Data Sources for the CROSSROADS Study
The next part of this report describes procedures of the study in detail, referring frequently to numerous documents, records, and sources with which many readers will not be familiar. Here, therefore, we briefly enumerate the sources and describe their general limitations and assets.
Study staff, as well as Department of Defense (DoD) staff and contractors, made strenuous attempts to identify the existence of any relevant records, to acquire those records, and to corroborate information from multiple sources. Data related to personnel movements, radiation exposure, and vital status relevant to this study 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: DoD (not organized as such during the time of CROSSROADS), including the Navy, Army, Air Force, and Marines, and the Defense Nuclear Agency's Nuclear Test Personnel Review; the Department of Veterans Affairs, including its benefit and health organizations; the Department of Health and Human Services, including the Health Care Financing Administration and the National Center for Health Statistics, which maintains the National Death Index; and the National Archives and Records Administration's National Personnel Records Center, Federal Archives Records Centers, and Office of the National Archives. Central office and regionally maintained records were reviewed by study staff, agency staff, and contractors.
We requested cause-of-death information directly from state (and some municipal) vital statistics offices.10
Differences in record ownership affect the rules that apply to the use of data. Ease or difficulty with access, privacy, fees, time frames, and definitions (conceptual and operational) were all factors in determining collection of the data upon which this report is based. Furthermore, among the federal agencies and private sources pursued, there are different interests (some overlapping, some in seeming conflict) behind the collection of data. These purposes might include understanding science, assuring that the government honors reasonable claims for compensation, learning lessons for application to future public policy decisions, cost accounting, advocacy support (legislative, regulatory, and emotional), and assigning responsibility for past decisions and their consequences.
The completeness and accuracy of the various data collections vary widely, as does the quality of the data themselves. Certain data may be perfect for their intended purpose but not otherwise useful. Other data—such as those which we worked to retrieve for this study—may have fortuitous uses other than those for which they were collected. Trade-offs exist in data sources: accurate but limited in time frame or population considered; broadly relevant but of variable and difficult-to-judge accuracy; and relevant but lacking linked information to other lifetime experiences that might affect health status. Some sources are based on information about individuals (e.g., military personnel records), some on direct measurement (e.g., radiation dose measurements), and others on expert-derived best estimates (dose reconstructions).
Among those sources viewed were handwritten paper logs, microfilm/fiche, computer files, medical records, work orders, transport orders, memoirs,
interoffice memoranda, testimony, secondary compilations of primary sources, letters from spouses, death certificates, burial notices, film badge records, computer programs, and benefits and compensation claims.
While any one type or source of data may be biased; taken together, these biases sometimes offset each other, at least partially. And, when independent sources corroborate specific facts, confidence in the accuracy of the information increases.