DOD Response to Gao Report Gao/Rced-86-15 Operation Crossroads: "personnel Radiation Exposures Should Be Improved."

Page 1

OFCUE OF THE SECRETARY OF DEFENSE TO : UNDER SECRETARY OF DEFENSE FOR RESEARCH 6 ENGINEERING THROUGH: DIRECTOR, PROGRAM CONTROL AND ADMINISTRATION CROSSROADS: Personnel Radiation Exposures Should PURPOSE: The attached memorandum (TAB A) transmits a proposed 0 o GAO prepared subject report for Senator Cranston, a member W SUBJECT: DOD Response t o GAO Report GAO/RCED- 86-15 “OPERATION of the Senate Veterans Affairs Committee o The GAO recommended that DoD adjust CROSSROADS o The DoD, in its response to the draft GAO report, stated participants exposure estimates non-concurrence with the GAO conclusions and recommendations that exposure estimates be adjusted. o In the final report GAO published, along with the DoD comments on the draft report, a rebuttal of DoD comments o This rebuttal d i d not present any new arguements and the DoD still non-concurs uith the conclusions and recommendations that exposure estimates be adjusted o The proposed response addresses each of the 34 comments in the GAO rebuttal in turn. It provides further detail to support and explain the DoD positions. separate letter to the Secretary of Defense, has requested that independent review of the DNA exposure estimates be conducted. Responses to these letters are currently being o Senator Cranston, in a letter to President Reagan and a RECOHMENDATION: That y o u sign the memorandum at TAB A. COORDINATION: Army, NJVY, DNA Prepared by p . Tunnle/OATSD(AE)/S7956/6 Feb 86

Page 3

THE UNDER SECRETARV OF DEFENSE MEMORANDUM FOR THE DEPUTY ASSISTANT INSPECTOR GENERAL FOR GAO SUBJECT: General Accounting Office (GAO) Final Report Radiation Exposure Estimates Should Be Improved", dated November 8, 1985, (GAO Code 301675) Attached is the proposed DoD response to the subject GAO report. The Department non-concurs with the conclusions and recommendations of the report.

Page 5

THE UNDER SECRETARY OF DEFENSE ler Genera- of the United Sta tates General Accounting Offi This letter is the Department of Defense (DoD) response to the General Accounting Office (GAD) Report. "OPERATION CROSSROADS-Personnel Radiation Exposure Estiaates Should Be Improved," dated November 1985 (GAO Code 301675) OSD Case 6797. non-concurrence with the conclusions and recommendations of the report. A detailed explanation of the DoD positions on each point uas provided in that response. also presented their rebuttal of the DoD position as a set of These comments have been evaluated and judged to have no The DoD, in its response to the draft report, stated its In the final report the GAO included the DoD positions and impact on the DoD positions. The attached DoD response, at attachment 1, addresses each comment in turn and presents supporting data for the DoD positions. The DoD has calculated radiation exposures in a responsible manner. A National Research Council (NRC) review of the methods used for determining radiation doses "found no evidence that the NTPR teams had been remiss in carrying out their mandate. If any bias exists in the estimates, it is a tendency to overestimate the most likely dose, especially for internal emitters or when the statistical procedure for assigning dose is used". The executive summary from tb NRC report is at attachment 2.

Page 7

The dosimetry study conducted by the National Bureau of physics groups. Some of these groups performed very we11 in the test and others had some difficulty. Eight of the fifteen groups in t h e 1953 study admitted that they lacked the necessary equipment, calibration sources, or experience to accurately report the doses. The exact methodology used by GAO to develop their uncertainty factor of +loo\ was not given in their report but a brief analysis of the data from the group with the best results in the NBS study gives an average of +17.21 with all of the errors being overstatements of the dose. Table 1. GROUP D Results

Page 8

GAO suggests that a f i l m badge recorded as 0.0 could have been an exposure of 0.1 rem. This conciusion is based on two points. First, is the incorrect GAO assumption that the minimum detection level for CROSSROADS f i l m was 0.05 rem and that error in this range was 1001. Second, GAO cites a 1953 study of 1 5 film badge processors in uhicb two processors reported a 0.1 rem exposure as 0.0 rem, and suggests that the f i l m badge processors in the study worked under better conditions than d i d the processors at CROSSROADS. The affidavit, Appendix 1, from Dr. Gerhard Dessauer, Chief of Photographic Dosimetry (Film Badges) at CROSSROADS states quite clearly that the minimum detection level for CROSSRAODS f i l m badges was 0.01, not 0.05 rem as GAo suggests. The GAO, in page 104 of their study, states that Dr. Dessauer wrote a report which defines the minimum detection level for the film as 0.05. However, GAO later admits on page 106 that this report i s really excerpts from a lecture given by I)r. Dessauer in 1947. GAO describes i t as a ”generic discussion of film badge dosimetry with SOW discussion of what transpired at CROSSROADS.” As Dr. Dessauer states unequivocally in the affidavit a t Appendix 1 , “hro thirds of the 10.000 file badges issued during July and August showed less than 10 m i l l i r (0.01 r e n j and were recorded as indicating zero exposure.l* The recently released National Research Council study, Review of Methods Used to Assinn Radi am to Ser US s~ personnel at Nuclear WeaDons Test& I a l s o concluded that the

Page 9

minimum detection levels for the film used at CROSSRAODS vas Approximately, 131 of a11 the CROSSROADS f i l m badges had readings of 0.01 rem to 0.04 rem. According to Dr. Dessauer, the uncertainty around exposures this low was 0.01 rem. As he states in his affidavit: "The reproducibility of readings for f i l m s exposed to the sane dose was better than 10 mill it (0.01). Hence, in the range of interest [around 100 milli r) the reproducibility was better than 10 percent. At insignificant doses such a5 20 milli r , i t would be 5 0 The degree of consistency of results when stated in terms of percent has been misleading. should be quoted in terms of the corresponding uncertainty of dose. In our case i t is 10 milli r, an insignificant amount at any dose level." At doses of 200 milli r,it would be 5 per&t. Lack of reproducibiity In theory, an exposure below 0.01 rea could ko undetected. An exposure below this level is biologically inconsequential and is comparable to the background radiation that everyone receives. G A O ' s suggestion of elevating these 0.0 rem exposures to 0.100 rem is unsound and without precedent. The health physics profession has always reported anything below the minimum detection level as 0.0 rem. In short, GAO is faulting DoD f o r following tbe long establisbed procedures used by the professional community.

Page 10

The 1953 study cited b y GAO shows two groups which reported a zero rem exposure when the actual exposure was 0 . 1 rem. One laboratory admitted that i t "does not have a very extensive service and had to perform special calibrations to arrive at the reported answers." The other group made no attempt to determine the energy of the radiation being estimated, which was the primary cause of the error. In short, there was a reason why each group made these mistakes and neither of these reasons applies to CROSSROADS. Dr. Dessauer's affidavit demonstrates that he had the proper calibration equipment, the best personnel in he rica, and could determine exposures as low as 0.01 rem gamma. Dr. Dessauer also strongly disagrees with the suggestion that he worked under harsh conditions. As he states: "The GAO report is full of conjecture and error. The part of the Report dealing with the Photometry Section is misleading particularly on two aspects: 1 . The report speaks of "harsh CROSSROADS Conditions ." The entire X-ray department of the air-conditioned hospital ship S.S. HAVEN was made available to m y Section. No interfering medical activities vere permitted as there was another hospital ship in the lagoon carrying out such normal functions. The space and equipment available was superb and was probably never approached before or after BIKINI by any other film badge processing laboratory".

Page 11

The groups in the NBS study of 1953 cited by GAO d i d have difficulty determining radiation exposures to mixtures of high and low energy radiation but the difficulty was a n artifact of the test. The calibration factor for f i l m exposed to high energy radiation is higher than the calibration factor for film exposed to low energy radiation. radiation exposure to a mixture of both higb and low energy radiation would be somewhere between the calibration factors for pure high or pure low energy radiation exposures and would depend on the ratio of high energy radiation t o low energy radiation. In the NBS test, s0.e of the films were exposed t o low energy radiation, some to higb energy radiation, and others to various mixtures of both h i g h and low energy radiation, but the groups were not given any data to identify the exposure conditions. This made i t difficult for them to decide what calibration factors to use and led to increased errors. The calibration factor for a Environmental damage to a film badge can occur that would not cause any error in reading the film badge. The 125 films cited by GAO are examples of this. The environmental damage t o most of these films was localized to small areas and did not occur over the whole film area. The dose measurements taken using these badges would not have been affected by the

Page 12

~t was possible to have environmental damage that affected large areas of the film and caused darkening to occur. This excess darkness that was not caused by radiation but b y environmental damage would have been interpreted as due to radiation and this would have resulted in an overestimation of the radiation dose. While i t is true that film badges subjecteo to high humidity for long periods of time (several months) can fade and cause an underestimation of the dose, this was not a factor at CROSSROADS since most badges were processed the same day that No claim was made that the accuracy of a film badge reading of 1.0 rem has an uncertainty of ZI. An example was given to show that for a given uncertainty, in this case . 0 2 0 rem, the percentage of error in a film badge reading of .020 rem would be 1 0 0 % while the percentage error in a 1.0 rem reading would only be 2%. This example was used to demonstrate that care must be used in discussing errors as a per cent of the dose, especially for small doses. To minimize the uncertainty caused by variations in the emulsion at CROSSROADS, films were exposed to a radiation source of known spectral distribution and intensity and their response was measured.

Page 13

response was measured. In a sworn statement, Appendix 1, Dr. Gerhard Dessauer stated: sources or high energy x-ray machines producing photons in the energy range of the expected BIKINI gamma rays." The energy spectrum a t CROSSROADS was known with sufficient detail for fila badge calibration. Dr. Gerhard Dessauer, Chief of the Photographic Dosimetry (Film Badgej Section a t CROSSROADS, in his affidavit, Appendix 1, states: ,!all devices at BIKINI had been calibrated with radium sources or high energy x-ray machines producing photons in the energy range of the expected BIKINI gama rays.I* Mr. Don Collins, the Chief of Radiation Detection (radiac) Instruments and Dosimeters at CROSSROADS, and Hr. Charles Could, the senior engineer and repairman f o r radiac devices at CROSSROADS described in a taped interview the radiation spectrum expected and encountered at CROSSROADS. They stated in the interview that any errors in the calibration of the equipment as a result of the spectrum would tend to result in an overestimated dose. Hr. Collins provided an affidavit, Appendix 2, in which he discussed the radiac

Page 14

The fact that film processing conditions at BIKINI and KWAJALEIN might have been similar is irrelevant. Environmental damage usually occurs when the f i l m badge is being worn not during processing and different types of work were conducted b y the personnel wearing the film badges at BIKINI and KWAJALEIN. The occurrence of environmental damage o n a film badge does not necessarily mean that measurement errors would be caused. In the case of most of the badges from KWAJALEIN, the damage would not have interfered with reading the badges. On August 7 , 1946 the dosimetry section did not consist of one enlisted man, with the capability of processing only 100 f i l m badges per day. Dr. Gerhard Dessauer, the Chief of Photographic Dosimetry (Film Badgelsection at CROSSROADS, states in his affidavit, Appendix 1, that: “The number of personnel a t my disposal was about twice what I could use effectively, hence only the more intelligent a n d responsible enlisted men were allowed to Moreover, from August 9 to August 27, the number of film badges processed never dropped below 100 per day except for Sundays when boarding activities were reduced. In fact on August 17, a total of 733 badges were processed.

Page 15

the day. As Dr. Dessauer states in his affidavit, Appendix 1, "Colonel Warren succeeded in obtaining the services of the country's leading experts in film badge design, manufacture, shipping, handling, and developing. All operations were performed under the most carefully controlled conditions. The operating crews in the Photometry (Film Badge) Section were trained by physicists and radiologists in academic positions, and by graduate students. Training took place daily from early June 1946 when the S.S. Haven sailed from the Oakland Navy Base, to the early part of July, when the first parties reentered the lagoon after ABLE. Thereafter, the Photometry activities occurred mostly at night, in supervised shifts. The results of the previous day's operation became available in the morning of the next day. People with reading over 100 milli r (0.1 rem) for the previous day were prevented from participating in the next mission(s)." We cannot verify that each film was checked in all eight locations as prescribed. The ledgers list averages for the areas of the film under the lead and in the corners. Some of the ledger entries also have the eight readings that constitute the averages for the areas under the lead and in the corners, but most simply have the two averages for the readings without noting the individual readings. Thus i t cannot be verified that eight readings were always used to calculate the averages,

Page 16

but most simply have the two averages for the readings without noting the individual readings. Thus i t cannot be Verified that eight readings were always used to calculate the averages, only that …