Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
CHAPTER 6 CONCLUSIONS AND RECOMMENDATIONS PURPOSES AND PROCEDURES OF THE ONGOING STUDY Amidst growing concern with environmental hazards and the resulting health effects and costs, Congress passed Public Law 95-623. Section 7 of the law calls for an ongoing study to provide information on sources of environmental hazards, on the health effects attributable to these hazards, and on the costs associated with these health effects. In particular, Congress requested a study that would provide estimates of the reduction in health costs which would result from each incremental reduction in environmental hazards resulting from human activity. The law also calls for reports every two years on these issues, and for advice on improving the data available to obtain the required estimates. Because of the law's broad scope, the Department of Health and Human Services (DHHS) asked the Institute of Medicine to engage in a planning study to consider how the ongoing study should be conducted and to assess the feasibility of meeting the requirements in Section 7. This committee recognizes the desirability of developing the kind of information requested in P.L. 95-623, but finds that the scope of the congressional charge far exceeds what is feasible in our present state of knowledge. Major deficiencies in data bases, methodologies, administrative procedures, and, most important, theoretical knowledge for relating environmental hazards to health problems and their associated health costs have precluded earlier straightforward assembly of the information needed. Therefore, the committee proposes a series of phased activities for the ongoing study, including case studies, methodological and operational research, and evaluative efforts, to develop the required data bases and strategies for analyzing them. Only by starting now to collect needed data will the base become available for deriving the estimates requested by Congress. As the initial phases of the ongoing study proceed, it will become clearer to what extent the estimates Congress has requested can be provided. The ongoing study can serve many functions. It can enhance data collection and coordination. It can draw attention to the need to -133-
relate environmental data to health data to develop hypotheses or evidence about environmental hazards. It can define areas needing basic research and encourage development of new methodologies and models, including predicting human risks from animal data and estimating costs in ways that policymakers can use. Many efforts are under way to determine health effects attributable to or associated with environmental hazards. Far fewer attempts are under way systematically to estimate attendant health costs, such as costs of medical care and loss of output due to morbidity or premature mortality. A decrease in health costs attributable to some action can be considered an increase in benefits. Policymakers have a pressing need for estimating benefits of proposed actions in order to make more informed choices among alternative actions. In attempting to estimate the health costs mentioned in the law, the ongoing study can provide a focus for developing methodologies and data for assessing benefits of steps that could reduce environmental hazards. P.L. 95-623 requests only some of the information that agencies require in doing cost benefit analyses. The ongoing study is to address only health benefits and omit other possible benefits, such as improved recreational facilities. It also will omit costs associated with pollution abatement, such as the costs industry incurs to decrease worker exposure or to decrease harmful emissions. Although the ongoing study need not specifically consider pollution abatement costs or non-health benefits, these factors, in addition to others, need to be included in decisions related to environmental hazard control. The ongoing study is to address those health problems of current concern. Many health problems in the United States, including major diseases and illnesses, such as heart disease, respiratory illness, and cancers, develop over long periods and result from the interaction of environmental factors with factors related to individual susceptibilities. These health problems are costly in terms of mortality, medical care costs, lost productivity, and associated pain and suffering and psychosocial problems. Long-term epidemiologic studies, and other kinds of studies and analyses, are needed to obtain information that will be useful in ameliorating or preventing these health problems. The ongoing study can encourage appropriate studies. Much relevant information is now collected both by the federal government and other groups. However, coordination of these efforts needs to be improved and committments of appropriate resources to long-tenm studies assured. Existing data collection systems need to be assessed for quality and relevance. While future analyses of existing information may yield useful results, substantial new data collection systems are also required to provide answers to current questions about environmental health hazards. -134-
The ongoing study as envisioned by this committee, will consist of two major parts. One encompasses all the data collection and analyses and related activities mentioned in the law. The other is made up of the biennial reports and any other written materials deemed appropriate. As indicated later in this chapter, the committee suggests that DENS be responsible for obtaining the data for the biennial reports from the various agencies that would be responsible for collecting such data. This committee has identified several goals and characteristics the ongoing study should have in order to work towards providing the estimates that Congress has requested. The first four items below are broad goals for the ongoing study; the last three are characteristics the ongoing study needs to achieve these goals. 1. Integrating available data The ongoing study should help to generate, coordinate, and integrate environmental and health data for use in identifying relationships among health problems and their environmental causes, and to enable evaluation of the consequences in terms of health costs. At present, health data and data related to the environment generally are collected independently, and it is difficult to integrate them. However, further exploration of extant data sets may show associations between illness in populations and environmental factors that suggest further study. In addition, there may be ways of linking various records about individuals in order to more effectively associate exposure to environmental factors and resulting health effects. Although the planning study focuses primarily on improving epidemiologic data, it is necessary also to consider data from animal and other experiments in order to prevent unnecessary human exposure and to extend the information obtained from human s tudies . 2. Coordinating statistical policy The ongoing study should encourage various groups, especially in the federal government, to explore the administrative and policy questions that hinder coordination of statistical policy. Questions of confidentiality and date compatibility might be addressed. The effort would help prevent unnecessary duplication of effort and would maximize the usefulness of data that are collected. Data linkage in this report refers to the process of combining a variety of sources of data about an individual in order to analyze them for suspected relationships. Thus, one might link data from occupational records with data from Social Security files and from death certificates to see if workers in certain occupations developed particular diseases or illnesses. -135- \
3. Developing cost estimates Many of the costs associated with illness cannot reliably be measured with existing data. The ongoing st dy should improve the methodology and data used to evaluate and expres_ these costs in dollar terms, as well as in dimensions other than dollars, from which the dollar f igures are derived. These other dimensions include, for example, morbidity measures, loss of workdays, and measures^of utilization of med ical servi ces . 4. Encouraging research Efforts to associate environmental factors with resulting health problems quantitatively, and to estimate all the associated health costs, are in early stages of development. Further research is needed to understand basic biological mechanisms, including individual variability, as well as to develop methods of analyzing data. Thus, the ongoing study should highlight areas needing further research and methodological development. These efforts should lead to better analyses of data and should improve the effectiveness of data collecting efforts. 5. Assuring continuity A mechanism to assure continuity and coordination is required for the ongoing study. The National Academy of Sciences (NAS) and DUBS can serve as the center of the coordinating effort, which will require cooperation of many groups inside and outside the federal government. 6. Providing funding A short-tenm effort would be of limited value in achieving the goals of the ongoing study, although useful insights could result. A continuing and adequate source of funding will be required to allow the ongoing study to begin to provide the estimates Congress has requested, and the ongoing study will need periodically to assess its costs. New funding sources may be necessary to implement the required long-tenm effort. 7. Providing flexibility Finally, the ongoing study must be flexible. It must respond to new developments. It must focus attention appropriately as new analytical methods become available, new information develops about disease etiology, and new problems assume major importance. Approach for the Ongoing Study P.L. 95-623 specifies an extremely broad charge. Although the committee wishes to plan an ongoing study that responds to that -136-
charge as closely as possible, it will not be possible to cover all areas mentioned in the law every two years. Therefore, the committee recommends a more modest approach initially. The committee considered which environmental factors should be considered at first, given that the law specifies "pollution and other environmental conditions resulting from human activity...." These environmental conditions include almost all extrinsic influences on humans. They include activities often classified as "life-stylet' or "voluntary," such as habits of exercise, nutrition, alcohol consumption, and smoking, as well as exposures to noise, chemicals, air pollutants, radiation, and related factors. The committee recognizes that "voluntary" and "involuntary" exposures span a continuum with no sharp distinction between the two.* Nevertheless, the committee recommends that the ongoing study initially focus primarily on the health effects and costs resulting from relatively involuntary exposures to physical and chemical agents deriving from human activity. Any studies to assess the health effects of the involuntary exposures to physical and chemical agents, however, must take life-style and related factors into consideration when they affect a person's response to harmful agents. For example, studies on respiratory illness among miners or asbestos workers should include smoking history. Thus, data related to life-style and other factors such as socioeconomic status should be collected when possible. The remainder of this chapter presents several types of recommendations. First are recommendations for implementing the initial phases of the ongoing study, including preparation of the first biennial report (after the planning study). Following this are more general recommendations designed to improve our ability to provide the estimates Congress requested. These recommendations concern ways of improving data and analyses related to health, the environment, and costing; improving coordination among various concerned groups and maximizing usefulness of the reports; and providing the required backup activities needed to carry out the earlier recommendations. Most of these general recommendations are long-term continuing efforts. They could begin in the near future with appropriate support, but would be unlikely to become fully implemented or to yield results in the near future. In making these recommendations, the committee is building on previous efforts and expects that the ongoing study will remain Differences of definition and usage can blur the distinction even further. For example, some people question whether smoking is a "voluntary" activity, and others argue that asbestos workers have "voluntarily" assumed their exposures. -137-
cognizant of related activities, especially those that the National Center for Health Statistics (NCHS) is carrying out in complying with other portions of P.L. 95-623. The Early Years of the Ongoing Study Conceptual Framework The committee recommends the use of a conceptual framework for identifying and organizing the data requirements for the ongoing study. The framework consists of a series of tables which, if they could be completed, would provide the data required to meet the congressional charge in Section 7(e)~1)E of Pit. 95-623.* (See Appendix A for the full text of Section 7.) For simplicity, the framework is expressed in teems of substances, but could apply to noise and radiation. To trace the chain of impacts, four types of tables are required. The first relates sources of hazards to ambient concentrations or total amounts of these substances in the environment,** a second relates amounts of hazardous materials in the environment to human exposure, a third relates human exposure to health effects,*** and a fourth puts a value on these health effects. To determine the reduction in health costs from incremental reductions in a particular environmental hazard, relevant entries from the four types of tables would be appropriately multiplied together and additional necessary computations performed. This scheme is idealized; much of the needed information is not available now. Even attempts to provide the required information for such well studied cases as asbestos or lead leave many questions unanswered, and require monographs rather than simple tables. In addition, the information that exists has been collected in a variety of forms for many different purposes. It is, therefore, difficult to mesh the various sources, including literature sources, to yield in a consistent manner information required to fill in the framework. However, this framework may be helpful in organizing available This Section asks the ongoing study to Quantify (i) the present and projected future health costs of the diseases and effects identified under subparagraph (B), and (ii) the reduction in health costs which would result from each incremental reduction and change referred to in subparagraph (~(ii)." **These tables would correspond to Sections 7(e)~1)A and C of P.L. 95-623. *The second and third tables correspond to Sections 7(e)~1)B and D of P.L. 95-623. -138-
information and seeing where information is available and needed. Any numerical estimates would have much uncertainty. Initial Activities The early phase of the ongoing study wit 1 have four re lated and intertwined activities, which may continue in modified form throughout the ongoing study. These are: o reviewing what is known o assessing the importance and promise of available information o deve loping i 1 lus trative case studies o encouraging development of required methodologies. In recommending these activities, the committee recognizes that the state of the art may be such that some of these activities may not be possible to complete at this time, but the committee regards the activities themselves as valuable and informative for developing the congressionally mandated program. These beginning activities of the ongoing study would not require new data collection. A. Reviewing What Is Known The conceptual framework described earlier will be used to identify and assess information on the sources and amounts of pollutants, levels of human exposure to these environmental hazards, the health effects resulting from such exposures, and costs of the health effects of such pollutants. Topics would be chosen by applying the general criteria for setting priorities identified in Chapter 1 of this report. The ultimate objective of this approach to classify the various agents and their adverse health effects in teems of: their known, suspected, or potential health impacts, taking into consideration the amount of hazardous material in the environment o the reliability of existing knowledge about them and the potential usefulness of additional information in minimizing harmful impacts the public perception of their importance o the need to study them further for policy purposes. 139-
Use of these criteria in setting priorities will require continuing refinement as more experience is gained with efforts to examine specific cases. B. Assessing the Importance and Promise of Available Information From the exercise of seeking needed information and evaluating the data available, data needs will be identified, promising areas for research outlined, and additional recommendations made to assemble the needed data. This assessment will provide a perspective for priority setting and a way of mandating feedback to improve future studies. C. Developing Illustrative Case Studies These case studies would not attempt to duplicate the efforts of other groups, such as the International Agency for Research on Cancer or the National Academy of Sciences, but would go further in attempting to integrate the many kinds of information needed to meet the congressional request. Until concrete cases have been developed, the appraisal of health costs will continue to be a theoretical exercise of little or no value for action. When many case studies are available, the resulting data on comparative health costs will help policymakers decide where the national effort can best be distributed. By choosing and appraising cases likely to be of economic importance, the ongoing study can add substantially to the growing body of knowledge so that, in addition to being a valuable illustration, it also contributes to the long-term goals stated by Congress. However, even for instances that do not have major economic implications, a case study may be valuable if reduction of ill effects is likely to be readily attainable at modest cost. A full case study shows the users of the report the complexity and the uncertainties raised by these cost and health questions. At each step in the process, the uncertainties and gaps in the data would be identified and the reasons for them described. From this activity, recommendations for changes could be derived that would make more satisfactory execution of the task possible in the future. The conclusions drawn from case studies would be presented with appropriate qualifications and their usefulness for the purposes of setting policy analyzed. It would be particularly important to attempt to extract from this exercise any generalizations that might apply to methodology, economic, political, or ethical issues. Two case studies should be undertaken at first. One would focus on a health problem and consider exposures as environmental determinants. The other would focus on a pollutant or source of -140-
pollution and consider all health consequences. This dichotomous approach would requi re re searchers to cons ider an array of environmental determinants and health consequences. Cost estimates would be attempted in both case studies. The goal would be to choose cases where the etiologies are relatively well understood so that data gaps could be identified, methodologies tried, and cost estimates attempted. Examples that might serve these purposes are: Health problems: chronic respiratory disease; lung cancer; angiosarcoma; chronic disorders of the nervous system Pollutants: asbestos; vinyl chloride; lead; nitrosoureas. During the first two years of the ongoing study, it should be feasible to report back to the Congress for a disease, or group of diseases, of known etiology. Relevant questions include: o What is the national estimate of exposure to the agents of concern? o How is this distributed among demographic subgroups of the popu let ion ? o What proportion of it is due to direct or indirect occupational exposure, such as by family contacts ? 0 How does community exposure, stratified, for example, by contact with production sources, compare with general exposures ? o Are there particularly sensitive populations at risk? Similarly, if antecedent epidemiology and animal studies have been done, it should be feasible to estimate the proportion of the disease attributable to environmental exposures of various types. Attempts to carry cut these case studies should indicate methodological and data gaps that require attention. D. Methodological Proarams In addition to the studies indicated above, the ongoing study needs a program of methodological development. Once systematic attempts are made to assess current information, it will become clear which areas require new methods to get quantitative results. The ongoing study, therefore, will need to explore the comparative merits of various data sources and methods for estimating health effects of the environment and their costs. Original appraisals of various methods will be required, because, as usually executed, these methods would not contribute much to the data of the ongoing study. Consequently, the appraisals would need~to consider not only the 141
of the problem every two years, areas of attention should vary, perhaps on a rotating basis. The reports would include the parts of the framework described earlier, that is, data on sources of hazards, human exposure, health effects, and health costs, as well as in-depth analyses of some aspect or aspects of environmental health cost problems. The reports also would discuss methodologies and make recommendations. The committee believes that these reports would be responsive to the congressional charge to "identify deficiencies and limitations in the data on the matters under consideration," "recommend actions which may be taken to eliminate these deficiencies and limitations," and "include recommendations for facilitating studies of the effects of hazardous substances on humans...." The law also requires that the reports list various regulatory and related activities that federal agencies-are taking to reduce the identified health costs. A section of each biennial report can include these. Because the committee views methodological development as a particularly crucial part of the ongoing study, it suggests that periodically, as part of the ongoing study, a systematic review of the methodologies available to do the work of the ongoing study be collated, summarized, and published. Monographs prepared by the National Center for Health Statistics, which puts out its methodological papers periodically, may be appropriate models. The proposed review would involve statistical, epidemiologic , toxicological, decision-making, econometric, and other methodologies, insofar as they are relevant to the work and reports of the ongoing study. The ongoing study could further encourage and publicize methodological developments by additional means, such as holding conferences and publishing papers. The First Biennial Report The first biennial report will reflect the work of the ongoing study during the first two years. Thus, it will consist of tables of data related to the conceptual framework, indicating in which areas data are available and how dependable the estimates are; two case studies; and a discussion of some specific methodological problems and data needs, particularly those discovered while attempting to do the case studies and fill in the framework. Recommendations arising from these activities will be noted. -143-
Expressing Uncertainty Estimates of uncertainty should be attached to each estimate of amount of pollution, health effect, or health cost. Almost all estimates derived in the ongoing study and published in the reports will have substantial uncertainty,~that is, the actual values could be quite different--even by orders of magnitude--from the estimates. The interpretation of data produced and the subsequent usefulness of these data will depend in large part on how successfully the uncertainty is expressed. The committee believes that if the degree of uncertainty is clearly expressed and explained, the estimates can be useful in making decisions and in guiding efforts to decrease the level of uncertainty. The committee makes the following recommendations about expressing uncertainty: Whenever possible, the reporting system should favor numerical estimates rather than use of imprecise teems such as "high," 'medium," and 'llow.'t In this way, the data potentially will be most useful to the various groups that might use it, and will avoid the misunderstanding that may result from using words' that are not clearly defined. Numerical reporting is not possible in all cases, and, where appropriate, various classification schemes should be tried for reporting uncertainties. A description of the major sources of uncertainty should be provided. In estimating risk, it is important to indicate not only the magnitude of the risk and its uncertainty, but also the basis for that uncertainty. Where the uncertainty is due to questions about which of several sets of assumptions applies, estimates should be reported conditional upon the different hypotheses. If an estimate is volatile, that is, likely to change in the near future as information is gathered, this fact should also be indicated, so that policymakers can take the potential for learning into account in making near-tenm decisions. In some cases, a final decision can await better data. Interim decisions should preserve flexibility for the future. Administrative Arrangements for the Ongoing Study P.L. 95-623 (Section 7(e)~1~) specifies an ongoing study to be carried out by the Secretary of the (now) Department of Health and Human Services and the National Academy of Sciences "jointly and -144-
in cooperation" with various other federal agencies. DHHS and NAS have interpreted that section of the law to indicate that they are jointly responsible for the ongoing study. However, the NAS is a private organization chartered to provide independent advice to the government. Several officials of the NAS believe it is inappropriate for the organization to undertake a study jointly with a government agency because such an arrangement would compromise the NAS' ability to provide independent advice. This planning committee strongly favors an administrative arrangement that would fix responsibility and authority for the ongoing study in one clearly defined body, presumably DHHS. This TOM planning committee recommends that the law be amended so that the NAS could serve in its customary advisory role, and DHHS could have primary responsibility for carrying out the ongoing study. The committee believes it is highly desirable for the NAS to participate actively in the ongoing study on a continuing basis. Details concerning the continuing relationship between the lead federal agency and the NAS and internal arrangements the NAS might need to assure continuity and the involvement of appropriate expertise should be taken up by officials of the respective organizations. Within the constraints of the present law, the committee describes what it believes to be the best of several alternative arrangements it considered in attempting to delineate some areas of joint activity while preserving the NAS' role as an independent advisory body. Most of the particulars of this arrangement could apply whether NAS served in a preferred advisory capacity or in an arrangement involving joint responsibility. DHHS is assumed to be the lead government agency in the proposed arrangement, which entails formation of two committees, one organized by NAS and the second by DHHS, each with different responsibilities for carrying out specific tasks of the ongoing study. The Process of Administration The NAS would establish a permanent committee with rotating membership to carry out its obligations for the ongoing study. To insure that health considerations remain paramount, the Institute of Medicine (IOM) should play a leading role in forming this committee. However, other groups within the NAS, such as the Environmental Studies Board, the Board on Toxicology and Environmental Health Hazards, and the Committee on National Statistics, have a great deal of expertise and experience in the various areas of the ongoing study, and these groups should play major roles in the ongoing study. -145-
The Secretary of Health and Human Services would create a parallel committee. It is proposed that the two committees have different responsibilities. - The DHHS committee would have primary responsibility for operation of the ongoing study, for intergovernmental relationships involving the study, and for generating and assembling the data. The NAS committee would have lead responsibility for developing research goals, for much of the data analysis, and for overall evaluation of the program. The actual collection of the data would be the responsibility of individual agencies. The two committees would jointly develop priorities for each biennial report and the specifics of its preparation, and would recommend priority areas for study, the scope of such studies, subjects for special monographs on relevant methodologies, and the agencies responsible for providing the reports or parts of the reports. The two committees may meet both jointly and separately and, as they find it desirable, form special task forces that include representatives of additional government agencies. The committees would separately and jointly review the reports and provide evaluations of the reports to the Secretary. The Secretary, after consultation with cooperating agencies, would forward the evaluations to the Congress with or without additional comment. ~ Both DHHS and the NAS/IOM committee would carry out such analyses as they deem fit. The NAS/IOM committee may from time to time issue formal recommendations for data collection, especially on matters of importance or on which it is difficult to reach consensus, and these recommendations would be subject to the usual review procedures of the NAS. The Secretary of HIS would engage various parts of the Department and would cooperate with other government agencies, especially those designated in Section 7 (e)(l) of P.L. 95-623. Within DHHS, the National Center for Health Statistics, which has statutory responsibility for the collection and dissemination of statistics on the health of the nation, would be an integral part of the ongoing study. To insure cooperation of agencies outside DHHS, the Secretary might request the President to create by Executive Order an interagency committee that would link the work of DHHS with the other agencies concerned. The work of committees and groups already in -146-
existence, such as the Regulatory Council and the Task Force on Environmental Cancer, Heart and Lung Disease, should be tied in appropriately with the work of the new committee. The various more or less formal coordinating committees with which the Task Force has links could be developed into a network to give valuable cooperation to the ongoing study. The ongoing study may In part use statistics and other data gathered for other purposes, but it may also require cooperation of the participating agencies in redirecting their data collection and research programs. New appropriations and significant additions to the budgets of the various agencies may be necessary. These efforts require the support of various parts of the Executive Office of the President in order to be effective. In particular, the Office of Management and Budget has powers which are required to be exercised for the coordination of statistical programs throughout the government, and OMB must support the provision of funds that will be required for the ongoing study. Commentary If DUES were not designated as the lead government agency, other groups within the federal government might conduct the ongoing study in conjunction with the National Academy of Sciences. These possibilities are considered briefly in Chapter 5 under three headings: 0 the lead agency approach o the interdepartmental committee approach o the Executive Office of the President approach. Because P.L. 95-623 specifies DUBS as the lead government agency, this committee did not consider other possibilities in sufficient depth to choose among them. However, they are listed as possible options if a different organization for the ongoing study is considered. Several advantages result if the Secretary of HUS is the government representive responsible for the ongoing study. The Secretary is a member of the President's Cabinet, with the status and access to presidential support that the position implies. Agencies within DHHS have heavy responsibilities in matters relating to the ongoing study, including the Centers for Disease Control, the National Institute of Environmental Health Sciences, the National Center for Health Statistics, and others. This planning committee emphasizes that whatever administrative arrangement is chosen should provide for participation by various -147-
federal agencies as well as others, such as non-federal agencies, public interest organizations, industry, and labor. DATA AP~ METHODOLOGIES TO ASSOCIATE EXPOSURES AND HEALTH EFFECTS Concurrent with the early part of the ongoing study would be its efforts to improve the available data and stimulate new data collection. Major efforts should be taken to improve our ability to relate environmental hazards to health outcome. This recommendation underlies virtually all the remaining statements in this section. Existing data on disease, disability, and death should be used to best advantage, and ongoing data collection systems modified to facilitate discovering exposure-outcome associations among groups and individuals, some of whom may be particularly susceptible. In designing new data collection systems, their ability to relate environmental exposure to health outcome should have high priority. In some cases, data that relate health consequences of specific populations to environmental variables may prove useful; in other cases, it may be possible to develop ways to link data about individuals while assuring that standards of confidentiality and privacy are not breached. Because of the way current data systems are set up, major needs for new data lie in the general areas of: 0 actual exposure of individuals and groups o more complete reporting of morbidity as well as mortality. Occupational records may be particularly useful in deriving suspected associations. Recommendations related to improving data and analysis for associating environmental exposures and health effects follow. Although many can be initiated in the near future, most require a long-term commitment. Existing Data and Data Systems Data systems may exist within the various agencies that could be analyzed more fully. Analysis of existing data, if adequate for the purpose, is generally less expensive than collection of new data. The committee recommends that, whenever possible, existing data and data systems be used to the fullest extent. -148-
Existing data systems and surveys should be expanded, where feasible, to aid discovery of associations between exposure and health outcome Information may be "piggy-backed" onto ongoing studies and surveys. As examples, future cycles of the National Occupational Hazard Survey (NOHS), conducted by the National Institute for Occupational Safety and Health (NIOSH), might be Expanded to provide duration and intensity of exposure rather than data only on the number of people potentially exposed. NCHS could add questions about occupational exposures to its National Health Survey. In addition, ongoing studies and systems can be used for follow-up studies, in which a particular group is studied again, or for followback studies, where the past expert enc e o f a group is analyzed. Integrating and Linking Data Systems Data integration among current large data collection activities should be carried out so that health outcome data can be related to environmental exposure data. However, all such efforts must be consistent with the need to protect confidential information and individual privacy, while providing data on possible differences in response to environmental factors because of individual susceptibility. Various agencies collect data that are useful for the ongoing study and could be shared with other agencies without introducing problems of confidentiality. However, the data often are not compatible . To encourage sharing and analysis of these data, the committee recommends that data collection be standardized so that data sets are compatible. This can be done by: o developing methodologies that will increase the capacity to integrate data sets in the future o collecting and reporting health statistics in suitable geographic units o working toward commonly accepted definitions, minimal data sets, and techniques to be used in environmental epidemiologic research. Guidelines that NCHS is preparing and those prepared by the Interagency Regulatory Liaison Group may be useful in achieving these goals. Lists of key words and characteristics of the various data systems available among the agencies should be prepared and disseminated. Such efforts will help ensure that the various agencies and groups are aware of efforts being undertaken by other -1 49 -
groups. The various data baser should be evaluated for their usefulness in associating exposures and health outcomes. Even if data are compatible, use of data from different source should be undertaken with care. Each agency collects data for its own purposes, and the data systems were set up to achieve those purposes. However, the committee notes that linkage of data files created for other purposes should be encouraged when there is good reason to believe that results of a specific linkage program will b sufficiently complete for the purpose at hand and that the biases _ limitations of linkage studies will not be so severe as to vitiate results. It is recommended that federal agencies consider ways to mask specific raw data so that such data can be made available to various groups, including non-government groups, to analyze and evaluate. The committee recognizes that there are often competing interests and objectives among such groups as industry, individuals, and public interest groups regarding confidential information. The need to protect the proprietary interests of companies and the privacy of individuals may conflict with the needs of researchers or other groups. Where possible, data should be released in a form that is useful but still preserves the necessary confidentiality. New Data and Data Systems In setting up new data collection systems related to health or the environment, experts in relevant fields should work together to make the resulting data as useful as possible for deriving relationships among environment, health effects, and costs. Health professionals, environmental professionals, economists, and others should contribute to and monitor the data collection and analyses. For example, if a system to monitor health effects of a particular occupational exposure is set up, the data collected should be sufficiently detailed to permit reliable cost estimates. Use of broad-based groups to plan for data collection also will increase the consistency of data and the likelihood that various data systems will be compatible. -150-
Longitudinal Studies Longitudinal studies will be required to obtain dependable, useful, population-based information. Such studies and development of appropriate data bases should be undertaken, and the needed long-term commitments of funds and personnel assured. Longitudinal studies, in which a population is followed through time, can record systematically changes in health status and environmental exposures. Careful evaluation of such groups, such as was done in Framingham to identify risk factors for cardiovascular disease, is useful for generating hypotheses that relate environmental factors to health outcome. Population groups should be selected to serve as "population laboratories." They may be groups with special exposures, or that are particularly susceptible to certain illnesses or hazards, or are normal. Information collected should include potentially important and confounding epidemiologic variables such as income, socioeconomic status, occupation and industry, education, race, smoking and alcohol consumption history, and occupation. Physiologic measures and analysis of body fluids might be included. These types of studies are expensive to set up and maintain, and, therefore, require a continuing commitment of resources. Designing these studies should be undertaken as soon as possible, although the results will be most valuable over the long term. Morbidity Reporting Systems for reporting morbidity, as well as mortality, should be improved, at least for certain indicator diseases. Some diseases almost always indicate that a certain exposure has occurred, for example, hepatic angiosarcoma is strongly associated with vinyl chloride. In most situations, however, an illness or disease is only loosely associated with certain kinds of exposures, and the rate of occurrence increases in populations that are exposed. As examples, the rate of occurrence of chronic pulmonary disease may be considered a crude indicator of air pollution, or bladder cancer an indication of possible exposure to certain classes of chemicals. Changes in the rate of occurrence and in the geographic distribution of these and other indicator diseases could provide useful clues about possible environmentally associated factors. Every effort should be made to link existing morbidity data -151-
for follow-up studies. ~ framework for making these decisions should be established and should include scientific and social criteria. Acute events also should be used to conduct research on resulting psychological and mental effects. For some episodes, such as the accident at the Three Mile Island nuclear reactor in March 1979, the resulting psychological effects may be more debilitating and costly than any detectable physical illness. Measuring Human Exposure Improved methods and measurements are necessary for determining human exposure to environmental hazards more accurately. Investigations should focus on determining ambient concentrations of various substances in locations where people spend time, and in determining individual exposures . The usefulness of environmental specimen banks should be explored. People are exposed to a wide variety of substances from many sources, including air, food , and water. However, quantitative information on actual exposure and absorption is scarce. Unti 1 now, most monitoring has focused on regulatory needs and the determination of ambient levels or emission levels of pollutants. But what is needed is information about the materials to which people are exposed and the amounts they absorb. Greater efforts should be made to monitor the various environments in which people spend the most time, particularly indoor environments, such as homes and offices. Methods for assessing individual exposure should be explored, including increased use of devices such as portable air samplers that can be carried around to monitor the air a person breathes. These devices are already used in various occupational settings and for research in monitoring individual daily exposure to a variety of air pollutants. Broad-based environmental sampling should be developed to identify the emergence of new or previously unrecognized environmental hazards. Specimen banks provide a way of storing samples so that they are available for later analyses. Samples from the environment, such as atmospheric particulates or soil, as well as from human and animal tissues, should be stored. Systematic examination of human tissues and body fluids may yield useful information on body burdens of various substances. Analysis of human blood, urine, teeth, hair, fingernails, adipose tissue, and breast milk can provide information about the absorption and metabolism of various substances. Relatively persistent biologically stable materials, such as heavy metals and organohalide compounds (DDT, PUB, PBB), are most easily assayed in this way. -153-
Occupational Surveillance High priority should be given to the continuing assessment of occupational exposures and consequent health effects. Cooperative efforts among universities, labor, management, and other interested parties should be encouraged. Occurrence of occupational illness or disease often constitutes the first warning of a new environmental hazard. Because workers are a relatively well-defined population and receive higher dose levels of various substances than does the general population, the possibility of associating exposure with health outcomes is high. Various organizations, including labor organizations and companies, may have records related to exposures, illness, and death among workers that would be useful for the ongoing study. However, difficulties arise in using worker records. Workers are exposed to many different materials and may hold a variety of jobs during their lifetimes. Records of unions and industry may be incomplete. In addition, some relevant information may be held confidential and be unavailable for study. Many of the previous recommendations, particularly those related to longitudinal studies and studies of acute exposure episodes, apply to the occupational environment. In general, occupation and industry information should be obtained in any epidemiologic study, where applicable. Specific recommendations related to occupational studies follow. Efforts to improve information on occupational exposures should be enhanced. A National Occupational Hazards Survey should be conducted at regular intervals and the results disseminated widely. Although workers are exposed to many substances, there is inadequate quantification of such exposure. The first NOHS survey, conducted in 1972-1974, found that many of the materials in the workplace were designated only by trade names without identified ingredients. Efforts should continue to identify the composition of materials that workers use and to which they are exposed. The continued assistance of industry and labor should be sought in establishing a taxonomy of probable exposures in various industry/occupational groups. Programs for collecting occupational exposure data should be systematized to permit linkage with data on disease, disability, and death, and with other kinds of health outcome data. -154-
For example, programs for air and blood lead monitoring mandated by the Occupational Safety and Health Administration (OSHA) should be related to the individual employee's health records. When new industries, technologies, and potentially toxic agents enter commerce, special efforts should be made to monitor the people who will be working under the new conditions to detect early problems. Development of continuous lifetime work records for an individual would be useful in reconstructing lifetime occupational exposure and health histories. Efforts should be made to insure that union and industry records have exposure data, relevant personal data, and provision for follow-up as the individual changes jobs. It should be possible also to link NOPS and Social Security data, including the Longitudinal Employer/Employee Data (LEED) File and the Disability Award File, to determine if adverse health outcomes are associated with certain occupational hazards. Uniform recording of occupation and industry on death certificates should be further encouraged to help ascertain associations between cause of-death and occupation. A list of sentinel health events for occupational disease and illness should be developed to alert people to potentially preventable occupational health problems. For some diseases, a single case is a "danger" sign; for other diseases, a change in rate of occurrence warns that a problem may exist. The feasibility of instituting a national decennial occupational mortality series in the United States, such as that in Great Britain, should be explored. Such a series would highlight occupations with increased mortality, and might prove helpful in efforts to lower mortality in high-risk occupations. ESTIMATING COSTS A major function of the ongoing study is to determine the costs of environment-related health effects, including the reduction in health costs that would result from stepwise reductions in pollutants and other hazards, and to suggest areas for future research in data sources and methodologies. This committee recognizes that costs may be expressed in terms other than dollars and that, ultimately, dollar estimates are derived from these other measures of the burden of illness. Many of the following recommendations are directed toward the expression of costs -155-
in monetary terms whenever possible, but it is assumed that the ongoing study will also present estimates for these other measures of the burden of illness when appropriate. Existing methods for estimating costs are neither complete nor exhaustive. As a fundamental recommendation, the committee emphasizes the need to assess the suitability of available methods and data for providing the cost estimates Congress~has requested and to foster development of improved methodologies and data. The ongoing study should attempt to determine the proportion of health costs of a disease or illness that should be attributed to environmental factors if a certain proportion of the cases could reliably be estimated as caused by these factors. An illness or disease may result from several independent or interacting causes, including environmental factors . Particular environmental factors may increase the incidence of a disease beyond that which would otherwise occur and may also increase the severity and consequent-costs associated with some or all of the cases. For example, air pollution and smoking may exacerbate asthma. Thus, even when there is a known association between an environmental factor and a disease, it is often difficult to estimate the share of the total costs of the disease that should be attributed to the environmental factor. Health costs should be calculated by the use of both output-accounting and willingness-to-pay methods, to the extent feasible. In the first few biennial reports of the ongoing study, estimates of monetary losses due to morbidity and mortality associated with environmental factors should be presented using a range of fixed values drawn from existing studies for cost per death, cost per day of illness, and cost per year of life lost. This type of presentation will indicate the range of possible results within present limits of accuracy of the available methods. Output accounting and willingness-to-pay are the two standard methods~now used for valuing the costs of health effects in monetary terms. Output accounting* measures losses in the output of the economy due to morbidity and death. Willingness-to-pay measures the resources that individuals would be willing to forgo to reduce a . Lithe output-accounting approach is often called the "human-capital" approach. Output accounting includes direct costs, which are primarily medical expenses, and indirect costs, which are a measure of lost output. -156-
risk. Both methods seek to quantify aspects of the burden of illness in monetary terms, but from different perspectives. Neither method is appropriate in all applications, but each can contribute to a greater understanding of the costs of health effects of environmental hazards. Chapter 4 of this report discusses these methods in greater detail and their theoretical and practical limitations. The ongoing study should compile a list of data systems that would be useful for developing the required cost estimates. Estimates using either output-accounting or willingness-to-pay approaches depend on available data. The list of environment and health data systems, prepared by NCHS,* could serve as a useful model for preparation of a comparable list of sources for cost data. The compilation might include evaluations of existing data sources and suggestions for modifying and linking them to make them more useful. Recommendations follow for estimating: o direct health costs: expenditures for prevention, detection, treatment, and rehabilitation o indirect health costs: earnings lost due to illness and death o willingness-to-pay: individual valuation of changes in the risk of illness and death o costs of pain and suffering. Direct Costs Efforts should be undertaken to improve estimates of direct costs by incorporating additional data in the calculations. Additional data sources may improve direct cost estimates. Utilization data is relatively good for hospitals and relatively adequate for physicians and nursing homes; the reliability of data on drugs and drug sundries by diagnosis has not been evaluated. Medicare hospital records provide good price data for the elderly, but other price data, such as those for physicians' services, require CONCHS. Environmental Health: A Plan for Collecting and Coordinating Statistical and Epidemiologic Data. DHHS Publication No. (PHS~80-1248. Washington, D.C. : U.S. Government Printing Office, 1980. -157-
substantial improvement. Promising sources for additional cost and utilization data include Medicare Part B records, the National Medical Care Utilization and Expenditure Survey (NMCUES), the National Medical Care Expenditure Survey (NMCES), the National Ambulatory Medical Care Survey (NAMES ), and third party payers, especially the data of the Federal Employees' Health Benefits Program. These data sets have, however, important limitations . Medicare applies only to individuals over 65 or disabled; the NMCUES and NMCES may include too few people to follow the costs of specific diseases, though they provide good estimates of aggregate costs; the NAMES includes only office visits by ambulatory patients to physicians engaged primarily in office practice. An investigation should be made of the difficulties of analysis introduced by the presence of multiple diagnoses. Although many patients have multiple health problems, the usual method in studies of costs of illness by diagnosis has been to attribute all costs to the primary diagnosis. For more accurate allocation of direct costs by disease category, in certain cases a portion of costs should be attributed to conditions that aggravate, bring about, or simply occur simultaneously with the condition that is the principal diagnosis. A comparison should be made between prevalence- based and incidence-based direct cost estimates for chronic diseases. Prevalence-based costs are more easily estimated than are incidence-based costs.* In many respects, however, incidence-based costs are more logical and would give more appropriate information, as, for example, in estimating future health costs of environmental hazards and stepwise reductions in costs that could be achieved by stepwise reductions in hazards. Few incidence-based health cost studies exist, but a comparative analysis of prevalence- and incidence-based studies would provide evidence as to how closely prevalence costs approximate incidence costs in those instances that logically call for incidence costs. Net direct costs of illnesses should be calculated In addition to the total direct costs now reported. Prevalence-based costs for a disease during a given year are the costs incurred during that year by all existing cases of the disease, both new and old. Incidence-based costs for one year are the total present and future costs (from discovery until cure or death) of all cases of the disease that began during that year. 158
Total direct costs are direct costs as conventionally estimated. Net direct costs take into account future direct costs, appropriately discounted, * that would have been incurred if the person had not died. Net direct costs are the total direct costs minus the discounted value of the health costs that would have been incurred if the person had survived. Net direct costs are the appropriate measure of direct cost reductions that could be expected if the incidence of an illness or disease decreased; use of total direct costs instead of net direct costs would exaggerate the savings in medical expenditures that could be realized by eliminating some d isease s . To obtain profiles of medical care costs over time incurred by individuals with specified diseases, a longitudinal survey of patients in a prospective study of costs by diagnosis should be begun. Where possible, existing data sources likely to provide useful information should be used. Profiles of medical care costs over time could indicate the distribution of costs between signal events, such as between first diagnosis of a disease and death, and the distribution of costs among a cohort at various points in the course of the disease. These two kinds of information would provide important parameters relating to some costs of disease about which little is known at present. The Medicare and Federal Employees' Health Benefits programs are two promising sources of data for developing profiles of medical care costs by disease category. The value of a longitudinal sample in providing more accurate estimates would be realized only over-a period of years. Direct costs in this sample would be calculated by directly measuring expenditures on behalf of patients with specific diagnoses. Hospital charges, physician charges, and other health care expenditures should be determined for each patient, and average expenditures for new cases of each d isease shou Id be derived from the individual data. The prospective study should permit more accurate allocations of direct costs among disease entities and also should yield new kinds of information, especially the distribution of costs over time within an episode. Non-health sector direct costs should be included with other direct costs if they are found to be significant for pollution-related diseases. The value of the sample would be further enhanced if socioeconomic variables and life-style characteristics were entered for each patient. l win estimating future costs, economists use a discounted value of these costs, because resources that will be available or lost in the future have a decreased value compared with those available or lost now. 159
In addition to longitudinal surveys, cross-sectional surveys of patients to determine utilization and costs of medical care should be considered as a source of information about direct costs and health effects. A cross-sectional survey does not provide information on medical care utilization and costs over extended time periods for individuals, but it can include a larger number of respondents at lower cost than can a longitudinal survey. As examples, NMCUES and NMCES are two surveys that will provide data, starting in 1981, on costs of disease and illness incurred by participants during one-year periods. Indirect Costs The development of indirect cost estimates poses problems that differ from those presented by direct costs. Direct costs measure expenditures that occur, whereas indirect costs seek to measure output or earnings that would have occurred, but didn't. The recommendations below are designed to improve estimation of indirect costs. New data sets containing personal and job characteristics for workers should be developed. They should include personal characteristics relevant to earnings and health, health histories, and work histories, including occupation, industry, and firm. Health problems should be categorized, at a minimum, into fatalities, permanently disabling injuries and illnesses, and temporary disabilities. There should also be measures of worker's compensation coverage. Non-market activities should be entered, if possible. If feasible, these histories should be combined with the longitudinal study of direct costs. These data will be valuable for measurement of both indirect costs and willingness-to-pay (the latter by improving the estimates of risk premiums in wages; in particular, failure to correct for worker's compensation causes a downward bias in the imputed wage premium for risk or injury). Improved methods should be developed for estimating earnings lost because of death. The estimates might be improved by using occupation and educational characteristics as well as age, sex, and race as -160-
predictors of earnings. Biases in the estimates of lost earnings may occur because health status and earnings are correlated ; this possible source of bias should be investigated. Improved estimates of the value of non-market time should be developed. The value of time for homemakers, retirees, and the unemployed needs further consideration. Estimates of productivity losses due to illness should be developed. These include direct loss of time due to absenteeism, the effects of chronic illness on productivity, and the effect of lost on-the-job training due to prolonged illness. The basis for choosing a particular discount rate for calculating present discounted values needs study and justification. There is strong theoretical justification for discounting, but there is little basis for choosing a particular value of the discount rate. However, the calculation of indirect costs (and net direct costs) depends critically on the discount rate chosen. Willingness-to-pay Estimates The ongoing study should assess the suitability of willingness-to-pay values drawn from existing studies for valuing losses due to morbidity and mortality resulting from environmental hazards. Further research on methods and data needs for estimating willingness-to-pay to avoid specific health outcomes should be undertaken. Continued use and development of relations between wages and risks as one way of estimating willingness-to-pay for avoiding death is recommended. The theoretical justification for willingness-to-pay is already well developed, and few methodological breakthroughs can be expected. However, there may well be considerable payoff from improvements of data. Wage data used in willingness-to-pay estimates should be improved. 161
According to economic theory, the appropriate measure for risk premiums should be after-tax wages rather than before-tax wages. However , gross wages should also be available for other types of ca lculations . Surveys should supplement market-based studies in estimating risk premiums paid to workers. Surveys, with all their difficulties, may be able to elicit information that alternative approaches cannot. For instance, wage premiums may inadequately reflect the cost of long-ter~ exposure to toxic agents, because workers inadequately perceive the risk. The committee proposes that, after identifying to the extent possible the diseases and symptoms caused or exacerbated by pollution, the ongoing study include surveys to assess willingness-to-pay to avoid consequent health problems, including death from; particular diseases. Symptoms of concern might include acute and chronic respiratory problems, angina, and so on. Any such survey should include important socioeconomic variables, such as age, sex, occupation, wages, education, wealth, and health status variables, especially a listing of the respondents' chronic diseases. These data should enable researchers to derive a function relating willingness-to-pay values for each of the health effects to personal characteristics of the individuals at risk. Costs of Pain and Suffering Surveys of willingne~s-to-pay seem at the moment the most promising method for measuring costs of pain and suffering. It is recommended that questions be added to surveys to determine if they yield consistent and reasonable answers for valuing pain and suffering. Any approach to the measurement of costs of pain and suffering must be regarded as a research problem at this time. IS SUES RELATED TO COORDINATION AND PLANNING Efforts should be made to coordinate data resources among the many groups that have an interest in environment-related health effects and ensuing health costs, because such efforts will be mutua 1 ly bene f i c ia 1 . As an early activity of the ongoing study, mechanisms of communication should be established with the many groups that can either provide information re levant to the ongoing study, or use the data and reports provided by the study. Such groups include federal officials, non-federal government officials, labor groups, industry, 162
health professionals, private foundations, insurance companies, voluntary health organizations, public interest groups, research institutions, universities, and international health groups. Information from the various groups should be sought about data collection and establishing principles, procedures, and standards for collecting, analyzing, and disseminating raw data. For example, public interest groups find that they need access~to raw data in order to support their views on policy issues. The ongoing study should maintain close liaison with non-federal government officials. Such contact should lead to more effective data collection at the state and local level. Such continuing contact also may guide the ongoing study towards providing data and analyses that more readily meet the needs of state and local government officials, many of whom require environment and health data in order to set regulatory standards. The various groups, especially those in the federal government, should explore the administrative and policy questions that hinder the coordination of statistical policies and the sharing of data. Feasible solutions should be sought. This is a long-term effort that includes questions of data compatibility, confidentiality, the purposes of the various agencies, and funding arrangements for collecting data that could be used by more than one agency. State and federal officials concerned with environmental health problems should increase efforts to communicate their needs to each other. Many issues need to be resolved about what kinds of data can be exchanged among agencies, and under what conditions and safeguards. Among its general recommendations, the committee would give high priority towards continuing efforts to coordinate statistical efforts at all government levels. Planning P.L. 95-623 calls for a broad study to determine the reductions in health costs that would result from incremental reductions in environmental hazards. However, environment-related health costs can be calculated reliably for few, if any, illnesses at present. The first biennial report will present some of the available information needed to make such estimations and will attempt to develop estimates for specific cases. A major purpose of the ongoing study is to set in motion the activities needed to develop improved estimates. This chapter has provided a conceptual framework for the ongoing study, including recommendations for its administration and suggestions for improving the available data. In order for the ongoing study to proceed, it requires additional funding, trained -163-
personnel, and research efforts. The ongoing study must also incorporate sufficient flexibility and feedback to respond to change and maximize the usefulness of the biennial reports produced. Funding Adequate resources must be available for Carrying out a task as complex and important as the ongoing study. Additional funding will be required to encourage various agencies to collect and analyze data in ways that meet the goals of the ongoing study. Costs of carrying out the ongoing study will depend greatly on the extent to which the recommendations in this report are followed. As the ongoing study proceeds, it will periodically need to assess its associated expenses. Although some recommendations for improving analysis of existing data or for adding questions to existing surveys are not overly expensive, collecting new data and setting up new data systems require a long-term commitment of funds. In particular, longitudinal studies must be undertaken to study long-te`~ effects of environmental exposures, and such studies are expensive. They can involve thousands of persons who must be studied over decades. Stable funding is required to make such efforts feasible. The costs of obtaining additional data or different kinds of data and of doing additional analyses should be considered in relation to the potential usefulness of the new information. Setting up and maintaining a data collection system is expensive. Therefore, the possibility that an existing system could be modified to serve a particular function adequately should be investigated before establishing a new system. Trained Personnel Efforts should be made to increase the number of persons trained in the fields needed to carry out the goals of the ongoing study. The language of P.L. 95-623 specifies a sufficiently broad and concerted effort to deal with environmental health cost areas that additional trained people will be needed in various fields. These fields include epidemiology, toxicology, biostatistics, environmental health, occupational medicine, and economics, the latter particularly combined with extensive knowledge of environmental health issues. Training in dealing with uncertain information would also be useful. -164-
Interdisciplinary efforts should be- encouraged so that data collection and analyses reflect the best information available in the various fields. Opportunities for training people in the environmental health fields should be increased. Training grants, such as those in toxicology given out by the National Institute of Environmental Health Sciences, may be useful. In addition, continuing education programs can develop competence in the various fields among people whose primary training was in related fields, such as medicine or scientific research. Medical school and other health training curricula should emphasize potential relationships between factors in the environment and health consequences. Research Further research, including research related to the basic biological mechanisms associated with disease processes, is needed for deriving the estimates required by P.L. 95-623. The committee suggests particular attention to methods of estimating exposure, populations with special susceptibilities, extrapolation of dose/response relationships to human beings from animal and other laboratory studies, and, in the long teed, the social environment. People are exposed to many substances for varying lengths of time and at different stages of their lives. Research is needed to learn what the important measures are. For example, are illnesses that are correlated with high peak exposures di fferent from illnesses associated with many years of chronic low-level exposure? In addition, ways of expressing and measuring effects of exposure to a variety of substances must be developed; usually only one substance or kind of substance is considered at a time. The effects of genetic factors and other conditions that predispose individuals and groups to disease susceptibility need further exploration. Results of these studies could play a major role in strategies to prevent illness from exposure to environmental hazards. There should be a strengthening of methods for assessing impacts of the environment on human health by extrapolation from comparative toxicological data in laboratory animals and other experimental systems. Such studies serve a predictive function, warning of potent ial harmful e f fee ts to human beings and al lowing prevent ive measures to be instituted before illness or disease occurs. 165
Some environmental factors resulting from human activity need considerable further definition and study, and, therefore, the committee does not suggest studying them in the short term. Research on effects of some parts of the social environment, such as housing and employment, that affect mental health should be conducted so as to generate the type of quantitative information needed for estimating the health costs required by the ongoing study. The social environment could be a subject for one of the biennial reports. Evaluation and Feedback The committee recommends that there be provision for evaluating the ongoing study and its reports to determine how they can better meet the needs of the various groups using them. The ongoing study must incorporate sufficient flexibility to respond to changing needs and conditions. The ongoing study and biennial reports should serve a variety of purposes. Regulators and policymakers may use the reports to obtain data for setting standards; public interest groups need data to help decide where to focus their efforts with limited resources; educators can use the reports as teaching tools. If widely distributed--and the committee hopes they will be--these reports could help keep the general public informed on such issues as pollution chemistry, health effects, economic analyses, and public policy issues related to environmental affairs. These reports also should alert readers to new methods relevant to studying costs of environment-related health effects and newly available data and data sources. For example, the reports could identify clearinghouses for certain kinds of information. The reports should present data and information in forms that are easy to use, with limitations, uncertainties, and sources of information clearly indicated. The various groups that may use the reports should have ample opportunity to indicate their needs. The ongoing study is seen as a continuing and changing effort, and it should be responsive to the various groups using it. CONCLUSION The cost estimates that Congress has requested cannot be obtained with precision in the near future. However, this committee concludes that if the assumptions, limitations, and uncertainties of any estimates are clearly indicated, attempts to obtain such estimates can help decisionmakers to make informed judgments. Much of the information Congress asks for is not now being provided, but much could be provided with a little more effort and -166-
some organizational changes. Some requires new attitudes or new methods of data collection. For example, change in data collection systems is required to remedy the current lack of information on the relationship between exposure to various materials or processes and health outcomes. During the past decade, increased public awareness has drawn attention to the role of environmental factors in health and disease. Many efforts are under way to understand, measure, and control these effects. The ongoing study provides a unique opportunity to evolve new efforts and methods and to focus and coordinate existing efforts. The ongoing study will encourage improved methods of estimating costs of health effects and of expressing them in a variety of ways. These estimates will permit a better evaluation of benefits of potential programs to reduce environmental hazards. Used with other necessary information, such benefit estimates can be used to better allocate resources for improving health and the environment. If the spirit of the foregoing recommendations is followed, the nation's ability to prevent illness and disease will be enhanced. Not only will information about relationships between environmental hazards and illness increase, but also analysts will be better able to assess the efficacy of actions designed to decrease harmful environmental health effects. By showing where efforts to modify the environment will yield the greatest health benefits, the ongoing study could provide a major impetus for preventing health problems. -167-
