EXECUTIVE SUMMARY
At the request of the Center for Environmental Health and Injury Control, Centers for Disease Control and Prevention, the National Research Council appointed the Committee on an Assessment of CDC Radiation Studies in the Board on Radiation Effects Research of the Commission on Life Sciences to undertake a critical review of the April, 1994 drafts of two documents prepared by the Hanford Environmental Dose Reconstruction (HEDR) project. This project is reconstructing radiation doses received by populations in areas surrounding the Hanford nuclear facility near Richland, Washington. The drafts, submitted to the Committee in June 1994, are
-
Columbia River Pathway Dosimetry Report (PNWD-2227 HEDR), by W.T. Farris, B.A. Napier, J.C. Simpson, S.F. Snyder, and D.B. Shipler.
-
Atmospheric Pathway Dosimetry Report, 1944-1992 (PNWD-2228 HEDR), by W.T. Farris, B.A. Napier, T.A. Ikenberry, J.C. Simpson, and D.B. Shipler.
PNWD-2227 HEDR describes the methods used to estimate doses received by representative individuals who ingested water, fish, or waterfowl from the Columbia River or who swam in or boated on the river. It uses information on radionuclides released from Hanford reactors (Heeb and Bates, 1994), on transport of radionuclides in Columbia River water (Walters et al., 1994), on accumulation of radioactivity in aquatic organisms (Thiede et al., 1994), and on dose-calculation methods and human-exposure parameters (Snyder et al., 1994) in addition to other sources. It is an update of an earlier report (Pacific Northwest Laboratory, 1991a) but is more complete and includes additional data collected by the HEDR project since 1991.
Doses were calculated for five radionuclides that together contributed over 94% of the total dose: sodium-24, phosphorus-32, zinc-65, arsenic-76, and neptunium-239. Doses are presented as the effective dose equivalent (sum of the committed effective dose
equivalent from internal deposition of radionuclides in the body and the effective dose equivalent from external radiation) and dose equivalent (determined by multiplying the absorbed dose in rads by a quality factor) for the red bone marrow and lower large intestine.
PNWD-2228 HEDR focuses on estimating doses received by the thyroid of people who were exposed, directly or indirectly, to radioactive iodine-131 released into the air. A principal exposure pathway for many of these people was milk from cows that ate vegetation contaminated by iodine-131 released into the air from the Hanford site. This document describes the methods used to estimate doses received by representative individuals who consumed contaminated foodstuffs, inhaled contaminated air, or were directly exposed to radioactive materials released from the Hanford site, and it presents re-estimates of many of the doses that were given in an earlier report of a feasibility study for the atmospheric pathway (Pacific Northwest Laboratory, 1991b). The use of new, more-detailed calculation methods resulted in estimated doses that are in some instances half those presented earlier.
Critical portions of the project—such as the definition of source terms, model validation, and demographic details—are not included in these two reports even in summary form. Instead they refer the reader to other reports; in the absence of any discussion of these crucial topics, the committee is forced to take a great deal on trust. However, this review is required to judge the completeness of the work, its accuracy, the justifications for assumptions that underlie the inputs to the various models and the thoroughness and quality of the calculations, and thus the overall credibility of the results. The committee views it as essential that the model calculations clearly demonstrate not only consistency and an appropriate choice of parameters, but also satisfactory agreement with past measurements and observations, a reasonable representation of actual exposure conditions of “real” people, and a capacity to serve as a suitable basis for epidemiologic studies.
While the work that went into these reports is impressive, the reports are not adequate in their present form in that some serious questions remain. The reports should
be edited to show which portions of the calculations have been validated to an acceptable degree of accuracy as needed in future epidemiologic studies, to indicate which segments represent projections (and projections should be justified), and to define dose ranges received by representative, well-described individuals. Although average doses might be acceptable, doses received by individuals might be unreliable for a variety of reasons, which should be discussed. It is apparent that very few people might have lived in the high-dose areas. The committee suggests that a more conventional approach should have been used to estimate doses such as was recommended by the ICRP (International Commission on Radiological Protection, 1985)—to estimate the dose to the “critical group.” A major advantage of the critical group approach is that it avoids having a dose estimate being prejudiced by a small number of individuals with unusual habits. Potential doses received by American Indian populations are not discussed in detail, and the preparation of the “one hundred realizations” for the computer codes is not discussed. A more thorough discussion is also needed to explain the implications of the various periods selected for different exposure pathways.
In general, the presentation of the uncertainty analysis suffers from the lack of a table of specific sources of uncertainty and sizes and shapes of the assumed distributions of these sources and their components. Until some type of documentation is given, the uncertainty analyses have to be regarded as questionable.
Briefly summarized, the committee’s overall impression is that the dose-reconstruction effort uses up-to-date methods and is creative. However, an enormous emphasis was placed on modeling, sometimes at the expense of other, more-empirical approaches. The stated intention of the authors of the reports was to use historical measurements supplemented by modeling. However, the extent of the historical measurements that are available and used is never made clear. It would have been helpful too if the report contained a fuller definition of source terms, model validation, and demographic characteristics ascribed to the three representative individual types on whom doses are computed should be included. The details are presented in other reports, which
are cited, but a brief discussion of these crucial topics here would make the reports more self-contained and comprehensible and make the results more credible to the average reader.
The committee suggests that worker body-burden data be considered as a means for validating models. In addition, the committee recommends that the authors emphasize that the model shows that the dose from the river pathway is not important from a public-health point of view. The committee also believes that the authors have overemphasized the “maximally exposed individual;” dose maps presented in the reports might mislead the public if only a few people live in high-dose areas.
It is important that the readers of the HEDR reports understand the difference between modeling and dose calculation for environmental-impact assessment and for retrospective dose assessment. In environmental-impact assessment, as has been done at Hanford in the past, a model is developed with a hypothetical target individual and a hypothetical population in mind, and a “representative” source term is developed as input to the model to calculate projected exposures. In contrast, for a dose-reassessment study, exposure to actual individuals or population groups must be estimated by using documented releases and specific environmental transport conditions at the time of releases. The HEDR modelers have not always appreciated that distinction nor the fact that dose-assessment estimates for a nonexistent individual or population are meaningless and can mislead the public.
It is also important to stress that the dose-distribution estimates must be suitable as a basis for epidemiologic assessments and hence that appropriate demographic data should be developed in parallel. Accordingly, the committee recommends that the HEDR study focus more on actual doses received by “real” people and on the use of realistic, rather than worst-case, exposure pathways. The HEDR reports stop short of estimating the collective doses that have occurred—a number that will be required if population impacts are to be estimated. By concentrating on doses to individuals, the results of the reports imply that
each individual living near the site can, through his or her personal dose estimate, predict or project their chances of developing a cancer as a result of their exposure. It would be beneficial to provide estimates of the total number of excess cancers and cancer deaths that might be expected. Such an estimate of the “societal impact” would also have a benefit in terms of placing the risk estimates in perspective by contrasting them, for example, to similar estimates based on doses received from natural, background radiation.
