levels, warrants an aggressive approach toward identifying lead hazards in the environment, determining the population at risk for lead exposure and developing strategies to prevent exposure. In 1991, the Centers for Disease Control and Prevention (CDC) issued recommendations that supported near universal blood-lead screening for children under the age of 6 years. This recommendation is under revision due to criticisms that universal screening is ineffective and unnecessary. Recently, the CDC issued an updated lead screening guidance document recommending the evaluation of individual and residual exposure risks to target screening efforts in areas where lead risks are significant.
Geographic Information System (GIS) technology can be used to create a spatial model of known risk factors from which a numeric risk value can be calculated representing the relative degree of risk of exposure to lead. Data layers representing risk factors are ranked and graded based on the estimated degree of contribution to overall risk. Health professionals can obtain an individual's exposure risk directly from an Internet Web site simply by entering a residential address. The model returns an overall risk value for the specific location by incorporating these lead risk elements. The use of a standard Web browser allows for the cost-effective delivery of accurate and current information. The use of an Internet server also allows for the data and application to be updated as needed at a central lcoation without the need to redistribute the data.
Childhood exposure to lead remains a critical health issue confronting the U.S. population. Although the exposure to lead, as measured by blood-lead levels, has fallen dramatically since the 1970s,1-2 significant numbers of children continue to be exposed to toxic levels of lead. Over-exposure to lead, however, is not equally distributed in all segments of the population. A 1988 report by staff of the Agency for Toxic Substances and Disease Registry (ATSDR) found that childhood blood lead levels were significantly associated with race, family income, residence inside or outside of a metropolitan central city, and the size of the metropolitan area.3 These findings are supported by the results of the recent National Health and Nutrition Examination Surveys (NHANES) which found elevated blood lead levels disproportionately distributed among children who are "poor, non-Hispanic black or Mexican American, living in large metropolitan areas, or living in older housing."
Within the past two decades, research on childhood lead poisoning has identified profound, adverse physiological and cognitive affects from low level exposure.4-10 The neurotoxicity of lead is of considerable concern. Several prospective studies have identified significant dose-dependent relationships between lead exposure and impaired neurobehavioral and psychological functioning.11-13 Specifically, the Cincinnati study and the Port Pirie study reported an inverse association between postnatal blood lead levels and IQ. Follow-up studies of adults with asymptomatic exposure to lead as children has demonstrated that children with early elevated lead exposure were at risk for later educational deficiencies including failure to graduate and poorer reading abilities.14-15 The findings of these and other studies suggest that the deleterious effects of early childhood exposure to low levels of lead results in profound and long-lasting impacts on learning and behavior.
In response to these data, in 1991, the Centers for Disease Control and Prevention (CDC) issued revised guidelines for management of lead poisoning which significantly lowered the levels at which children are considered at risk for lead toxicity.16 This recommendation greatly increased the number of children considered to be lead poisoned. Additionally, research regarding childhood lead exposure and poisoning determined that the problem of lead poisoning extends beyond the population of inner-city children, traditionally considered to be the primary population at risk. These new guidelines led to the recommendation that virtually all children under the age of 6 be screened. The CDC also recommended that physicians administer a five-part questionnaire in order to classify children as being at high risk or low risk for lead exposure.
The recommendation of near universal screening for lead toxicity of children between the ages of 12 months and 72 months has proven to be very controversial. Critics of universal screening have challenged the necessity of the recommendation citing the dramatic reduction of blood lead levels in the nation's preschool children since 1976. Two additional reasons frequently cited are the low prevalence of elevated blood-lead levels in much of the population as well the high costs of universal screening.17
As a result, this recommendation has not been followed by many primary health care providers, resulting in relatively few children in any region of the country having been screened for lead exposure. A nationwide survey conducted by the CDC in 1994, reported that only about one fourth of parents reported that their young children had been screened. The screening rate for children considered to be high risk for lead exposure was only marginally better. Approximately one third of the children living in housing built prior to 1960 and one third of children in families making less than $20,000 were reportedly screened.18
To address the concerns raised regarding the 1991 recommendation of universal screening, and to further improve the use of screening to identify and prevent childhood lead poisoning, the CDC issued an updated lead screening guidance document in November 1997, entitled: Screening Young Children for Lead Poisoning: Guidance for State and Local Public Health Officials. This document describes methods of achieving more effective screening efforts in areas where lead exposure risks are significant. The guidance document refers to the use of census tract data regarding age of housing, the administration of a personal-risk questionnaire, the determination of the socioeconomic status of the individual and the use other data, to identify children who are at greatest risk for lead exposure to help ensure that they receive the proper services.
Dakota County staff have developed a Geographic Information System (GIS) application that incorporates data from several sources, to assist in the identification of populations at high risk for exposure to lead hazards. The application combines information representing risk factors for lead exposure including the age and distribution of housing, location of lead emitting sources, distribution of lead contaminated soils, case mapping of lead poisoning and demographic data. An individuals overall lead exposure risk can be modeled and classified by incorporating the various contributors to lead over-exposure as it relates to their home address. The application is accessed using a standard Web browser, allowing for the cost-effective delivery of accurate and current information. Dakota County is in the process of applying for funding from the Minnesota Department of Health through a grant from the State and Community-Based Childhood Lead Poisoning Prevention Program administered by the CDC. If received, the funding will be used to assess the usefulness of the model and to evaluate the risk factors employed in the model.
Setting
Dakota County is located on the south side of the Minneapolis/St. Paul, seven county metropolitan area of Minnesota. Originally settled in the mid-1850s, the County's 1997 population is estimated at 330,000. It is the second fastest growing county in the 87 county state, and its population is younger than the Statewide and metropolitan average. Once largely rural, the second half of this century has seen the intensive suburbanization of the northern half of the County. Initially bedroom communities furnishing labor to Minneapolis, St. Paul and inner suburban businesses, the County's dozen major cities have grown substantially in the past two decades attracting a significant number of commercial and industrial businesses. Employment growth in the County has increased at a rate that exceeds the population growth. The largest employment sector is retail trade, followed by services and then manufacturing. The County is home to 40 regulated solid waste facilities, including two large sanitary landfills. The County also licenses and inspects nearly 1300 hazardous waste generators and facilities, including a large secondary lead smelter. The housing stock is considerably newer than that of the Twin Cities, with approximately 10% of the homes constructed prior to 1950. However, concentrations of older housing are found in cities developed when streetcars were the primary mode of transportation as well as in communities that historically served as agricultural centers.
Overview of the Lead Exposure Risk Assessment Application
The GIS application, developed in Dakota County, is intended as a tool to help determine when blood lead testing is necessary. The model does not however, account for all the potential risks that contribute to lead poisoning and the professional analysis and opinion of a physician or health care provider are still required to determine whether or not to perform blood lead testing.
The GIS application was developed in Visual Basic and, as noted above, is accessed using a standard Web browser through an Internet Web site. The front-end used by this application is Netscape. The Web page contains graphic images, text and input fields. A map of the County is depicted as a graphic image that can be clicked with the mouse to perform basic map viewing functions. An address can be entered and a search initiated to determine the cumulative risk of exposure to lead for the resulting location (figure 1). A detailed report of all risk factor contributors is available by clicking the 'Details' button (figure 2).
Client / Server Application
Lead-based paint
Lead-based paints are considered to be the most significant and widespread sources of childhood lead exposure. The lead content of residential paints can be generally related to their period of production and it is estimated that approximately three-fourths of the housing built before 1980 contains lead-based paint. The lead content of paint was largely unregulated prior to enactment of voluntary restrictions by the paint industry in the mid-1950s. Prior to the 1940s, paints contained up to 50% lead by dry weight. In response to increased regulatory pressures, the lead content of paints was gradually reduced. By 1978, lead had been eliminated from paints produced for residential application. For the purposes of the model, it is assumed that a correlation exists between the year of a building's construction and the presence and concentration of lead in the paints.
The influence of lead-based paint is most often related to an individual's place of residence. Other places such as relatives' homes, daycare centers and schools can also serve as a source of exposure for children.
For the purposes of tax collection, the County assessor's offices collects parcel information that includes the property's address, value, ownership and building data. Information regarding a home's year of construction can be determined from the child's home address.
Lead Contaminated Soils
Lead contaminated soils are also important sources of lead exposure, primarily in older, urban residential areas. Its overall influence on lead exposure however, is considered to be less than the influence of lead-based paint. To reflect this lesser influence, a weighting factor is applied to this risk feature. The soil component of the risk model is actually derived from several sources, each which may have an individual relative score based on the assumed degree of influence on the overall potential risk.
Soil is contaminated by lead from various sources including weathered lead-based paint and the historical deposition of lead fallout from combustion sources such as incinerators or automobiles fueled with leaded gasoline. Since lead does not decay in the environment, deposits of lead from these sources accumulate over time in the upper 5 cm of undisturbed soils. Urban residential soils are generally more lead contaminated than rural soils (EPA 1989). This is due to the greater concentration and longer operational history of emission sources as well as the higher traffic density and larger stock of older housing most often associated with urban environments.
The distribution of lead contaminated soils can be approximated for urban residential areas based on the extent of urban development at a time where the inputs from the various sources would be anticipated to be the greatest. For the purposes of the model, urban areas developed prior to 1970 are assumed to be associated with greater soil lead contamination than urban areas developed after that time. The cutoff date of 1970 reflects the enactment of regulations that resulted in reduced inputs from lead-based paints and atmospheric fallout from automobile exhaust and industrial sources.
Industrial Point Sources
Soils contaminated by air-borne lead emissions from industrial sources can be characterized as a release point with a subsequent downwind zone of influence representing an atmospheric fallout area. A release point would be used to represent known industrial emission sources such as secondary lead smelters, sewage sludge incinerators and coal-fired power plants.
The extent of the plume and the degree of overall risk assumed to be associated with a source would be based on the type of facility, the nature and quantities of its emissions, the prevailing winds and its history of operation. The areal extent and the level of risk associated with these facilities are approximated due to the lack of detailed environmental studies of air and soil contamination adjacent to these industrial emission sources.
Industrial and Solid Waste Disposal Sites
Industrial and other solid waste disposal sites can serve as significant sources of childhood lead exposure. Among Superfund sites, lead is the most frequently identified hazardous substance found in completed exposure pathways. The notable presence of lead at disposal sites is due to its persistence in the environment and its wide usage and dissemination in industrialized countries.
Dakota County has created a countywide inventory of known solid and hazardous waste dump sites. The inventory contains approximately 1600 sites ranging from farm dumps to extensive industrial/hazardous waste dumps, including those associated with the operation of a large secondary lead smelter. This component of the model is represented by the known or suspected boundary of the disposal site. The types and quantities of waste disposed and the site characteristics that influence exposure are used to classify dumps. Dump sites creating a high potential for lead exposure, such as battery casing dumps, are assigned an associated level of risk.
To varying degrees, the areas surrounding a site may also be impacted through a gradual dispersal of contamination. Factors influencing the extent of this zone include the current land-use of the site and adjacent properties, and the presence of physical barriers serving to reduce human contact with contaminated wastes or soils. Sites possessing an increased potential for dispersal are assigned a buffer to account for these exposure characteristics.
Case Mapping
Reported cases of elevated blood-lead are mapped and included in the model as points and buffered polygons around those points. This assumes a zone of influence around the identified cases due to factors already in the risk model or may be suggestive of factors not yet identified.
Proximity to known cases may be represented as concentric polygons around each residence location with varying degrees of offset representing the degrees of related estimated risk. In situations where a given location falls within multiple zones of influence of known cases, a higher weighting factor can be applied to the model for overall risk. This may aid in the discovery of unknown sources of contamination or otherwise aid in determining a relationship between the cases.
Influence of Non-Source Factors
Other factors that influence overall risk, but that are not directly associated with sources of lead, can be included in the model. Demographic factors, such as distribution of poverty, can be used in combination with other risk factors to target areas for screening, education or environmental assessment.
Determining lead risk using the GIS model
The model was developed to allow easy access to detailed information regarding individuals' risk of lead exposure. The model runs on a Web server and is accessed through a Web page. A query is initiated by entering a standard street address in the box next to the appropriate prompt. A certain amount of flexibility is allowed for matching misspelled street names. Since the matching process, or geocoding, uses a street centerline containing segments with assigned address ranges; the address does not need to match an actual property address. This technique establishes a geographic location for any address number that falls within the address range of a given street segment.
The Web page also includes several yes/no questions derived from a sample questionnaire developed by the CDC that are answered by checking the corresponding box for those statements that apply. These questions are included to assist health professionals account for other potential sources of lead exposure, or high-risk activities, not included in the model.
The button labeled 'Search' is pressed to initiate a search on the server for the address location. If a location is found, it is used to search all risk data sets using pre-established techniques for spatially related features. A total risk factor is calculated by adding the risk factors of individual risk features. For example, the parcel associated with the address entered is found and its year of construction retrieved. The following risk values are associated with the residence's year of construction:
| Year of Construction | Degree of Risk | Risk Value |
| Less than or equal to 1940 | High | 7 |
| 1941 - 1956 | Medium | 5 |
| 1957 - 1978 | Low | 3 |
The model also retrieves the average building construction date for all parcels within a 200' radius of the target location and returns the count of the structures constructed prior to 1941. This value is not directly incorporated into the risk calculation but may be useful in identifying exposure risks, from improper paint removal for example, in cases where the target residence is a newer home that is surrounded by older homes.
Other polygon features such as dumps, lead dumps, residual soil lead, and poverty within 200' of the target location are identified and incorporated into the overall risk value. These features are assigned a corresponding risk factor, which is applied directly to the total risk as follows;
| Degree of Risk | Risk Value |
| Extremely High | 5 |
| Very High | 4 |
| High | 3 |
| Medium | 2 |
| Low | 1 |
Risk values for each feature are assigned on an individual basis. These values can be easily changed to incorporate new information regarding a site or to reflect modifications to a site that affect exposure risks.
When the application completes the analysis, a dot is displayed on the map at the target location and the total risk value is displayed. The map extents are changed to be centered on the location at the same scale. The 'Zoom To Location' button is provided to change the map extents to include a 200' radius around the selected location.
Basic Map Functions
A list of options is provided that specify the operations performed when the mouse is clicked in the map graphic. These include basic functions for zooming and panning. An option is also provided for identifying locations interactively. This allows a total risk value to be determined for specific locations, such as street intersections, or for a general area if an exact address is unknown.
An index map is provided to help determine where the current view lies in relation to the entire County. Municipal boundaries are displayed in a small index map that is highlighted with the outline of the approximate extent of the current map view.
All map, search, and analysis operations are performed on a centralized back-end Web server using a custom application, which responds directly to the user's requests. This client/server application model has distinct advantages over typical methods of installing software and data on individual computers.
Data
Lead risk features are modeled as points and polygons. Points are used to represent the locations of known cases of elevated blood lead levels. These points are not shown graphically due to concerns of protecting the privacy of individuals. The actual case locations are only used to provide a summary of reported test results within a defined distance from the target location. Polygons are used to spatially represent exposure risk features such as dumps, lead dumps, residual lead, and poverty. Parcels are also represented as polygons from which the year built is retrieved to determine degree of risk.
Polygons are also used to represent regions of influence around a feature either by buffering the feature itself or by manually representing a risk area such as a neighborhood or park service area. Each feature is modeled independently. This allows each to contribute individually to overall risk through overlapping features. Data sets are represented with a unique hatching pattern in which color gradation is used to represent degrees of risk. This method allows overlapping polygons to be visually distinguished and ranked.
Actual results of blood lead testing need to be applied to the application on a regular basis to modify and improve the model. This information will provide a mechanism for verifying the validity of the model as well as providing insight toward identifying unrecognized contributors to increased exposure risk. As patterns of elevated blood lead levels become apparent, corresponding physical features may be identified. The model can be improved by incorporating the newly identified features, or by assigning an increased risk weighting to an existing feature, thus ensuring that others in the vicinity are also screened.
Field verification of risk features is very important since many features may not have been evaluated directly for actual lead exposure characteristics. These unverified exposure-risk features can still be incorporated in the model; however, efforts must be made to better define their risk attributes or characteristics, as resources permit.
Software
The software used in this application consists of a Web server, custom Visual Basic executable and a client browser. Web servers are available from several vendors and are used to host a Web site and execute application software. The Web server in this case is Microsoft Internet Information Server.
The application executable was developed using Microsoft Visual Basic and runs on Microsoft NT Workstation. Map Objects is an Active X control supplied by Esri. This has been incorporated with Visual Basic in the application to provide the basic mapping and querying capabilities. Internet Map Server is another Active X control from Esri, which has been used to provide communications with the Web server and for image file format conversion.
Any Web browser supporting HTML 2.0 or greater should be compatible with this application however, each browser can have its own unique characteristics. This may result in slight variations in the physical look and the interactive performance of the application. The use of standard HTML maximizes the number of compatible variations of computers, operating systems, and browsers that may be available to users for accessing this application. Minimal configurations including low-end PC's and Windows 3.1 are assumed to be reasonably compatible.
Summary
A Web-based lead exposure risk analysis application provides many benefits by modeling real world contributors to lead exposure and makes this information readily accessible. Physicians and public health professionals will no longer have to rely solely on an individual's self-reported assessments of environmental lead exposure risks, and can now use objective information to target screening efforts on those individuals who are at greatest risk of exposure.
As blood-lead results become available, and risk feature attributes are better defined, it is our belief that the widespread use of the model will serve as an effective tool in the early identification of lead-poisoned individuals, while realizing improvements in cost effectiveness of lead screening efforts.
Supplying this application on the Internet provides health professionals with near universal, on-demand access to data that would otherwise not be available. This data can be maintained regularly and frequently without burdening the users with file maintenance. The data can also be analyzed and presented to end-users in a consistent manner irrespective of computer resources. Furthermore, access can be given to data derived from sensitive patient information without compromising privacy, and research, supporting documentation or related Web sites can also be assembled and easily provided.
Screening and early detection of lead exposure are effective means of preventing cases of severe lead poisoning. Presently, however, many children exposed to toxic levels of lead are not being identified. The use of GIS and Internet technology can assist in the identification of children at greatest risk of lead exposure and ensure that exposed children receive the necessary services.
References
Author Details:
Stephen A. Scott
Environmental Specialist
Environmental Management Dept., Dakota County
14955 Galaxie Ave.
Apple Valley, MN 55124
(612) 891-7537
fax: (612) 891-7588
Steve.Scott@co.dakota.mn.us
Randal D. Knippel
GIS Specialist
Survey and Land Information, Dakota County
14955 Galaxie Ave.
Apple Valley, MN 55124
(612) 891-7080
fax: (612) 891-7127
Randy.Knippel@co.dakota.mn.us