A National Research Council Committee on an Assessment of CDC Radiation Studies was asked by the Centers for Disease Control and Prevention to review the final results of the Fernald Dose Reconstruction Project. The purpose of that dose-reconstruction project conducted by Radiological Assessments Corporation (RAC) was to provide to members of the public scientifically sound radiation-dose estimates resulting from the operations of the Feed Materials Production Center (FMPC) at Fernald, Ohio, from 1951 through 1988. Dose reconstruction was necessary because the radiation doses had not been directly measured. The most-accurate method for measuring the doses would have been to measure periodically the uranium or other radionuclides in the bodies of members of the public and to measure the radon in their homes, schools, and places of work. However, direct public dose monitoring was not done at Fernald, for a variety of reasons, including the assumption that a monitoring system would have been intrusive, impractical, expensive, and unnecessary.
To protect the health of the public, an alternative approach, environmental monitoring, has been routinely used. If the radionuclides in the environment are measured and kept below a predefined limit, excessive doses to members of the public can be avoided. That approach has been generally successful, and the risk associated with additional radiation exposure of members of the public has in many cases been smaller than risks associated with large industrial facilities in their vicinities. However, to gain perspective on risks, concerned members of the public have raised questions about what their actual radiation exposures were. How does one know what the doses were if they were not measured? How does one know that all the facts have been disclosed? A limited answer to such questions can be provided by a scientific dose reconstruction that is based on analyses of radionuclide-release data and modeling of exposures of typical members of the public. RAC's final report reviewed in this document represents such a study for the FMPC.
By using dose reconstruction, it is possible to fill some of the gaps in the data. When environmental-monitoring data do not exist, likely values can be reconstructed on the basis of measurements obtained elsewhere under similar conditions or at different times in the same
general area. How various radionuclides move through the environment can be estimated by using complex mathematical models that attempt to account for many important physical and chemical properties of the radionuclides. Ultimately, actual or potential doses to people can be approximated with these models and others that estimate how much and for how long the released radionuclides come into contact with members of the public in a given setting.
How good are the dose estimates? Whether the dose estimates are scientifically credible depends on the methods used and the care with which they were used. But even when the best methods are used carefully, dose estimates can be flawed. Gaps in the historical records lead to uncertainty. Occasionally, the gaps can be accurately filled because there is a clear relationship between missing and known data; more often, the relationship between missing and known data is complicated, and assumptions must be made. Such assumptions are necessarily imperfect and add uncertainty to the estimates of dose. Uncertainty is generally expressed not only by stating the best estimate of the dose but also by including a range of possible values. Scientists must acknowledge that uncertainty exists and try to estimate it. But it can be reduced only if more-relevant data are available. Even estimating the amount of uncertainty is subject to uncertainty. As with the dose estimates themselves, whether the uncertainty estimates are scientifically sound depends on the methods used and the care with which they are applied.
This committee concludes that the methods used in the Fernald Dose Reconstruction Project are appropriate and scientifically sound— indeed, often innovative. The uncertainty in the final dose estimates results mainly from gaps in the environmental-monitoring record and from the assumptions that are required to go from the reasonably well-known amounts of radionuclides released (the source term) to descriptions of how they were transported through the environment, potentially exposing members of the public. Added to those are the uncertainties in the cancer-risk coefficients that have been applied. Given the extensive record search, it is unlikely that important documents that would reduce the uncertainty or change the magnitude of the dose estimates remain to be discovered.
There are some differences between the approaches used by RAC and approaches that might have been recommended by the National Research Council committee. Those
differences are to be expected when complex tasks such as dose reconstruction are undertaken and there are significant gaps in the historic record. In the opinion of the committee, the RAC approach has resulted in an overestimation of doses to people from exposure to radon. This overestimation of the doses from radon is discussed in detail in the body of the present report.
The RAC dose reconstruction work reviewed by this committee was carefully done and covered all known pathways of exposure of the surrounding population. The committee is concerned that RAC's overestimation of doses received by the population from radon, although prudent, might produce some undeserved apprehension on the part of some citizens and should be carefully presented. While parts of the summary booklet of the report intended to present the risk estimations are well written, other parts do a poor job of overall risk communication. In particular, certain figures are misleading and should be revised. The committee recommends that RAC present separate estimates for some of the risk scenarios contingent on whether the person is a smoker or nonsmoker. In addition, the committee believes that the RAC report underestimates the uncertainties associated with its risk estimates. Although it is difficult to estimate the precise extent of overestimation, it might be a factor of 4 or 6 and needs to be corrected. The above issues might be important factors if CDC decides to evaluate whether an epidemiologic study of the exposed population should be proposed.