The National Research Council was asked by the Centers for Disease Control and Prevention (CDC) to review the draft report titled A Feasibility Study of the Health Consequences to the American Population from Nuclear Weapons Tests Conducted by the United States and Other Nations. The draft report, which was prepared at the request of the US Congress, presents a limited evaluation of the potential public-health impact of atmospheric nuclear-weapons testing and suggests ways in which a more detailed evaluation could be performed. To review the report, the National Research Council formed a committee consisting of members of its Committee on An Assessment of the Centers for Disease Control and Prevention Radiation Studies from DOE Contractor Sites and other experts.
The committee to Review the CDC-NCI Feasibility Study of the Health Consequences from Nuclear Weapon Test was asked by CDC to address the following specific questions:"
Are the methods and sources of information used in the technical report to estimate radiation doses and health effects from fallout appropriate for this study?
Are the methods and results clearly presented in the main text of the technical report?
Are the findings presented in the report supported by the data and analyses provided?
Do the Options for Future Work presented in Chapter 6 represent an appropriate range of options for public health activities that could be pursued as a result of this study?”
Before addressing those questions, the committee offers general conclusions and recommendations.
The committee believes that the CDC-NCI (National Cancer Institute) working group performed a very competent feasibility assessment of the geographic distribution of probable doses to the population, the projected risks associated with those doses, and a potential communication plan. However, the committee has identified some weaknesses in the feasibility study and the draft report and has a number of suggestions for improvements.
Implicit in the feasibility study is the question of whether it is worth while to perform a substantially expanded study of all radionuclides. The committee believes that although a more detailed study is technically possible, neither the data nor the consequences appear to justify it. The measurement data on which the dose estimates rest are limited;1 additional dose refinements are unlikely to transform the existing sparse data into precise exposure estimates. Improvements in methods are always possible, but the quality of the data and the very low doses involved do not justify such efforts. To obtain substantially better estimates of exposures from NTS (Nevada Test Site) fallout would require the location of a body of exposure data that is much more extensive and of better quality than the available measurements; it is doubtful that such data exist. However, if extensive additional exposure data are found, the question about performing a more detailed dosimetry and risk assessment should be revisited.
If further dose-reconstruction work were to be proposed, however, it should be evaluated according to specific scientific and public-health criteria. The first would be improvement in the accuracy of exposure estimation at the individual or geographic level (to the extent of making new epidemiologic studies feasible, for example). Given the measurement data available, there is little hope that further dose reconstruction can achieve that; compellingly better estimates of doses are unlikely to be possible unless a substantial new body of dose-measurement data is discovered. A second would be sufficient improvement in the dosimetry to permit substantial risks to be addressed with meaningful studies. There is no evidence that either external or internal fallout exposures other than to 131I (iodine-131) could be large enough to produce detectable increases in individual risk in any feasible epidemiologic study, and fine tuning of the dosimetry will not change that situation. There could be a third criterion: to improve the characterization of the uncertainty of individual exposure estimates. But the ability to make major improvements in that arena also is speculative. The lifetime risk of a death due to cancer is about 20% absent the fallout radiation exposure. The fallout putatively raises that risk to about 20.03% (with a credibility interval of about 20.01% to 20.09%), which is of little health consequence. However, if new exposure data were found to indicate that exposures were substantially higher, this could trigger a need for additional studies. Nevertheless, the committee believes that there is insufficient justification for a more detailed study of the amounts and effects of fallout radionuclides other than 131I.
The committee supports continuation of the identification, cataloging, and preservation of fallout-related documents, that is, the archival work described in the draft report. In addition to its value for evaluating exposure to fallout, the committee believes that such work is well justified as a record of a historically important period in the nation’s history. If substantial new knowledge of radiation biology and markers of radiation injury are developed, data that are of limited current interest may be able to support the testing of new hypotheses that relate low doses of radiation to human-health effects. The ability to collect, document, extract, archive, and effectively use fallout records depends on the knowledge possessed by the diminishing number of scientists from that era who are still available and who can help to identify and interpret the data. It is not that the type of data available cannot be analyzed now, but that the scientists of
that era have knowledge unobtainable except through experience of the instruments and methods used at the time, their practical limitations, the conditions under which they were used, the reasons for recording or not recording various observations, the interpretation of records that are ambiguous without first-hand knowledge, and the likely location of useful observations among large quantities of otherwise irrelevant records. Without the participation of those scientists, the possibility of deriving much value from archival material for future testing of hypotheses on radiation effects will be seriously compromised.
The committee’s findings with respect to the four questions previously enumerated are summarized in the paragraphs that follow.
Are the methods and sources of information used in the Technical Report to estimate radiation doses and health effects from fallout appropriate for this study?
Methods used to estimate doses and health effects
The dosimetric modeling for the feasibility study has been built on the substantial body of work that was previously performed to estimate the amounts of radionuclides released by the Nevada Test Site (NTS) atomic-bomb tests, the amounts deposited within 100 miles of the site, and the amount of 131I fallout throughout the coterminous United States. That work greatly facilitated the modeling for the feasibility study. Moreover, the estimates of exposure to global fallout given in the draft report are based on a large body of measurement data from locations around the world. Those data and estimates notwithstanding, considerable uncertainty remains in the estimates in the draft report. That uncertainty appears to be largely irremediable because historical fallout measurements were made at fewer than 100 US sites. It seems likely that only marginal improvements in the dose estimates could be made by further refinements in the modeling unless some important new data were found that would substantially improve the quantity and quality of dose information or unless an extensive program of soil sampling and analysis were carried out.2
As to the health hazards produced by the exposures, the authors of the draft report used conventional risk coefficients for lifetime cancer mortality in characterizing risk. Those coefficients are based on the linear non-threshold hypothesis that presumes that any exposure to radiation increases one’s risk of cancer. Specifically, the International Commission on Radiological Protection (ICRP) risk coefficient of 5% Sv−1 was used to derive total cancer-mortality risk. That coefficient includes a factor to indicate that fractionated or low-dose-rate exposures are assumed to be only 50% as effective in cancer induction as high-dose, acute exposures. This is consistent with current methods but still has its limitations, including the use of linear extrapolation to doses that are so low that no available data could either validate or invalidate the assumption. In particular, such an approach may overestimate cancer risk in that it excludes the possibility of reduced risk per unit dose with small exposures to radiation. The US population was treated as a whole without specific attention to more-susceptible subgroups or subgroups at risk of higher exposures, other than an analysis of children born in 1951 at the
beginning of the NTS testing. If a more detailed analysis of doses is undertaken, consideration should be given to estimating risk for subgroups that might be at higher risk, such as children who drank goat’s milk.
The feasibility study estimated health effects on the basis of external exposure from both NTS and global fallout for total cancer. In addition, it estimated exposures and risks of thyroid cancer and leukemia on the basis of both external and internal radionuclide exposure. It did not estimate the risk posed by internal radionuclide exposure for other cancers, because the contribution of internal exposure to cancer of organs and tissues other than the thyroid and bone marrow was generally only a small fraction of that from external fallout. The committee concurs with that approach.
The risk-estimation models considered uncertainties associated with several factors: dose uncertainties, uncertainties in risk estimates, extrapolation to low doses and low dose rates, transfer of risk estimates from the Japanese to US populations, uncertainties in relative biological effectiveness for various kinds of radiation, and temporal projections of risk. The authors of the draft report did not consider several other types of uncertainty, such as uncertainties in selection of the particular risk model (for example, the shape of the dose-response curve) used, interindividual variation in radionuclide uptake and metabolism, or the existence of sensitive populations. The approach used in the feasibility study for uncertainty characterization tacitly assumes that a cancer risk exists at very low doses and dose rates. The validity of that assumption is contentious; however, it seems justified in this instance because the result helps to guard against the underestimation of risk.
One limitation of the findings presented in the CDC-NCI draft report is that to calculate the plausible range of risk estimates, the authors relied on a general estimate of the “credibility interval” of risk estimates per unit of dose derived from other studies and the assumption of a “credibility interval” for the dose estimates. The basis of assuming a factor of 3 for the “credibility interval” for the dose estimates was not explained, but it should be, because it is a key factor in generating the ranges of putative risks given in the report. Nevertheless, this is a report of a scoping study, and it was not feasible to conduct a detailed uncertainty analysis based on the specific uncertainties that went into the dose estimates of the study. If a more comprehensive study is performed, a thorough uncertainty analysis of both doses and risk will be needed.
To summarize, the methods used to estimate doses and health effects are reasonable for a feasibility study, although there are some limitations.
Sources of information
The draft report identifies many sources of information that were used in the feasibility study in the calculations of dose from weapons tests, and the methods used. It also identifies other sources of potentially informative data that were unavailable when the feasibility study occurred. Some of those data are in well-archived repositories, but many are poorly stored and cataloged. Cognizant of the possible loss of useful information on the magnitude of exposure to
ionizing radiation, the Department of Energy has ordered a moratorium on the destruction of energy-related documents of potential epidemiologic utility and has shipped all such documents to Federal Records Centers. Other relevant agencies, including the Department of Defense, have not declared a moratorium on the destruction of possibly relevant records; the Navy and Air Force, for example, have extensive documents of potential importance if additional analyses are undertaken. We recommend that CDC urge Congress to prohibit the destruction of all such records relevant to fallout and to permit appropriate access to them.
Are the methods and results clearly presented in the main text of the Technical Report?
The committee believes that the particular findings that the authors of the draft report chose to emphasize and the manner of presentation were not optimal and were in some cases misleading. The primary emphasis with respect to risk is on the population-summed risk, that is, the total number of expected excess cancer deaths in the exposed population over a lifetime, on the basis of the collective dose to that population. The number is presented with little context. A valuable context would be comparisons of the magnitude of the risk (the expected number of lifetime excess cancer deaths) posed by fallout with the corresponding risks associated with natural background radiation and with the total number of cancer deaths expected in the lifetime of the exposed population. In addition, the expected number of excess cancer deaths in the population is not readily translatable into an individual’s excess risk posed by fallout. The risk to individuals and the associated uncertainty are the most important objective pieces of information for members of the population in interpreting the personal implications of their fallout exposure. They are also key pieces of information for public-health purposes in that individual risk, and not population risk, drives considerations of whether special preventive screening programs would be recommended.
Most other aspects of the methods and results are clearly presented, with a few exceptions. The committee notes, however, that the report contains only an abbreviated presentation of the dosimetry methods used in the feasibility study. Such brevity may be appropriate for a general audience, but those with technical dosimetry expertise need fuller specification of the methods to permit evaluation and critique of credibility. It is therefore recommended that the authors of the draft report improve the completeness and clarity of the dosimetry chapter and its linkage to the four dosimetry appendixes.
A table and discussion outlining the sources of uncertainty in the dose estimates should be presented, including the basis for using a factor of 3 to define the range of uncertainties in dosimetry. If a more detailed study is to be conducted, the various sources of uncertainties in dose estimates should be incorporated into the modeling of risk, but the committee views such modeling as beyond the compass of this feasibility study.
The potential merits and limitations of a more detailed study of fallout dosimetry and risks are not adequately described in the draft report. It does not state the expected types and magnitude of changes that would occur if a more detailed study were conducted and the public-health value of a more detailed study also was not evaluated.
Are the findings presented in the report supported by the data and analyses provided?
The dosimetric findings of the feasibility study presented in the CDC-NCI draft report are generally supported by the data and the analytic methods used. It was beyond the National Research Council committee’s resources to check the models and the use of data in detail. However, given that the overall results are in general agreement with the body of measurement data available, the committee concludes that the dosimetry estimates are acceptable for the feasibility study commissioned by Congress. The committee is also in general agreement with the findings and conclusions in the draft report. The report could be construed as more responsive to the original request if it had included in the risk estimates the risks associated with other internally deposited radionuclides, but such contributions to risk would probably have been relatively small compared with external exposures.
The draft report indicates that the dose and risk estimates stemming from the feasibility study are population averages and therefore should not be used to estimate risks to specific individuals. The committee concurs with that caveat; to generate adequate estimates of individual risks, one needs to account for a person’s age, dietary sources and habits, geographic locations at particular times, and other factors.
Do the options for future work presented in Chapter 6 represent an appropriate range of options for public health activities that could be pursued as a result of this study?
The draft report identifies five options for future study, which can be summarized as follows:
Do no additional fallout-related work.
Retrieve and archive the historical documentation related to radioactive fallout from nuclear-weapons tests conducted by the United States and other nations.
Conduct a more detailed dose reconstruction for radioactive fallout from global nuclear-weapons tests for 131I, the most important radionuclide identified in the feasibility study.
Conduct a more detailed dose reconstruction for multiple radionuclides in radioactive fallout from both NTS and global nuclear-weapons tests.
Conduct a detailed study of the health effects of nuclear-weapons test fallout, including in a single project dose estimation, risk analysis, and communication of the results to interested parties.
The feasibility study considered an appropriate range of options, albeit it did not identify the benefits associated with each option or what policy, health, or scientific decisions would be affected by each.
The five options above intermix three separate issues requiring decisions:
Should effort be devoted to retrieving and archiving additional documentation concerning fallout? If so, how much effort?
The committee recommends that documents from a certain number of the sites most likely to have valuable information be retrieved, examined, and archived; the selection of the sites should be based on the knowledge of dosimetrists with experience from the 1950s. What the effort produces will determine whether it should be continued. The current destruction of old records for inventory-reduction purposes and the attrition of dosimetrists who are knowledgeable about the 1950s era mean that the need for document archiving and interpretation is urgent.
Should an expanded study of the magnitude of fallout-related doses and health effects be mounted? If so, what should its scope be?
The committee favors a prompt publication of the results of the feasibility study, to be followed by a reanalysis of the 131I fallout data and the data relating 131I dose and thyroid-cancer risk. The justification for redoing the 131I analysis is that it could include new dosimetry and uncertainty information available from recent research from Chernobyl and other sources and could correct errors in the earlier NCI study that have come to light. Those alterations would probably not make large changes in the key results, but they would lend more credibility to the risk estimates. Apart from the 131I-related risk of thyroid cancer, the draft report indicates that the likely risks associated with the other fallout radionuclides are small, so the increased precision gained by performing a full-fledged dose and risk estimation will probably be of little added value. Developing more precise estimates than given in the feasibility study about total collective dose to the American people from all radionuclides, and resulting total risk of all cancers may be reasonably motivated by academic and historical interest in this extraordinary chapter in our nation’s history, and to show that the US government is not attempting to keep veiled in secrecy the true consequences of the era of nuclear testing. However such an interest in collective dose would not imply that a full dose reconstruction for all radionuclides (to give dose estimates at a county-by-county level as was done for 131I) is scientifically justifiable. In the view of the committee the very low levels of exposure to radionuclides other than 131I, combined with uncertainties concerning their temporal and geographic distribution, imply that the resulting estimates would be without scientific (e.g. epidemiologic) or public-health consequence.
What should be the scope of a communication plan associated with the present NCI-CDC draft report or with a possible future report?
An effort must be made to communicate the feasibility study’s important findings to interested Americans. The goal is to inform the public on fallout and health-related issues. The committee recommends that a comprehensive and understandable public summary be developed and made part of the final report of the feasibility study report and that it be disseminated widely and in particular be placed on the NCI and CDC Web sites. The two agencies should also send the public summary as a separate document to previously identified stakeholders and others who have expressed interest, issue a press release about it, and announce its availability on various e-mail lists of interested organizations.
The public summary should be developed with all due speed. Stakeholders were frustrated about delays in the release of the draft report, and they remain frustrated about delays in releasing materials from the NTS 131I communication effort. Long delays in releasing reports of radiation studies have contributed to distrust in those expressing opinions to the committee.
Information from the feasibility study should be included in the NTS 131I communication effort.
The committee commends the NTS 131I communication plan for involving stakeholders and other interested members of the public on many levels. In its efforts to involve stakeholders, it set an example that should be followed for public dissemination of information from the feasibility study. The NTS 131I communication plan sponsored a well-developed risk-communication conference to which it invited various stakeholders and many distinguished US leaders in health communication, risk perception, and risk communication. The conference provided excellent information about the communication effort being planned. The committee recommends that a risk-communication conference be conducted by CDC to help it develop effective means of involving the public in formulating and communicating information about radiation and health issues.