Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 148
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program 6 Expanding RECA Eligibility: Implementation The committee was charged with making recommendations to the Health Resources and Services Administration (HRSA) as to whether the Radiation Exposure Compensation Act (RECA) should cover additional geographic areas. This chapter discusses some of the issues related to the current RECA downwinder areas, which include only people living in specific counties near the Nevada Test Site (NTS). It confirms the need to include additional geographic areas, and it proposes a procedure based on risk and probability of causation/ assigned share (PC/AS) to determine which people would be eligible for compensation under RECA. Eligibility for compensation through RECA depends on many factors. The committee believes that compensation for loss of health caused by exposure to radiation resulting from the United States government’s nuclear-weapons program should be the first priority of any amendment to modify eligibility for compensation under RECA. Thus, we determined that epidemiologically based methods should be developed to identify additional populations and geographic areas that could be considered by Congress for inclusion in RECA. PROBABILITY OF CAUSATION/ASSIGNED SHARE Background An important health detriment after exposure to ionizing radiation is the increased risk of several types of cancers, referred to as radiogenic, that can be caused by radiation. Such radiogenic cancers can also be caused by other agents
OCR for page 149
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program or factors. No symptom or marker has been found that can identify a specific cancer as having been caused by radiation. That creates a problem when one is trying to establish eligibility for compensation for radiogenic cancer. A method using probability of causation/assigned share (PC/AS) has been developed to address this issue. PC/AS was introduced and discussed in general terms in Chapter 5. We consider here how PC/AS could be used in determining eligibility under RECA to avoid some of the problems described in Chapter 5. In Chapter 5, we noted several possible schemes for dealing with compensation that differ from the classical PC/AS or use a PC/AS with various adjustments with reference to the threshold for compensation and credibility interval. These include compensation based on years of life lost (YLL), compensation proportioned according to the posterior probability that PC/AS exceeds 0.5 or some other cutoff value, and compensation based on the values of the payment schedule weighted by the probability of having each value of PC/AS as determined with respect to the credibility distribution. The committee recognizes the merit of those alternatives and how they address shortcomings in the classical PC/AS approach. After considerable discussion, however, the committee crafted its recommendations in terms of solely PC/AS (including credibility interval). Two positive factors affecting the committee’s decision to adopt PC/AS are its widespread use in current compensation programs and the availability of user-friendly tools designed to implement the PC/AS approach. In addition, a compensation scheme based on YLL would neglect most individuals suffering from radiation-induced papillary thyroid cancer, the primary disease shown to be related to 131I dose, because of thyroid cancer’s small effect on longevity. A modified YLL method based on incidence rather than mortality, such as quality-adjusted years of life lost, was also considered and may be more reasonable. A compensation scheme based on YLL would be difficult to implement because the calculation of an individual’s YLL would be a very uncertain projection. Compensation proportional to the area of the upper tail of the PC/AS distribution and compensation weighted by the PC/AS distribution are also attractive alternatives. Such approaches fall within the bounds of the committee’s suggestion about PC/AS-based eligibility and merely adjust the amount that an individual would be awarded in compensation, which is not, in itself, a scientific decision. The committee notes that compensation schemes based on other approaches may ultimately be preferable if the infrastructure is developed to support them. The committee would endorse such approaches if they provided for a more equitable distribution of compensation than would be possible with a PC/AS system. In Chapter 1, we noted that the values of the threshold for compensation and the associated credibility interval result from societal decisions rather than scientific decisions. Later in this chapter, we recommend that Congress establish criteria for awarding compensation on the basis of computed distributions of PC/AS for any persons making such a claim. This chapter presents several examples
OCR for page 150
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program using different thresholds for compensation; these are only examples and are not meant to suggest any specific value for the threshold for compensation or the credibility interval. Use of PC/AS to Identify Geographic Areas for RECA Compensation RECA explicitly defines the geographic areas in which people must have lived to be eligible for compensation for a cancer that could have been caused by their exposure to radiation as downwinders relative to nuclear tests at the NTS. It specifies portions of Utah, Nevada, and Arizona as areas whose residents may be eligible for compensation. The committee considered whether other geographic areas should be added to the previously defined areas on the basis that residents had been at similar or higher risks from exposure to fallout from United States nuclear tests. We considered a range of possible expansions of the downwinder geographic areas currently included under RECA. In the next sections we will develop and present our recommendation that compensation be fundamentally based on a consistent PC/AS-based process rather than a priori solely on residence in specific geographic areas. We will propose that PC/AS criteria be used for both: 1) a preassessment to guide potential claimants and the implementing agency regarding which diseases, groups of individuals and geographic areas may satisfy RECA eligibility and 2) subsequently, a determination of individual claimant eligibility for RECA compensation. We then describe the underlying reasons that led the committee to this conclusion. We arrived at this conclusion, in part, on the basis of the results of the 1997 National Cancer Institute’s (NCI’s) iodine-131 (131I) report and the 2001 Centers for Disease Control and Prevention (CDC)-NCI draft feasibility study discussed in Chapter 4. The NCI 131I report shows that estimated doses and associated risks for radiation-induced thyroid cancer from ingestion of 131I depend not only on location but also on diet and age at exposure. Areas in which estimated thyroid doses were increased are not only near the NTS but also in such distant locations in the United States as the Midwest and upstate New York and Vermont (primarily the result of precipitation-borne fallout after test Simon). The distribution of those estimated doses is difficult to define geographically, especially because they also depend on diet and age at exposure. In addition, cancer risks from exposure to radiation depend not only on dose but also on other factors, including on the risk coefficient (see Chapter 5), which itself also depends on age at exposure. The risks to two individuals of different ages living in the same county would differ not only because their estimated doses may be different but also because their risk coefficients may be different. Similarly, the 2001 draft feasibility study by the CDC-NCI shows the importance of factors other than location in the estimate of doses that individuals may have received from other radionuclides in fallout. The associated radiation risks
OCR for page 151
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program of many radiogenic diseases also depend on such factors as age at exposure and age at diagnosis. Simple expansion of RECA eligibility based on location alone would not be equitable in that for some cancers it may fail to compensate higher-risk people in ineligible areas, such as those exposed to 131I in fallout as newborns, and may compensate lower-risk people in eligible areas, such as those exposed to fallout at an advanced age. It should also be recognized that, with the currently estimated levels of radiation and associated risk coefficients for cancer, it is unlikely that a very large number of individuals with cancer, even thyroid cancer, would be newly eligible for compensation. The actual number will depend on the threshold criteria established by Congress. EXPOSURE TO FALLOUT RADIATION Exposure of the Thyroid to NTS Fallout Radiation Chapter 5 showed the distribution of the estimated thyroid dose from 131I in fallout across all counties in Utah and showed which counties are compensated under RECA 2000 from NTS operations. As an example, the committee considered a male in each Utah County, born on January 1, 1948, residing in the same county during the nuclear-test period, and consuming an average amount of milk daily. That example showed that persons in one county (for example Joab County), who later developed thyroid cancer, would currently not receive compensation under RECA, whereas persons in other counties (such as Iron County) would receive compensation, even though the estimated doses in some of these other counties were substantially smaller (see Figure 5.1). This section discusses additional variations in a person’s risk of radiogenic thyroid cancer that the current RECA scheme does not take into account. Estimated thyroid doses vary not only between counties but also can vary considerably within a given county. The risk of thyroid cancer will vary even more markedly within a county than the thyroid dose, because the thyroid risk coefficient also varies with age at exposure and other factors. For the RECA-compensable cancers, the current PC/AS tables and calculators do not take coincident factors into account except for lung cancer (smoking). As noted earlier, the committee has chosen to use the current PC/AS approach because it is available and practical; it recognizes that improvements and refinements could be and perhaps should be made. The 1997 NCI 131I report contains an extensive series of maps showing all counties in the continental United States and the estimated thyroid doses that would be received by county residents born from January 1, 1930, to January 1, 1962, from 131I intake from NTS weapons tests given four milk-consumption scenarios (no, average, and high cow’s milk intake from store-purchased milk, and high milk intake from a backyard cow). The Institute of Medicine-National
OCR for page 152
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program Research Council committee that reviewed the NCI 131I report noted that the within-county variability in many cases exceeds the between-county variability. Age at exposure, sources of milk, and amount of milk ingested are particularly important risk factors expected to influence the occurrence of thyroid cancer in a person exposed to radiation. As an example, the present committee considered the thyroid doses from 131I for a person living in Custer County, Idaho, one of the counties identified in the 1997 NCI report as having a relatively high per capita thyroid dose from NTS operations. Although that county does not have the highest per capita thyroid dose (the county with the highest estimated per capita dose is Meagher County, Montana), it represents the general issue. Our committee noted that people who lived only in counties in Idaho or Montana during the nuclear-testing era are not eligible for compensation under RECA, even though some of these counties were among the ones that received the highest per capita thyroid dose. Figure 6.1 shows the best estimate and upper and lower 90% bounds of the thyroid doses from ingestion of 131I from all NTS tests for a male living in Custer County. The figure shows the estimated doses by year of birth and for average milk consumption (the upper 90% bound was over 1 Gy for birth years 1932-1953 and is not shown in the figure). We calculated the thyroid doses for male residents of this county with NCI’s on-line thyroid-dose calculator, assuming that a person was born on January 1 of each year (doses for persons born on other days of the year may be different) and resided in the same county during 1951-1971. For example, the 0.2 Gy estimated dose shown for a person born on January 1, 1953, would be the estimated dose received from the aboveground tests conducted from 1953 through 1962, when aboveground testing at the NTS ended. The estimated dose also includes a small component from in utero exposure to fallout from tests in 1952, which would affect individuals born in the early part of 1953. Residents of Custer County received most of their total estimated dose in 1952 from the Tumbler-Snapper series of tests that were conducted that year at the NTS, when the county was in the path of the fallout. As seen in the figure, the highest estimated dose for this set of birth dates occurred to people born on January 1, 1952, who were newborns during that test series. People born in 1948-1951 were infants and young children in 1952 and therefore had higher estimated thyroid doses than those born before 1948 or after 1952. Estimated dose decreases as the year of birth moves earlier from 1952; people born in 1930 (or before) would receive an estimated dose less than 10% that received by people born in 1952, even though they were residing in the same county. The estimated dose is also sharply lower for people born after 1952 because the major fallout events for this particular county occurred before their births. That illustrates further a problem in defining a geographic area for future RECA compensation: any such area would include people who received estimated doses that varied over a wide range, some of which would not be significant.
OCR for page 153
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program FIGURE 6.1 Thyroid dose (Gy) that the NCI dose calculator estimates a male resident of Custer County, Idaho, born on January 1, would have received from all NTS tests vs. year of birth for average milk consumption. The “step-like” structure of the graph is due to age-interval-based assumptions about variables, such as food consumption, and dose-conversion factors in dose calculator. The highest thyroid dose in this set of people was estimated by the NCI thyroid dose calculator to be 0.61 Gy, for a person born on January 1, 1952. The uncertainties in the dose estimates are large; the 90% uncertainty limits are 0.11 to 8.4 Gy. By comparison, most residents of the United States would receive a thyroid dose from background radiation of about 0.001 Gy per year (estimated from NCRP, 1987). In the roughly eight years covered by atmospheric nuclear testing, they would receive a cumulative dose of 0.008 Gy to the thyroid from background radiation. Several factors combine to increase the estimated 131I thyroid dose to infants and young children, including the relatively higher uptake of ingested iodine because of increased thyroid metabolism at those ages and the smaller size of the thyroid, which increased iodine concentration relative to older children and adults (NCI, 1997).
OCR for page 154
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program As noted earlier in this chapter, the variability of fallout-related cancer risks among residents of the same county reflects the variability of the estimated thyroid doses, which depend on age at exposure and diet, and also the variability of the thyroid cancer risk coefficients, which also depend on age at exposure. As discussed in Chapter 5, the risk of thyroid cancer posed by a given thyroid dose is highest in newborns and decreases until the risk is small after the age of 20 years. As an example, Figure 6.2 reproduces from Figure 6.1 the estimated thyroid doses that residents of Custer County received from fallout from atmospheric tests. Assuming that a person developed thyroid cancer that was diagnosed in 2000, PC/AS can be found for that person’s cancer depending on his or her estimated thyroid dose and age at exposure. Figure 6.2 shows these PC/AS values, and also the PC/AS compensation thresholds of 0.5 and 0.3 (two of the values used in other compensation schemes) and the effect on eligibility for compensation. The PC/AS curve in Figure 6.2 is the 50th percentile value of PC/ FIGURE 6.2 Estimated thyroid absorbed dose (Gy) from all NTS tests, and 50th percentile PC/AS, for a person living in Custer County, Idaho, born on January 1, by year of birth. Assumptions include average milk consumption, diagnosis of thyroid cancer in 2000, and exposures from NTS operations. The “step-like” structure of graph is due to age-interval-based assumptions about variables, such as food consumption, and dose-conversion factors in dose calculator. NOTE: PC/AS is not calculated for 1953 and 1958. Dose to individuals born on January 1 in those years included dose from the previous year because of in utero exposure. IREP is not currently designed to evaluate doses from in utero exposure.
OCR for page 155
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program AS by year of birth. (The 50th percentile value of the PC/AS value of 0.5 or 0.3 was chosen for convenience. Other percentile values, such as 99th percentile used in other compensation programs, can also be used.) The figure shows that PC/AS values for people born before the maximum dose period in 1952 decrease even more rapidly than dose with age at exposure. PC/AS and estimated thyroid dose are highest for a newborn and decrease with age at exposure. If, for example, the compensation criterion is that the diagnosed thyroid cancer has to be “as likely as not” due to fallout radiation exposure, PC/AS would have to be at least 0.5. The figure shows that if a PC/AS threshold of 0.5 were used for compensation, people who were born in 1946-1952 would be eligible for compensation. Others would not be eligible. The age-at-exposure dependence of the thyroid-cancer risk coefficient complicates further the use of geographic areas as the sole criteria for RECA compensation. Different groups of people may be eligible for compensation if other PC/AS thresholds for compensation are used. For example, if a scheme similar to that used by British Nuclear Fuels Ltd. (BNFL) were adopted, partial compensation would occur at PC/AS of 0.2, 0.3 (shown in Figure 6.2), and 0.4. The above scenario represented by Figure 6.2 uses consumption of purchased cow’s milk to illustrate how the general outline of the dose distribution within a county depends on age at exposure. Other scenarios—such as using goat’s milk, obtaining milk from a backyard cow, or consuming no fresh milk—are possible; they would give higher or lower estimated doses depending on which scenario is examined. Another perspective on the relation of age at exposure and PC/AS is shown in Figure 6.3. The committee used NCI’s Interactive RadioEpidemiological Program (IREP) version 5.3 to estimate the thyroid doses that gave a PC/AS of 0.5 at the 50th and 95th percentiles for geometric standard deviations (GSD) of the thyroid dose of 2 and 4 (the GSD is used because dose distributions are found to be skewed to the right and assumed to be lognormal). Those doses, which depend strongly on age at exposure, are shown in Figure 6.3. The curves for GSD equal to 2 and 4 for the 50th percentile coincide, because the 50th percentile value, or median, of the distribution is insensitive to the spread of the distribution given by the GSD. The figure shows that a person would have a PC/AS = 0.5 at the 50th percentile if the person received 0.15 Gy as a newborn, or 0.35 Gy as a 10-year old. If the person only received 0.15 Gy as a 10-year-old, the PC/AS value would not reach 0.5. For the 95th percentile, the GSD = 4 curve lies below the GSD = 2 curve. As discussed in Chapter 5 and illustrated in Figure 5.5D, the wider spread of the distribution for GSD = 4 causes the tail of the distribution to cross the 0.5 PC/AS value for lower values of the estimated dose than for GSD = 2. Consequently, as seen from the figure, the PC/AS of a 30-year old with an estimated thyroid dose of 0.15 Gy and GSD = 4 would exceed 0.5 for the 95th percentile, while that of a 30-year old with the same estimated dose but a GSD = 2 would not. An estimated
OCR for page 156
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program FIGURE 6.3. Estimated thyroid dose (Gy) for which PC/AS = 0.5 for different ages at exposure. Dose is given for 50th percentile and 95th percentile. The geometric standard deviation was assumed to be 2 and 4. Curves for the 50th percentile GSD = 2 and GSD = 4 coincide with each other. dose of about 0.26 Gy would be needed for the PC/AS of the second person to exceed 0.5. As an example of a possible PC/AS-based approach, to determine PC/AS for a particular person born in Custer County, Idaho, on January 1, 1950, the NCI dose calculator would be used to estimate the thyroid dose received from 131I each year. This illustrative use of the calculator yielded the doses shown in Table 6.1. The thyroid doses for the years not shown in the table are equal to zero. If the person had a diagnosed thyroid cancer, PC/AS could be calculated. The estimated doses are then entered into the NCI IREP program with the years of exposures, the date of birth, and the year of diagnosis to give the results presented in Table 6.2. The 50th percentile value of PC/AS is 0.67 (66.55%), indicating that it is more likely than not that the person’s cancer was due to exposure to 131I in fallout. If the threshold for compensation were set at the 95th or 99th percentile, rather than the 50th percentile, the person would of course also qualify for compensation. If one wants to ensure that people whose thyroid dose from NTS operations contributed substantially to the development of their thyroid cancer are properly compensated, a possible mechanism is to determine the estimated thyroid dose to a
OCR for page 157
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program TABLE 6.1 Estimated Annual Thyroid Doses Estimated Using the NCI Dose Calculator for a Person Born on January 1, 1950, and Living in Custer County, Idahoa Year of Exposure Estimated Thyroid Dose (Gy) Best Estimate Uncertainty Rangeb Low Estimate High Estimate 1951 0.0014 0.000093 0.028 1952 0.33 0.028 5.2 1953 0.0048 0.0011 0.028 1955 0.0024 0.00057 0.023 1957 0.03 0.0074 0.18 1962 0.000062 0.0000012 0.0037 1965 0.000018 0.00000019 0.0024 aDoses do not add to total dose shown in Figure 6.1; best estimates are medians of lognormal distributions and are not additive. bUncertainty range given by NCI dose calculator corresponds to “credibility intervals” used in this report. person who has a diagnosed thyroid cancer on the basis of the person’s date of birth, milk-consumption pattern, and place and time of residence. The estimated thyroid dose and age at exposure would be used to determine PC/AS. If a designated PC/AS percentile equals or exceeds a threshold for compensation (such as 0.5), the person would be compensated. That procedure could be used for any person in the United States, and not be limited to particular geographic areas. Exposure of Other Organs or Tissues to NTS Fallout Radiation The committee also examined PC/AS for radiogenic cancers in addition to thyroid cancer for people in areas receiving NTS fallout. We used NCI’s IREP version 5.3. For estimating doses, we relied on the information in the CDC-NCI draft feasibility study (CDC-NCI, 2001). The committee agrees with the authors of the draft feasibility study that the dose estimates would have to be defined TABLE 6.2 PC/AS for Person Who Received Thyroid Doses Shown in Table 6.1 and Had a Thyroid Cancer Diagnosed in 2000a Percentile PC/AS 1st 15.00% 5th 26.83% 50th 66.55% 95th 92.10% 99th 95.67% aFor example, the 50th percentile shows the PC/AS value (66.55%) that exceeds 50% of the calculated values of the PC/AS distribution.
OCR for page 158
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program better before they could be used for individuals, so we are using the information here only for illustration. If the PC/AS approach is adopted, the dose estimates for other organs and tissues, for example, red bone marrow, will need to be completed and finalized. An important difference between the thyroid and other cancer sites with respect to the radiation dose from fallout is that the thyroid is the only organ exposed to radiation from NTS tests for which the NTS-derived internal dose exceeds the external dose (CDC-NCI, 2001). For all other organs and tissues, the external dose from fallout is higher than the internal dose. The external dose is relatively uniformly distributed over the body. The most important organs for consideration are those most susceptible to cancer induction by evenly distributed whole body radiation. The committee specifically examined the estimated dose to red bone marrow because of its susceptibility to induction by radiation exposure of all types of leukemia except of chronic lymphocytic leukemia (CLL). The draft feasibility study concluded that the maximum total (internal + external) estimated dose from NTS fallout to red bone marrow was 3-10 mGy (CDC-NCI, 2001) (see Figures 4.3 and 4.4 of the present report). CDC-NCI noted, however, that differences in such variables as rainfall in a county might have resulted in higher estimated doses. Because of the importance of external radiation from fallout deposited on the ground, the time a person spent outdoors would also affect the estimated dose. In the draft feasibility study, the uncertainties in estimated dose were not thoroughly studied, but we considered a higher-upper bound dose to investigate the possible range of PC/AS for leukemia. We only considered median values of PC/AS in this example because the uncertainty intervals for estimated dose were not determined. To estimate the possible higher values of PC/AS for leukemia, the committee used the NCI’s IREP version 5.3 calculator to estimate PC/AS for the highest estimated dose to red bone marrow (10 mGy), as determined by the draft feasibility study, and to estimate an upper bound on the maximum estimated dose (40 mGy). For simplicity, the committee assumed that the entire dose occurred in one year (1952), although in actuality people would have incurred doses from the longer-lived radionuclides well beyond that date. Assuming that all the dose occurs in one year maximizes PC/AS because of its dependence on age at exposure. In contrast to radiation-induced thyroid cancer, the age at diagnosis of radiation-induced leukemia is an important factor in addition to age at exposure and dose. PC/AS values for a newborn and a 10-year-old are shown in Figure 6.4 for different years of diagnosis. As noted above, the PC/AS values were calculated for two different values of the estimated dose to the red bone marrow, 10 mGy and 40 mGy. Figures 4.3, 4.4, and 6.4 indicate that people who received the typical maximum estimated red bone marrow dose of 10 mGy would not have a PC/AS over 0.5. However, individual variations may exist within the average values used for the dose as-
OCR for page 161
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program The study included estimates of organ doses and cancer risks from those materials. As noted earlier in the case of dose estimates to organs and tissues other than thyroid from NTS fallout, the global-fallout dose estimates, although they have been released in a predecisional draft for peer review and public comment, are tentative because the report has not been finalized. In addition, as the authors of the draft feasibility study pointed out, the actual dose estimates themselves, even when they become finalized, are preliminary and support a feasibility study that shows that more detailed and accurate dose estimates are possible, but still must be developed (CDC-NCI, 2001). Considerable work would need to be performed to obtain dose estimates suitable for individual PC/AS calculations. For example, the authors of the feasibility study noted that only a “crude model was developed to describe the geographical variation in Sr-90 deposition density …” Determination of the 90Sr deposition density in each county across the United States, which was calculated from each county’s precipitation and its estimated 90Sr concentration, was essential in estimating the radiation dose to residents of those counties from most of the other fallout radionuclides. As a result, the study’s authors stated that “the specific county estimates or estimates for years prior to 1958… may be quite uncertain and should be used with discretion” (CDC-NCI 2001). They also had serious reservations about current dose estimates from the small amount of 131I that does return to the earth’s surface in global fallout and the estimated doses from global fallout of 3H and 14C due to their entry into the hydrological and carbon cycles, respectively. Once debris from a particular test enters the upper atmosphere, it mixes with debris from some of the other tests conducted by the US and other nations. Global fallout is a mixture of debris from all those tests. The United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR) has estimated the fraction of 90Sr deposition due to the nuclear testing programs of each country, including the United States, on the basis of their fission yields (UNSCEAR, 2000). The CDC-NCI draft feasibility study also indicated that the relative contribution of a country’s tests to global fallout could be roughly determined from the relative fission yield of its tests compared with that from tests conducted by other countries. In principle, then, the fraction of global fallout due to the US tests can be estimated, albeit roughly, from the fission yields of its tests. Because RECA was established to compensate people for harm incurred as a result of the US nuclear-weapons program, radiation doses and associated risks from exposure to the US fraction of global fallout should be included in RECA. However, the discussion above indicates that the information and methods needed to estimate individual doses and risks from global fallout are still in a preliminary stage, and a more detailed understanding of global fallout needs to be developed before they can be determined. Consequently, the committee will recommend in the next section that a detailed preassessment of fallout doses be performed. That includes determining and finalizing the global fallout doses, and identifying geographic areas in the US in which the estimated doses from fallout for people with
OCR for page 162
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program diagnosed RECA-specified cancer could result in a PC/AS value high enough to qualify for compensation on the basis of PC/AS criteria established by Congress. Recommendations for Expanding Eligibility for People Exposed to Radiation from Fallout The committee recommends that Congress establish a process using probability of causation/assigned share (PC/AS) to determine the eligibility of any new claim for compensation for a specified RECA-compensable disease in people who may have been exposed to radiation from fallout from US nuclear-weapons testing. The committee further recommends that Congress establish criteria for awarding compensation on the basis of computed distributions of PC/AS for any persons making such a claim. Prior to implementation of the revised compensation program, the National Cancer Institute (NCI) or other appropriate agencies should perform a population-based preassessment of all radiogenic diseases using PC/AS to provide guidance to individuals who might apply for compensation by determining the likelihood any individuals in a given population of being compensated. This analysis would be determined by disease identified, places of residence at the time of exposure, ages at the time of exposure and at diagnosis, and other demographic factors using the PC/AS criteria (including consideration of the upper credibility intervals) established by Congress. The calculation would use data for the maximal doses that such individuals may have received from fallout. In settings where variability is important in evaluating risk, there may be several such defined populations, and each would be evaluated on its own merits. The criteria for evaluating such population-based preassessments should be the same as those established by Congress for compensation of claims under RECA.1, 2 The preassessments should be made for the following two purposes: 1 Although agreeing with the general PC/AS and preassessment approach, committee member Kathleen N. Lohr wishes to emphasize one aspect of the committee’s suggested plans for the preassessment activities of particular concern to her. The acceptability of the PC/AS approach and the proposed equivalency of RECA compensation criteria and population-based preassessment criteria rests on the assumption that the preassessments will be done using values for variables in the calculations that will in fact mimic “worst-case” scenarios and produce the highest possible dose estimates for the population groups in question. In this way, possible bias against individuals with specific characteristics different from the majority of the population group under consideration can be minimized. 2 Although concurring in principle with the PC/AS approach and with the concept of population-based preassessment to enhance efficiency, committee member Stephen G. Pauker points out that if the criteria for evaluating population-based preassessments are quantitatively the same as the criteria for compensation, a bias may be introduced against some individuals whose PC/AS is substantially greater than the majority of the population. Rather, the criteria for evaluating preassessments might need to be quantitatively less strict than the compensation criteria to avoid such bias.
OCR for page 163
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program To provide guidance to potential claimants and the implementing agency as to which diseases may satisfy the compensation criteria established by Congress. To provide guidance to potential claimants and the implementing agency as to which population groups or geographic areas may satisfy the compensation criteria established by Congress. The recommendation applies to residents of the continental US, Alaska, Hawaii, and overseas US territories who have been diagnosed with one of the specified RECA-compensable diseases and who may have been exposed, including exposure in utero, to radiation from US nuclear-weapons testing fallout. Both Nevada Test Site fallout and the US fraction of global fallout should be considered. PC/AS for any individual should be obtained from an estimate of the radiation dose resulting from US nuclear-weapons testing and the risk estimate associated with such dose. Uncertainties in PC/AS cannot be avoided and may be part of the compensation decision process. Because of substantial gaps in the existing data, the uncertainties in estimated doses3 incurred by people exposed to radiation from fallout, and consequently the uncertainties in the associated PC/AS estimate, are large. This emphasizes the need to choose compensation criteria carefully. For example, a PC/AS value associated with a high percentile of uncertainty could exceed the criteria for compensation even for some very small median doses. The challenge Congress faces will be to decide if it is best to define criteria that avoid rewarding compensation in cases in which there is very low risk, but the uncertainty associated with its PC/AS is very large, because the connection of these cancers with radiation is not well established or the estimated doses are not well known. To support the use of the PC/AS process for compensation, The Centers for Disease Control and Prevention (CDC) and the NCI or other appropriate agencies should complete dose estimates for all significant radionuclides in fallout from US nuclear weapons testing to the population groups identified above. This should include all the major sources of dose related to US nuclear-weapons tests considered to have potential health consequences that the CDC-NCI 2001 draft feasibility study described. An updated dose calculator, similar to the existing NCI dose calculator for 131I, should be developed for determining dose to the thyroid and other important 3 The dose estimates depend on the measured deposition of radionuclides taken at the time of the nuclear weapons tests. Given the very small number of monitoring stations, most estimates represent interpolations over very large areas. Among the 3000 plus counties in the continental United States, fallout monitoring in areas other than a limited region in Nevada and its neighboring states occurred at never more than 95 stations through the years of aboveground US nuclear-weapons testing. See Chapter 5 and sections of this chapter for further discussion.
OCR for page 164
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program organs from fallout. Such an updated dose calculator should be directly coupled to a risk calculator similar to IREP Version 5.3 that can compute PC/AS and propagate uncertainties for establishing credibility intervals. NCI or other appropriate agencies should maintain and revise the parameters in the models or calculators for estimating PC/AS based on risk estimates recommended by the National Research Council Committee on Biological Effects of Ionizing Radiation, report number 7 (BEIR VII). Over time, the agency should update the PC/AS calculators with the latest risk parameters. IMPLEMENTATION AND ANTICIPATED IMPACT OF THE RECOMMENDATION TO EXPAND THE SCOPE OF RECA TO ADDITIONAL GEOGRAPHIC AREAS The recommendation would remove the a priori requirements based solely on geographic location (but would continue to require the presence of a RECA-specified disease) to establish eligibility for compensation. However, it is critically important to provide procedures to ensure that claims that are likely to be eligible for compensation could be processed efficiently and rapidly. Thus, before this modification to RECA is implemented, the NCI or other appropriate agencies is expected to perform population-based preassessments of PC/AS for different diseases, geographic areas and population groups in those areas. A population-based dose reconstruction is considered part of a detailed preassessment in this sense. The preassessments would provide information to potential claimants and the implementing agency as to whether a person might be eligible for RECA compensation under the new criteria that Congress establishes. For example, the previous discussion noted the strong dependence of thyroid-cancer risk on age at exposure. The preassessment may show that people diagnosed with thyroid cancer who were children in some areas of the United States when testing occurred have PC/AS values eligible for compensation but that persons with the same diagnosis who were adults in those same areas during the same time period do not. RECA implementation would then encourage claims from people who were living in those areas while they were children. Thus, the preassessment would provide information about conditions based on age at exposure that would encourage submission of claims that have a chance of success while discouraging other claims. That conclusion is especially true for diseases other than thyroid cancer. As discussed earlier, the CDC-NCI draft feasibility study found that estimated radiation doses to organs and tissues other than the thyroid were less than 10 mGy in a typical exposure scenario. If a person who received that dose developed a cancer other than thyroid cancer, the PC/AS value would be low no matter where the person lived at the time of exposure. Although the person’s specific eligibility for RECA compensation would depend on the criteria for compensation that Congress establishes, the radiation risks of almost all the nonthyroid cancers at
OCR for page 165
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program those low doses suggest that eligibility for RECA compensation for these cancers would be infrequent. For example, the previous discussion in this chapter showed that even for exposures of highly radiosensitive bone marrow, doses considerably greater than 10 mGy were needed for PC/AS values to satisfy the PC/AS criterion of 0.5 or above at the 50th percentile for induction of leukemia and only if the disease had been diagnosed within a few years after exposure. The preassessment, based on these criteria, would demonstrate that there could be large regions of the country where people could not have received doses (with their associated uncertainties, as may be applicable under the criteria) sufficient to be eligible for compensation. Preassessments would, therefore, greatly streamline implementation of the modified RECA program. The committee also evaluated several other cancer types. We used, as an example, a PC/AS of 0.5 at the 50th percentile. (Other criteria would lead to different results.) As noted in the paragraph above, bone marrow is highly radiosensitive, and a dose greater than 10 mGy is needed for the PC/AS to exceed 0.5 for leukemia. Because other organs and tissues are not as radiosensitive as bone marrow, that finding suggests that other organ doses must be considerably greater than 10 mGy for the PC/AS in those cases to exceed 0.5. If the criterion for compensation were a PC/AS of 0.5 at the 50th percentile, applying for compensation for the cancers that may have resulted from these other organ doses would be unwarranted. But, of course, if Congress established the criterion to be based on the upper limit of the 99th percentile credibility interval, compensation of some populations for some other radiogenic cancers might be possible. Population-based preassessments would greatly simplify implementation of the RECA program in those areas. The simplification would be essential for the PC/AS procedure to be useful for determining eligibility for RECA compensation. Without it, a person diagnosed with a RECA-compensable disease would not know whether he or she was eligible for compensation until his or her application was reviewed and the person’s PC/AS determined. The preassessment is intended to encourage potential claimants to submit only viable applications for compensation. This will improve the efficiency of the administrative process and reduce delays in awarding compensation to people who are eligible. Hence, preassessment information would be communicated in advance by the implementing agency to the affected public. The outcome of the preassessment could be presented in several ways. One could be a set of tables for each compensable disease, population group or geographic area that would be likely to satisfy the criteria for compensation. Another possibility would be a series of maps that show areas of eligibility, similar to those presented in the NCI report on 131I. It might also be useful to develop a web-based program, similar to the NCI dose calculator or IREP that would allow a potential claimant or agency staff member to enter variables from the claimant’s
OCR for page 166
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program history. The results would then indicate if an applicant should consider submitting a formal application. Uncertainties in Dose and PC/AS Estimates The committee notes an important difference between the precision of the dose determination of people exposed to radiation from fallout with the precision of dose determination in other compensation programs using PC/AS, such as the compensation programs that apply to military veterans and the Energy Employees Occupational Illness Compensation Program (EEOICPA). These programs award compensation if a person’s PC/AS exceeds 0.5 at the upper 99th percentile level. The NCI and CDC determined doses for the people exposed to fallout from dose-assessment models based on environmental measurements taken at the time of the tests. In some cases, the estimated doses were based only on atmospheric-transport models, which tracked a fallout cloud from the point of detonation to the location of the downwind resident and included no environmental measurements. Estimates of deposition of fallout based on these models have considerable uncertainty. In addition, even when measurements had been made, fallout deposition in large areas of the United States was estimated by interpolation among up to 95 fallout monitoring stations (Beck, 1980). Estimating fallout across the entire United States with such a small number of stations also involves large uncertainties. Dose estimates based on such methods are not as precise as those based on measurements more directly associated with the person. Data from measurements are routinely available for many of the claimants in the other compensation programs. In the EEOICPA program, for example, many workers wore personal dosimeters from which to estimate the external radiation dose and participated in bioassay programs to estimate the internal dose. People exposed to fallout rarely had similar personal dosimetry. As a result, the uncertainties in the estimated doses incurred by people exposed to radiation from fallout, and consequently the associated PC/AS estimates, are generally much larger than many of those for the other programs. Geometric standard deviations (GSD) of 4 are not unusual for some of the fallout dose estimates; thus, for example, an estimated dose of 10 mGy could range from 0.6 to 160 mGy at the 95th percentile level (i.e., 2 standard deviations). If Congress were to adopt compensation criteria for RECA that include some high percentile of the uncertainty, as in the other compensation programs, the PC/AS value at this high percentile value could exceed the criteria even for some very small estimated doses. That obviously could make the compensation program very large and difficult to administer. This emphasizes the importance of choosing compensation criteria carefully.
OCR for page 167
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program We have chosen an example to demonstrate the combined effects of uncertainty in the dose estimated and the selected criteria for compensation. It is meant to illustrate some consequences of implementing PC/AS, but should not be interpreted as an endorsement of any specific criterion for compensation. For this case, a person at age 25 is exposed to 131I resulting from fallout. The person is diagnosed with thyroid cancer at age 35. A dose assessment is made and yields a geometric mean estimated dose of 80 mGy (8 rads) that is distributed log normally. The distribution of PC/AS is determined using the methodology described in IREP and presented in Chapter 5. Figure 6.5 shows the results of this computation in terms of the value of PC/AS corresponding to the median (central value) of the distribution, as well as the 75th percentile, 95th percentile and 99th percentile, as a function of increasing uncertainty based on the geometric standard deviation, GSD, in the estimate of thyroid dose. The PC/AS associated with a true dose of 80 mGy is 0.06. This corresponds to the median value of PC/AS shown in Figure 6.5. The median value remains constant at 0.06 even when the GSD in the estimate of dose increases from 1.5 to 4.0. The 75th percentile for PC/AS increases with GSD, but remains below 0.2. The 95th percentile for PC/AS increases as the uncertainty of the estimated dose increases and eventually exceeds 0.5 when the FIGURE 6.5 Values of PC/AS based on the median, 75th percentile, 95th percentile, and 99th percentile of the distribution for PC/AS as a function of the uncertainty based on the geometric standard deviation for a distribution of dose having a geometric mean of 80 mGy (8 rads).
OCR for page 168
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program GSD reaches 4.0. The 99th percentile of PS/AS is even more sensitive to uncertainty in the estimate of dose; it exceeds 0.5 when the GSD reaches 2.0. If the latter case were selected as the criteria for compensation, many claims would satisfy the criteria because of uncertainty in estimated dose rather than the central estimate of dose. Other examples using fixed cut-off criteria, as well as sliding scales for compensation, were presented in Table 5.2 using the complete distributions of PC/AS shown in Figure 5.5. Uncertainties in PC/AS cannot be avoided and will be part of the compensation decision process. The challenge Congress faces will be to decide if it is best to define criteria that avoid rewarding compensation in cases in which there is very low risk, but the uncertainty associated with its PC/AS is very large, because the connection of these cancers with radiation is not well established or the estimated doses are not well known. URANIUM MINERS, MILLERS, AND ORE TRANSPORTERS IN OTHER GEOGRAPHIC AREAS The committee considered whether any circumstances warrant the extension of RECA to include workers who were employed in uranium mining and milling in geographic areas not now covered by RECA. The committee noted that RECA already covers uranium miners, millers, and ore transporters who worked in the RECA-designated uranium mining areas in 1942-1971, regardless of where they are now employed (see Figure 2.1). RECA has a provision that allows individual states not covered under the uranium mining and milling sections to apply for inclusion in RECA if uranium mining occurred in the state during the period January 1, 1942-December 31, 1971. The committee recommends that the provision which allows individual states not currently covered under Section 5 of the Radiation Exposure Compensation Act to apply for inclusion under RECA if uranium mining occurred in the state during the January 1, 1942 to December 31, 1971 period be expanded to include not only uranium mining but also uranium milling and ore transportation occurring during that period in support of the US nuclear-weapons program. Probability of Causation/Assigned Share for White and Navajo Uranium Miners The 2002 amendment to RECA added the duration-of-exposure option to the criteria for compensation. Uranium miners may now be compensated if they have accumulated 40 WLM or worked one year or more in uranium mines. The committee investigated how that decision affected the PC/AS for lung cancer
OCR for page 169
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program associated with exposure to radon daughters in the mines. The committee used the relative-risk model developed in BEIR VI (NRC, 1999), which included a factor for duration of exposure. Data from the National Institute for Occupational Safety and Health study of 4,102 underground uranium miners were used to obtain all the estimates that study included 3,347 white miners and 755 Navajo miners who worked in the mines during the period 1950 to about 1970 (Hornung et al., 1998; Roscoe, 1997). The BEIR VI model was applied to four strata defined by Navajos and whites, each divided by duration of mining (less than 1 year vs. 1 year or longer). The BEIR VI model provides different estimates of relative risk for lung cancer for five increasing durations of exposure. Relative risk appears to increase for a fixed level of WLM as duration of exposure increases. The model is similar to that developed specifically for the Colorado Plateau uranium miners (Hornung et al., 1998). Accordingly, miners who have worked one year or more are estimated to have higher relative risk per WLM than those working less than one year. Table 6.3 presents the results of the BEIR VI lung cancer model applied to the four strata described above. Results are provided for the median and the 5th, and 95th percentiles of the exposure distribution in each stratum. The exposure distributions are similar for Navajo and white miners. It is clear from examination of the table that most of the miners who worked more than one year had PC/AS greater than 0.50. The typical miner who worked less than one year (as defined by the median exposure) had PC/AS less than 0.50. However, about 10% of the short-duration miners had PC/AS over 0.50. By contrast, PC/AS for the RECA requirement of 40 WLM for compensation is 0.18 for short-duration miners and 0.70 for workers who accumulated 40 WLM in working more than five years underground. CONCLUSION The committee calculated thyroid doses and associated PC/AS values by using the on-line NCI dose calculator and IREP programs. The committee found that estimated thyroid doses and associated radiation risks and PC/AS values varied substantially both among different counties and among individuals in the same county. Age at exposure and diet were important factors affecting the within-county variability. Identification of RECA eligibility solely on the basis of geographic area would be scientifically problematic because a given area would contain residents having both relatively high and virtually nonexistent risks of radiation-induced thyroid cancer, depending on such factors as age at exposure. Consequently, the committee has recommended that compensation be based on diagnosis of thyroid cancer in an individual and the associated PC/AS, which takes those factors into account. The committee also recommended extending the PC/AS-based compensation system to other eligible cancers in downwinders. However, performing an
OCR for page 170
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program TABLE 6.3 PC/AS of Diagnosed Lung Cancer in White and Navajo Uranium Miners in Colorado Plateau White miners Duration: < 1 year N = 515 Exposure (WLM) RR PC/AS Median 46.5 1.26 0.20 5th percentile 2.1 1.01 0.01 95th percentile 262.4 2.44 0.59 Duration: ≥ 1 year N = 2,832 Exposure (WLM) RR PC/AS Median 563.5 11.40 0.91 5th percentile 67.9 3.8 0.74 95th percentile 3059.8 49.6 0.98 Navajo miners Duration: < 1 year N = 101 Exposure (WLM) RR PC/AS Median 43.7 1.24 0.19 5th percentile 2.1 1.01 0.01 95th percentile 271.0 2.49 0.60 Duration: ≥ 1 year N = 654 Exposure (WLM) RR PC/AS Median 513.9 10.64 0.91 5th percentile 66.3 3.79 0.74 95th percentile 2604.7 42.6 0.98 initial population-based PC/AS preassessment is essential if the approach is to be efficient and timely. The population-based preassessment would provide guidance concerning the extent of the problem and identify locations and population groups in which compensation would be viable. Estimated radiation doses to organs other than thyroid and the corresponding PC/AS values for the tissues at risk are substantially lower than those for the thyroid. Because of the low radiation doses estimated for organs other than the thyroid, the committee suggests that the preassessment would enable the implementing agency to focus on selected diseases, areas and population groups. The results of the preassessment, however, would depend greatly on the PC/AS criterion for compensation that is eventually adopted. We presented an example showing the importance of the PC/AS criterion, especially the influence of the percentile of the PC/AS distribution that is specified for awarding compensation. If Congress chooses a 95th or 99th percentile, compensation would be awarded for very low values of estimated
OCR for page 171
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program dose and PC/AS. This outcome is due to the large uncertainties in the dose estimates, which generally depend on an environmental measurement grid that had a limited number of fallout sampling stations to cover the majority of the United States. If the 50th percentile is chosen as the criterion, compensation would be based on the median value of the PC/AS distribution. As a consequence, the decision to award compensation would not be dominated by the large uncertainties in the measurements. The committee also reviewed the geographic areas and population groups that RECA now covers for uranium mining, milling, and ore transporting. The committee recommended that the RECA provision allowing states not currently covered for uranium mining to apply for coverage be expanded to include uranium milling and ore transporting as well. The committee evaluated the PC/AS values for uranium miners meeting current RECA eligibility requirements and found that the miners who worked for more than 1 year generally would have PC/ AS values greater than 0.5. The PC/AS values for miners having exposure of 40 WLM would depend heavily on the duration of the exposure.
Representative terms from entire chapter: