such as environmental exposures, that might be used to identify potential health effects associated with exposure to airborne hazards and other environmental agents, serve as a basis for future medical surveillance, and provide improved care for Gulf War veterans. The voluntary registry was modeled after VA’s Agent Orange and ionizing radiation registries (VA, 2017). As of July 2021, the initial startup costs for the registry as well as ongoing maintenance costs were unavailable (VA, 2021b). There is no sunset provision for the registry.

Population and Methodology

The Gulf War Registry covers all 1990–1991 Gulf War veterans who were deployed to the Southwest Asia¹ theater of military operations, and it also covers a substantial portion of post-9/11 service members and veterans (see Table 4-1). Post-9/11 eligibility is extended to service members and veterans who deployed to those Southwest Asia locations in support of OIF, OND, and other missions including Operation Inherent Resolve (OIR) in Syria and Iraq. Those service members and veterans who were deployed only to Afghanistan, Djibouti, Syria, or Uzbekistan, and not to any of the other eligible countries, waterways, or airspaces are excluded from participating in the Gulf War Registry.

To participate in the Gulf War Registry, an eligible individual requests an examination, which is scheduled by a VA environmental health coordinator (VA, 2017). Potential exposures to oil-well fire smoke or munitions destruction at Khamisiyah were strong predictors of registry participation (T. C. Smith et al., 2002). A veteran may participate in the registry regardless of whether he or she is enrolled in VA health care. Participation is a two-phase process. In the first phase, a veteran meets with his or her local VA environmental health coordinator and completes the Gulf War Phase I worksheet. This worksheet covers military status, deployment locations, deployment exposures (including smoke from oil-well fires, burning feces/waste, and diesel and petrochemical fumes; there is a separate section on exposure to sodium dichromate at the Qarmat Ali Water Treatment Plant in Iraq), birth data for any pregnancies and offspring, and any recommendations for consultations with other health care providers (VA, 2021i).

The second phase consists of a comprehensive health examination, including medical history, exposure history, laboratory tests, and physical examination, as well as follow-up examinations for new or unexplained health problems (Metzger-Smith et al., 2017). A worksheet is completed to indicate what tests and examinations were conducted, what further consultations were recommended, and what diagnoses were made (VA, 2022a).

Several standardized forms are used to record examination information, including a complete medical history that collects family, occupational, psychosocial, and behavioral (i.e., tobacco, alcohol, and drug use) histories; exposure history to possible toxic agents encountered as a civilian; and a review of body systems. Also captured are time of onset and intensity of symptoms or conditions, the degree of physical incapacitation, and details about any treatments for those symptoms or conditions. Basic laboratory tests are also administered. Examination data are recorded in the veteran’s health record as is done for any other medical examination. If the individual has not enrolled in VA health care, a record is created for him or her (VA, 2017). Participants are informed that the data collected through the Gulf War Registry will be used to track patient demographics and medical diagnoses and “for analysis to assist in budget planning for care of Gulf War Veterans and to generate hypotheses for future research or epidemiological assessments of Gulf War Veteran medical conditions and health status” (VA, 2017, p. 4). Individuals may complete the Gulf War Registry examination but still decide to opt out of having their information included in the Gulf War Registry database. Follow-up letters including the results of any tests or need for additional testing are sent to participants within 2 weeks of their completing the Gulf War Registry examination. When both phases are completed, the veteran is considered a participant in the registry.

If a follow-up medical examination is needed after the initial examination to evaluate complex or unexplained illnesses, it follows the Uniform Case Assessment Protocol. The Uniform Case Assessment Protocol addresses Gulf War veterans’ health concerns, particularly those with unexplained illnesses (VA, 2017). The follow-up examinations are not routinely conducted, but for those that are done, any new exposures and health conditions

___________________

¹ Specified as service in one or more of the following areas: Iraq, Kuwait, Saudi Arabia, the neutral zone (between Iraq and Saudi Arabia), Bahrain, Qatar, the United Arab Emirates, Oman, waters of the Persian Gulf, Red Sea, Arabian Sea, and gulfs of Aden and Oman (VA, 2017).

that are identified are documented in the veteran’s health record. VHA Directive 1325 (VA, 2017) provides guidelines for the minimum diagnostic workup using validated and available tests for the most frequent complaints experienced by Gulf War or OIF/OND veterans. The directive states that the “goal of the protocol is to identify previously unrecognized major diagnostic entities to explain symptoms commonly reported by veterans who have unexplained illnesses” (VA, 2017). If a diagnosis is still not made, the veteran may be given a referral to the nearest regional tertiary care center or a War Related Illness and Injury Study Center (VA, 2017). These centers offer comprehensive multidisciplinary consultation, including an environmental exposure assessment, usually via telehealth (VA, 2022c). Although not commonly conducted, veterans are eligible for further follow-up registry examinations if such an examination is requested by an environmental health clinician or if the individual reports new symptoms. During the follow-up examination, any new exposures and health conditions are documented in the veteran’s health record (VA, 2017).

The Gulf War Registry examinations are conducted and documented by VA clinicians, who counsel participants about the examination, its results, and its limitations. Data collected by the Gulf War Registry are regularly linked to VA electronic health records. The environmental health coordinator conducts administrative aspects of the examination and completes the registry worksheet in the VA web portal for Environmental Agent Service Registry data (VA, 2017). There is no template for the comprehensive registry health examination, although some clinicians have designed their own (VA, 2022a). About 400 VA clinicians may access the Gulf War Registry, and they include medical doctors, doctors of osteopathic medicine, physician assistants, nurse practitioners, and advanced practice registered nurses (VA, 2021b).

In 1994 the Department of Defense (DoD) implemented a two-phase clinical diagnostic program similar to that of VA’s Gulf War Registry called the Comprehensive Clinical Evaluation Program. DoD’s Comprehensive Clinical Evaluation Program closed on June 1, 2002 (DoD, 2002). Active-duty service members who wish to apply for a Gulf War Registry examination must make an email request to VA’s Health Outcomes Military Exposures (HOME; formerly the Post-Deployment Health Services), which coordinates Gulf War Registry examinations for active-duty service members. Results of the examination for an active-duty service member may be released to that service member’s military health facility with a signed release from the individual. Of the almost 700,000 1990–1991 Gulf War veterans who are eligible to participate in the Gulf War Registry, 181,914 had completed the initial examination by the end of 2016; 1,472 had completed follow-up examinations in the nearly 30 years since the Gulf War Registry was established (VA, 2021d). Since eligibility was expanded to include OIF and OND veterans, as of May 2022, 60,816 eligible post-9/11 veterans have participated in the Gulf War Registry; 3,589 registry participants served in both the 1990–1991 Gulf War and OIF/OND conflicts (VA, 2022b). In a VA analysis of 200,757 veterans who had joined the AH&OBP Registry, 18,555 (9.2%) were also in the Gulf War Registry. Of those who were in both registries, 13,951 had not received an AH&OBP Registry health evaluation, and 4,604 had received both an AH&OBP Registry health evaluation and a Gulf War Registry examination (VA, 2022b). The 2022 VA Office of Inspector General report noted, “Although some information on health conditions in the registries overlap, the Gulf War registry addresses a broader range of exposures and [health] conditions, while the airborne hazards and open burn pit registry focuses more on the length and proximity of exposure to specific types of hazards” (VA OIG, 2022).

VA stated that, at present, registry information is used on a veteran-by-veteran basis for claims (VA, 2021b). The committee has no further information on how many claims have used Gulf War Registry information.

Strengths and Challenges

The Gulf War Registry relies on voluntary participation, but its strength is that it requires a comprehensive health examination conducted by a VA provider for participation. The health examination offers diagnostic testing and specialty referral(s) as clinically indicated. Because the registry collects data from VA medical centers across the United States, the examination and clinical data have a low risk of bias from geographic location (Gray et al., 1998). Publications have examined patterns of health care use (i.e., Miller et al., 2006), specific diagnoses (i.e., Murphy et al., 1999; T.C. Smith et al., 2004), or clusters of symptoms and diagnoses (i.e., Hallman et al., 2003). Results from analyses of Gulf War Registry data are shared with key audiences in quarterly newsletters, however, data from the Gulf War Registry are not shared with the AH&OBP Registry (VA OIG, 2022).

nonrepresentative sample of the entire eligible population about whom there is little information, especially those who were in transit or on the street the morning of September 11, 2001 (Farfel et al., 2021). WTCHR enrolled 71,463 of the more than 400,000 people (about 17%) who were estimated to be eligible for the registry; coverage was as high as 34% for responders and 25% for residents (Murphy et al., 2007).

To date, there have been five survey waves. Enrollment and the baseline interviewer-administered survey (wave 1) were completed in 2003–2004. About 30 minutes were required to complete the survey, which focused on self-reported exposures, including proximity to the World Trade Center and duration of exposure (Cone and Farfel, 2011). Health outcomes chosen for surveillance included mortality, cancer, respiratory conditions, and mental health conditions, including posttraumatic stress disorder (PTSD), among others. More extensive information on World Trade Center exposures was collected in wave 2, such as use of personal protection equipment, evacuation and residential cleanup, and workplace conditions for area workers (Cone and Farfel, 2011). Information on health outcomes, social support, and unmet health care needs were also collected prospectively with four follow-up surveys: 2006–2007 (wave 2), 2011–2012 (wave 3), 2015–2016 (wave 4), and 2020 (wave 5) (NYC DHMH, 2021). Attrition from wave 1 was about 15,000 people (22% never responded to another survey); about 38% skipped at least one follow-up survey; and about 39% completed the first four survey waves (Farfel et al., 2021). Participants may also contact WTCHR to update their health diagnoses (NYC DHMH, 2022a).

Based on wave 3 responses, Yu et al. (2015) found that attrition from the WTCHR waves did not result in serious bias in associations between 9/11 disaster exposures and key health outcomes. Sixty-eight percent of wave 1 participants responded to wave 2, and 63% of them responded to wave 3 (Yu et al., 2015). Analyses of the WTCHR participants compared with nonparticipants showed that nonresponse bias did not have a strong effect on associations between 9/11 disaster exposures and health outcomes (Yu et al., 2015). The serial survey design allows for etiologic research on temporal relationships, including survival analyses, trends, and mediation and trajectory analyses (Farfel et al., 2021).

Use of WTCHR Data

WTCHR is the largest of four New York City–based cohorts of exposed survivors and first responders of the World Trade Center disaster (the other three are the open responder clinical cohorts of the Fire Department of New York, the general responder cohort, and the New York City Health and Hospital’s survivor cohort, all of which are under the World Trade Center Health Program). Additionally, the World Trade Center Health Program supports a national provider network for eligible responders and survivors outside of the New York City area (Santiago-Colon et al., 2020). Research on the World Trade Center disaster has resulted in over 940 publications as of 2020, including 291 funded by the World Trade Center Health Program (Santiago-Colon et al., 2020). Analyses detect new and emerging symptoms and conditions and their persistence. Most registry research has focused on characterizing the public health burden resulting from and the etiology of those health conditions associated with the disaster (Santiago-Colon et al., 2020). Many studies using registry data rely on matching with other administrative data resources, such as the National Death Index, state cancer registries, New York City vital records, and the Statewide Planning and Research Cooperative System for data on hospitalizations and emergency room visits in New York City (Cone and Farfel, 2011). To date, some notable research findings on registry participants include increased risks of asthma, stroke, gastroesophageal reflux disease, pulmonary fibrosis, PTSD, depression, certain cancers, and early retirement (WTCHR, 2019).

WTCHR is also a valuable information source about people who need treatment for health conditions associated with the disaster. In 2008 the New York City health commissioner recognized that the registry needed to do more than count the numbers of exposed people who had symptoms. He encouraged WTCHR to address smoking cessation because many of the World Trade Center exposures have potential effect on the lungs. As a result, registry staff began to do personalized outreach to help with smoking cessation among the registry cohort (Farfel et al., 2021). WTCHR serves as a source of information and statistics about participants with unmet health care needs and provides personalized outreach to people with such needs through its Treatment Referral Program, which began in 2009 (Welch et al., 2014). Participants are identified by the symptoms and conditions they report on registry surveys that could be related to World Trade Center disaster exposures. Dedicated staff, including

nurses and social workers, who are trained in motivational interviewing reach out to each enrollee. The staff aims to remove barriers and motivate people to enroll in the World Trade Center Health Program, partly by providing an application and by reporting on the applicants’ survey responses about exposures and symptoms. As of 2015, over 8,000 enrollees had been contacted through the Treatment Referral Program and more than 2,400 enrollees subsequently applied to the World Trade Center Health Program (Petrsoric et al., 2018).

WTCHR data and research also contribute to World Trade Center Health Program policies, for example, by identifying emerging exposure-related conditions that may be eligible for treatment and compensation. The WTCHR staff participate in scientific meetings and in responder and survivor steering committee meetings with community and labor advisors, city, and state agencies to share information as policies are developed for post-disaster care and compensation. The creation of the World Trade Center Health Program also established the Survivor Steering Committee and Responder Steering Committee in 2011–2012. The WTCHR’s public affairs liaison attends monthly Survivor Steering Committee meetings and the WTCHR director or medical director attends monthly Responder Steering Committee meetings.

Future research using WTCHR will consist of additional survey waves and will rely heavily on linkages to other data sources such as the World Trade Center Health Program, Medicare, Medicaid, and regional health organizations. Linking data to conduct research is gaining importance as registry participants’ age and attrition increases. The registry is actively reaching out to participants who were enrolled by their parents when they were children to consent them to remain in the registry as adults (Farfel et al., 2021). In addition, external researchers may submit proposals for new 9/11-related studies using de-identified survey data or request that the registry assist with recruiting enrollees into a researcher’s 9/11-related study. WTCHR participants have also been queried about their experiences following more recent adverse events such as Hurricane Sandy, which struck New York City in 2012 (Garrey et al., 2021), and the COVID-19 pandemic (NYC DHMH, 2022b).

Several papers have shown how the Treatment Referral Program might serve as a model for other post-disaster registries by contributing to long-term disaster response. For example, WTCHR served as a model for the Flint water registry in Michigan (Concannon et al., 2021). WTCHR research also contributes to occupational and emergency response science, including the use of personal protective equipment, hearing loss, PTSD, and high-rise building evacuation (Farfel et al., 2021).

Strengths and Challenges

Several factors contribute to the success of WTCHR. Flexible information technology systems and staff have allowed WTCHR to respond to new challenges quickly by tracing participants or developing and processing web or paper surveys. Because it is housed within the New York City Department of Health and Mental Hygiene, WTCHR is able to interact with and benefit from the department’s communications, information technology, and research and public health expertise. WTCHR has a good retention rate (about 60% for the wave 4 survey), which is attributed to the timely release of survey findings through publications, regular communication with registrants, maintaining an up-to-date dedicated website, and ensuring transparency through prompt communications with labor and community advisory committees (Farfel et al., 2021).

In addition WTCHR has several design features that enhance its usefulness. First, it has prospectively followed participants for more than 20 years, regularly collecting and monitoring information about physical and mental health and unmet needs for care through multiple survey waves and nested in-depth surveys and studies. Participants can also update their health information at any point. Second, the WTCHR cohort complements several other cohorts, which together have been used to develop an evidence base about the health impacts and care needs of the exposed population. Third, enrollment was restricted to a set time period (2003–2004) after which the registry became a closed cohort, which in contrast to open or continuous enrollment designs limits recall bias about exposures. Fourth, regular outreach and engagement with the community have enabled long-term follow-up (Cone and Farfel, 2011).

WTCHR leadership credits its success to several factors, the most prominent of which is long-term dedicated funding that is congressionally mandated through the James Zadroga 9/11 Health and Compensation Reauthorization Act (PL 114-113) (Concannon et al., 2021; Farfel et al., 2021). Such funding ensures resources to maintain a

diverse group of participants, information technology infrastructure, long-term engagement, and rigorous research. In particular, WTCHR supports a team of public affairs, information technology, and administrative staff as well as about 14 research scientists. WTCHR uses numerous approaches to communicate its activities and findings and to receive information from participants, stakeholders, and the public (Concannon et al., 2021), which registry staff believe has contributed to continued participation and interest in the registry (Farfel et al., 2021).

Although WTCHR is a model exposure registry in many ways, it is not without its limitations. Because participation in the registry is voluntary and derived from lists of potentially exposed people as well as self-identified people, some populations may be under-represented, and selection bias may influence the results, although some analyses do account for the source of enrollment. However, analyses assessing the effects of attrition have found that there is little evidence of bias between exposures and health outcomes among WTCHR participants (Yu et al., 2015). The five survey waves have been conducted using multiple modes (paper, web, and computer-aided telephone or in-person interviews), which means that the potential for mode effects needs to be considered in any analyses of the data (Cone and Farfel, 2011); analyses have attempted to adjust for biases introduced by the source of enrollment and mode of survey (Jurek and Maldonado, 2016).

Exposure information may be particularly affected by recall bias as there are few objective measurements that allow for validation, and the exposure and the initial health outcomes information were collected 2–3 years after the disaster. Some health information can be validated by other sources such as mortality, cancer, and hospitalization data (Cone and Farfel, 2011; Li et al., 2016). Furthermore, the 9/11 World Trade Center disaster was a single, discrete event compared with the multiple exposures experienced by service members in the Southwest Asia theater, which were highly variable in their magnitude, duration, and frequency. Similarly, the eligible population for WTCHR is closed and limited to people who lived, worked, or went to school near the World Trade Center and rescue and recovery workers who responded to the incident, whereas the eligible population for the AH&OBP Registry is constantly changing and covers active-duty, reserve, and National Guard service members and veterans who served over a 30-year period (see Chapter 3). Similarities between the two registries include their reliance on voluntary participation and self-reported health and exposure information.

INDIVIDUAL LONGITUDINAL EXPOSURE RECORD

The Individual Longitudinal Exposure Record (ILER) is an information technology effort created by a DoD/VA collaboration that links and compiles, in a consistent manner, data on locations, events, hazard exposures, and health information spanning a service member’s military career. ILER was approved in January 2013 in response to White House Presidential Review Directive-5, A National Obligation: Planning for Health Preparedness for and Readjustment of the Military, Veterans, and Their Families after Future Deployments, and to DoD Instruction 6490.03: Deployment Health (effective June 19, 2019; DHA, n.d.). VA and DoD intend that ILER will:

• Improve health care for veterans and service members with health outcomes related to deployment exposures;
• Identify cohorts (based on specified location, date, and event or common experience) for epidemiologic research studies;
• Identify exposed individuals for medical, occupational, and environmental health surveillance; and
• Provide information on likely exposures to support veteran disability and compensation applications and to assist in VA disability evaluation and benefits determinations (VA, 2021e).

ILER is neither a database nor a data warehouse;² rather it is an on-demand, web-based compiler of exposure-related data. It can link an individual to occupational and environmental exposures experienced while in garrison and during deployments to health data spanning his or her military career in a single record (see Figure 4-1). Although most of these data are linked at the time of a search request, some data sources, such as the periodic

___________________

² A data warehouse periodically imports data from databases and systems, formats and curates them into an organized structure, and stores the processed data so that they are readily available for analysis.

occupational and environmental monitoring summary (POEMS) reports, are stored within ILER because of their size and the relatively large amount of network consumption necessary each time these reports are requested. As of July 2021, ILER provided access to more than 50 million records describing the military service, exposures, or health of approximately 10.3 million service members and veterans.³ Several records may be associated with an individual. See Table 4-2 for a summary of ILER data sources, including the dates for which data are available, the frequency of updates, approximate number of records, and description of each source (location, exposure, or health information).

Because ILER functions as a compiler, it depends on the individual source or database developers and managers to perform quality control assessments. It does not validate, clean, or edit data, so errors or incomplete data that are found within a source or database are maintained in ILER (VA, 2021e). Similarly, ILER does not prioritize sources or attempt to reconcile discrepancies between them. For example, self-reported exposures (as well as other data) in the AH&OBP Registry can be accessed and compiled in ILER along with exposure information based on Defense Occupational and Environmental Health Readiness System–Industrial Hygiene (DOEHRS-IH) reports, and the user determines how the information from each source should be used (VA, 2021f).

The primary users of ILER are VA and DoD clinicians and VA claims and benefits administrators, who can conduct individual-level searches by entering as much information on the individual as possible, such as name, date of birth, last four digits of Social Security number, and electronic data interchange personnel identifier (a unique record identifier) or DoD identification number. Currently, ILER uses a two-step query process. In step 1, a VA or

___________________

³ Personal communication, Larry Vandergrift, Defense Health Agency, chief information office/deputy assistant director, Information Operations. July 13, 2021.

DoD user enters all available identifiers for an individual using the Joint Legacy Viewer;⁴ ILER then determines whether exposure information matching that individual is available. In step 2, a PDF of an individual’s exposure summary is generated if available (VA, 2021f).

ILER is not designed to communicate or conduct outreach activities with service members or veterans, and these individuals cannot access ILER; however, they can access their exposure and health information through a query performed by a DoD or VA clinician. PL 116-283 § 9105 requires that VA provide veterans with read-only access to their printable ILER documents through a portal on a VA Internet website. VA and DoD representatives indicate that future development plans include online access for individual service members and veterans. Exposure information contained in ILER may assist veterans who are applying for VA benefits (VA, 2021e).

Location Information

DoD’s Defense Manpower Data Center (DMDC) is the central source for identifying, authenticating, authorizing, and providing information on all individuals who currently serve or previously served in the military (active-duty, reserves, or National Guard). Approximately 35 million personnel records exist in DMDC’s Person Data Repository (DMDC, 2021). ILER compiles declassified DMDC records from January 2001 to the present with continuous updates. As of July 2021, there were 24,512,112 DMDC deployment history records available in ILER, representing 10,101,368 unique individuals.⁵ DMDC data are used to identify an individual’s location, both in garrison and during deployments, as well as internationally and within the United States and territories. As required by DoD Instruction 6490.03: Deployment Health (DoD, 2019), DMDC contains 28 individual location data elements, including location start and end dates, latitude and longitude coordinates, military occupational specialty, assigned and attached unit identification codes, component code, and operation name (Shuping et al., 2020).

The quality and completeness of location information varies, with limited electronic data available prior to 2000 (VA, 2021f). Some deployments may be missing if they were to classified locations (until the location is declassified; some locations, such as special operations, may never be declassified), were miscoded, or lack a GPS geocode annotation. Although each service is required to report the location (by latitude and longitude) of its members daily, such reporting may be logistically difficult during combat operations, potentially leading to miscoded or unreliable data (VA, 2021e). VA stated that ILER “is very sensitive but not specific. It is a system designed for moving forward vs. capturing information from the past” (VA, 2021e).

ILER also has a map feature tied to Google Earth that allows the user to look at the weather conditions or other geographic satellite data for a deployment. For example, satellite data down to the minute may be accessed, if available, for a particular exposure event on a specific date and time (VA, 2021f).

Exposure Information

ILER compiles several sources of exposure data which vary with respect to the dates they cover (see Table 4-2). ILER indicates that the date range for exposure data from DOEHRS-IH is January 1, 2006, to the present, although some exposure data may be available as far back as 2001 (VA, 2021f); however, as retrospective exposure information is added, data for earlier years may become available. Most exposure data are at the unit level, although the sparse individual-level exposure data from dosimeters, wearable devices (e.g., silicon wristbands), and blast overpressure gauges are being added to ILER. Most exposure data are for occupational and environmental conditions at a specific geographic location, such as a military base or deployment site, as documented in

___________________

⁴ The Joint Legacy Viewer was developed in 2013 as a joint effort by VA and DoD to further support interoperability between the two departments. Its browser-based, graphical user interface provides an integrated, read-only view of electronic health record data from VA, DoD, and Veterans Health Information Exchange community partners within a single application, and it eliminates the need for VA and DoD clinicians to access disparate viewers. The graphical user interface retrieves clinical data from several native data sources and systems, then presents it to the user, each corresponding to a clinical data domain (VA, 2019).

⁵ Personal communication, Larry Vandergrift, Defense Health Agency, chief information officer/deputy assistant director, Information Operations. July 13, 2021.

DOEHRS-IH and POEMS reports (VA, 2021e). Much of the exposure information in ILER is routinely collected by environmental and industrial hygiene surveys that include garrison industrial hygiene reports and routine air, soil, and water sampling assessments (based on DoD regulations) conducted when establishing, monitoring, and closing base camps (VA, 2021e).

The primary source of exposure data in ILER is the Defense Health Agency’s DOEHRS-IH, which captures military occupational and environmental health risk data and actively tracks biologic, chemical, physical, and other health hazards. For example, DOEHRS-IH contains specific area exposures, such as the Khamisiyah weapons depot demolition and oil-well fire locations in the 1990–1991 Gulf War (VA, 2021e). Garrison reports in DOEHRS-IH are generally more reliable than deployment site reports for individual exposure information (VA, 2021f). Industrial hygiene reports will indicate if any exposure exceeded the acceptable level, and those that do are highlighted in red in ILER compilations. DOEHRS-IH data are updated daily with approximately 200 new reports and can be queried in ILER as uploaded (VA, 2021f).

The completeness of exposure information in DOERHS-IH reports varies according to the site surveyed and the known hazards at the time of the survey. While all bases should be assessed in the same manner and environmental sampling should follow DoD industrial hygiene regulations that include quality assurance checks (and the laboratories that analyze the collected specimens also have quality assurance procedures in place), differences in sampling and frequency exist among industrial hygiene reports. The time between an exposure and when that information is available in ILER depends on several factors, including when the samples were collected, shipped, processed, and analyzed (including quality assurance checks) and when the data were uploaded into DOERHS. Typically, there is a greater lag in uploading data collected during deployments than in uploading data collected during routine garrison assessments. Ideally, all service members’ exposures over the course of their military service would be captured, but in practice this information is likely to be incomplete because not all required industrial hygiene surveys are completed, not all completed surveys are digitized and added to DOEHRS-IH, and not all surveys are linked to an individual at a particular location and time (VA, 2021e,f). With the most recent release of ILER v1.2.3 (July 2021), DoD estimates that there are more than 3 million individuals who have at least one exposure-related report associated with at least one of their deployments.⁶ See Figure 4-1 for an example of an individual exposure summary.

DOEHRS-IH reports include the publicly available POEMS reports (APHC, 2018) which provide environmental monitoring data for specific locations and times (Shuping et al., 2020). POEMS reports summarize general population-level exposure estimates (air, soil, water, endemic diseases, heat/cold stress, noise, and unique occupational or military hazards [e.g., depleted uranium]) in relation to potential acute and long-term adverse health effects to inform veterans, military public health officials, health care providers, and decision makers of potential health risks at a location. The quality, completeness, and variability of the POEMS reports depend on the availability of environmental and personnel sampling for each site, which varies by base size, troop population, duration of operation, and base resources to conduct exposure monitoring. However, a 2019 review of POEMS reports concluded that “inconsistent collection, assessment, and presentation of available data seriously weakens the potential utility of the POEMS files as the primary source of information on toxic exposures” (Williams and Fahy, 2019). Even with robust monitoring, it is difficult to draw conclusions about an individual’s exposures and health risks from POEMS reports (Heller et al., 2019; Williams and Fahy, 2019). Part of the challenge with ILER is integrating current exposure information, such as DOEHRS-IH data, with other data sources, such as National Aeronautics and Space Administration satellite observations. Combining these data sources may yield a more complete picture of an individual’s exposures and help refine POEMS. Mapping individual location data to exposure sources may improve the confidence of data in ILER.

Additional exposure information in ILER can come from other routine environmental surveys, such as exposure incidence reports that document specific events that can be linked to individuals. These reports may not have had a known exposure or defining exposure event. Sampling is determined by industrial hygiene and environmental

___________________

⁶ Personal communication, Larry Vandergrift, Defense Health Agency, chief information officer/deputy assistant director, Information Operations. July 13, 2021.

health procedural guides, subject-matter experts, associated working groups, and known or suspected exposures of concern. For area-level sampling, exposures are not extrapolated to the individuals in that area. Other ILER exposure sources include the Military Exposure Surveillance Library (which was a distinct source before being migrated and integrated into DOEHRS-IH in 2017; see Table 4-2); blast and overpressure pilot data; and the six VA exposure registries, including the AH&OBP Registry. Individual exposure summaries are presented in the Joint Legacy Viewer (VA, 2021f).

Health Information

Although ILER compiles selected health information from DoD and VA electronic health records, it is not designed or intended to be a health record itself. ILER’s advantage from a health effects research perspective is that DoD and VA health records can be linked to individuals’ military service locations and their potential exposures at those locations. A long-term goal of ILER is to have an individual’s exposure summary reflected in the electronic health record in the interoperable DoD (Military Health System GENESIS) and VA (Millennium) health records systems when the interoperability feature is completed (VA, 2021f).

Information from VA electronic health records is available in ILER through the Veterans Affairs Informatics and Computing Infrastructure (VINCI) and the Corporate Data Warehouse (CDW), but there are no plans to add VA electronic health records directly into ILER. As of April 2021, over 210,000 completed AH&OBP questionnaires were available in ILER (VA, 2021f). ILER will eventually contain self-reported health information from other VA exposure registries, including the Gulf War Registry, Ionizing Radiation Registry, Toxic Embedded Fragments Registry, and Depleted Uranium Registry (VA, 2021f).

DoD health information compiled in ILER includes Medical Data Repository data and pre- and post-deployment health assessments from the Armed Forces Health Surveillance Branch. Electronic health records information from the Military Health System’s GENESIS, the combined DoD and VA Health Artifact and Image Management Solution, electronic deployment health assessments, and DOEHRS–Hearing Conservation data are expected to be available in ILER in 2021–2023 (VA, 2021f). Clinical encounter records accessible through ILER include about 1.7 million records from the DoD’s Comprehensive Ambulatory/Professional Encounter Record and 11.4 million records from the Theater Medical Data Store; records from October 2012 to the present are available, and they are updated monthly. The Defense Medical Surveillance System (DMSS; see Table 4-2), through the Armed Forces Health Surveillance Branch,⁷ is a compiler of service branch information to support specific research requests. At present DMSS provides ILER with annual periodic health assessments, pre-deployment health assessments, and post-deployment health assessments from September 2012 to the present, accounting for approximately 9.45 million records (an individual may have multiple records). An automatic data feed is in development for this source with a goal of backfilling data from before September 2012. The Defense Enrollment Eligibility Reporting System is listed as an ILER source, but it serves only as an identifier to link to the Medical Data Repository. There are no plans to link biospecimen collections, such as the DoD Serum Repository, with ILER (VA, 2021e). At this time, external health information sources, such as Medicare, Medicaid, and private insurers, are not available in ILER and there are no plans to include such sources in it (VA, 2021e).

ILER was considered fully operational for clinical use at the end of 2020. VA environmental health coordinators and clinicians are the target users of ILER. Beginning January 1, 2021, each VA medical center is required to have at least one person trained in and able to access ILER (Shuping et al., 2020). As of October 2021, 118 of the 131 VA medical centers had at least one ILER-trained clinician or environmental health administrator, and 330 VA

___________________

⁷ The Armed Forces Health Surveillance Branch (AFHSB) operates the Defense Medical Surveillance System (DMSS), a continuously expanding relational database that documents military and medical experiences of service members throughout their careers. DMSS is the central repository of medical surveillance data for the U.S. Armed Forces, and it contains up-to-date and historical data on diseases and medical events (e.g., hospitalizations, ambulatory visits, reportable medical events, HIV tests, and casualty data) and longitudinal data on personnel and deployments. AFHSB routinely publishes summaries of notifiable diseases, trends of illnesses of special surveillance interest and field reports describing outbreaks and case occurrences in the Medical Surveillance Monthly Report. Through DMSS, AFHSB provides the sole link between the DoD Serum Repository and other databases (DHA, 2022).

staff had been trained in its use.⁸ A small number of VA researchers also have ILER accounts. DoD directed all services to identify the initial users of ILER and to encourage its use by the DoD clinical and research communities in order to improve clinical care and solicit feedback to enhance ILER functionality. As of October 2021, DoD had 96 registered ILER users.⁹ Each user is required to request an account, receive standardized training, and follow HIPAA (Health Insurance Portability and Accountability Act of 1996) and cybersecurity requirements. Users must log on to ILER at least once every 35 days to maintain an active account (Shuping et al., 2020).

The ILER dashboard (see Figure 4-2) is customizable and contains system notifications, lists of data sources (including dates available and last refresh), resources for exposure and general reports, and frequently asked questions. Searches can be initiated based on an individual, a specific type of exposure or event, or by location and date range. An ILER report is not intended to be a health diagnosis or to replace one, nor is it a definite confirmation of exposure; rather, it is an indication of likely exposures based on dates and location.

For example, a clinician seeing an individual for breathing difficulties could search ILER and see an individual’s deployment history, any associated POEMS reports (which include risk estimates for exposures and known acute and long-term health effects), and post-deployment health assessments (available from September 2012). Based on the POEMS report, ILER would list potential exposures and summarize their likely sources (e.g., air, water, soil), frequency, duration, what was sampled, the means of sampling, when the sample was taken, analytes, and personnel groups involved; pictures might be included, if available. The distance between an individual and a hazard can be generalized by data source. A broad source such as POEMS can put an individual in the vicinity (within about 100 meters) of the hazard, whereas an incident report will put them in a particular building or location with confirmed or probable exposures within about 10 meters. Individual monitoring device information will also be shown if available, and it would be expected to provide the most accurate exposure information

Although complete health records are not available in ILER, clinical encounter summaries (dates, type of encounter, International Classification of Diseases [ICD]-9 and ICD-10 codes, and disposition) for an individual are displayed. Establishing better links between health conditions (represented by ICD-10 codes) and known exposures is an ongoing long-term effort by VA and DoD (Shuping et al., 2020). Clinicians can also view postdeployment health assessments and reassessments (from DMSS) to identify behavioral risk factors, such as alcohol and tobacco use, and self-reported deployment exposures (VA, 2021e). Additionally, if the individual has completed the AH&OBP Registry questionnaire, the clinician can access his or her responses.

Continued Development for Use in Research

One ILER goal is to provide researchers with a tool to allow greater access to hazard and exposure events information and to more efficiently develop research cohorts. Such research cohorts may inform clinical care or identify areas where further investigation is needed. Over the next few years, the first studies using ILER data to characterize exposures of military cohorts are expected to be published. As stated in the most recent National Academies report on this topic, “These initial efforts are likely to uncover strengths and weaknesses in the application of [ILER] as a tool in epidemiologic studies that should be identified and disseminated throughout the research community” (NASEM, 2020, p. 209).

ILER will use a four-phased process to support research initiatives (VA, 2021f). Phase I (2020) defines what happened (exposure information) and where it happened (location data). Phase II, which had full functionality in October 2021, identifies individuals at a location when a specific exposure occurred (i.e., search by location and refine by date). Phase III, also available as of October 2021, permits searching by an exposure; for example, ILER can produce a list of every location that had exposures to chromium or an open burn pit.¹⁰ A user can filter the exposure information provided by DOEHRS-IH by a specific route (e.g., skin contact, inhalation, digestion), source (e.g., soil, dust, sand, animal, fire, smoke), environmental media (soil, air, water, other), priority of the exposure

___________________

⁸ Personal communication, Dr. Eric Shuping, director, Post-9/11 Era Environmental Health Program, Health Outcomes Military Exposures, VA. October 12, 2021.

⁹ Personal communication, Dr. Eric Shuping, director, Post-9/11 Era Environmental Health Program, Health Outcomes Military Exposures, VA. October 12, 2021.

¹⁰ The committee notes that searches by burn pit exposure would be limited to those acknowledged by DoD as having certain conditions, such as size and operation during certain specified time frames.

(used to indicate a level of confidence that the exposure poses a risk [low, moderate, high, and extremely high] to the mission or requires medical surveillance of exposed personnel), and start and end dates of exposure (Shuping et al., 2020)—to generate a roster of individuals who were associated with a particular incident or exposure event. The roster may then be used to begin to assemble a cohort of service members with a common exposure or experience. However, as of early 2021 information on how many service members had a particular exposure was not readily available. A researcher or clinician would need to go to each individual item and click on the exposure of interest. That search capability is planned for the future and would be dependent on the availability of a data mining tool that can search across sources and reports. ILER does not have such analytic capabilities (although some capability may be added in the future); rather, the compiled data can be exported to a researcher’s preferred system for analysis. Phase IV of ILER development will include the capability to search by health effects, which will allow researchers to identify where individuals with a particular health effect had been located and to examine common exposures for those locations (VA, 2021f).

Major updates to ILER are made about every 3 months, but other updates are released as a new or improved capability becomes functional (VA, 2021f). Once a capability is built, it is tested and vetted through the Defense Information Systems Agency before being released to users. Following each major ILER update and release, a user-testing event is conducted, involving approximately 35–65 participants, to ensure that the added capabilities have value (VA, 2021f).

Although ILER was considered fully operational for clinical use at the end of 2020, it is expected to have full operating capability (including use for building research cohorts) by September 2023. The current focus is on developing, expanding, and refining ILER’s capability to search by exposure. This capability would allow researchers to compile cohorts by location, time, hazard, or event and, eventually, by health effects (VA, 2021e).

ILER’s predominant challenge is compiling retrospective electronic data from specific sources. For example, post-deployment health assessment data are only available beginning in 2012 because prior to 2012 each service had its own method of collecting and storing this information. There are plans to integrate more data collected before 2012, but such data must be available in a compatible electronic format. Because ILER is not a data warehouse, databases and information systems to be compiled by ILER must be active and maintained; if a system is no longer active, ILER cannot access it.

Cost

The operation and continued development of ILER are supported by funding streams from both DoD and VA, with DoD, as the project business owner, responsible for most of the cost. Cumulative funding from 2014 through 2020 totaled $10.17 million from DoD and $8.79 million from VA.¹¹ After ILER is completed, it is estimated that its annual maintenance and sustainment costs will be approximately $2.5–$3.0 million.¹²

Strengths and Challenges

Over the last several decades, many reports—including several National Academies reports—have recommended enhanced collaboration and data sharing between VA and DoD; ILER is an example of greater collaboration between the two agencies. PL 117-168, signed in August 2022, requires that VA establish a working group to make “recommendations to the Secretary [of Veterans Affairs] on corrections needed in the Individual Longitudinal Exposure Record to better reflect veterans and dependents” who may have been exposed to toxic substances during active duty and that VA and DoD provide briefings to Congress yearly on the quality of the databases that are compiled by ILER. The committee recognizes that ILER is an ambitious and challenging undertaking. Its major benefit is providing a platform for accessing disparate data sources from both VA and DoD, but its usefulness as a tool for toxic exposures research, conducting population heath surveillance, improving health care, and aiding

___________________

¹¹ Personal communication and spreadsheet, Dr. Eric Shuping, director, Post-9/11 Era Environmental Health Program, Health Outcomes Military Exposures, VA. October 12, 2021.

¹² Personal communication, Larry Vandergrift, Defense Health Agency, chief information officer/deputy assistant director, Information Operations. October 13, 2021.

claims adjudication has yet to be determined. ILER’s application to each of those areas is discussed in the following chapters.

ILER has several strengths as a data collection system compared with a voluntary exposure registry. First, it systematically collects exposure information and clinical records from multiple sources for all nonclassified service members and veterans (including active duty, reserves, and National Guard). This approach minimizes the recall and selection biases that can affect surveys and questionnaires that are based on self-reported information. Second, because ILER automatically pulls data from existing sources, there is no burden on service members or veterans, which maximizes participation. Third, ILER’s near-real-time access to multiple data sources enables the user to conduct timely queries, in contrast to working with cross-sectional or retrospective data that may not reflect current circumstances.

Although ILER is continuously adding new data sources and expanding coverage of existing sources, it does not include all service members and veterans, nor does it include all service members and veterans who are eligible to participate in the AH&OBP Registry. ILER will continue to be limited by the varying quality and completeness of the disparate sources on which it depends, especially for individual-level information.

EPIDEMIOLOGIC COHORTS OF SERVICE MEMBERS AND VETERANS

Information on the exposure and health outcomes of service members and veterans who may have been exposed to burn pit emissions in the Southwest Asia theater may come from many sources in addition to those discussed earlier in this chapter. Those sources include several epidemiologic studies conducted by VA and DoD, such as the Millennium Cohort Study, the National Health Study for a New Generation of U.S. Veterans, and other large cohort studies. Although these studies cannot entirely substitute for the information collected by the AH&OBP Registry, they do provide a wealth of similar and relevant exposures and health information on service members and veterans deployed to Southwest Asia. This section is not intended to be a compendium of every epidemiologic cohort study conducted on exposures to open burn pit emissions or other airborne hazards exposures, but rather an overview of these major epidemiologic cohort studies.

Millennium Cohort Study

Study Origins

In response to concerns about the health effects of veterans deployed to the 1990–1991 Gulf War, the Institute of Medicine¹³ recommended that DoD conduct prospective epidemiologic research to evaluate the impact of military exposures, including deployment, on long-term health outcomes of service members (IOM, 1999). In response, in 2001 (prior to the September 11, 2001 terrorist attacks), DoD launched the Millennium Cohort Study—a large, population-based, longitudinal epidemiologic study of U.S. service members from all branches and components (Chesbrough et al., 2002; DoD, 2021a; T. C. Smith et al., 2009). This ongoing study is designed to follow veterans who used VA health care after separation from the military.

Population and Methodology

Service members from all branches (Army, Navy, Air Force, Marine Corps, and Coast Guard) and all service components (active duty, reserve, National Guard) were randomly selected to participate (Chesbrough et al., 2002; Porter et al., 2020; B. Smith et al., 2008). Women, reserve, and National Guard service members, along with personnel with past deployment experience, were oversampled to ensure adequate power for statistical analyses (B. Smith et al., 2008, 2009). As of 2016, 54% of the cohort had deployed at least once during their military service and 33% were reservists or National Guard members (DoD, 2021c); by April 2021, more than 201,000 service

___________________

¹³ As of March 2016, the Health and Medicine Division of the National Academies of Sciences, Engineering, and Medicine continues the consensus studies and convening activities previously undertaken by the Institute of Medicine.

members have taken part in the study of whom 65% were ever-deployed, 73% were separated from the military, 1.5% were deceased, and 31% are female (Rull, 2021).

The vast majority of study participants (over 90%) have enrolled in the VA health system (Boyko, 2021). A 2009 analysis of participants concluded that “[Millennium] Cohort members well represented the U.S. military, that health prior to enrollment did not influence participation, and that cohort questionnaire data were reliable” (B. Smith et al., 2009, p. 1434).

The Millennium Cohort Study was designed to have multiple groups (termed panels) of participants who were enrolled every 3 to 4 years and followed up at multiple times—on average, every 3 to 5 years—at the same time that the baseline assessment is conducted for the next panel (Rull et al., 2020). For the first panel, 214,338 individuals randomly selected from DoD rosters in October 2000 were contacted and invited to participate; of those, 77,047 participants enrolled beginning in July 2001—a 36% enrollment rate (Ryan et al., 2007). Four additional panels were recruited in 2004, 2007, 2011, and 2020. A fifth panel was completed in 2021 and enrolled 58,173 participants (Millennium Cohort Study, n.d.). Initial enrollment rates for panels 2, 3, and 4 were also low, ranging from 20% to 28%, although each panel’s follow-up survey response responses were higher (see Table 4-3). The study will follow participants until at least 2068 (Rull et al., 2021).

The study’s self-administered questionnaire for the first panel consisted of more than 450 questions and could be completed online or via hardcopy and returned by mail; the median time for a participant to complete the online questionnaire was 29 minutes (B. Smith et al., 2007a). The questionnaire covers occupational exposures, health behaviors, health conditions, health care access and use (outpatient and inpatient visits), the use of complementary and alternative medicines, DoD and VA facility use, vaccinations, military life, and other health and well-being concerns at each 3-year interval (DoD, 2021b; White et al., 2011). Standardized instruments were incorporated whenever possible and included the Primary Care Evaluation of Mental Disorders Patient Health Questionnaire, the Medical Outcomes Study Short Form-36 for Veterans, a VA Gulf War survey of specific war-time exposures, the CAGE questionnaire for the detection of alcohol abuse, and the PTSD Checklist–Civilian Version (Ryan et al., 2007). Free-text fields are also available for information not captured by the survey questions (Leleu et al., 2011). Some questionnaire data are linked to administrative records and military databases to collect or verify information in areas such as mortality (Hooper et al., 2010), demographics (T. C. Smith et al., 2007), medical histories (B. Smith et al., 2008), vaccine history (B. Smith et al., 2007b; LeardMann et al., 2007), health care use (Wells et al., 2008; White et al., 2011), and deployment status. These linkages allow investigators to study additional associations, adjust for potential confounding, or factor additional covariates into the analyses of health outcomes. The

initial questionnaire did not ask participants about exposure to burn pits specifically or about airborne hazards in general, except for one question about exposure to “occupational hazards requiring protective equipment, such as respirators or hearing protection” (B. Smith et al., 2007a, p. 1350). At present there are no links between the Millennium Cohort Study database and the information collected by the AH&OBP Registry, although it is expected that the Millennium Cohort Study will eventually be compiled in ILER as is currently done for the AH&OBP Registry (Rull, 2021).

To determine whether Millennium Cohort Study enrollees were representative of the target population, almost 350,000 individuals, who were invited to participate in the study between 2004 and 2008, were assessed (Horton et al., 2013). Study enrollees were more likely than nonenrollees to have deployed either during or immediately before a study enrollment period. Active-duty participants were also more likely than nonenrollees to have had an outpatient medical visit in the year before being invited to join the study. The study found enrollees were less likely than nonenrollees to have had a hospital stay of 2 days or longer in the previous year or to have a documented mental disorder.

Use of Cohort Data

To date, more than 120 papers using data from the Millennium Cohort Study have been published by the team and their collaborators (Belding et al., 2022). Research findings are used to inform prevention efforts, interventions, clinical training, and policy (Rull, 2021). However, only a few publications examine airborne hazards exposures encountered in the Southwest Asia theater (Rivera et al., 2018; Rull, 2020; B. Smith et al., 2009, 2012). Exposure data have been limited by the classification of deployment locations and records of where and when burn pits were operational. As of 2021, the Millennium Cohort Study has used DMDC data to identify 22,300 Air Force and Army service members who were deployed within 3–5 miles of three bases in Iraq with known burn pits between 2003 and 2008: Joint Base Balad, Camp Taji, and Camp Speicher. About 3,500 of these service members were deployed within 3 miles of one of these burn pits. These service members with potential exposure to a burn pit in Iraq were included in panel 1 (enrolled 2001–2003) and panel 2 (enrolled 2004–2006) (Rull, 2021). Thus far, only a few studies of burn pits using the Millennium Cohort Study and DMDC exposure data for these service members have been published (Jones et al., 2012; Powell et al., 2012; Sharkey et al., 2016; B. Smith et al., 2012). Additional information about burn pits and other exposures may be available and be linked to the Millennium Cohort Study when ILER becomes fully operational (Rull, 2021).

The Millennium Cohort Study has additional components, such as the family assessment survey that may contribute to or corroborate information on service members’ exposures. The study database can also be linked to a number of complementary data sources (T. C. Smith et al., 2011) including health care data from DoD and VA, and VA benefits data. Approval for data sharing among DoD, VA, and the Millennium Cohort database was received in 2018, and the merging of files is just beginning (Boyko, 2021). Links to VA health and benefits data expand the Millennium Cohort Study’s ability to follow participants and increase the information available to researchers (such as prescription and laboratory data). An ongoing collaboration between the Millennium Cohort Study and VA is focused on comparing health measures between survey self-reports and electronic health records among participants receiving VA health care (Rull, 2021). Other collaborations are focused on mental health, homelessness, and mortality.

Millennium Cohort Study data have been used extensively for research on health outcomes and health care use for both service members and veterans, and numerous studies have been published. For example, T. C. Smith et al. (2014) compared the prevalence of chronic multisymptom illness in veterans of Iraq or Afghanistan with its prevalence in 1990–1991 Gulf War veterans. The cohort surveys also provide insight into behavioral health and use of preventive services. Matsuno et al. (2020) used study data to examine cervical cancer screening among U.S. military service women between 2003 and 2015. Welsh et al. (2014) compared the responses of Millennium Cohort Study participants regarding depression screening with their responses on the DoD post-deployment health assessment and found the participants were more likely to report symptoms of depression on the study survey than on the assessment. The risk of suicide among current and former service members was evaluated using data from three waves of the Millennium Cohort Study (LeardMann et al., 2013), and the results were used to inform suicide

prevention programs in both DoD and VA. Other Millennium Cohort Study research has examined alcohol use and binge drinking and found that among participants deployment was associated with a greater risk of drinking problems, especially among younger service members and National Guard or reservists (Jacobson et al., 2008).

Current research efforts are examining self-reported respiratory outcomes including chronic bronchitis, sinusitis, asthma, emphysema, persistent or recurrent cough, and shortness of breath. Future research will include assessing tobacco use as an effect modifier for deployment-related environmental exposures, the effects of specific military occupations on health outcomes, the frequency and severity of asthma symptoms, and all-cause mortality and cancer outcomes (Rull, 2021). Additional studies are being conducted to validate the self-reported health conditions using VA health care data (Rull, 2020). Although the Millennium Cohort Study is able to link to VA health care databases, not all veterans receive their health care at VA, which makes it difficult to assess all health care use and validate health outcomes in veterans who are not enrolled in or who do not receive all their health care from VA.

Communications

The Millennium Cohort Study maintains a website (https://www.millenniumcohort.org) which may be accessed by the general public, study participants, and researchers. The website contains facts about the study and separate information for study participants and researchers. Communication tools for participants include several infographics on specific health topics designed for lay audiences, postcards that are sent to all participants each Memorial Day and Veterans Day to thank them for their continued participation in the cohort, occasional newsletters (i.e., 2011, 2012, 2014, 2015, and 2019), a list of study publications (including a link to a report on 20 years of research findings for Millennium Cohort Study participants [DoD, 2021c]), information on the survey instruments and what they assess, and a page on VA’s Million Veteran Program. Similar materials are available for researchers, including study topics (i.e., topics covered by the questionnaire such as health conditions and potential exposures, including to burn pits or smoke), press coverage, presentations and briefings given by study staff, and a comprehensive list of publications using the cohort. The website currently contains a notice to spouses of service members about the Millennium Cohort Family Study and provides several infographics on findings from that study. Other than the postcards and the periodic surveys sent to participants, there does not appear to be frequent or consistent outreach or communication activities from the study researchers.

Strengths and Challenges

The Millennium Cohort Study offers several advantages for the study of deployment-related airborne hazards and burn pits health effects research. First, it provides a large and representative sample of military service members from all service branches and service components. It oversamples women and service members who were deployed, and is the only large-scale prospective study of current and retired military members who served during a time of high operational tempo. The large sample size increases the power of the analyses and may result in a more generalizable representation of the applicable military and veteran population. Second, the Millennium Cohort Study can link to numerous DoD and VA databases containing information on inpatient and outpatient health care, mortality, prescriptions, vaccinations, births and infant health, vital status records, serum and other biological samples, exposures, and contingency and deployment data (T. C. Smith et al., 2009). Third, its prospective and longitudinal design allow for both longitudinal and serial cross-sectional investigations (Rull, 2020). Because the study began before post-9/11 operations, it allows for the prospective study of service members deployed to Iraq or Afghanistan. The longitudinal study design with follow-up surveys may be used to identify changes in health outcomes over time, and multiple health outcomes from one or more exposures may be examined concurrently (T.C. Smith et al., 2011). Furthermore, the observational prospective design reduces recall bias commonly seen in retrospective or cross-sectional studies and allows for the ability to determine the temporal sequence in exposure and outcomes.

There are also several challenges to using the Millennium Cohort Study. Like all prospective studies, the Millennium Cohort Study is susceptible to bias resulting from the loss of study participants (unable to contact, death, or refusal to participate) as well as bias between survey panels and in the overall study, but analytic methods are

available to minimize these. Response rates to follow-up panels have been fair to good (see Table 4-3); loss to follow-up increases the potential for selection bias. Consistent follow-up surveys and the linking of medical and exposure databases and deployment records with questionnaire responses help to minimize the recall and reporting biases typically seen in self-report survey studies. As noted earlier, although the amount of recall or reporting bias in the Millennium Cohort Study has been investigated extensively and found to be low, it cannot be eliminated completely.

Nevertheless, standard biases that affect many observational studies such as self-report, self-selection, generalizability, and loss-to-follow-up persist. For example, digital military medical records are available only for active-duty military personnel or those personnel seeking care while on activated status, making it difficult to link survey data with military databases for service members who have separated from the military (Rivera et al., 2018). The fact that the surveys are conducted on average every 3 years has serial cross-section implications, as it may not be possible to pinpoint the time between exposure to a hazard and the onset of a symptom or illness. Additionally, because there are only nine questions on hazardous exposures and responses are self-reported and validated by deployment status or proximity, the Millennium Cohort Study does not have accurate information on the magnitude, frequency, or duration of the exposures. As with most large databases, the variety of information sources on active-duty service members and veterans, such as electronic health records, personnel files, pharmacy records, and deployment status and locations, each of which will have some level of data errors, may affect the strength of any associations between exposure and health outcomes.

National Health Study for a New Generation of U.S. Veterans

The population-based cohort National Health Study for a New Generation of U.S. Veterans (NewGen) began as a pilot study in 2009 with data collection concluding in January 2011. It was initiated by VA to assess the health of post-9/11 veterans, help identify VA health services that veterans may need, and improve the quality of VA health care (VA, 2018, 2021g). A total of 60,000 veterans were randomly selected; 30,000 veterans who were deployed to Iraq or Afghanistan between October 2001 and June 2008 were compared with 30,000 veterans who served elsewhere during the same time. The sample population of deployed veterans was selected from DMDC, and members of the comparison group were selected from the VA/DoD Identity Repository (VADIR) database. Of the 60,000 veterans selected and contacted, 20,563 participated in the study (response rate 34.3%; 55.1% deployed) with representation from each service branch and component; female veterans were oversampled to make up 21.1% of respondents (Eber et al., 2013). Although the response rate was low, the overall number of survey participants was more than 20,500 with representation from all branches and components of the military (NASEM, 2020).

Data collection included a 72-item self-administered questionnaire, which collected information on potential environmental exposures (including smoke from oil fires, burning trash and feces, and diesel, kerosene and petrochemical fumes; there is no question about burn pits although respondents are asked to list any other exposures they consider harmful), health behaviors, health conditions, general health, and health care use (including use of VA facilities). No biological samples were collected; however, medical records were reviewed for about 2,000 veterans to validate certain health conditions. The study participants included veterans who accessed VA health care and those who did not, which is more generalizable to the total OEF/OIF and OEF/OIF-era veteran populations than studies of only veterans who use VA health care services (Eber et al., 2013).

As with any cohort study that uses self-report questionnaires, this study was likely subject to recall bias in the survey responses, leading to both over-reporting and under-reporting of exposures and health conditions. The extent of recall bias may be related to time since deployment, and it would be expected to be greater for veterans who deployed many years before the survey. An analysis of the NewGen data by VA researchers examined misclassification and nonresponse biases (Yoon et al., 2013). To account for these issues in the NewGen data as well as the stratified sampling design, publications by VA researchers have weighed the results and presented frequencies in subsequent publications. At least 13 publications have analyzed NewGen data on topics ranging from health-related quality of life among deployed and nondeployed veterans (Díaz Santana et al., 2017), to calculating the population prevalence of respiratory diseases (Barth et al., 2014, 2016), positive PTSD screens (Dursa et al., 2014), and reproductive outcomes (Katon et al., 2014, 2017). Data on smoking and other behaviors allow models

to address these factors where appropriate, particularly when looking at health outcomes. Although the study took into account the number of deployments for OEF and OIF service members, it does not contain data on the total time spent in the military or the start date of the participants’ entry into the military; both omissions are likely to lead to inaccurate exposure assessments for airborne hazards (Barth et al., 2014, 2016). Furthermore, the survey asked about exposures during military service in general, not specifically during OEF/OIF deployments. Analyses cannot adjust for health conditions—especially respiratory and cardiovascular conditions—that occurred before or after joining the military or before or after deployments, which may result in spurious associations between reported exposures and health outcomes. There have been no new publications on the NewGen cohort since Schneiderman et al. (2018).

Other VA Cohort Studies of Respiratory Health

VA researchers have assembled several cohorts and conducted epidemiologic studies focused on health effects (particularly respiratory outcomes) associated with deployment to the Southwest Asia theater in general or particular exposures that service members and veterans encountered while on deployment. These include:

• Comparative Health Assessment Interview Study, a survey of over 15,000 OEF, OIF, and OND-era veterans and 4,600 civilians that was intended to help VA better understand the effects of military service, deployment, and combat on the health and well-being of these veterans. Analyses of study data published to date have focused on the risk of suicide attempts and ideation (Hoffmire et al., 2021), but no analyses of respiratory health outcomes have been published (Schneiderman, 2019; Schneiderman et al., 2020).
• Study of Active-Duty Military for Pulmonary Disease Related to Environmental Deployment Exposures (STAMPEDE) is a series of three prospective studies that examined deployment-related respiratory symptoms and pulmonary disease among U.S. service members returning from Iraq, Afghanistan, Kuwait, or Qatar in support of OIF and OEF. Although the three studies have overlapping information, the questionnaires and overall methods used for each differ. In brief, STAMPEDE I asked 50 active-duty military personnel to complete a post-deployment questionnaire of new-onset respiratory symptoms to determine potential etiologies (Morris et al., 2014). STAMPEDE II was a prospective study of active-duty soldiers, 843 of whom completed both the pre- and post-deployment questionnaires and an examination that focused on lung function (Morris et al., 2019). STAMPEDE III consisted of 380 active-duty and retired military personnel who had deployed to Southwest Asia for a minimum of 6 months and had an onset of chronic respiratory symptoms after deployment. Participants completed a post-deployment questionnaire, answered additional questions taken from the AH&OBP Registry questionnaire, and underwent a comprehensive history and physical examination that included pulmonary function testing (Morris et al., 2020; Powell et al., 2020). The three STAMPEDE studies are detailed in Respiratory Health Effects of Airborne Hazards Exposures in the Southwest Asia Theater of Military Operations (NASEM, 2020).

Because these cohort studies are generally stand-alone descriptive analyses, they are not summarized in detail, nor are they considered viable alternative information sources for the AH&OBP Registry.

National Health Survey of Gulf War Era Veterans and Their Families

This study has been detailed in several reports as part of the National Academies’ Gulf War and Health series (e.g., NASEM, 2016). In 1994, PL 103-446 directed that VA conduct a population-based study to identify the incidence and nature of health problems in veterans of the 1990–1991 Gulf War. In response, VA has conducted a set of studies that began with a longitudinal survey of 30,000 veterans known as the National Health Survey of Gulf War Veterans and Their Families (NHS). It was designed to be representative of the nearly 700,000 Gulf War veterans and the 800,680 veterans who served in the military between September 1990 and May 1991 but who were not deployed. The mental and physical health status of a randomly selected cohort of 15,000 Gulf War veterans from each branch of the military, including active-duty, reserve, and National Guard, was compared with 15,000

of their nondeployed military peers, with some oversampling of women (approximately 20% of study population), National Guard (25%), and reservists (33%). Three survey waves have been conducted: wave 1 in 1993–1995 had an overall response rate of 70% (Kang et al., 2000), with physical examinations in 1999–2001 (Eisen et al., 2005); wave 2 in 2003–2005 had a response rate of 34% (Kang et al., 2009); and wave 3 in 2012–2013 had a response rate of 50% (Dursa et al., 2016). The characteristics of those who did not respond to the mailed survey were also examined. The self-administered structured health questionnaire contained a 48-symptom inventory (somatic and psychological symptoms) and asked about sociodemographic factors, chronic medical conditions, chronic fatigue syndrome-like illness and unexplained multisymptom illness, mental health, health behaviors, general health, functional limitations, women’s health, environmental exposures, the use of medical services, and VA health care use and satisfaction (Kang et al., 2000).

As part of wave 1, medical records were obtained for a random sample of 4,200 respondents to validate self-reports of clinic visits or hospitalizations within the previous year. Medical record reviews verified more than 90% of self-reported reasons for clinic visits or hospitalizations. The third phase was a comprehensive medical examination and laboratory testing of a random sample of 2,000 veterans drawn from the deployed and era groups (Eisen et al., 2005). In the analysis of wave 1 data, exposure–symptom relationships were not conducted, but rather the percentage of veterans who reported each of 23 environmental exposures and 9 vaccine or prophylactic exposures (such as to pyridostigmine bromide) was calculated (Kang et al., 2000). The five most common environmentally-related exposures—reported by more than 60% of survey participants—were diesel, kerosene, or other petrochemical fumes; local food other than that provided by the armed forces; use of chemical protective gear; smoke from oil-well fires; and burning trash or feces (Kang et al., 2000).

Several derivative studies of the NHS and its successor research efforts have been reviewed as part of the National Academies’ Gulf War and Health series of reports (IOM, 2006, 2010; NASEM, 2016, 2020). Three studies analyzed respiratory symptoms collected from wave 1 (Eisen et al., 2005; Kang et al., 2000; Karlinsky et al., 2004); they are described in detail in Respiratory Health Effects of Airborne Hazards Exposures in the Southwest Asia Theater of Military Operations (NASEM, 2020).

APPLICATION TO THE COMMITTEE’S STATEMENT OF TASK

In its Statement of Task, the committee was asked to “look at other means or methods to obtain similar or better information by VA and to achieve the goals of burn pit exposures research. Such means might include the Individual Longitudinal Exposure Record (ILER) and the Millennium Cohort Study.” Thus, this chapter presents an overview of the various data sources (surveys, studies, registries, and compilers) to provide background for the more detailed discussions of them in the following chapters. Where appropriate, these data sources may serve as models of how the AH&OBP Registry might achieve some of its stated goals, as potential alternatives to the current AH&OBP Registry, or as examples of how specific registry goals and activities have been accomplished by other governmental organizations. As not all of the data sources are applicable to all of the goals of the AH&OBP Registry as described in chapters 5 through 9, only those that have particular relevance to a chapter’s topic are considered in that chapter. This chapter demonstrates that there are challenges in expecting any one data source to fulfill all DoD and VA requirements for assessing health military exposures and potential health consequences in veterans.

REFERENCES