The national exposure registry: history and lessons learned.
The National Exposure Registry (NER) was created as a comprehensive group of data repositories that sought, over time, to relate specific environmental chemical exposures to registrants' health conditions. The NER's primary goal was to facilitate epidemiologic research by establishing multiple databases (subregistries) that contained demographic, environmental, and health information on large populations exposed to selected chemicals. The databases created by the NER were intended to provide information that would generate appropriate hypotheses for future activities, such as epidemiologic studies.
This article reviews the decisions and events that led to the formation of the NER. It cites the history, operations, accomplishments of the NER, and the les sons that can be learned from this enterprise. The lessons learned will, hopefully, serve as a guide for more meaningful and useful data collection and interpretation for all future exposure registries.
Formation of the National Exposure Registry
In the late 1970s and early 1980s, public concern was increasing about the potential impact of human exposures to hazardous substances from waste sites and chemical spills on the health of communities (Harris, 1983; Levine, 1983). The 1978 Love Canal, New York, incident of soil contamination with over 200 hazardous substances; the 1983 Times Beach, Missouri, dioxin soil contamination; and the 1984 Battle Creek, Michigan, trichloroethylene (TCE) well water contamination were seminal events in the American public's perception of hazardous waste sites as a major environmental hazard. In response to these and related concerns, Congress passed the Comprehensive Environmental Response, Compensation, and Liability Act of 1980 (CERCLA). This law provided for liability, compensation, cleanup, and emergency response for hazardous substances released into the environment and the cleanup of inactive hazardous waste disposal sites. This legislation also created the Agency for Toxic Substances and Disease Registry (ATSDR). Congress created ATSDR to specifically address the potential public health problems associated with environmental exposures to hazardous substances from waste sites and chemical spills. ATSDR was given two mandates relating to registries. They were first, to establish and maintain a national registry of persons exposed to hazardous substances, and second, to establish and maintain a national registry of persons with serious diseases or illnesses resulting from these exposures.
In April 1981, a subcommittee of the U.S. Department of Health and Human Services (DHHS) Committee to Coordinate Environmental and Related Programs (CCERF) released a report on the use of longterm registries for health follow-up of populations exposed to toxic hazards. This report outlined several conditions necessary to establish and operate such exposure registries, e.g., a scientifically acceptable and feasible rationale; sustained institutional commitments for funding, analysis, and reporting; ethical constraints in recruiting and following such populations; and the kinds of information that needed to be collected. Two years later, in 1983, DHHS published in the Federal Register a statement of organization, functions, and delegations of authority for ATSDR (The Public Health and Welfare Chapter 103--Comprehensive Environmental Response, Compensation, and Liability Subchapter I--Hazardous Substances Releases, Liability, and Compensation, 1983).
FIGURE 1 Goals and Objectives of the National Exposure Registry Primary Goals: To fill data gaps for long-term-health effects, defined as persistent adverse health outcomes, lasting more than one year, which may be manifested at any point in time. Those health outcomes will be focused on medically confirmed diagnoses potentially related to exposures to hazardous substances and their mixtures. Goal 1: To facilitate epidemiologic or health studies. Objective 1.1: To create data files that will be of value in research related to verifying what are thought to be known adverse health outcomes (hypothesis testing). Objective 1.2: To create data files that will be of value in research related to identifying previously unknown, undetermined adverse health outcomes should they exist (hypothesis generating). Goal 2: To facilitate state and federal health surveillance programs by enabling the identification and tracking of participants in the programs and by maintaining information on the participants. Objective 2.1: To create data files that can be used for follow-up surveillance activities if needed. Goal 3: To provide information that can be used to assess the burden of the effects of an exposure or health outcome on a population. Objective 3.1: To provide, to appropriate health officials, health information necessary for any contingency planning (such as information on appropriate screening plans, available diagnostic or therapeutic facilities, and educational materials for health care providers and the general public). Objective 3.2: To provide appropriate officials with environmental information necessary for any contingency plans to reduce exposures to hazardous substances. Supplemental Goals: To provide service and to be of direct benefit to the registrants, in keeping with the dictates of a public health agency. Goal 4: Working with state and local health officials, to provide exposed persons with current relevant information about the hazardous substances to which they have been exposed. Objective 4.1: To provide important toxicologic and epidemiologic findings that become available. Objective 4.2: To provide medical information that encourages registrants to seek medical attention, if needed. Goal 5: To enable local, state, or federal health officials to notify registrants rapidly of the adverse health effects of an exposure, preventative measures, or therapeutic advances that were not understood or known when the registry was established. Objective 5.1: To continue to provide state-of-the-art scientific information related to the exposure(s). Objective 5.2: To promptly provide new information on screening, diagnostic, or therapeutic activities related to the exposure(s). Blue text indicates process not achieved in practice.
In 1984, a document was produced titled, Criteria and Methods for Establishing and Maintaining Exposure and Outcome Registries for Environmental Public Health Problems (Center for Environmental Health, 1984) that proposed criteria for the establishment of nonoccupational environmental disease and exposure registries to be established under CERCLA. It also presented discussions of the strengths and limitations of exposure and disease registries. The top priorities for ATSDR registries, as stated in this document, were as follows:
(1) Persons exposed outside the workplace to
A. persistent, measurable levels of hazardous agents in which animal studies or other evidence predicts significant adverse effects in humans; or
B. hazardous agents for which current methods exist to prevent an adverse outcome; or
(2) Persons with outcomes of interest where measurements of exposure to hazardous agents are available.
Of the two original CERCLA mandates published in 1980, the second one relating to the establishment and maintenance of a national registry of persons with serious diseases or illnesses was deferred in order to first concentrate resources on the establishment of an exposure registry. This second mandate was, in fact, never acted upon. Therefore, this review is concerned with only the first mandate--the formation of a national registry of persons exposed to hazardous substances--the NER. An additional mandate was created by the Superfund Amendments and Reauthorization Act of 1986 (SARA), which stated that the agency must consider establishing a registry as a follow-up action to a health assessment when a potentially significant risk to human health is indicated. Two exposure registries, the World Trade Center Health Registry and the Tremolite Asbestos Registry, were recently established by ATSDR, but neither was created in response to the SARA mandate and neither is part of the NER. They will, therefore, not be discussed in this review.
Defining the NER
The stated goals and objectives of the NER are shown in Figure 1. The purpose of the NER was to assess and evaluate the potential relationships between environmental exposure and adverse health effects for an exposed population, particularly the relationship between chronic health effects and long-term, low-level chemical exposures. Of particular concern when considering environmental health issues is the lack of information about the effects of low-level exposures of long duration on general populations (versus worker populations), such as those typically surrounding hazardous waste sites. Registries can be a valuable tool for assembling the information necessary to address the potential health outcomes of these types of exposures.
The NER's primary goal was to facilitate epidemiologic research by establishing multiple databases (subregistries) that contain demographic, environmental, and health information on large populations exposed to selected chemicals. The processes for establishing, maintaining, and terminating a registry are shown in Figure 2.
It is important to differentiate between an exposure registry and an outcome registry. An exposure registry, such as the NER, is a registry in which people with particular exposure histories are followed forward from a period of exposure to determine whether they develop one or more adverse health outcomes. An outcome registry is a registry in which people with particular health outcomes, such as a specific disease, are studied backward in time from the date of their first such outcome, or onset, or diagnosis, to determine whether they may have been exposed to a hazardous agent of interest.
When the NER was launched, an exposure registry was a relatively new approach, in contrast to the more established disease or outcome registries, e.g., cancer registries. The purpose and function of an exposure registry has been defined as follows (Schulte & Kaye, 1988):
1. Delineation of the population at risk.
2. Follow-up of a cohort to ascertain exposure-disease associations.
3. Follow-up of a cohort to ensure the institution of appropriate primary and secondary prevention and medical surveillance.
4. Follow-up of a cohort to allow for appropriate social, legal, and economic support.
5. Demonstrate societal concern for the cohort and provide a base for political action relevant to the exposure.
6. Notify a cohort of an exposure, preventive measures, or therapeutic advances, which were not understood or known at the time that the registry was established.
[FIGURE 2 OMITTED]
Establishing an NER Subregistry
Subregistries were established for dioxin in 1988, TCE in 1989, benzene in 1991, and trichloroethane (TCA) in 1992. The establish ment of an NER subregistry required several well-defined steps. They were (1) primary contaminants selection, (2) site selection, (3) population selection, (4) data collection, and (5) file creation and maintenance. Detailed descriptions of these steps are described elsewhere (Agency for Toxic Substances and Disease Registry [ATSDR], 1988, 1994; Burg, 1989; Burg & Gist, 1995, 1997; Gist & Burg, 1995; Gist, Burg, & Radtke, 1994).
ATSDR compared the NER data for people exposed to the hazardous substance to data for people in the U.S. who were not exposed to the hazardous substance. The comparison group for the NER subregistries was drawn from the National Health Interview Survey (NHIS). Five health conditions were reported in excess in each subregistry in at least one data collection period for some group of registrants. They were (1) anemia and other blood disorders, diagnosed in the past year; (2) all cancers, diagnosed in the past year; (3) skin rashes, eczema, or other skin allergies, diagnosed in the past year; (4) urinary tract disorders, including prostate trouble, diagnosed in the past year; and (5) stroke, ever diagnosed. Table 1 shows the health effects reported in excess when compared with national data (baseline or follow-up).
It is important to note that despite the many health effects observed to be in excess when compared with national norms, NER staff believe that the excess reporting of these health conditions might be explained by several factors including methodological differences in data analysis and reliance on self-reported, nonverified data. They further believe that these results do not identify relationships between the registrants' reported adverse health effects and their exposure to the chemicals studied in the four subregistries: TCE, benzene, dioxin, and TCA.
Many questions arise when designing an exposure registry. What constitutes adequate exposure measurement? How can exposed persons be adequately differentiated from nonexposed persons? What group could serve as a comparison group so that the disease experience of the exposed group can be evaluated? For how long should the group be followed? These questions can become quite technical, but often, even the most basic questions are the hardest to resolve. At least four important issues were never satisfactorily resolved in the design of the NER. They were adequacy of exposure measures, unverified self-reporting, appropriate control groups, and the use of biomarkers.
Inadequacy of Exposure Measures
The most important design deficiency of the NER was its inability to adequately assess exposure. Yet accurate exposure assessment is central to the utility of any exposure registry and is critical in establishing an association for health outcomes to a particular chemical of interest.
It is generally accepted that the four primary characteristics for the determination of exposure are (1) the nature of the agent, (2) the intensity of exposure, (3) the duration of exposure, and (4) the frequency of exposure. In the NER, the inability to measure the intensity, duration, and frequency of exposure to the chemicals of interest has been the most critical limiting factor in its design. In the TCE subregistry, for example, the exposure assessment was limited to the measurements obtained at the time of the U.S. Environmental Protection Agency (U.S. EPA) investigations and was usually based on a single measurement from a well. The probable day-to-day and seasonal fluctuation of TCE levels in well water also makes a single measurement questionable. A more adequate exposure assessment would have documented the level of TCE in a registrant's well; the routes of exposure, e.g., ingestion or showering; and the frequency and duration of exposure. These measurements would be the most important variables in quantifying an individual's exposure to TCE.
An accurate exposure history is also important in the identification of any other major sources of exposure to the chemical in question, notably occupational exposure. Asking for a full occupational profile would have been difficult, but it might have been helpful to have added a question to ascertain whether the individuals worked in the industry responsible for the contamination, in addition to living in the neighborhood.
Another concern in exposure assessment is the issue of multiple contaminants. Many sites have other chemicals present in addition to the chemical of concern. To address the possible roles of other contaminants, registrants might have been grouped as (a) persons exposed to specific chemicals, (b) persons exposed to chemically related compounds, or (c) persons exposed to chemicals known to produce similar effects by in vitro or animal data. Such a grouping might have improved the prediction of specific health outcomes of interest.
In order to deal with the problem of inadequate or erroneous exposure measurements, future exposure registries should use as extensive a battery of exposure measurements as possible. Multiple environmental measurements would improve the accuracy of assessment and direct human exposure measurements should be used whenever possible. The initial NER questionnaires should have had questions to assess the likely extent of exposure by route, duration, and any other relevant factors. In addition, the questionnaires should have included questions that addressed other possible risk factors for the health outcomes that may be associated with the chemical of interest, such as medication use, lifestyle choices, heredity, and exposure to other chemicals.
Appropriate Control Groups
It should be noted that a subregistry of unexposed persons was not routinely created for a given site concurrently with the subregistry of exposed persons. The basis for this decision was that at most sites an appropriate control group could not be identified at the time the exposure subregistry was created. The hypothesis to be tested either may not have been known or defined well enough to enable the identification of a valid control population at that time. An example of an inappropriate control group might be a control subregistry of older persons when the adverse outcome is later identified as a birth defect. In some instances where subregistries might be combined for research purposes, an existing source of national statistics may have sufficed. At some sites, an exception to not creating a control subregistry could have been made (e.g., where extensive information is known about the exposure and the expected outcome of this exposure). For most sites, however, the identification of an appropriate comparison group was considered part of the future research activity.
The comparison group for the NER subregistries was drawn from the National Health Interview Survey (NHIS). Both NER and NHIS are self-reported and nonmedically verified data collections. The validity of comparisons between the subregistry populations and the NHIS was dependent on comparability between the two populations and between the data collection instruments. Comparability between the two instruments may have been compromised since some of the questions had been phrased differently in the two surveys. In addition, comparisons might have been misleading if the subregistry population was more likely to be aware of and to report health conditions because of their exposure or because of being in the registry. "Dose response" was another issue. It was not known whether health outcomes occurred more frequently in groups with the highest exposure because the available data did not provide gradients of exposure based on duration, routes, frequency, and intensity. An issue of potential variation in exposure from site to site also arose, creating potential problems of comparability within exposed populations. All of these issues brought into question the adequacy of the control populations for the NER subregistries.
The validity of self-reported data was an important issue in the design of the NER. Verification of reported outcomes with medical records could improve the quality of health information; however, verification was not part of the NER design. Verification adds expense and is relatively difficult to interpret unless an appropriate control group is available. Inherent shortcomings of verification are also present--not contacting the correct medical provider, faulty memory of where and when a diagnosis was made, problems in getting old records, and accessibility to records of a retired physician.
Issue of Biomarkers
Biological markers, or biomarkers, are cellular, biochemical, or molecular measures that are obtained from biological media such as human tissues, cell, or fluids and are indicative of exposure to environmental chemicals. In addition to indicating exposure, biomarkers can also signal events in biologic systems indicative of susceptibility or effect. Biological markers of exposure have been used to study environmental contaminants such as lead levels in blood, teeth, and hair from the inhalation and ingestion of lead. Unfortunately, markers of exposure currently exist for only a small number of hazardous substances.
Lessons can be learned from both the strengths and the weaknesses of the NER experience. Some parts of the NER were well conceived, whereas other parts were not. Substantial thought and documentation went into creating the parameters for primary contaminant selection, site selection, population selection, and data collection. Less thought and documentation went into defining methods to measure adequately exposure to the chemicals of interest. This was the key missing element and the Achilles heel of the NER program. Questions about the adequacy of self-reporting and the absence of biomarkers cast doubts about the program's design. Another weakness of the NER program was that it did not adequately focus on minority and ethnic populations that are disproportionately impacted by exposure to the chemicals in question. Also, there were inadequate efforts directed to the scientific and public health communities to get their participation in follow-up studies based on NER's findings.
An abundance of excess health outcomes existed for each of the chemical entities studied. Five health conditions were reported in excess for the dioxin, TCE, benzene, and TCA subregistries during at least one data collection period among some group of registrants. Additional health conditions were reported in excess for some, but not all, of the subregistries during at least one data collection period among some group of registrants (Table 1). Many of the excess health outcomes are backed up by biologically plausible hypotheses gained from occupational medicine, clinical studies, and laboratory studies. Yet, despite the many health effects observed to be in excess when compared with national norms, NER staff believed that the excess reporting of these health conditions might be explained by methodological differences in data analysis and reliance on self-reported nonverified data. The NER staff further believed that these results do not identify relationships between the registrants' reported adverse health effects and their exposure to the chemicals studied in the four subregistries. Other risk factors for disease such as heredity, lifestyle factors (such as smoking), medication use, and exposure to other chemicals were not taken into account in the subregistries and may be responsible for observed disease.
The positive lessons learned from the NER experience have been the progenitor for the design of two recent exposure registries, the World Trade Center Health Registry and the Tremolite Asbestos Registry. Both of these exposure registries combine well-documented exposure information with personal health information. They are considerably simpler than the NER. In the case of the World Trade Center Health Registry, exposure was limited to one time and place and the public had a much greater awareness of that event than of the NER. In the case of the Tremolite Asbestos Registry, only one substance is of concern.
Creating and maintaining an exposure registry is a more difficult challenge than creating and maintaining an outcome registry, such as a cancer registry. The former looks prospectively, whereas the latter looks retrospectively. Hindsight is generally more accurate than foresight. In addition, expectations may have been set too high for the NER. These notions were expressed two decades ago by Dr. Vernon Houk, thendirector, Center for Environmental Health, Centers for Disease Control, when he addressed a workshop that was held in Atlanta about the proposal to establish the NER (ATSDR, 1987):
Registries cannot be used to give quick, conclusive answers to the many questions that people ask about the effects of toxic chemical exposures to their health, the health of their families, or the health of their future children. Even the most active and aggressively pursued studies using registries may require decades to produce meaningful results. It is also true that even when the results of long-term studies are made available, the findings are likely to be hotly debated by scientists with honestly different points of view. Other important limitations of registries are the infeasibility due to mobility of the population, privacy, ethics, and the inability to detect conditions with a low attack rate after exposure. Even cost must be considered. I point out these inherent limitations to registries to call your attention to the fact that expectations for registries being used as tools to answer in the near future the questions before us today may be unrealistically high.
In addition to the inherent limitations of exposure registries cited by Dr. Houk, the flawed design and methodologies of the NER resulted in data that could not be used to reliably answer questions about the relationships between registrants' exposures and adverse health outcomes that would lead to hypothesis-generated health studies.
The NER was one of the most comprehensive data repositories tracking specific en vironmental chemical exposures and registrants' health conditions over time. The methods developed by NER for primary contaminant selection, site selection, population selection, and data collection are worthy of emulation. The methods used to compare registrant health data trends to national trends have evolved and improved over time. The lack of adequate individual exposure measures has been the most limiting factor in the NER design. Other design issues of the NER that has limited the utility of the results are the competency of selfreport health outcomes, the lack of an adequate control population, and the absence of biomarkers. The NER experience will, hopefully, contribute to more meaningful and useful data collection for all future exposure registries.
Agency for Toxic Substances and Disease Registry. (1987, March). Public workshop on a national registry of persons exposed to toxic substances. Atlanta: U.S. Department of Health and Human Services, Public Health Service.
Agency for Toxic Substances and Disease Registry. (1988). Policies and procedures for establishing a national registry of persons exposed to hazardous substances. Atlanta: U.S. Department of Health and Human Services, Public Health Service.
Agency for Toxic Substances and Disease Registry. (1994). National Exposure Registry trichloroethylene (TCE) subregistry baseline technical report (Revised) (NTIS Publication No. PB95-154589). Atlanta: Author & U.S. Department of Health and Human Services, Public Health Service.
Burg, J.R. (1989). Policies and procedures for establishing the National Exposure Registry. Journal of the American College of Toxicology, 8(5), 949-954.
Burg, J.R., & Gist, G.L. (1995). The National Exposure Registry: Procedures for establishing a registry of persons environmentally exposed to hazardous substances. Toxicology and Industrial Health, 11(2), 231-248.
Burg, J.R., & Gist, G.L. (1997). The potential impact on women from environmental exposures. Journal of Women's Health, 6(2), 159-161.
Center for Environmental Health. (1984). Criteria and methods for establishing and maintaining exposure and outcome registries for environmental public health problems. Atlanta: Agency for Toxic Substances and Disease Registry.
Comprehensive Environmental Response, Compensation, and Liability Act, 42 U.S.C. [section] 9601-9657 (1980).
Gist, G.L., & Burg, J.R. (1995). Methodology for selecting substances for the National Exposure Registry. Journal of Exposure Analysis and Environmental Epidemiology, 5(2), 197-208.
Gist, G.L., Burg, J.R., & Radtke, T.M. (1994). The site selection process for the National Exposure Registry. Journal of Environmental Health, 56(6), 7-12.
Harris, J.S. (1983). Toxic waste uproar: A community history. Journal of Public Health Policy 4(2), 181-201.
Levine, A. (1983). Psychosocial impact of toxic waste dumps. Environmental Health Perspectives, 48, 15-17.
The Public Health and Welfare Chapter 103--Comprehensive Environmental Response, Compensation, and Liability Subchapter I--Hazardous Substances Releases, Liability, and Compensation, 42 U.S.C. [section] 9604 (1983).
Schulte, P.A., & Kaye, W.E. (1988). Exposure registries. Archives of Environmental Health, 43(2), 155-161.
Myron G. Schultz, MD, FACP
James H. Sapp II, MS
Caroline D. Cusack, MSPH
Jennifer M. Fink, MPH, cHES
Corresponding Author: Myron G. Schultz, Senior Medical Officer, Division of Health Studies, NCEH/ATSDR, Centers for Disease Control and Prevention, 4770 Buford Highway NE, MS F-57, Atlanta, GA 30341. E-mail: email@example.com.
TABLE 1 Health Effects Reported in Excess Compared With National Data TCE Registry Anemia Cancer Skin rashes Urinary tract disorders Stroke Diabetes Liver disease Kidney disease Hearing impairment Speech impairment Benzene Registry Anemia Cancer Skin rashes Urinary tract disorders Stroke Diabetes Liver disease Kidney disease Respiratory allergies Arthritis GI ulcers Dioxin Registry Anemia Cancer Skin rashes Urinary tract disorders Stroke TCA Registry Anemia Cancer Skin rashes Urinary tract disorders Stroke Diabetes Liver disease Respiratory allergies
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||ADVANCEMENT OF THE PRACTICE|
|Author:||Schultz, Myron G.; Sapp, James H., II; Cusack, Caroline D.; Fink, Jennifer M.|
|Publication:||Journal of Environmental Health|
|Date:||Feb 27, 2010|
|Previous Article:||General public health considerations for responding to animal hoarding cases.|
|Next Article:||The intern, or: a tale of the economy.|