Click for next page ( 20


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 19
4 Data Sources The assembled information for this epidemiologic study comes from more than 100 distinct sources. Handwritten paper logs, microfilm or microfiche, computer files, medical records, work orders, transport orders, memoirs, interof- fice memoranda, testimony, secondary compilations of primary sources, letters from spouses, death certificates, film badge records, computer programs, and benefits and compensation claims represent a diverse sample. In this chapter, we describe the sources of data and their general limitations and assets. These data formed the basis of efforts to (1) identify individual members of the two study cohorts, (2) ensure the comparability of these cohorts, (3) ascertain vital status and mortality information, and (4) compare the mortal- ity experience of those cohorts while controlling for characteristics of individu- als, military service, or time period that might influence mortality. Study staff, as well as DoD staff and contractors, made strenuous attempts to identify the existence of any relevant records, to acquire these records, and to cor- roborate information using multiple sources. Data related to personnel movements, radiation exposure, and vital status proved to be dispersed across the nation in cartons, computers, and file cabinets under the authority of many federal, state, and local agencies. The following federal agencies and facilities maintain collections that the study staff used: the Department of Defense, including the Navy, Army, Air Force, and Marines, and the Nuclear Test Personnel Review Program of the De- fense Threat Reduction Agency (DTRA); the Department of Veterans Affairs The organizational locus of the Nuclear Test Personnel Review Program within the Department of Defense has been the Defense Nuclear Agency (until June 1996), the De- fense Special Weapons Agency (until October 1998), and, currently, the Defense Threat Reduction Agency. 19

OCR for page 19
20 THE FIVE SERIES STUDY (VA), including its benefit and health components; the Department of Health and Human Services, including the National Center for Health Statistics, which main- tains the National Death Index; and the National Archives and Records Admini- stration's National Personnel Records Center, regional records centers, and the National Archives. Table 4-1 displays the relationships between the various sources and the data elements they yielded. COHORT IDENTIFICATION Nuclear Test Personnel Review Program The source of information on participant identification and radiation expo- sure is the database maintained by DTRA the NTPR database. The nature of this database and its implications for the design of the present study are de- scribed in this section. In 1978, shortly after the Defense Nuclear Agency (now DTRA) became the executive agency for matters pertaining to the participation of DoD personnel in atmospheric nuclear tests, it officially established the NTPR program. The pri- mary purposes of the NTPR were threefold: (1) to identify DoD personnel pres- ent at each test site and estimate their radiation exposures; (2) to identify the radiation monitoring measures that were in effect at the time of the tests; and (3) to develop a history of every atmospheric nuclear event that involved DoD per- sonnel (Johnson et al., 1986~. Initially, DNA directed the individual military services to conduct the NTPR research pertinent to their respective services, but in 1987 it consolidated the individual efforts into a single team effort. The principal sources of information for the NTPR teams were the various military records available for review. Each branch of service has historical records, although not all in the same format. The Navy has deck logs that list officers and muster rolls that list enlisted personnel; Marine information comes from personnel rosters and daily diaries. Army and Air Force records are morning reports and personnel rosters. Personnel records from the National Personnel Records Center (NPRC) in St. Louis, Missouri, were also examined, where appropriate, to aug- ment individual identifications. Not all Army and Air Force personnel records were available, however, because many were destroyed in a 1973 fire. Another source of information for the NTPR program has been a nationwide toll-free call- in program set up by DNA for veterans of atmospheric nuclear tests to report their participation (1-800-462-3683~. The NTPR-provided data tapes included name; military service numbers; date of birth; Social Security numbers; sex; paygrade, rank, or rating at series; unit membership during participation; permanent unit; and dates of entry into and separation from the service, among others. Names and service numbers are the primary concern of the NTPR program; the availability of the additional pieces of information is limited. For example, Social Security numbers are missing for the majority (55.6%) of personnel and dates of birth for approximately one-third.

OCR for page 19
DATA SOURCES 21 National Archives and Military Collections To select military units as potential sources of comparison cohort member- ship, JAYCOR, a DTRA contractor, reviewed Station Lists. Data on the indi- viduals in the units chosen were obtained from military records such as deck logs and morning reports, as described earlier. CHARACTERISTICS OF COHORT MEMBERS, INCLUDING DATE OF BIRTH AND VITAL STATUS To permit vital status ascertainment of the military record-identified members of the participant and comparison cohorts, date of birth is essential and Social Se- curity number is valuable. The main sources of these pieces of information are the VA Beneficiary Identification and Records Locator Subsystem (BIRLS), the VA Master Index (VAMI), and individual military personnel records. Beneficiary Identification and Records Locator Subsystem The sole source of mortality ascertainment in this study was the VA BIRLS database. It contains, among other things, identifying information on individuals who have submitted claims for veterans' benefits. BIRLS data were used in this study to verify information from military rosters, such as spelling of names, and to acquire date of birth, date of death, and the location of the claims record folder (from which the death certificate is retrieved). The key identifiers used in the BIRLS search are first and last names and military service number. A veteran's death is noted in the VA records system if a claim is filed for death-related benefits, such as reimbursement for burial expense or burial in a national cemetery. Eligibility is determined by various factors including time of service, service-connected disability, cause of death, and financial resources of the veteran's estate. The eligibility rules were modified by legislation in 1981, making benefits more restrictive than earlier, which may have affected the num- ber and characteristics of veterans whose deaths are reported to VA and re- corded in BIRLS. When a death benefit is claimed for a veteran, the VA requires a copy of the death certificate for claim processing. The death certificate then becomes part of the veteran's claims folder, which eventually is retired to the federal archives records centers.

OCR for page 19
22 V, Ct - ., o o ~s O. , Ct C~ o .= Ct o Ct .= ~ m =~0 Oe ~ pL, 1 o U. ~0 O V, . ~ ~ C.) O ~0 - > S - , ~ ._ _ ~ o ._ CC C~ :~ ._ ~ U) P~ U' ce '~,D ~C~ V) z o CQ a' C) o V) .- ~ ~ _ ~ ~ ~ z C) z o c) ~ Ct O ~ V) ~ + + : + + + : + + + : + + + + + + + ~:+ + + + + + + + + + + +: + + :: + + + :~:::: + + + x + x : : :: ~, z z ~ m ~ ~ ~ ;f <: ~ z ~ ~ ~:e Z

OCR for page 19
23 5 C) U. C) ~ ~ C,0 C) O > U) ._ Ct ~ Ce 50 ~ - ~ O C<5 - .C) Z .= ~- ^ _94 _ X ~ ~ p . ~ ~ Z ~ ^ C)^ A:} , ~ Cat ~ca ~ O O ce ce 5 0 Cry ~ ? Z Z ~ z z

OCR for page 19
24 THE FIVE SERIES STUDY VA l\Iaster Index The VA Master Index (VAMI) was the predecessor of BIRLS. From 1917 to January 1972, VA created an index card for each veteran who applied for any benefit, including insurance, education and home loans, health care, and disabil- ity compensation. These VAMI cards have since been transferred to microfilm. Because the BIRLS database was not created directly from VAMI, some refer- ences to pre-1972 deaths can be found in VAMI but not BIRLS. Therefore, when no record for an individual could be found in the computerized BIRLS database, the microfilm copy of VAMI was searched. VAMI was also the source of additional identifying information such as an alias or a military service num- ber that allowed a more accurate repeated search in BIRLS. Military Personnel Folder The military maintains a personnel record folder for each service member. The folders for those who served in the 1950s are archived at the National Per- sonnel Records Center in St. Louis, Missouri. These records contain personal identification data, such as date of birth, military service numbers, and some- times Social Security numbers, in a standardized format. As mentioned previously, the availability of personnel records is limited; the fire at the NPRC in 1973 destroyed about 80 percent of the records for Army personnel discharged between November 1, 1912, and January 1, 1960, and about 75 percent of the records for Air Force personnel with surnames from "Hubbard" through "Z" discharged between September 25, 1947, and January 1, 1964. For most individuals whose personnel records were destroyed, their medi- cal records, which were filed in the same folder, were also lost. Social Security numbers were not routinely used in the military records sys- tem during 1951-1957; hence, a large yield was not expected from these sources. Because of concerns regarding completeness of death reporting by BIRLS and differential characteristics between the deaths found and not found by BIRLS (Boyle and Decoufle, 1990; Page, 1992), the National Death Index, maintained by the National Center on Health Statistics, a non-VA source, was searched to vali- date BIRLS-based vital status ascertainment (see Chapter 9 for details). CAUSE OF DEATH Veteran's Claim Folder Once a veteran's death had been identified through a BIRLS search, a copy of the death certificate was requested from the VA regional office (VARO) or the regional federal archives records center (FARC) noted as the claims folder location in BIRLS. Obtaining death certificate copies from these sources is a time-consuming process. MFUA staff has estimated that about 70 percent of the

OCR for page 19
DATA SOURCES 25 death certificates are obtained within six months after the initial request is sub- mitted. For the remaining 30 percent, however, the process may take years. National Death Index When, after reasonable effort, a death certificate could not be obtained from the VARO or FARC, information on the individual was submitted to the Na- tional Death Index (NDI) with a request for cause-of-death information. NDI is a computer database maintained at the National Center for Health Statistics. Compiled from data tapes submitted by each state's vital statistics office, it contains identifying information on all U.S. deaths since 1979. Researchers can get the state and death certificate number of a known death and then request the death certificate from the state. Since 1998, through the NDI-Plus program, re- searchers can request the death certificate information directly from NDI. The NDI-Plus computer tape includes name, date of birth, date of death, and under- lying and associated causes of death, as recorded on the death certificate. Although we used NDI cause of death in the study's analysis, we relied on itsfact-of-death ascertainment only as a validation tool (see Chapter 8~. POPULATION MORTALITY RATES FOR COMPARISON For the calculation of standardized mortality ratios for each cohort, the Uni- versity of Pittsburgh's Mortality Data and Population Statistics program created cause-specific mortality rates for the ages and calendar time of interest for each of the study cohorts, participant and referent. Although race and sex information was not available to us for the study cohort, we determined that of the cohort deaths less than half a percent were female and between 8 and 9 percent were black. With that data, along with historical anecdotal information about the military and the nuclear weapons program in the 1950s, we decided to use white male rates as an approximation to calculate expected mortality rates.