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 65
The Utility of Proximity-Based Herbicide Exposure Assessment in Epidemiologic Studies of Vietnam Veterans 4 Data for Epidemiologic Studies of Vietnam Veterans Chapter 3 examined the components of the Stellman team’s herbicide exposure assessment model, including the geographic information system (GIS) and the Herbicide Exposure Assessment–Vietnam software tool. Chapter 3 also assessed the model’s strengths and limitations in providing surrogates for exposure to herbicides sprayed in Vietnam. The model and its exposure metrics are welcome contributions to the effort to assess veterans’ herbicide exposure, but other information must also be available to carry out epidemiologic studies. This chapter addresses two categories of needed information: data on the likely location of individual study participants at particular times during their service in Vietnam, and health outcome information for these individuals. Important considerations and challenges are associated with gathering both types of information. Examining the availability, quality, and usefulness of existing information on Vietnam veterans permits consideration of some important factors that should be taken into account in planning for studies. AVAILABLE TROOP LOCATION DATA As described in previous chapters, the exposure assessment model makes use of a GIS consisting of two primary components: (1) a cleaned and documented georeferenced database abstracted from multiple types of military records providing information on herbicide spray missions during the Vietnam conflict, and (2) a database engine allowing information on locations of military units and their personnel to be entered and linked
OCR for page 66
The Utility of Proximity-Based Herbicide Exposure Assessment in Epidemiologic Studies of Vietnam Veterans with the locations of spraying missions. To generate an exposure measure (a “hits” count, exposure opportunity index [EOI] score, or other exposure measure) for subjects in an epidemiologic study, it is necessary to have data on the location, in longitude and latitude, of study subjects for the duration of their deployment in Vietnam. However, because the locations of individuals were not recorded during their deployments, it is necessary to rely on the recorded locations of the units to which they were assigned. Location Data Assembled by the Stellman Team As a result of the work that has already been done, the committee found that at least some location-tracking information is available for more than 3,300 of the 4,778 U.S. Army units included in the Stellman team’s databases. These databases also list 1,106 Air Force units, 22 Coast Guard units, 617 Marine Corps units, and 558 Navy units deployed to Vietnam. Location data have been tracked for all of the Air Force units, many Marine Corps units, and some Navy construction battalions and “brown water” units1 (Stellman, 2007b). Location information for the remaining units and time periods is likely to be available but will have to be gathered through research of archival records (discussed below). The Stellman team has made a distinction between stable units with missions that required only infrequent movement from base camp and mobile units that were frequently away from a base camp. In addition, some units were primarily stable but had elements that were mobile. Stellman and colleagues, with the assistance of the U.S. Army and Joint Services Records Research Center (JSRRC), developed location histories for 3,017 stable Army units, representing most of the combat arms support and combat support units assigned to Vietnam (Stellman and Stellman, 2003). According to the Stellman team (Stellman and Stellman, 2003), personnel serving in these units accounted for 80 percent of the troops in Vietnam. Compiling location histories for mobile combat units is more challenging. Daily records for combat battalions may contain indications of multiple locations that reflect field positions of several companies and occasionally even platoons. The Stellman team has carried out preliminary cleaning and analysis of location data for more than 200 mobile Army combat units that were part of 55 battalions assigned to Military Region III of South Vietnam (also referred to as III Corps) between 1966 and 1969, a period and location of intense spraying (Stellman and Stellman, 2003). These location data had been gathered in the 1980s by the predecessor of 1 “Brown water” Navy units operated in the rivers and deltas of Vietnam; “blue water” units operated in the open sea off the coast.
OCR for page 67
The Utility of Proximity-Based Herbicide Exposure Assessment in Epidemiologic Studies of Vietnam Veterans JSRRC on behalf of a study that had been planned (but was later cancelled) by the Centers for Disease Control and Prevention (CDC). The Stellman team plans to further clean and confirm location data for these units as part of a 3-year project that began in 2007 (see below; Stellman, 2007a). They also note that similar location data were collected for a set of Marine Corps units in conjunction with a Department of Veterans Affairs (VA) mortality study, and that with appropriate review those data could be added to the location databases (Stellman and Stellman, 2003). The Stellman team (2003) carried out extensive quality control checks on their location data. For stable units, this included consultation with military experts and checks for erratic locations or movements that were considered too frequent. Major location changes detected in database entries were verified to avoid typographical or data-entry errors. The location data assembled for the mobile combat units were also evaluated in an effort to detect errors. When records showed multiple locations for a mobile unit on a single day (specified by Universal Transverse Mercator [UTM] values), the Stellman team calculated a weighted average (a “center-of-mass”) and standard deviations for the UTM values. They paid special attention to cases in which the standard deviation was greater than 5 kilometers. In most cases corrections could be made, but some records were deleted from the database because evident errors could not be corrected. When their project ended in 2003, the Stellman team had not been able to complete their review of all suspect records or their review of data gaps, but they expressed confidence that many data problems could be resolved (Stellman and Stellman, 2003). Concerns Regarding Troop Location Data Although the locations of battalions may have little error, the aggregation of information at the battalion-level can lead to errors if those locations are attributed to individuals. Some uncertainty surrounds the assignment of a location to an individual on the basis of the reported location of the battalion-level unit (500 to 1,000 soldiers) or at best, a company-size unit (75 to 200 soldiers) to which he was assigned. Small units such as platoons were not always at the exact location of the larger units to which they were assigned, and the deviation of their location from that recorded for the larger unit (or the accuracy with which the unit’s location was recorded) cannot now be determined. On the other hand, individuals’ absences from their units are documented in records such as morning reports, which were generally prepared daily. Given the exigencies of combat, however, there are likely to be instances when a report was not created on a given day. Individuals’ personnel files also may contain information about absences from the units to which they were assigned.
OCR for page 68
The Utility of Proximity-Based Herbicide Exposure Assessment in Epidemiologic Studies of Vietnam Veterans Some have raised other concerns about the location data for military units (e.g., Young et al., 2004; Ross and Ginevan, 2007). These critics claim that many relevant military records may be missing, incomplete, or reflect poor recordkeeping practices. They also point to military policies and Department of Defense (DoD) guidelines as evidence that the spray missions were planned with care to ensure that friendly forces were not in spray areas. However, a report by the General Accounting Office (GAO, 1979) estimated that for the period 1966–1969 about 5,900 marines were assigned to units that had been located within 0.5 kilometer of areas sprayed with Agent Orange on the same day. These troops were 2.7 percent of the 218,000 marines in Vietnam during 1966–1969. About 16,100 marines (7.4 percent) were within 0.5 kilometer of sprayed areas before the official DoD 4-week reentry period was over. The GAO (1979, p. 8) conclusion was that “DoD’s contention that ground troops did not enter sprayed areas until 4 to 6 weeks afterward is inaccurate; the chances that ground troops were exposed to herbicide orange are higher than DoD previously acknowledged.” OBTAINING ADDITIONAL TROOP LOCATION DATA FROM MILITARY RECORDS Even with the work already done and planned by the Stellman team, researchers using the exposure assessment model will inevitably need to abstract additional information from military records. The task is likely to include determining which individuals were assigned to units of interest, their dates of service and assignment histories, and, for units for which location information has not yet been gathered, following the units’ movements over the period of interest. The Stellman team reported to the committee on their own experience with participants in a case-control study on amyotrophic lateral sclerosis (ALS) (Stellman, 2007c): Of the 384 units in which study participants served, complete tracking information had already been collected for 152 units and partial information for 102 units. For the remaining 130 units, they estimated that collection of unit location information would take roughly 30 days of work. Access to Military Records Military records from the Vietnam era include a mix of publicly accessible material and records still controlled by the individual military service branches. In either form, the records include handwritten notes and files, often stored on outdated media requiring specialized hardware to read. Most of the records are in paper form or on microfilm. Three basic types of military records are likely to contain the kind of
OCR for page 69
The Utility of Proximity-Based Herbicide Exposure Assessment in Epidemiologic Studies of Vietnam Veterans information that will be needed to develop location histories for military units or individual veterans: The programmatic and operational records of military field organizations at various levels of the organizational hierarchy. These records (e.g., Army daily journals, Navy deck logs) contain reports on the locations and activities of troop units. The personnel strength accounting records of military units, generally at the lowest level of the organizational hierarchy. These records include Army morning reports and rosters and Navy and Marine Corps muster rolls. Parallel records for Air Force units are only available through June 1966. Personnel strength accounting records identify the individuals assigned to the units that are the subjects of the programmatic and operational records. The Official Military Personnel Files (OMPFs) of individual service members. These records contain demographic and other personal data on individuals and can provide dates of their unit assignments over the course of their military service. The operational and programmatic records created by the U.S. Army in Vietnam have been accessioned by the National Archives and Records Administration (NARA). They are fully accessible to the public, with no administrative barriers or impediments to using them. For the most part, similar records created by the other branches of the armed services remain in the legal custody of those agencies and are generally available to researchers pursuant to a Freedom of Information Act (FOIA) request. Exceptions include deck logs of Navy ships and the Air Force Operation Ranch Hand records, both of which series have been accessioned by NARA. Some records, however, such as those created by intelligence organizations, may still retain security classifications and remain unavailable to those not having the appropriate security clearance. Records in the second and third categories remain in the legal custody of the military services. Access to these records is restricted by provisions of the Privacy Act of 1974 (5 U.S.C. 552a) and the administrative rules promulgated under that act by each of the service departments. In general, the statute and rules limit access to the records to the individual subject of the record; his or her immediate next-of-kin (in the event of the subject’s death); anyone specifically granted access by the subject of the record or, if the subject is deceased, by the next-of-kin; and the agency that created the record. Access also is available to “routine users” of the record, which are entities designated by the creating agency and identified as having a legitimate need to access the record to conduct official business on behalf of the public.
OCR for page 70
The Utility of Proximity-Based Herbicide Exposure Assessment in Epidemiologic Studies of Vietnam Veterans Even with a routine user designation, it is necessary to request and receive permission for access to records for the specific purposes of any given study. Privacy rules also allow for access to the records by contractors of the creating agency, by other federal agencies and their contractors, and by private organizations “for purposes of conducting personnel and/or health-related research in the interest of the federal government and the public.” These latter arrangements are ad hoc in nature, with access approval being granted on a case-by-case basis and inevitably involving a degree of uncertainty that is avoided when access is sought by an entity specifically designated a routine user of the records in the agency regulation.2 Access to personnel strength accounting records and OMPFs may be crucial for the completion of epidemiologic studies using the exposure assessment model. In the short term, based on the best information available from DoD and the military service departments, it would appear that investigators who seek access to personal data from records that document the service of Vietnam-era veterans would be best advised to contract directly with DoD or VA, contract with an organization that has been specifically designated by each of the service departments as a routine user of military personnel records, or enter into a collaboration with an investigator affiliated with any of these organizations. However, the committee sees benefit in providing access to military records that remain in the legal custody of the services to support independent research, including work by investigators who are not under contract to DoD or VA, contingent on establishing and maintaining safeguards for privacy and for appropriate research use of the records. The committee urges VA to work with its sister agencies to make access to the relevant records feasible for independent researchers. If changes in the law are necessary to make this possible, the committee urges congressional attention to modifications in the law to improve appropriate access to records for research purposes. 2 The committee notes that the National Academy of Sciences is cited by the Army, Navy, Marine Corps, and Air Force as an example of a “private organization” to which records containing personal data can be made available “for purposes of conducting personnel and/or health-related research in the interest of the Federal government and the public” (5 U.S.C. 552a; Department of the Air Force, 2000; Department of the Army, 2000; Department of the Navy, 2000; U.S. Marine Corps, 2000). The Navy regulation further names the National Research Council as a routine user of Navy privacy system records (5 U.S.C. 552a[b]; Department of the Navy, 2000).
OCR for page 71
The Utility of Proximity-Based Herbicide Exposure Assessment in Epidemiologic Studies of Vietnam Veterans Using Military Records to Gather Troop Location Information Gaining legal access to military records is a necessary first step, but the challenges of gathering information from these records extend beyond that to the availability, quality, and usefulness of the information in the records. Table 4-1 lists the kinds of information that researchers may need, sources for those various types of information, and where the relevant military records are located. The committee notes that a compilation of information about the content, form, and location of combat and combat-related records from the Vietnam era, which was assembled in the mid-1970s (Carter et al., 1976), may also be useful to researchers attempting to work with any of these types of records. Drawing upon committee expertise, input from experts provided at the committee’s information-gathering sessions, and the experience of the Medical Follow-up Agency (MFUA) of the Institute of Medicine (IOM) in using military service records for epidemiologic studies, the committee identified several considerations to bear in mind in planning studies. Personnel Strength Accounting Records The Stellman team has done extensive work on identifying units that served in Vietnam, but epidemiologic studies will require information on the individuals who served in these units over time. This information can be gathered from the personnel strength accounting records of military units, which identify the individuals assigned to each unit. Navy and Marine Corps muster rolls are stored at the National Archives facility in College Park, Maryland. Army morning reports and unit rosters are stored at NARA’s National Personnel Records Center (NPRC) in St. Louis, Missouri. Also stored at NPRC are Air Force morning reports through June TABLE 4-1 Types and Sources of Information for Developing Location Histories for Military Units or Personnel Information Type Source Location Individuals serving in a unit Personnel strength accounting records of military units NARA, College Park, Maryland NPRC, St. Louis, Missouri Individual’s history of unit assignments Official Military Personnel Files NPRC Locations of units during deployment in Vietnam Programmatic and operational records of military field organizations NARA NOTE: NARA, National Archives and Records Administration; NPRC, National Personnel Records Center.
OCR for page 72
The Utility of Proximity-Based Herbicide Exposure Assessment in Epidemiologic Studies of Vietnam Veterans 1966. If the services grant permission to access the records at NPRC, records must be located and retrieved by NPRC staff, typically according to a schedule that is requested in advance through a formal contract. The records are on microfilm and must be reviewed in the NPRC Research Room. Alternatively, copies of the film maybe ordered from NPRC and shipped to researchers. Official Military Personnel Files When the identities of individual study subjects are known, researchers will need to access OMPFs to determine each individual’s dates of entry into and exit from units during Vietnam service. Veterans’ military personnel records are stored at NPRC in St. Louis, and as with the other records held there, they must be accessed through NPRC personnel and viewed on site. Personnel files may sometimes be unavailable because they have been misfiled or are in use.3 In carrying out data collection for their work, the Stellman team provided NPRC with a list of 6,393 names and Social Security numbers of veterans for whom they sought dates of service and unit assignment information. They received useful information on 91 percent of the individuals, but for 574 veterans the files were unavailable or the data abstracted by NPRC could not be used (Stellman and Stellman, 2003). A pilot study carried out as part of an IOM study of sailors exposed to test agents in the 1960s found that of a sample of 174 hard-copy personnel records requested, 32 of them (18 percent) were not available (IOM, 2007). Eight records were charged out for another use, and 24 could not be found. The location of information on unit assignments within a service member’s personnel file will vary with the branch of service. For Army files, the information in Vietnam-era records is summarized on a single card-stock form (DA Form 20) in the service member’s file. A soldier whose military career extended over several periods of service may have more than one DA Form 20, but the forms are readily identifiable and are generally filed together in the record. Even so, a review of the entire file may be useful to ensure that a complete location history is compiled. Information on unit assignments for Marine Corps veterans may be available on a summary form. For those who served in the Navy, the information is likely to be clearly identified but distributed through the record. Some of the sources of information on assignment histories have been identified by the Stellman team (Stellman et al., 2003). Careful plans should be made for the data abstraction process to ensure that information can be collected from the personnel records efficiently and in a usable form. Unit assignments can be designated in various ways, and 3 A 1973 fire at the National Personnel Records Center had little impact on records of Vietnam-era veterans (see http://www.archives.gov/st-louis/military-personnel/fire-1973.html).
OCR for page 73
The Utility of Proximity-Based Herbicide Exposure Assessment in Epidemiologic Studies of Vietnam Veterans they will need to be recorded in a manner that will ensure that a consistent set of unit assignments can be generated. If researchers intend to make use of the unit location histories that the Stellman team has compiled, they will need to ensure that the information collected on individuals’ unit assignments can be linked to the unit nomenclature used in the exposure assessment model. Researchers should also consider whether there is other information in the personnel records, such as demographic data or occupational duties, that they would want to have collected at the same time. The Stellman team’s review of data abstraction work done by others suggested that errors were made with 9 to 10 percent of the records (Stellman and Stellman, 2003). As with any records-based research, it will be important to ensure that the personnel performing the data abstraction are well trained and that verification procedures are part of the data collection process. Programmatic and Operational Records of Military Field Organizations Although the Stellman team has assembled at least partial location histories for many military units, researchers may find it necessary to collect additional unit location information. As noted, information on Army units can typically be found in military records stored by NARA in College Park, Maryland, and information on other units has to be obtained through the individual service branches. Researchers seeking information at College Park must request specific unit records from the Archives staff, and only a limited number of archive boxes can be requested at one time. Retrieved records can be reviewed, scanned, or photographed onsite; records cannot be taken from the Archives building. Abstracting information about locations of units over the time of their deployment is likely to be a challenging task to those unfamiliar with military recordkeeping. It can greatly benefit from a familiarity with military activities in Vietnam and typical means of recording those activities. Guidance in this task is available from such sources as a description of standard operating procedures prepared by a predecessor of the Army’s JSRRC (Department of the Army, 1985) and a guide to the use of military records prepared by the Stellman team (Stellman et al., 2003). The Army document describes daily journals as probably providing the most accurate record of unit activities, with a hierarchy of other information sources that can be consulted, including situation reports, intelligence summaries, lessons learned reports, after-action reports, and morning reports (Department of the Army, 1985). Many of these documents were produced at the battalion, brigade, and division levels. Similarly, the Stellman team’s document describes the different types of primary information sources available and suggests appropriate strategies for finding unit locations. Figure 4-1 shows an excerpt from a daily journal.
OCR for page 74
The Utility of Proximity-Based Herbicide Exposure Assessment in Epidemiologic Studies of Vietnam Veterans FIGURE 4-1 A Daily Journal excerpt. The excerpt illustrates a record of the activities of A, C, and D companies of the Headquarters and Headquarters Company, Third Brigade, First Air Cavalry Division, over a 4-hour period on February 4, 1968. Locations are specified using UTM coordinates, which are given as two letters followed by six numerals, for example YD 644312. SOURCE: Department of the Army, 1968.
OCR for page 75
The Utility of Proximity-Based Herbicide Exposure Assessment in Epidemiologic Studies of Vietnam Veterans The size of the unit for which detailed location information is available varies, and the quality of units’ recordkeeping is not consistent (Shaughnessy, 1991; Boylan, 2007). Larger units (e.g., battalions) are generally more likely to have records of their daily locations than are smaller units such as companies. The Stellman team (Stellman et al., 2003) notes that for Army infantry and cavalry units, which are likely to be highly mobile, detailed locations can be found only for battalions or squadrons (100 to 1,000 men). However, these records often can be used to determine the locations of units’ component companies or troops and, in some cases, also can be used to determine the locations of their component platoons. Determining the locations of independent companies or troops is more difficult and time-consuming, and many times not possible. Other Sources of Location and Personnel Information The Stellman team (Stellman and Stellman, 2003) suggests that Vietnam veterans themselves be considered for sources of information such as maps and photographs that may aid in locating individuals or units. In addition, several military units and veterans organizations have compiled histories that may include troop locations, combatant names, and dates of service. Although these data vary in completeness, coverage, and accuracy, they may serve as a source of confirmatory and supplementary information. Committee Comment on Collection of Troop Location Data Despite some of the limitations of the available data resources for troop locations and the challenges in assembling additional data, the committee considers the approach the Stellman team took toward development of their unit location database to be reasonable, and it has some confidence that their approach would be successful. However, as the committee gained an understanding of the procedures necessary to gather location data for military units and the inherent challenges of those procedures, it had increasing concerns regarding the ability of typical academic researchers to carry out this work without the benefit of expertise in military records. The committee agrees with the guidance given by the Stellman team that “it is essential that those with specific experience and knowledge of military records and military terminology be involved in the research” (Stellman et al., 2003). In their work the Stellman team benefited from interaction and collaboration with what is now JSRRC. But the capacity of JSRRC to respond to requests may be limited (see GAO, 2006), without additional resources. The Stellman team also cautions that for the high volume of records likely to be needed for a major epidemiologic study, the current method of record retrieval available to researchers using the NARA facility may not be prac-
OCR for page 76
The Utility of Proximity-Based Herbicide Exposure Assessment in Epidemiologic Studies of Vietnam Veterans tical. The Stellman team urges consideration of an arrangement that can draw upon the experience and skills of the JSRRC in using military records (Stellman et al., 2003). The committee concurs. SOURCES OF HEALTH OUTCOME INFORMATION To carry out a study examining potential associations between exposure to herbicides and health problems in Vietnam veterans, researchers will need information on individuals’ health outcomes as well as troop locations. This section discusses sources of health information for Vietnam veterans and related access issues. Many of the challenges involved in gathering health outcome information for veterans are not unique to this population and should be familiar to epidemiologists carrying out studies on members of the civilian population. Mortality Data Information regarding mortality for an epidemiologic study is available from several sources. Complete information on deaths since 1979, including cause of death, is available for the U.S. population through the National Death Index, or NDI. Maintained by the National Center for Health Statistics, NDI is a computerized file of the death information available through each of the states’ vital statistics offices (NCHS, 2007). Lists of potential study subjects can be compared with NDI entries to determine which subjects are deceased, their dates of death, and the states in which they died. Users are charged $0.15 for each year searched for each study subject. For example, records for 10,000 study subjects searched against 10 years would cost $15,000 ($0.15 × 10,000 × 10), plus a small user’s fee. For large studies covering many years, the costs can be significant. For slightly higher fees, NDI (through “NDI Plus”) can also be the source of information on cause of death, as recorded on death certificates. A potentially useful source of mortality information specifically for U.S. veterans is VA’s Beneficiary Identification and Record Locator Subsystem (BIRLS). The automated system identifies veterans and their beneficiaries who have received compensation, pension, education, and other VA benefits (Kang, 2007a). A 1996 study of deceased male veterans born between 1936 and 1955 (a proxy for Vietnam-era service) found BIRLS death ascertainment to be roughly 90 percent complete (Page et al., 1996). A more recent study, however, suggested that BIRLS death reporting may be only 77 percent complete (Sohn et al., 2006) and concluded that BIRLS mortality ascertainment is, by itself, too incomplete for research use. Information on veterans’ deaths can also be obtained from the Social Security Administration’s Death Master File (SSA DMF), which contains
OCR for page 77
The Utility of Proximity-Based Herbicide Exposure Assessment in Epidemiologic Studies of Vietnam Veterans records for some 65 million deaths reported to the Social Security Administration (NTIS, 2007). The file can be purchased by researchers, along with quarterly updates. A recent study suggests that ascertainment of veterans’ deaths is roughly 92 percent complete in the SSA DMF (Sohn et al., 2006). Another source of mortality information for older veterans is the Medicare Vital Status file, which is maintained by the Centers for Medicare and Medicaid Services (CMS). The primary source of information for this file is the SSA DMF, but it is updated with information from other sources, including Medicare claims data (Sohn et al., 2006). Sohn et al. (2006) found the Medicare Vital Status file to be 83.2 percent complete in their sample of veterans, and for veterans enrolled in Medicare it was the most accurate and complete (99 percent) of all single sources of death information evaluated. Studies assessing mortality are likely to benefit from using more than one of the information sources. Access to the NDI, SSA DMF, and Medicare Vital Status file is readily available to researchers, while access to VA’s BIRLS database is more limited. Morbidity Data Obtaining information for studying morbidity of Vietnam veterans is more problematic than obtaining information about mortality. Because of the dispersed nature of health care in the United States, Vietnam veterans receive care from many types of providers and care systems. The committee focused its attention on the strengths and weaknesses of VA and Medicare data and data systems, which it saw as two promising sources of morbidity data for Vietnam veterans. Researchers could also explore other resources such as state cancer registries and hospital discharge datasets. In addition, if data are needed on the cumulative incidence of a condition, it may be necessary to obtain information on veterans’ medical histories and medical care during military service. VA Databases Morbidity data are available from VA from three primary sources: the Patient Treatment File (PTF), which contains data on persons who received inpatient care at VA facilities; the Outpatient Clinic File (OPC), which similarly contains records for outpatient care given at VA facilities; and a set of health registries (see Box 4-1). VA researchers have used the PTF as the starting point for case-control studies of Vietnam veterans (e.g., Kang et al., 1986; Dalager et al., 1991; Mahan et al., 1997). The study populations assembled for these studies are discussed in Chapter 5 as potential popu-
OCR for page 78
The Utility of Proximity-Based Herbicide Exposure Assessment in Epidemiologic Studies of Vietnam Veterans BOX 4-1 VA Health Information Databases BIRLS The Beneficiary Identification and Record Locator Subsystem (BIRLS) is an automated system identifying veterans and their beneficiaries who have received compensation, pension, education, or other VA benefits (Kang, 2007a). The database has more than 44 million records and includes an initial record for each person discharged from military service since the early 1970s. The database contains information that includes veterans’ names, social security numbers, dates of birth and death, gender, periods and branches of service, and current location of claim folders. One of the most frequent research uses of the BIRLS database is in identifying deceased veterans for mortality studies (Page et al., 1996; Boyko et al., 2000). Patient Treatment File The Patient Treatment File (PTF) contains information on inpatient records from each discharge from a VA hospital facility since 1970. Data are updated every 2 weeks. Currently, the database captures approximately 400,000 discharges annually, down from a peak of about 1 million annually in the late 1980s (Kang, 2007a). The PTF’s demographic data is complete except for address, but information on military history is not considered reliable. For each discharge, up to 10 different diagnostic codes are recorded, including an indication of the diagnosis most responsible for the length of stay. The PTF can be used as a sampling frame to identify potential subjects for case-control studies. It can also be used to assess health care utilization or morbidity for selected cohorts of veterans. However, there is no quality control on the data, which are subject to the errors that are typical in medical records (Kang, 2007a). Outpatient Clinic File The Outpatient Clinic File (OPC) contains records from visits to outpatient VA clinics. Records prior to 1997 contain only basic utilization data. The newer records include information for each clinic stop during a patient visit and include demographic information and up to 10 diagnostic and 15 procedural codes. In 2004, 4.9 million unique veterans were seen at VA outpatient centers (Kang, 2007a). Health Registries VA maintains various health registries. The Agent Orange Registry includes records for nearly 409,000 veterans who have identified themselves as having concerns about medical problems related to exposure to Agent Orange and who have undergone physical examinations (Kang, 2007a). The Cancer Registry is compiled from cancer registries maintained at each VA facility. It contains records on all veterans receiving a diagnosis of cancer within the VA medical system. VA also has registries of veterans diagnosed within the VA system with other conditions, including ALS, multiple sclerosis, and hepatitis C.
OCR for page 79
The Utility of Proximity-Based Herbicide Exposure Assessment in Epidemiologic Studies of Vietnam Veterans lations for studies based on applications of the Stellman team’s exposure assessment model. It is estimated that 20 percent of veterans from the Vietnam War era receive at least a portion of their medical care through the VA system (VA, 2001). Among a sample of veterans between 55 and 64 years of age, for example, 7.6 percent reported receiving health care only from VA sources and 12.7 percent reported using both VA and non-VA health care sources (VA, 2001). Despite Vietnam-era veterans’ limited use of VA health care services, VA records are likely to provide the largest proportion of health information on Vietnam veterans available from any single source. Use of VA health care data for epidemiologic studies is problematic, as it is difficult to estimate the population at risk. For example, if there were no record of an individual having received medical treatment from VA, it could mean that person was healthy or was receiving medical care elsewhere. Users of VA data must consider that the Vietnam veterans who use VA medical care are probably not representative of the population of Vietnam veterans as a whole. An analysis of veterans overall found that those who did not use VA health care services had higher incomes, more education, and better health status than those who relied entirely or in part on VA health care (Nelson et al., 2007). Access to VA’s administrative databases, health care information, and medical records is generally restricted to researchers who are part of the VA system (employed at least 5/8 time by VA). Should access issues be overcome for other researchers, the VA resources described in Box 4-1 may be of interest to them, taking into account the constraints discussed above. Medicare Claims Data As more Vietnam veterans reach the age of 65 and become eligible for Medicare health insurance, Medicare data may prove increasingly useful as a source of information on health status for this population. A 2001 survey of veterans found that 96 percent of those who were age 65 or older were enrolled in Medicare (VA, 2001). Table 4-2 shows that something approaching 30 percent of the nearly 3 million veterans listed in VA’s Vietnam Veteran Roster were eligible for Medicare in 2007. Within the next 10 years more than 90 percent of them will be eligible. Under a data-sharing agreement with CMS, VA receives Medicare data for enrolled veterans each year (Sohn et al., 2006). Other researchers can obtain data from the Medicare system on the demographics of enrollees and on billing claims for beneficiaries’ Medicare-eligible fee-for-service encounters. About 85 percent of Medicare coverage is accounted for by fee-for-service claims (as opposed to managed care) (CMS, 2007). The claims data contain information on diagnoses, treatments, costs, lengths of stay,
OCR for page 80
The Utility of Proximity-Based Herbicide Exposure Assessment in Epidemiologic Studies of Vietnam Veterans TABLE 4-2 Distribution of Veterans Included in the Vietnam Veteran Roster, by Year of Birtha Birth Year Number of Veterans Percent Cumulative Percent 1900–1910 1,029 < 0.1 < 0.1 1911–1915 2,584 0.1 0.1 1916–1920 15,985 0.6 0.7 1921–1925 36,415 1.3 2.0 1926–1930 97,253 3.5 5.4 1931–1935 186,874 6.6 12.1 1936–1940 200,061 7.1 19.2 1941–1945 475,397 16.9 36.1 1946–1950 1,539,984 54.7 90.8 1951–1955 252,917 8.9 99.7 1956–1957 7,532 0.3 100.0 aBased on the 2,816,031 veterans in the roster with a date of birth recorded. SOURCE: Modified from Kang, 2007b. and dates of service for every ambulatory visit and hospital stay covered by Medicare. Demographic data can be obtained through the Vital Status File, as noted above, and a Denominator File, which includes all beneficiaries enrolled in a given year. Claims data are stored separately for different settings in which care is provided, including inpatient, outpatient, skilled nursing facility (SNF), and hospice care. A separate record exists for each service for which a payment claim was filed. Records on inpatient and SNF care are also available in a form that links all services for a given stay. Access to individually identifiable CMS data is limited by Privacy Act provisions. Data can be made available to researchers upon approval of formal requests for specific study purposes (ResDAC, 2007b). To obtain data on specific individuals, researchers must submit Social Security numbers or other specified identifying information that will allow CMS to select records of interest. Although the claims files are complex and their primary purpose is not to support research studies, various academic groups have developed expertise in their use. In addition, the Research Data Assistance Center is funded by CMS specifically to aid government, academic, and nonprofit researchers in using Medicare (and Medicaid) data (ResDAC, 2007a). CONCLUSIONS From its review of the sources of data on troop locations and health outcomes that will be needed in order to apply the herbicide exposure
OCR for page 81
The Utility of Proximity-Based Herbicide Exposure Assessment in Epidemiologic Studies of Vietnam Veterans assessment model in epidemiologic studies of Vietnam veterans, the committee reached the following conclusions: It is generally possible to obtain useful data on individuals’ unit assignments and unit locations. However, the processes of gaining permission for access to relevant military records and of collecting data for individuals is likely to be administratively difficult for many researchers, as well as time consuming and costly. Assistance from experts in the location and interpretation of Vietnam-era military records is likely to be essential for effective collection of data from these sources. With appropriate identifying information, mortality data for Vietnam veterans are readily and reliably available through the National Death Index and the Social Security Administration’s Death Master File, while access to VA’s Beneficiary Identification and Record Locator Subsystem is more limited. However, no single source of comprehensive morbidity data currently exists for these veterans. Within the next 10 years it should be possible to use Medicare records to obtain morbidity information for most Vietnam veterans. Although VA has several databases that could contribute information about the health status of some Vietnam veterans, only about 20 percent of these veterans are receiving care from the VA, and access to VA data is typically restricted to VA researchers. REFERENCES Boyko, E. J., T. D. Koepsell, J. M. Gaziano, R. D. Horner, and J. R. Feussner. 2000. U.S. Department of Veterans Affairs medical care system as a resource to epidemiologists. American Journal of Epidemiology 151(3):307–315. Boylan, R. 2007. Accessing military unit records at the College Park Archives. Oral presentation to the IOM Committee on Making Best Use of the Agent Orange Reconstruction Model, Meeting 2, April 30–May 1, Washington, DC. Carter, G., J. W. Ellis, A. H. Peterson, J. H. Pierce, and J. S. Reiley. 1976. User’s guide to Southeast Asia combat data. Document Number: R-1815-ARPA. Santa Monica, CA: RAND Corporation. http://rand.org/pubs/reports/2005/R1815.pdf (accessed November 1, 2007). CMS (Centers for Medicare and Medicaid Services). 2007. 2007 Statistical supplement. Health Care Financing Review. http://www.cms.hhs.gov/MedicareMedicaidStatSupp/LT/list.asp (accessed October 19, 2007). Dalager, N. A., H. K. Kang, V. L. Burt, and L. Weatherbee. 1991. Non-Hodgkin’s lymphoma among Vietnam veterans. Journal of Occupational Medicine 33(7):774–779. Department of the Air Force. 2000. Privacy Act of 1974; Notice to alter a system of records. Federal Register 65(199):60916–60918. Department of the Army. 1968. Daily staff journal or duty officer’s log. A, C, and D companies of the Headquarters and Headquarters Company, Third Brigade, First Air Cavalry Division. Photocopy.
OCR for page 82
The Utility of Proximity-Based Herbicide Exposure Assessment in Epidemiologic Studies of Vietnam Veterans Department of the Army. 1985. ESG standard operating procedures: Steps in researching and tracking exposed combat units. Washington, DC: U.S. Army and Joint Services Environmental Support Group. Department of the Army. 2000. Privacy Act of 1974; Notice to alter a system of records. Federal Register 65(199):60918–60921. Department of the Navy. 2000. Privacy Act of 1974; Notice to alter a system of records. Federal Register 65(199):60923–60925. GAO (General Accounting Office). 1979. U.S. ground troops in South Vietnam were in areas sprayed with Herbicide Orange. FPCD-80-23. Washington, DC: U.S. Government Printing Office. GAO (Government Accountability Office). 2006. VA can improve its procedures for obtaining military records. GAO-07-98. Washington, DC: U.S. Government Printing Office. IOM (Institute of Medicine). 2007. Long-term health effects of participation in Project SHAD (Shipboard Hazard and Defense). Washington, DC: The National Academies Press. Kang, H. K. 2007a. Data resources within VA for an epidemiological study of Vietnam veterans. PowerPoint presentation to the IOM Committee on Making Best Use of the Agent Orange Reconstruction Model, Meeting 2, April 30–May 1, Washington, DC. Kang, H. K. 2007b. Vietnam roster veterans: Demographic/military service characteristics. Unpublished document submitted to the IOM Committee on Making the Best Use of the Agent Orange Reconstruction Model, July 11. Kang, H. K., L. Weatherbee, P. Breslin, Y. Lee, and B. Shepard. 1986. Soft tissue sarcomas and military service in Vietnam: A case comparison group analysis of hospital patients. Journal of Occupational and Environmental Medicine 28(12):1215–1218. Mahan, C. M., T. A. Bullman, and H. K. Kang. 1997. A case-control study of lung cancer among Vietnam veterans. Journal of Occupational and Environmental Medicine 39(8):740–747. NCHS (National Center for Health Statistics). 2007. National Death Index. Division of Vital Statistics, Centers for Disease Control and Prevention. http://www.cdc.gov/nchs/r&d/ndi/what_is_ndi.htm (accessed September 18, 2007). Nelson, K. M., G. A. Starkebaum, G. E. Reiber. 2007. Veterans using and uninsured veterans not using Veterans Affairs (VA) health care. Public Health Reports 122(1):93–100. NTIS (National Technical Information Service). 2007. Social Security Administration’s Death Master File. http://www.ntis.gov/products/pages/ssa-death-master.asp (accessed September 18, 2007). Page, W. F., C. M. Mahan, and H. K. Kang. 1996. Vital status ascertainment through the files of the Department of Veterans Affairs and the Social Security Administration. Annals of Epidemiology 6(2):102–109. ResDAC (Research Data Assistance Center). 2007a. About us. University of Minnesota. http://www.resdac.umn.edu/AboutUs/Index.asp (accessed October 8, 2007). ResDAC. 2007b. Medicare data available. University of Minnesota. http://www.resdac.umn.edu/Medicare/data_available.asp (accessed October 8, 2007). Ross, J. H., and M. E. Ginevan. 2007. Points for the committee to consider when evaluating the Stellman model. PowerPoint presentation to the IOM Committee on Making Best Use of the Agent Orange Reconstruction Model, Meeting 2, April 30–May 1, Washington, DC. Shaughnessy, C. A. 1991. The Vietnam conflict: “America’s best documented war”? The History Teacher 24(2):135–147. Sohn, M. W., N. Arnold, C. Maynard, and D. M. Hynes. 2006. Accuracy and completeness of mortality data in the Department of Veterans Affairs. Population Health Metrics 4:2. http://www.pophealthmetrics.com/content/4/1/2 (accessed September 5, 2007).
OCR for page 83
The Utility of Proximity-Based Herbicide Exposure Assessment in Epidemiologic Studies of Vietnam Veterans Stellman, J. M. 2007a. A data resource for the health and environmental consequences of the Vietnam War. Unpublished document submitted to the IOM Committee on Making the Best Use of the Agent Orange Reconstruction Model, July 31. Stellman, J. M. 2007b. Responses to IOM 091407. Unpublished document submitted to the IOM Committee on Making the Best Use of the Agent Orange Reconstruction Model, September 14. Stellman, J. M. 2007c. Separate modeling and reconstruction issues. PowerPoint presentation to the IOM Committee on Making Best Use of the Agent Orange Reconstruction Model, Meeting 3, June 13–14, Washington, DC. Stellman, J. M., and S. D. Stellman. 2003. Contractor’s final report: Characterizing exposure of veterans to Agent Orange and other herbicides in Vietnam. Submitted to the National Academy of Sciences, Institute of Medicine, in fulfillment of Subcontract VA-5124-98-0019, June 30, 2003. Stellman, J. M., R. Christian, S. D. Stellman, and F. Benjamin. 2003. Characterizing exposure to Agent Orange and other herbicides used in Vietnam: An epidemiologist’s guide to useful military records (version 1.1 June 26, 2003). In Contractor’s final report: Characterizing exposure of veterans to Agent Orange and other herbicides in Vietnam. By J. M. Stellman and S. D. Stellman. Submitted to the National Academy of Sciences, Institute of Medicine, in fulfillment of Subcontract VA-5124-98-0019, June 30, 2003. U.S. Marine Corps. 2000. Privacy Act of 1974; Notice to alter a system of records. Federal Register 65(199):60914–60916. VA (U.S. Department of Veterans Affairs). 2001. National Survey of Veterans, 2001. http://www1.va.gov/vetdata/docs/survey_final.htm (accessed August 28, 2007). Young, A. L., P. F. Cecil, Sr., and J. F. Guilmartin, Jr. 2004. Assessing possible exposures of ground troops to Agent Orange during the Vietnam War: The use of contemporary military records. Environmental Science and Pollution Research 11(6):349–358.