Click for next page ( 90


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



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 89
APPENDIX A A Review of the Dosimetry Data Available in the Nuclear Test Personnel Review Program This appendix contains a reprint of A Review of the Dosimetry Data Avail- able in the Nuclear Test Personnel Review Program: An Interim Letter Report of the Committee to Study the Mortality of Military Personnel Present at Atmos- pheric Tests of Nuclear Weapons to the Defense Nuclear Agency, which was delivered to the agency on 15 May 1995. 89

OCR for page 89

OCR for page 89
APPENDIXA hIr. D. Michael Schaeffer May 15, 1995 Page 2 Committee with broad expertise in biostatistics, uncerta~ntf analysis, environmental dosimetry dose assessment, and health physics. The Five-Series Comrriittee has carefully scrutinized Me information presented to it by the Working Group and has prepared this interim letter report to document its findings and conclusions. The objective of the Working Group was to determine whether doses that have been assigned lo the Atomic Veterans and entered into the database of the NTPR program can be used as a basis for dose-response analysis in the Five-Series Study. The Working Group reviewed dose-related fields contained in the NTPR database and the methods used to estimate doses. The review considered four criteria that, ideally, should be applied when generating dose data for an epidemiologic study: (1) consistency in the technical approach, (2) nondifferential methods of dose assignment, (3) quality assurance, and (4) application of uncertainty analysis. The Worlcing Group held discussions with NTPR staff who are involved in the dosimetry and reviewed printouts of the NTPR database, as well as files of specific veterans for whom individualized doses have been determined. The Working Group concluded that there has been a lack of consistency over time in NTPR dose estimation methods and, in particular, in the methods of assigning "high-sided" doses, that is, doses in which uncertainties are resolved in favor of assigning higher doses rather than lower doses. In some cases, because of Tic existing compensation program, procedures for assigning doses have been different for those who did and did not file a claim for a radiogenic cancer. Neither Me dose assignment methods nor the database itself are thoroughly documented. In addition, uncertainties have not been estimated in a consistent manner and do not incorporate all potential sources of variability inherent in the dosimetry. The Working Group, therefore, concludes that the NTPR dose data are not suitable for dose-response analysis. This conclusion is based on the fact that the NTPR dose data were based on incomplete records, were developed primarily for the purpose of ensuing appropriate follow-up for participating veterans and do not meet the particular standards needed for use in epidemiologic research. The Worldog Group believes, based on its review, that comprehensive dose reconstructions Inay be feasible for a limited subset of veterans who participated in Me above ground nuclear test program. If doses on this population are required for epidemiologic purposes, they should be recalculated according to the fundamental principles described in this report. OBJECTIVES OF 1~; DOSI1METRY WORKING GROUP The Medical Follow-up Agency KAFKA) of the IOM is underwing a 5-year study to evaluate the mortality cxpericocc of military personnel, the atomic veterans, who participated in at least one of Five~cries of atmospheric nuclear weapons tests during Me period 1951- 1957. The principal purpose of the study is to ascertain whether morality from lcuicemias, other cancers, or any other diseases has occurred at a higher rate among participants in 91

OCR for page 89
92 THE FIVE SERIES STUDY Mr. D. Michael Schaeffer May 15, 1995 Page 3 atmospheric nuclear weapon tests compared with a similar group of veterans who were not participants. For additional background on the study and the IOM committee overseeing it, see Appendix A. As initially planned, the study would utilize participant identification and radiation exposure data provided by the Defense Nuclear Agency (DNA) of the Department of Defense (DoD). In designing the study protocol, the committee recommended that the dosunetry assignments for the Five-Series personnel be evaluated for the purpose of this epidemiologic study. To conduct this review, the committee constituted a working group with expertise in dose reconstruction, environmental transport of radionuclides, uncertainty analysis, measurement and dosumetry techniques, general health physics, and statistics. The Dosimetry Working Group roster is provided as Appendix B. The Working Group recognizes that dosimetry data could yield valuable insights into a dose-response relationship for an epidemiologic analysis, if they were derived from specific information that characterizes the veteran's duties at the time of his participation in the weapons tests and if they were estimated in a consistent and well-documented manner. The Working Group also understands that the NTPR database was developed primarily for the purposes of responding to veterans' inquiries about radiation exposures and as a basis for providing appropriate follow-up and settling claims for compensation in accordance with federal regulations. Furthermore, it iS apparent that the NTPR database has evolved over a period of time (Appendix C) that has seen improvements in the state of the art of dosimetric methods, advances in uncertainty analysis, and further discovery and review of historical records. The objective of the Working Group, therefore, was to evaluate NTPR dosimetry with an eye toward its applicability in an epidemiologic study. The Working Group did so by reviewing relevant documentation; tracing the origin of several individual dose assignments; and comparing the methods used with those generally acceptable for epidemiologic analysis. PREVIOUS NRC REPORTS ON NTPR DOSE - Y The National Research Council (NRC) previously published two reports on dosimetry related to exposures of participants in atmospheric nuclear weapon tests. The first report, "Review of the Methods Used to Assign Radiation Doses to Service Personnel at Nuclear Weapons Tests,. (NRC, 1985a) advised the DNA on whether the methods used by N1TR to assign doses of radiation arc comprehensive and scientifically sound, and recommended improvements. The second report, Film Badge Dosimetry in Atmospheric Nuclear Tests,. (NRC, 1989) was an in-depth evaluation of film badge dos~metry practices used during the weapons testing period, recording and record-keeping of dosimetric data, and overall uncertainties associated with the film badge readings. Neither of these reports, however, judged the feasibility of using NTPR dose data as the basis for epidemiologic analysis.

OCR for page 89
APPENDIX A Mr. D. Michael Schaeffer May 15, 1995 Page 4 EVALUATION CRITERIA EMPLOYED BY 1~ WO~G GROW In its review, the Working Group considered four criteria that, ideally, should be met if dose estimates are to be useful in an epidemiologic study. Criteria used included (1) consistency in the technical approach, (2) nondifferential methods of dose assignment, (3) quality assurance, and (4) uncertainty analysis. A brief description of these criteria follows. Consistency in the Technical Approach Ideally, consistent methods should be applied to assigning doses to all subjects contained in the database. Algorithms used to estimate doses should be uniform from one subject lo the next. Assumptions made to permit calculations of dose in cases in which no physical dosimetry data are available should be applied uniformly among the study population. If there is a tendency to bias doses in either direction, the bias applied should also be incorporated into each dose assignment. This consistency is crucial to successful merging of individual dosimetry with individual effects data for use in an epidemiologic study. Nondifferential Methods of Dose Assignment Dose assignment methods should not differ for individuals who were made known to NTPR because they or their surrogate fiend a clarion for compensation versus those who did not. If veterans who developed leukemia and filed a claim, for example, were assigned doses by methods that systematically differed from those used to assign doses to veterans who have never filed a claim, then this could produce serious bias In e~ratuat~g the dose-response relationship between radiation exposure and leukemia mortality. Quality Assurance Ideally, there should be comprehensi~c documentation of both the methods used to determine doses and the individual dose assignments. Ash dose should be traceable and capable of being recalculated through donunentation. It is important to document methods as actually applied, such as the algorithms applied (for dosimetry and urine), pua~s used (and associated distributions), assumptions regarding scenarios of e~c, and default values included when data arc not available, as wail as medication of data espy. A carefill audit trail would allow any corrections that had been made over talc to the persons assigned 93

OCR for page 89
94 THE FIVE SERIES STUDY Mr. D. Michael Schaeffer May 15, 1995 Page 5 dose to be retraced. Mathematical models estimating doses should also be tested to compare predicted results to measured values. Uncertainty Analysis Significant advances have been made over the past decade in the analysis of uncertainty for environmental dose analysis. Current state-of-the-art environmental dosirnetry requires quantitatively deriving best dose estimates coupled with associated uncertainties. Uncertainties should account for all possible sources of bias, including modeling bias, parameter bias, and parameter variability. To be most useful for the Five-Series Study, unbiased best estimates of dose should be bounded by a range that indicates the degree of subjective confidence. REVIEW PROCESS The Working Group met twice, once on 12-13 April 1994 and once on 16-17 May 1994. Both meetings were held in Washington, D.C. Prior to its first meeting, the Working Group was provided with background material on the NTPR program (Appendix D). On 12 April, the Working Group received briefings by Mr. D. M. Schaeffer of DNA on the NTPR database and Dr. W. J. KIemm of Science Applications International Corporation (SAIC) on dose assessment methods. The Working Group devoted considerable time to understanding the methods used in assigning doses and building and maintaining the NTPR database. On 13 April, the Working Group made site visits to JAYCOR and SAIC to review pertinent records. This review included documentation of methodologies, dose assignment policies, records from the NTPR database, and files of individual participants. At its second meeting, the Working Group reviewed additional information and drafted its report for the Five Series Committee. Upon receiving the report, the hill committee carefully scrut~n~d the information presented to it by the Working Group and prepared this interim letter report to document its findungs and conclusions.

OCR for page 89
APPENDIXA 95 Mr. I). Michael Schaeffer May 15, 1995 Page 6 OBSERVATIONS The NTPR Database The NTPR database contains the following categories of information: personal identification and information related to claims, cause of death, etc.; records of correspondence sent by the DNA; dud assignments and intervals of participation for specific test series; dose assigrunents obtained from dosimetry records; dose assignments derived by reconstruction; and, total doses by test series and summed across all series combined. Individual fields contain codes that identify results, explanations, sources of information, and so on (Appendix E). The dose estimates entered Into the database may be derived from film badges, dose reconstructions, a combination of the two, or from other sources. The ideal dose estimates are those based on undamaged badges actually worn by the individual during that person's entire time of participation in atmospheric tests. Such instances, however, are rare for series conducted before 1955. In the more typical case, the film badge results do not account for the participant's complete exposure and doses are reconstructed, if a reconstruction has been requested. When the badge records are available, the results of the separate badge readings are itemized in the database, along with the corresponding dates and identifiers for each badge. The corresponding identifiers provide a link to the badges stored at Reynolds Electrical and Engineering Company (REECo) in Las Vegas, Nevada. Following this variable-length record, which accounts for the separate dose contributions, the estimated total gamma dose assigned to a person is given as the sum across badges catered for each series. In this way, We database allows entry of separate badge doses as well as a total dose for all external gamma dose contributions within each of the series in which that person has participated. When a dose reconstruction has been carried out, the reconstructed doses are stored in the same manner, accounting for interval-specific,and total doses. No uncertainty estimate is given and there is no data field for upper or lower bounds on the estimated total dose. Although data fields for the internal dose estimates exist in the NTPR database, most individuals do not have internal dose estimates entered into these fields. As described earlier, undividual records contained in the dose field are also accompanied by Wee fields explaining how the dose was derived (Appendix F): data soured, dose explanation, and dosimetry type. The dosimetry type field has 55 numeric codes for diffcrcat methods of dose assignment, including, for example, .~0-no dose assigned,.

OCR for page 89
96 THE FIVE SERIES STUDY Fir. D. Michael Schaeffer May 15, 1995 Page 7 "423-extremity TLD dose assigned by investigation," and "455-exposure assigned to cohort member based on more than another cohort film badge, wearer ur~own." In this field, there are also alphabetic codes that indicate the minimum detection level (MDL) for the specific type of film badge used. These possible MDLs range from 10 rnrem lo 120 rnrem. It is not clear whether this information on the MDL is ever provided. Ambiguities in interpreting the dose data may arise because of missing information in the NTPR database. Appendix G illustrates such an example by using a case presented to the Worlcing Group at its site visit to JAYCOR. In this case, four separate film badge readings are listed for the veteran's participation in Operation CASTLE. It is not clear, however, whether these readings were based on individual or cohort badges, since the dosimetry type field (T1) is left blank for all four badge readings. After reviewing the SAIC dose reconstruction report, however, it became clear that the first of Me four readings is identified as a cohort badge reading, whereas the remaining three readings are identified as being individual badges issued to the veteran. In cohort badging, film badges were often issued to representative personnel in units with common activities and equivalent relationships to the radiation environment. It is important to note, however, that when the dosirnetry type code corresponds to 34, "exposure assigned to cohort based on average of several film badges,. it is difficult to know exactly what dose assignment method was used. Sometimes, a dose was assigned Mat is equal to the mean of the cohort badges plus two standard deviations (97.5 percentile). This is in accord with the methods to be used for assigning radiation doses as described In the Federal Register (50 FR 42258 October 21, 1985). In other instances, the average dose was assigned. Because these instances are not identifiable in the database, it cannot be determined whether the former method was chosen or whether the average was actuady assigned. This ambiguity is further complicated by the fact that summing across a number of 97.5 percentile levels yields a total dose that is at an even higher percentile. Although the dose data have been updated over time, the history of dose estimate changes is not traceable in the NTPR database, since only the latest dose assigs~nes~t for each period of exposure is maintained, asked the reason for change is not specified. This is illustrated in AppendL'c G. It may be possible to trace the dose assignment history by using transaction record tapes maintains by REECo since 1988. Even if full transaction records were to exist, however, the policy decisions on which the changes were based may not be available. The NTPR database includes a field for on upadoslal code of the pardcspaot at the time of a test series. However, it appears that this field is usually vocal Even after a dose reconstruction that includes use of an ocalpatios)a1 title and grade, the corresponding code is not always entered in Tic database (see Appendices G and H). The NTPR database does not contain the veterans' military specialty codes for the period of participation in weapons tests, for subsequent periods of service, or at the time of discharge. Such information would be unportant for tracking potential occupational exposures to radiation subsequent to series exposure. For example, it is highly likely that veterans who were involved in radiological

OCR for page 89
APPENDLYA Mr. D. Michael Schaeffer May 15, 1995 Page 8 sciences during their service have continued to be involved as civilians in occupations related to nuclear power or radiological sciences. In those eases, subsequent radiation doses may outweigh the doses received during weapons tests. Film Badge Dosimetry The pnneipal source of film badge information is the REECo master file located in Ids Vegas. In reviewing the badge information, the Working Group found that correction factors recommended by the previous NRC committees had not been applied consistently to NIPR dose assignments. The two NRC eotsunittees, (1985a, 1989) noted that film badge readings were biased high for actual exposures. It was also noted that die deep dose equivalent in rem would be only 0.8 to 0.9 of die true exposure in roentgens (NRC, 1989). The latter NRC committee had recommended corrections for overall bias ranging from 1.1 to 1.4, Hat is the badge readings should be divided by these correction factors to obtain He best estimate. The DNA did not adopt these correction factors to modify film badge results. Instead, it has established policy that Film badge readings expressed in terms of roentgens (R) or its subunits shall be converted directly to dose equivalent in rem, i.e. 1.0 roentgen equals 1.0 rem. This factor allows an unequivocal traceability of film badge doses direedy to source records containing film badge readings. Conversion to deep-dose equivalent and the associated bias factor (1.3) shall not be applied (NIPR, 1992). In the ease of Operation REDWING, the NBC (1989) recommended no bias correction for environmentally damaged badges. By utilizing new findings on He REDWING badges, however, DNA concluded that Judging from recent dose re~onstruetions and film badge analysis by SAIC, it is evident that the NAS guidance is incomplete and appropriate action must be taken to portray REDWING doses more accurately (analysis of REDWING Film Badges," RARP/NTPR memorandum, 9 October 1992). Subsequent to dds policy decision, however, DNA decided Hat it will revise REDWING film badge doses only as required to support veterans applying for compensation from He Department of Veterans Affairs. The correction factors recommended by the NRC Committee (1989) do not account for additional biases introduced when an unbadged individual's dose is derived from cohort badge data. Originally, badging was not done in order to cshmate individual exposures for epidemiologic purposes, but rawer to verify that radiation safeq limits were not exceeded. During the GREENHOUSE and UPSHOT-KNOTHOLE series, and to some degree in the CASTLE series, cohort badging was oRen used to represent the entire unit or group. Dosuneters, however, also tended to be assigned to radiation monitors "d others who were expected to receive the highest doses. According to the DNA, the number of participants whose dose data were derived from film badges varied substantially for the Fivc~eries participants (see Table 1). The overall fraction of individual Five~erics participants with doses based on film badge data is 97

OCR for page 89
98 THE FIVE SERIES STUDY Mr. D. Michael Schaeffer May 15, 1995 Page 9 approximately one-half. This number of badged personnel, however, can be misleading. The primary reason is that it does not represent the number of participants who actually wore undamaged badges during their entire period of exposure, for whom good dosimetry data can be obtained. TABLE 1. Description of Approximate Numbers of Participants, Number of Personnel Issued Personal Dosimerers (film badges), and Dosimeter Correction Factors as Determmed by the NBC (1989). Test Series No. of No. of Badged Dosimeter Correction (B) Participants ' Participants ' and Uncertain (K) Factors b GREENHOUSE 7,723 (1951, Pacific) 2,317 (30%) B = 1.4; K = 2.0 UPSHOT-KNOTHOLE 17,062 2,282 (13%) B = 1.1; K = 1.5 (1953, Nevada) CASTLE 13,958 8,113 (58%) B = 1.3; K = 2.1 (1954, Pacific) REDWING (1956, Pacific) 13,540 11,044 (82%) B = 1.3; K = 1.5 PLUMBBOB (1957, 12,938 10,243 (79~) B = 1.3; K = 1.5 Nevada) Total 65,221 33,999 (52~) ~ NTPR data distributed to the Working Group by ONA on 12 April }994. b B = estimated bias correction (divide by this number to obtain the corrected dose). K = estimated geometric uncertainty factor, as recommended by NRC (1989). 9S ~e confidence limits on a single badge dose can be obtained by multiplying the badge reading, after correcting for bias, by (1/~ and K. The following reasons explain why these film badge data would not be a suitable subset of NTPR dosimetry information for use in epidemiologic studies: an undetermined number of these badges (REDWING) may have been environmentally damaged by high temperature, high humidity, water or light leaks (NRC, 1989);

OCR for page 89
APPENDIXA hIr. D. Michael Schaeffer May 15, 1995 Page 10 some of these individuals may have been assigned doses based upon cohort badging (CASTLE, GREENHOUSE, and UPSHOT-KNOTHOLE); some film badges lacked adequate identifying information to uniquely link them lo an individual (CASTLE); although an individual may have been badged in one series, he may have participated in other series for which he was not badged; in some instances, the dose is based solely on the individual's medical records, which listed film badge doses without clear explanation as to whether they covered the person's entire exposure period; and for certain series (GREENHOUSE), only one date (e.g., date of issue) was recorded for film badges, so that the actual interval of exposure is unknown. Reconstructed doses When film badge dose data were not available or were incomplete, or when there was reason to believe that these data did not adequately characterize the actual exposure, alternative approaches were sometimes used to estimate doses. All approaches commonly involve the investigation of individual or group activities and their relationship to the radiation environment. First, if it was apparent that personnel were not present in the radiation environment-that is, personnel were far distant from the nuclear test(s) and did not experience fallout or enter the fallout area-and had no other potential for exposure, then the assigned dose was zero. Second, if some members of a group had film badge readings and others who did not wear film badges had a common relationship with the radiation environment, NTPR used cohort badging to derive individual doses for unbadged personnel. Third, when sufficient badge readings or a common relationship to the radiation exposure did not exist, doses were sometimes reconstructed. Consistent application of these methods to assign doses did not begin, however, until after 1987, when the DNA consolidated the individual service databases. For example, doses to typical crew members on ships may have been reconstructed from radiation surrey measurements and assumptions of time spent topside and below decics (Thomas et al., 1982). In some instances, different approaches were taken. In the case of UPSHOT-KNOTHOLE, observers without badges were assigned the highest dose measured among the observers who wore badges, regardless of how long they were there. There was an apparent lack of consistency between the Army and Nary approaches in assigning reconstructed doses. When the DNA consolidated t.hc service NTPR teams, the Navy provided documentation of the lend of confidence associated with assigned doses, whereas the Army did not. The Working Group found no Air Force documentation on the rationale or confidence lc~cl for assigned doses. 99

OCR for page 89
142 1 o o ~0 D ~ ~IL . ~ ~ g ~ a ~ . ~ _ _.

OCR for page 89
143 . . . ~ C Z! 1' D D US U' ~ l ,[ W I I ~ ' I a t I ~ 1 1 ! ' ~8 a) - ~n Q C) Q it as .= U) 8 a: o n ~ n n ~ ~ n n n ~ n E O n nO 3 2 8 _ _ ~ N ~ 0 60 o o o o 3000 n ~O _ 0h ~ 8 8 8 , ~_ w Q Q ax: o o -

OCR for page 89
144 THE FIVE SERIES STUDY Mr. D. Michael Schaeffer May 15, 1995 Page H-3 H-2 Memorandum from SAIC to DNA-RAEM/NTPR Subject: Dose Reconstruction: 0perationCROSSROADS(1946), BUSTER-JANGLE, (1951), UPSHOT-KNOTHOLE (1953), CASTLE (1954) Date: 27 February 1994

OCR for page 89
APPENDIXA DNA-RAEM/NIPR (hi. Owais) Memorandum For Record Dose Reconstruction for Operations CROSSROADS (1946), BUSTER-JANGLE (1951), UPSHOT-KNOTHOLE (1953), CASTE (1954) Introduction: At the time of Operation CROSSROADS, (then) Motor Machinist's Mate 2nd Class was a crew member on the USS SYLVANIA (AKA 44). The Standard Engineering Dose applies, and is available from the data in References 9 and 10. Dunag BUSTER-JANGLE, UPSHOT-KNOTHOLE, and CASTLE, the veteran was a civilian working for For B-J and U-K, film badge data is available for all periods during which Tic veteran would have been exposed. For CASTLE, some film badge data is available, and dose reconstruction is used here for the reminder of He time. Operation CROSSROADS: The veteran was a Motor Machinist's Mate 2nd Class aboard the SYLVANIA, and is assigned Tic Standard Engineering Dose. From the methodology and data contained in References 9 and 10, for the SYLVANIA this dose is 0.723 rem. Operation BUSTER-.lANGLE. During Operadon BUS-~-IANGLE, the veteran was a civilian working for Few badge data for the veteran are available for the dates 26, 27, 28 November, and 1 December 1951. References 11, 12, and 13 show that was only involved at Shot UNCLE. Reference 14 indicates that personnel were not allowed into the display area before 1 Dee 1951, two days after the detonadon. Thus, it would appear Hat He available film badge data covers Al times Hat Tic veteran might have been exposed to radiation dunog Operation BUSTER-JANGLE. The first three film badge dates are for equipment SCt-Up prior to Shot UNCLE. The reason that there was non-zero exposure is that the UNO Few display area lay in the fallout field of the previous shot, Shot SUGAR The cu~nulatiYc total of Tic four badge readings Is 0.630 rem. Operation UPSHOT-KNOT}IOLE During Operation UPSHOT-KNOTHOLE, the veteran was t civilian worldog for Film badge data for the veteran are available for the dates 31 March and 12, 13, 14, 15, 16, 18, 25, 27, 29 May and 1 June l9S3. E'cpenence has shown that the UPSHOT-KNOTlIOI"E film badge data is reasonably complete, ant that the film badge records of any one individual usually covers the cadre dose that the individual received during the whole of UPSHOT-KNOTHOLE. Aftcrexa~nining the veteran film badge records, along with those of the other personnel involved with UPSHOT- KNO~IOLE, it appears that those badges dated 1 June were actually worn during Tic recovery operations following Shot CLrMAX on 4 June. Among the personnel film badge records, there is no record 145

OCR for page 89
146 THE FIVE SERIES STUDY Page 2 dated 4 June, when substantial recovery activity took place folIov~ng the CLIMAX detonation. All personnel, on the other hand, have records date 1 June, and all are in the range 255 to 285 mR. Reconstruction shows this to be about what they would have received on 4 June, and there is no possibility of anyone receiving this dose on 1 June. The total film badge dose of the veteran for U-K is 1.370 rem. the veteran accrued an additional 0.060 rem in November 1953. ()peration CASTING The attached table s =es the locations and doses used for the veteran dose reconstruction. There arc fit badge records for the veteran for most of the period in question. For the rounder of the time, the use of she Rcfcrencc 3 wformadon and reasonable assumptions is considered to lead to reliable results. For instance, Reference 3 could not establish whether or not the veteran remained on the IJSNS AINSWORTH (TAP 181) during the period of March 2 - 4, when the ship was anchored at Enewctak Atoll. Since the veteran was a civilian, not a crew member, one might normally assume that he Ad the other civilians spent those days ashore at Eneweta~ R. - ation safety is an additional consideration in support of this assumption. Comparing Reference 5, pages 80 and 119 shows that the radiation intensity aboard the AINSWORTH was higher than that on Enewetak by roughly a factor of 3 or so. Ibus, the assumption that He civilians stayed ashore during this period is reasonable. Rcfcrencc 3 shows the A~SWORTH skiing at Bikini Atoll at 0900 on 5 Mas 54. Rcfercncc 6 lists one boat mission (LCU #638) with Project 3.2 people for Hat date, but the names listed do not include He Deter. The radiation levels at Bikini locations Nan and Tare for that day a`,erage about 1 and 0.15 R/hr, respectively (Reference 7, p.76). These intensities arc much higher that the intensity on the AINSWORTH (Rcferenec 5, p.80), so that red-safe considerations would have dictated that personnel stay aboard ship except at times when necessary dunes dictated otherwise. Thus, when neither film badge records nor any other indication of the veteran' presence elsewhere can be found, it is assumed Hat he remained aboard ship at Bikini. Refcrencc 6 shows him on a helicopter on March 7, a day for which a film-badec reading is available. For March 11, Rcfercsce 6 shows the veteran and Coworkers on a helicopter mission. There is no badge reading assailable for the veteran for that day, but there is one for [coworicer3 (Refcrcoce 3). Therefore, [cowor~cer]'s badge reading is used. Rcfercacc 3 has the prc-KOON and prc-UNION surreys taking place prior to 1 April. In case these acti~vides did not fall on days covered by the VCtC~' known fit badge mc~ a septets day for each pre shot survey was assumed. The fallout field plot in Reference 7, p. 99, was used as a basis for derermilung the radiation intensity to which the veteran was exposed. A time of about 2 hours was assumed for He rime duration of each survey. The period of May 6 ~ 14 is without documentation. The assumption Hat the veteran spent this dmc at Enewetak is not critical, sines he would have recopied about the same dose had he remained aboard He A~SWORTH.

OCR for page 89
APPENDIX A Page 3 Dose Summary: Neutron: < 0.001 rem Gamma lime ocuod 31 Mar - 12 May - 13 May - 14 May - 15 May - 16 May - 18 May -25 May - 27 May 53 -29 May53 - 01 Jun 53 - 03 Nov 53 - 04 Nov 53 - llFebS4 1 Memamndu~ ~ Dose Reconstruction. . i i- Hi. . ~. Dosc(rem) Moths 01Jul 46- 07Dec46 0.723 26 Non 51 - 26 Nov 51 0.060 27 Nor 51 - 21 Nov 51 0.120 28 Nov 51 - 28 Nov 51 0.080 OlDec Sl-OlDec 51 0.370 31 Mar 53 - 31 Mar 53 0.020 12 May 53 - 12 May 53 0.240 13 May 53 - 13 May 53 0.000 14 May 53 - 14 May 53 0.2S5 15 May 53 - 15 May 53 0.190 16MayS3- 16MayS3 0.130 18 May 53 - 18 May 53 0.130 25 May 53 - 25 May 53 0.020 27 May 53 - 27 May 53 0.130 29 May 53 - 29 May 53 0.000 01 Jun 53- 01 Jun 53 0.255 03 Nov 53 - 03 Nov 53 0.020 04 Nov 53 - 04 Nov 53 0.040 11 Feb 54- 11FebS4 0.000 01 Mar 54 - 05 Mar 54 0.349 06 Mar 54 - 06 Mar S4 0.120 06 Alar 54 - 07 Mar 54 0.260 07 Mar 54- 08 Mar S4 test) 0.160 09 Mar 54 - 09 Mar 54 0.000 09 Mar 54 - 11 Mar 54 0.049 lOMar 54- lOMarS4 0.110 ~_~ 11 Mar 54- 11 Mar54 0.180 12Mar 54- 16MarS4 0.060 17 Mar 54 - 27 hIas 54 0.180 28Mar54- 31MarS4 0.509 30Apr54-OlMay54 0.150 01 May 54 - 05 May S4 O.OSO _ 06 May 54 - 14 May 54 0.034 Reconstrucdon Total = 5.0 rem (upper bound 6.7) Reference' Reconstrucdon Few Badge (651 F) Film Badge (203 G) Film Badgc (527 H) Film Badge (1189 n him Badgc (000017) him Badge (014318) film Badgc (01264 film Badge (029310) film Badge (013805) film Badge (013693) film Badge (012668) film B - ge (015985) Film Badge (016173) film Badge (016853) film Badgc (017070) film Badge (018329) film Badge (018014) him Badge (00485) Rccons~cton him Badgc (06639) him Badge (08076) Film Badgc (06795) film Badgc (09296) Rcconstrucdon Elm Badgc (09824) Elm Badgc (09292)(msh [cowo~erl) Film Badgc (17810) film Badgc (18951) Recons~on Elm Badgc C3?469) Flute Badge (3821O " I. Kcsselman, JAYCOR, 06 July 1993. 2. NUCLEAR JEST PERSONNEL REVIEW, Telephone Infom~adon Few, . 27 January 92. Be "Chronology of she Vcsc~an's Movements," unsigned but apparently Mom JAYCOR, 16 July 92. 4. "Operation CASTLE, Prioicot 3.2, Crater Survey," HQ Weld Command, AFSWC. Sandia Base, Albuquerque NM, WT-920 (ED), June l9SS. 147

OCR for page 89
148 THE FIVE SERIES STUDY Page 4 5. '`Analysis of Radiation Exposure for Naval Personnel at Operation CASTLE," DNA TR-84-6 Defense Nuclear Agency, 28 February 1984. . 6. "Bikini Daily Dianes," Operation CAST, 1 March 19S4 - 14 May 1954. 7. "Dis~nbudon Ed Intensity of Fallout," Project 2.5a, Operation CAST AL Steton, et al, WT-915, January 1956. 8. Letter lo the veteran, 98113FD/2008, from W. H. I~oeffler, Captain USN, Navy Nuclear Test Personnel Review, 17 February 1984. 9. "Analysis of Radiation Exposure for Naval Units of Operation Crossroads, Volume I Basic Repon," DNA-TR-82~5V1, Defense Nuclear Agency, 3 March 198~ 10. "Analysis of Radiation Exposure for Naval Units of Operation Crossroads Volume m (Appendix 8) Support Ships," DN\TR-82~5-V3, Defense Nuclear Agency, 3 March 198i 11. "Operation BUS1ER-JANGLE, 1951," Defense Nuclear Agency, DNA 6023F, 21 Junc 1982. 1~ "Shots ABIE to EASY, Tic First Five Shots of die BUS1ER-JANGLF Senes," DNA 6024F, Defense Nuclear Agency, 22 Junc 1982. 13. "Shore SUGAR and USA F. The Final Tests of the BUSTER-JANGLE Scnes," DNA 6025F, Defense Nuclear Agency, 23 June 1982. 14. "Analysis of Radiation Exposure for Military Participants, Exercises Desert Rock I, I:E, & m, Operation BUSTER-JANGlE'" DNA-TR-87-116, DcScasc Nuclear Agency, 22 December 1987. IS. "Analysis of Radiation Exposure for troop Observers, Exacise Desert Rock V, Operation UPSHOT- KNOTHOLE," DNA 5742F, Defense Nuclear Agency, 28 April 1981. 16. "Operation UPSHOT-KNOlHOI~, 1953," DNA 6014F, t 1 January 1982. 17. "Shots ANNE to RAY The Hat F~vc Tests of the UPSHOT-KNOTHOLE Scnes, 17 March - 11 April 1953," DNA 60171:, lit January 1982. 18. "Shot BADGER, A Test of the UPSHOT-KNOTHOI-E Series, 18 April 1953," DNA 6015P, 12 January 1982. 19. "Shot SIMON, A Test of tic UPSHOT-KNOTHOLE Senes, 2S April 19S3," DNA 6016P, 13 January 1982. 2Q "Shots ENCORE to CLIMAX The Final Four Tests of the UPSHOT-=OTHOLE Senes, 8 May - 4 June 1953," DNA 6018F, 15 January 1982. 21. "Operation UPSHOT-KNOTlIOLE, Radiological Safes Opmadon," Tom. D. CotLison, AFSWP, WT-702 (REF.), June l9S3.

OCR for page 89
z c - ~ A c - c, c c`: - Cs3 - cl: ~ ~3 c 'E ~ ''3~\ ~ ~X 4,) == ~ . ~ "X ~X== en, . ~$. m. ~ ~ ~I. l 4, ~ ;~ - G ~m I ~;~ ^ - arcs mm- ~lo- ~-~ it} o ~ 8 ~ 8 8 ~ ~ ~ ~ ~ ~ ~ t~ _ _ _ _____ _____ ___ o~o~o*~= c~= Orion == o. 5855 5855 55 ~ ~ ~ Is ~ ~ ~ ~ ~ ~ ~ =~=mmmm~mmmm=C~Xo:mmc: x '' V`_~0880~0080~0000o-- I_ O ~ ~ ~ ~ ~ ~-~ ~ ~ ~ ~ ~ , ~ 11 ~ o3BYB3~88_39~. t\d, ~ 1~iI]~]It]I]'I~ ~ ~6 149

OCR for page 89
150 THE FIVE SERIES STUDY Mr. D. Michael Schaeffer May 15, 1995 Page I-1 Potential for Bias due to Differential Methods of Dose Assignment Comparison of Doses Assigned to Atomic Test Paruc~pants Who Had Individual Dose Reconstructions and Those Who Did Not The Dosimetry Working Group of the Committee to Study the Mortality of Military Personnel Present at Atmospheric Tests of Nuclear Weapons (Five-Series Study) attempted to quantify the effect of differences in dose assignment methodologies between those who had an individual dose reconstruction and those who did not. It is impossible to do this directly, since exact individual doses are unknown. Adopting an indirect approach, the Working Group first assembled dose data on Five-Series participants who had individualized reconstructions. who: The Working Group used, as a benchmark, the doses that had been entered for these individuals in the original Five-Series Study (NRC 1985). Because these benchmark doses could have a tendency to increase or decrease over time within the Five-Series cohort as a whole, a comparison group was also assembled. Changes in dose assignments for the comparison group presumably reflect the updating and cleaning activities that have been carried out by the DNA and its contractors since the creation of the 1985 NRC analysis file. The Working Group identified 277 participants (members of the Five-Series cohort) had a dose in the 1985 data file, had been referred to SAIC for dosunetry, and could be matched to a person in the current (May 1994) data file. For comparison, the Worldag Group randomly selected 415 participants who had an entry in both 1985 and current data files, but whose dose had not been referred for Individualized reconstruction. In this set of individualized reconstructions and controls, a robber of individuals were fouled to have missing dose values. Those individuals were eliminated from further consideration, leaving l9S individualized reconstruction cases and 269 nonindividualized reconstn~ction cases. To he included in this comparison, an individual had to hare nonmissing external gamma dose data for each of the five test series in which he participated. These nonm~ssing components were sunned to give a total gamma dose. Doses from tests other than those included in the Fi~c~erics Study were not considered, because those data were not available in the 1985 data set. The differences between the paired dose entries (1985 data matched to Tic current data) arc summarized in Table 2. The currently assigned doses tend to be higher than their 198S

OCR for page 89
APPENDIX A SIr. D. Michael Schaeffer May 15, 1995 Page I-2 counterparts, and the tendency is highly statistically significant within each group (p < .0001, sign test). Also, the pattern of the changes is different between the individualized vs. nonindividualized doses, that Is the nvo categorized distributions of paired differences are dissimilar (%2(3) > 70, p ~ .0001). Despite the highly significant difference between the dose changes experienced by the individualized vs. nonindividualized groups, one can see Mat the primary source of the discrepancy is the relatively greater tendency for the non~ndividualized dose estimates to stay the same. If one repeats the comparison, removing participants whose doses stayed the same, thereisno difference atall(z2(2) = l.9,p > .3)inthepatternof the charges. Thus,these data suggest that when the dose was changed by the individualized reconstruction methods, there was not a tendency for the change to be greater or less than that for a participant whose dose also changed because of alterations in the unit-based assignments. Table 2. Numbers of Participants Categorized by the Difference, AD = (l)c Do), Between Their Current Dose (Dc) and Original Dose Assigned in 1985 (DO) for Each Dose Assignment Method (individualized and nonindividualized) 7~; ~Dose Increased, Dose Increased, Method AD < 0 mrem AD = 0 mrem 0 < AD < 1,001 AD ~ 1,000 mrem mrem . Individually 12 60 88 35 reconstructed (6 DO) (31 %) (45 O (18 O Not 7 191 44 27 individually (3 %) (71 %) (16 %) (10 %) reconstructed This evaluation was not an exhaustive study of potential biases in the NTPR dose database. In fact, there arc several notable caveats to consider. First, the NIPR database is constantly being updated. This study used dose data from the Fivc-Serics Study frozen at two moments in time, separated by about 10 years. Clean up of the current dose data for the Five- Senes Study is still in progress and will not be completed until 1996. The results presented here should not be interpreted as a demonstration that there are no systematic differences htwecn individualized and nonindividualized dose assignments. We do not know the true dose for any participant, and thus have no direct means to assess bias. Moreover, participants were to some extent self-selected to have an individualized reconstruction, and their real doses may be higher or lower on average than those for the remainder of the cohort. Also, doses assigned to the same participants for series other than Me five might have revealed a very different pattern, as the dose data for over series have not been cleaned as completely. These other series doses would be relevant to computing a tote Baja dose for analysis of potential health effects. 151

OCR for page 89
152 THE FIVE SERIES SlrUDY Mr. D. Michael Schaeffer May 15, 1995 Page I-3 The process of generating these data revealed some perplexing results for individuals in the database. These included significant differences between assigned doses in the current (May 1994) data file and corresponding SAIC dose reconstructions. There were numerous instances in which the database does not seem to reflect SAIC reconstructions completed several years ago. In some cases, the current dose data are missing despite the existence of individualized reconstructions. As noted above, these could be the result of Me incomplete cleanup of the dose values in the current data set or there may be other reasons for the disparities. The Working Group attempted to quantify possible bias in the FiYe-Series dose data that could have resulted from different treatment of those who had individualized reconstructions and those who did not. This was done by comparing the dose entered in the data file for the 1985 Five Series Study (NRC, 1985b) with that in the current data file, for all participants whose name had been referred to SAIC for an individualized dose reconstruction and for a comparison group of participants who had not had an individualized reconstruction. Participants with individualized reconstructions were more likely than those without individualized reconstructions to have had their doses modified between 1985 and 1994. In both groups, when there was a change in the assigned dose, it was much more likely to be an increase than a decrease. When individuals whose doses did not change over time were eliminated from consideration, no differences were found in the pattern of changes in dose between the individualized and nonm.dividualized groups. That is, between these two groups, no difference was detected in the proportion of participants who fell into categories defined by whether their doses went down, went up by a little (up to 1 rem), or went up by a lot (more than 1 rem). Thus, for those individuals whose dose did change, it did not appear to make a difference whether they received an individual reconstruction or not. This suggests that the individualized dose reconstruction methodology was not systematically biased relative to the generic "cleanings of the data. These findings do not, however, allay one of the Worldag Group's most serious concerns over differential dose assignment-that individualized doses could have experienced a significantly different pattern of change than nonindividualized doses. For those individuals whose doses did change, it did not appear to make a difference whether they received an individual reconstmction or not. Nevertheless, the patterns of change shown in the above table of differences (Table 2) do suggest that those with individualized reconstructions will tend to have higher assigned doses than those without individualized reconstructions. This could bias any dose-response analysis based on the existing data.