___________________

¹ The Centers for Disease Control and Prevention describes epidemiologic surveillance as “the ongoing and systematic collection, analysis, and interpretation of health data in the process of describing and monitoring a health event. This information is used for planning, implementing, and evaluating public health interventions and programs” (Klaucke et al., 1988, p.1). Monitoring, on the other hand, is the collection of data without intended action or response.

taken in response to a predetermined signal in the surveillance information. Identifying a signal (a change in the pattern of data over time or in comparison to a reference group) in adverse outcomes or events through surveillance can be a stimulus for follow-up or for embarking on etiologic research. Identifying what will constitute a signal and what will be done in response to one must be established early in the design of the surveillance process. Surveillance data may provide a signal to initiate research on potential etiologic relationships. When direct action or intervention is not necessary or intended, such as when the purpose is to establish a baseline level of exposure or track the incidence of an outcome, then monitoring—that is, data gathering without an intended action—may be more appropriate than surveillance.

Population Health Surveillance

Several federal agencies conduct population health or disease surveillance, including the Centers for Disease Control and Prevention (CDC), the National Institutes of Health (NIH), and the Agency for Healthcare Research and Quality. CDC is at the heart of surveillance for general population health in the United States, and it partners with numerous federal, state, local, and tribal agencies to accomplish this function. As of 2018, CDC maintained more than 100 surveillance systems for infectious diseases, noninfectious health conditions, occupational diseases, mortality, and risk factors and exposures (CDC, 2018). The Department of Defense (DoD) Armed Forces Health Surveillance Division conducts health surveillance on all service members (MHS, 2022).

A variety of methods are used to collect population-level information about exposures, risk factors, and health. Some surveillance systems rely on annual surveys administered to a sample of the population, such as the Behavioral Risk Factor Surveillance System (BRFSS), the Medical Expenditures Panel Survey, the National Health and Nutrition Examination Survey (NHANES), and the National Health Interview Survey (NHIS). Other systems collect data continuously. These include the National Death Index, which passively collects data about mortality from death certificates, and NIH’s Surveillance, Epidemiology, and End Results Program, which actively collects data about cancer cases. Examples of surveillance systems or surveys used to examine population health and well-being are listed in Table 6-1.

TABLE 6-1 Examples of National Population Health Surveillance Systems or Surveys

NOTE: AHRQ = Agency for Healthcare Research and Quality; CDC = Centers for Disease Control and Prevention; DoD = Department of Defense; NIH = National Institutes of Health.

SOURCE: Adapted from Appendix D: Table of Surveillance Systems in NASEM, 2019.

of stroke, pulmonary fibrosis, and respiratory disorders and symptoms (Santiago-Colón et al., 2020; WTCHR, 2019). Survey information about emerging health conditions is shared with registry enrollees and the public to inform their health decisions, with medical providers to guide the care they provide to exposed individuals, and with health researchers to help direct future research efforts (NYC DHMH, 2021).

Challenges for Conducting Environmental Health Surveillance

Compared with surveillance for infectious diseases, injuries, and occupational health, surveillance for health outcomes associated with environmental exposures poses additional challenges. These challenges include the need to identify and measure the specific environmental hazards, the difficulty of linking exposure to a hazard with a health outcome, and objectively assessing relevant health outcomes. These challenges are described next.

Identifying Specific Environmental Exposures

As discussed in the Chapter 5 section on Exposure Assessment of Adequate Quality, identifying and characterizing exposures to a specific airborne hazard is difficult because of the many challenging requirements, including the need to obtain adequate information on the composition, duration, frequency, and concentration of the airborne hazard. Furthermore, many airborne hazards are combinations of chemicals, and exposure to a hazard such as smoke from a burn pit may be only one of many concurrent exposures, including those in the ambient air, which may have its own hazardous components.

In some cases surveillance can be conducted using an established biomarker for the exposure. One example of a registry that effectively uses a biomarker for surveillance is the Beryllium-Associated Worker Registry (see Box 6-3). However, there are no known exposure biomarkers specific to burn pit emissions, although researchers are looking for biomarkers that are indicative of but not specific to such exposures (Mallon et al., 2019). This lack of biomarkers is an important challenge for current surveillance of and research involving service members and veterans.

The relationships between health effects and environmental hazards are not always clear or well-established, a fact that complicates the design of environmental health surveillance. (The same could be said of etiologic

research; a point discussed in Chapter 5). A strong association, as determined by magnitude of effect or precision of the estimate, between a health effect and exposure to a single environmental hazard (the relationship between mesothelioma and asbestos, for example) is relatively uncommon. Many health conditions develop from any number of factors, including environment, genetics, and lifestyle, which may interact to cause or exacerbate the condition. Furthermore, the current understanding the biologic effects and health outcomes associated with many environmental hazards or mixtures of hazards is incomplete (Hernandez et al., 2019; Thacker et al., 1996; Thomas et al., 2018). This is particularly relevant for deployed service members who are exposed to many different hazards that vary by frequency, location, and composition (e.g., emissions from burn pits vary depending on what is being burned) (IOM, 2011).

Linking Exposures to Health Outcomes

Exposure information collected for purposes other than environmental health surveillance or research may be insufficient for quantifying those exposures, may lack temporal identifiers (i.e., information on when the exposure occurred as well as the time between occurrence and report of those exposures), may not be representative of the entire exposed population, may have data quality issues, or might be missing information on other contributing or relevant factors (e.g., tobacco use). Because it is difficult to use existing datasets for applications for which they were not intended, multiple data sources may need to be combined (such as mortality data, recurrent surveys like NHANES, and environmental monitoring data) to link exposure and health outcomes (Thacker et al., 1996). Box 6-4 provides an example of combining data sources for occupational cancer surveillance.

Public Perception and Objective Assessments

The third challenge for environmental health surveillance (and etiologic research, as described in Chapter 5) is public attention, which can make it difficult to objectively assess the potential relationship between an exposure and a health outcome (Thacker et al., 1996). Public perception of the risks associated with an exposure may result in selective participation in a registry or survey and increased reporting of exposure or disease by those who believe that they have been adversely affected. This selective participation can result in spurious signals or

a failure to detect signals (Fleisher and Kay, 2006; Galea and Tracy, 2007). In some circumstances, legal and financial considerations may influence registry participation.

CURRENT USE OF THE AH&OBP REGISTRY FOR SURVEILLANCE

VA has a long history of conducting health surveillance through a variety of means. Surveillance of the long-term health impacts of military experiences is the responsibility of the Health Outcomes Military Exposures (HOME) office within the VA Office of Patient Care Services. Through its Epidemiology and Environmental Health programs, HOME focuses on military environmental and occupational health; its epidemiology program conducts surveillance and research studies on veterans’ health and health care. HOME administers a variety of programs, including exposure registries, cohort studies, and longitudinal surveys that focus on different service eras (VA, 2016, 2021a), and it also has oversight of the Airborne Hazards and Burn Pits Center of Excellence (AHBPCE), all of which seek to improve health care practices and policies for veterans who receive care from the Veterans Health Administration (VHA). HOME has also created a Post-Deployment Surveillance Report that uses VA health care data to describe the incidence and prevalence of a wide range of health conditions for users of VA health care, including results stratified by deployment era (VA, 2021b); these reports were published in 2015 and 2021.

As part of these activities, VA prepares and makes available online periodic descriptive reports and factsheets presenting select data from the AH&OBP Registry.² These reports and factsheets describe participation over time and cumulatively; demographic and military characteristics of the participants, including countries of deployment, cumulative days of deployment; self-reported exposures; select self-reported health conditions; the numbers of optional health evaluations completed by month and cumulatively; and VA health care use. The reports do not cover every health condition listed in the self-assessment questionnaire, do not show changes of reported health conditions over time, and do not attempt to examine patterns of health experiences in relation to exposure or deployment (VA, 2021c). Such reports of descriptive frequencies and statistics provides documentation, but without an overarching goal or action, cannot be considered surveillance. VA may regularly create reports for internal purposes, for example, by linking to medical records to provide additional health information; however, none are publicly released or otherwise communicated to veterans or other stakeholders.

The committee reviewed published abstracts and full text articles on the AH&OBP Registry data (described in Chapter 5), none of which focused on surveillance, but rather addressed disease etiology by providing descriptive statistics of characteristics and disease risk factors among participants. Although not explicitly population health surveillance, information on the frequency of participants’ symptoms or health conditions may help detect trends or identify emerging health care needs at a population level despite registry participants being self-selected. For example, analyses of AH&OBP Registry participants found a high prevalence of insomnia, sleep disordered breathing (snoring), and sleep apnea (Jani et al., 2020a,b). However, given the self-selected and nonrepresentative nature of the registry population, in which individuals with health problems may be more likely to participate than individuals without any health problems (see Table 3-4 for the number of health outcomes reported for each individual), simply providing descriptive prevalence counts for respondents has limited value for surveillance. Only presenting outcome frequencies or descriptive statistics, as is done in these abstracts, does not lead to specific actions to address the outcomes or potential underlying issues, as would be expected of a basic surveillance system (criterion 4).

Several ongoing activities have isolated features of a surveillance system and are described below, but none constitute an explicit attempt to meet all the criteria for a surveillance system. For example, some VA analyses and reports linked other data sources to the AH&OBP Registry. One VA-directed analysis linked registry participants (through August 2019) with the Joint VA/DoD Mortality Data Repository and the National Death Index to obtain the fact of death and cause of death for participants who died between 1997 and 2017. Of the 183,957 registry participants examined, 797 were known to be deceased as of 2019; 440 (55.2%) of them had an unknown cause of death. A descriptive analysis compared demographic and self-reported health information of all registry participants

___________________

²https://www.publichealth.va.gov/exposures/burnpits/reports.asp. No reports are available for 2016, 2017, or 2018. The most recent report is from quarter 1, 2021.

with those who were deceased, stratified by known and unknown cause of death (VA, 2020a). Because the cause of death was unknown for more than half of the decedents and because no comparisons were made to determine whether death rates in general and by cause differed between this group and the total population of deployed service members and veterans, linking the AH&OBP Registry with mortality datasets for surveillance was of limited use. Links to external data sources do not make the registry itself capable of prospective or longitudinal data collection, and therefore it is not possible to analyze trends over time or detect health conditions with long latencies.

Regular, periodic analyses of the AH&OBP Registry data might reveal patterns of morbidity and mortality among registry participants; however, in addition to issues with the quality of the registry data, the absence of longitudinal data on participants makes it impossible to analyze time trends or detect health conditions with long latencies. In addition to the mortality analysis, VA has looked at respiratory morbidity associated with burn pit exposure or country of deployment (VA, 2020b). These reports appear to be stand-alone analyses and are not part of any ongoing efforts to monitor for changes in health, morbidity, mortality, or VA health care use, nor to make comparisons with other populations to detect unexpected differences in health outcomes.

Similarly, quarterly reports of data from the Gulf War Registry and other VA exposure registries are used to inform health care providers, veteran service organizations, individual veterans, and others. Although these reports are still being produced, their publication is not systematic or regular. These presentations of descriptive results can, at most, offer a list of common or frequent exposures and health outcomes experienced by participants, but they are not associated with a comparison group, a particular exposure, or other measures of severity that might alert VHA clinicians about health conditions to be considered during medical encounters or VHA epidemiologists about priority areas for conducting etiologic research. Either of these functions would be expected actions of a population health surveillance system.

ASSESSMENT OF THE AH&OBP REGISTRY AS A POPULATION HEALTH SURVEILLANCE SYSTEM

In the sections below the committee considers to what extent the AH&OBP Registry currently meets each of the four criteria of a surveillance system. Where the criteria are not met, the committee discusses whether and how the registry might be modified to satisfy them. To reiterate, the four criteria of a public health surveillance system are (CDC, 2012; German et al., 2001):

1. Deciding on the overarching goals to address the prevalence or incidence of a potentially harmful exposure or occurrence of a disease in a defined population;
2. Regularly collect, analyze, and interpret the data that are acquired to look for predefined “signals” or unexpected occurrences of outcomes or association patterns that would motivate some type of action;
3. Disseminate data, results, and information to key audiences; and
4. Seek to improve health as a result of the collection and dissemination of that information.

Overarching Goals

The first criterion of a public health surveillance system is to decide on the overarching goals to address the prevalence or incidence of a potentially harmful exposure or occurrence of a disease in a defined population. The AH&OBP Registry’s purpose was determined by legislation (PL 112-260) and implemented by VA. The legislation specifies that the registry include any information VA “determines necessary to ascertain and monitor the health effects of the exposure of members of the Armed Forces to toxic airborne chemicals and fumes caused by open burn pits.” The AH&OBP Registry was not created as a population health surveillance system, but rather was developed as a mechanism to collect selected self-reported information about exposures and health from service members and veterans who had deployed to the Southwest Asia theater. Although neither the legislative language nor VA’s description of the registry’s purposes require VA to specifically conduct surveillance, the committee believes that this function is implied by both the legislative language and VA’s actions. In practice, one purpose of the registry’s ascertaining and monitoring health effects is to “study and treat conditions associated with exposure

to toxic airborne chemical and fumes caused by open burn pits” (PL 112-260 § 201). Thus, detecting a signal in the reported health effects might trigger or redirect research efforts to further diagnose or treat those health outcomes as discussed in the next section. Thus, the AH&OBP Registry has an overarching goal as directed by public law.

Regular Collection, Analysis, and Interpretation of Data

The second criterion of a registry for population health surveillance is to regularly collect, analyze, and interpret data to identify signals such as emerging or trending health issues. The AH&OBP Registry is unable to fulfill this criterion because as it lacks the following factors:

• regular data collection (data for each registry participant is collected only once, cannot be corrected or updated after the questionnaire is submitted, and the exposure questions are not granular);
• regular data analysis (there is no appropriate comparison group,³ the data sensitivity is poor, and data are not systematically analyzed but only summarized); and
• regular data interpretation (lack of a comparison group and an ever expanding population of registry participants makes it difficult to detect signals in health outcomes, and the lack of longitudinal data makes it difficult to identify signals or trends in the reported health outcome data).

Location and exposure information are useful for surveillance only if they can be linked to health data and there is a plan and process for actively monitoring and analyzing the data in real time. Linking data sources is not sufficient by itself for conducting surveillance activities, though it can provide a valuable component of a surveillance system. Electronic health records have several advantages for surveillance compared with actively querying individuals. These advantages include the lack of a burden on those who are in the system, the availability of a large population if the health data resource is comprehensive, the existence of high-quality data on health outcomes, and the fact that information is available rapidly because the data are entered electronically (Birkhead et al., 2015). Information about diagnoses, laboratory tests, prescriptions, and inpatient and outpatient visits is objectively ascertained from an external source as opposed to the self-reported information in the AH&OBP Registry. However, self-reported data can include experiences and symptoms that are not always available from administrative data.

The lack of an appropriate comparison group hinders the identification of changes in the expected occurrence of a health outcome in the registry population. A comparison group might be either a referent group (e.g., unexposed or less exposed veterans or service members) or part of a larger prospective cohort, such as all individuals deployed to a particular base. As shown in Table 3-3, nearly all registry participants (98.5%) report at least one exposure to the airborne hazards listed in the registry questionnaire. While there are variations in the frequency and intensity of the exposure, it is not possible to accurately distinguish these exposure differences solely on the basis of registry information.

It is unclear to the committee what algorithm or process VA might use to detect a signal in the health outcome data in the registry without knowing the incidence or prevalence of the health outcome in a comparison group. It is possible that VA could simply identify the number of self-reports of rare or critical health outcomes such as idiopathic pulmonary fibrosis or constrictive bronchiolitis among registry participants. It is evident from the published abstracts and the few internal reports of registry data linked to VA health care participants (VA, 2020a,b,c) that VA is using registry data to look for elevated risks of conditions or disease among registry participants, but it has not indicated how or what metric is being used to determine what constitutes a signal of an elevated risk and cannot do so without knowing the risk among the entire eligible population.

It is not possible to detect health signals simply by comparing registry participants who enrolled at different times (e.g., comparing early and late participant populations as described in chapters 3 and 5) as there are many variables that affect each population. For example, the eligible population for the registry is constantly expanding

___________________

³ As explained in Chapter 3, the vast majority of registry participants report exposure to airborne hazards. In reality some participants were exposed less frequently and to a lesser degree than others, but the registry is not designed to capture such differences, nor can deployment records reliably identify gradations in exposure. Therefore, comparisons among participants using those relatively less exposed as a comparator group are not possible.

as more service members are deployed or redeployed and new countries or deployment locations are made eligible for the registry, such as Syria, Egypt, and Uzbekistan. Data analysis is affected by data quality, i.e., the completeness and validity of the data in the system (see Box 6-2). All data on exposures and health outcomes (symptoms, diagnosed conditions, functional status, and behaviors) are self-reported and thus subject to information and recall biases, among other limitations (see Chapter 5 for a discussion on validating self-reported exposure and outcome data). The second issue with interpreting registry data as part of a population health surveillance system is the lack of a mechanism for prospective, ongoing data collection on participants. VA informed the committee that “there are no immediate plans to follow up with the entire cohort with additional questions at this time” (VA, 2021d).

Registry participants cannot update or amend their information once it has been submitted, nor are the data in the registry corrected in response to findings from the health evaluation (see Chapter 7 for a discussion of the registry health evaluation). As of March 29, 2022, VA has added a capability to the registry that allows participants to reenter the registry and add new eligible deployment segments (VA, 2022b)—a major change—although the committee has no information on how many participants have added new deployment segments. Thus, although the registry has data on newly enrolled participants, it does not contain any new or revised health outcome data on existing participants. Comparing early with late participants (see chapters 2 and 3) also does not provide information on the temporal course of health outcomes for individual participants or across the registry population. Furthermore, linking the AH&OBP Registry to external data sources, such as the National Death Index, cancer registries, and DoD and VA medical records, also would not provide prospective follow-up of participants.

The sensitivity of any would-be surveillance system is compromised by the lack of detailed information on exposures of concern and the restriction of health conditions to those included in the questionnaire. For example, participants diagnosed with rare diseases that were not included on the questionnaire are not captured and cannot be monitored by the registry. Similarly, the questions on exposures are nonspecific and standard definitions are not provided. For example, there is no indication of how a burn pit is defined (e.g., size, location, contents), nor are there questions on other deployment exposures, such as blasts from improvised explosive devices or refueling operations (these exposures are only asked about for anytime during military service, not for each deployment). Therefore, it is difficult to identify specific exposures to a particular substance or experience and link these exposures with health outcomes. Furthermore, the quality of the data may be affected by the lack of a time limit on when service members or veterans can enroll in the AH&OBP Registry relative to their deployments or exposures, which also hinders the registry’s capability for detecting signals. For example, a veteran of the 1990–1991 Gulf War who deployed in 1991 can complete the questionnaire in 2021—30 years after deployment. The majority of registry participants completed the questionnaire more than 2 years after the end date of their last deployment (see Table 5-2). A longer time between an event and its recording in the questionnaire can result in less accurate information.

VA’s analyses of AH&OBP Registry data are of little use for public health surveillance, given their lack of comparison populations, the poor representativeness of the eligible population, and the variable timing of when the analyses are conducted relative to when participants completed the questionnaire. Without the ability to detect signals in the exposure or health outcome data, the interpretation of these results is limited to describing the participants and summarizing their data.

Disseminate Data, Results, and Information

The third criterion of a surveillance system—disseminating data, results, and information to key audiences—is partially achieved by the registry. VA shares some registry data in reports and communications, including electronic newsletters with the public via its website (VA, 2021c). Electronic newsletters are also sent to anyone (registry participants or not) who has signed up for VA’s listserv on this topic. These communication activities are detailed in Chapter 9. The AH&OBP Registry–focused reports include information on the number of enrolled participants and their demographic distribution, military characteristics, and deployment information. The reports also include information on self-reported exposures and medical conditions among newly enrolled participants as well as on the number of health evaluations conducted and the distribution of VA health care use among registry participants. An advantage of the AH&OBP Registry is that it captures as participants not only those who use DoD or VA health care systems but also those who do not. Disseminating or sharing information in a purely descriptive manner,

without any analysis or interpretation, does not constitute surveillance, given the lack of a well-defined referent group or denominator (i.e., total number of exposed individuals). VA infrequently disseminates descriptive data from and about the registry with key audiences through targeted notices and periodic reports published on VA’s website, and therefore this criterion is not met.

Improve Health on the Basis of Registry Information

The fourth criterion of a surveillance system is to generate knowledge to improve the health of the target or eligible population. A key factor in determining whether this criterion is met—and also whether registry data are useful for informing decision making—is the timing of data collection, analysis, and interpretation (criterion 2). It is not clear how (or if) the data collected by the AH&OBP Registry have been or could be used to improve the health of the population of service members and veterans who served in the Southwest Asia theater. In principle, the AH&OBP Registry information could be used to help guide DoD’s environmental health programs, for example, by identifying hazardous exposures that might be mitigated, although this function is presumably conducted by other means, such as the Armed Forces Health Surveillance Division or the Defense Occupational and Environmental Health Readiness System. Although it does not qualify as population health surveillance, the registry may offer opportunities for health outcome monitoring and medical monitoring for certain service members and veterans for the purpose of improving clinical care—for example, by facilitating early disease detection via health evaluations and through the work of VA’s AHBPCE, as discussed in Chapter 5. The AH&OBP Registry does not appear to collect data that are or could be used to improve the health of service members and veterans who were deployed to the Southwest Asia theater at the population level.

Use of the AH&OBP Registry to Conduct Health Monitoring

The committee defined monitoring activities as those used to gather data, without intended action or intervention, or to determine the prevalence of a health outcome in the selected population. Monitoring in this sense implies a time component since health effects may be acute or chronic, immediate or latent, and may develop, resolve, or change over time. The only way to assess these factors is to periodically collect updated information on them, which cannot currently be done by the registry. Continuing to allow eligible individuals to enroll in the registry will not meet or serve a monitoring function as the core issue of individuals not being followed over time will still not have been addressed. Moreover, VA informed the committee that “there are no immediate plans to follow up with the entire cohort with additional questions at this time” (VA, 2021d).

If registry information were matched to DoD and VA health records for participants who use VA health care, the registry could potentially serve as a monitoring mechanism. With other changes (such as removing questions that are not relevant to exposure assessment and linking to DoD and VA health records) and by allowing for follow-up of individuals, the AH&OBP Registry might become a basic monitoring mechanism to document specific exposures, health experiences or outcomes, and other concerns (e.g., an open text field to capture a participants’ concerns not listed elsewhere in the questionnaire) of registry participants (as it is not representative of the full eligible population; see chapters 2 and 5). However, the committee questions the value of collecting this information (and spending participant and VA staff time and monetary resources) if the information will not or cannot be used other than to generate periodic summaries of participants’ responses to the registry questionnaire.

ALTERNATIVE SOURCES FOR POPULATION HEALTH SURVEILLANCE OR MONITORING

There are a number of ways to ascertain the health of service members and veterans who have been exposed to open burn pits and airborne hazards, and each method has its strengths and limitations. The AH&OBP Registry provides a standardized mechanism to create an ongoing roster of service members and veterans who are experiencing or are concerned about health effects that may be related to their deployment exposures to airborne hazards in Southwest Asia. However, creating a list or roster of concerned individuals is not a function of a population health surveillance system.

The committee considered other approaches that have been used to conduct health monitoring and population health surveillance, particularly of military populations. Some VA and DoD research efforts (such as the Millennium Cohort Study) may fulfill some of the requirements of such a system through longitudinal data collection from representative samples and by building on the other national service member and veteran epidemiologic surveys described in Chapter 4. But epidemiologic studies cannot serve as alternative surveillance mechanisms for health outcomes without major changes as they were designed for and are more optimally suited to etiologic research than for routine surveillance; consequently, they will not be discussed further as viable alternatives for population health surveillance.

The committee also considered whether other exposure registries of military populations might perform as population health surveillance systems. The registry that is closest to the AH&OBP Registry in terms of goals, design, and data collection is the Gulf War Registry (see Chapter 4 for a more detailed description of the registry). The Gulf War Registry covers a similar (though not precisely overlapping) population of veterans and service members who deployed to eligible areas in Southwest Asia and requires that participants complete a short list of questions on their deployment exposures and undergo a clinical health examination. The AH&OBP Registry’s health evaluation uses diagnostic procedures and tests similar to those of the Gulf War Registry (VHA Directive 1325 [VA, 2017] for the Gulf War Registry and VHA Directives 1307 and 1308 [VA, 2019, 2022c] for AH&OBP Registry).⁴ Results from analyses of Gulf War Registry data are shared with key audiences in quarterly newsletters, but those analyses do not appear to be used to improve VA health care. Therefore, the Gulf War Registry may be used to monitor its participants, but it is not designed to conduct population health surveillance, and thus the committee did not consider it to be an alternative to the AH&OBP Registry for this purpose.

The committee also considered whether several national health surveys—such as BRFSS, NHIS, and NHANES—conducted by CDC to collect information about the health and wellness of the U.S. population, might be alternative population health surveillance mechanisms. The surveys are designed to be nationally representative and to collect a wide range of information about demographics, risk factors, health, and functional status. Veterans are included in those surveys as a segment of the general population, but veterans with relevant exposures are only a small subset of the entire veteran population, and none of the surveys collect information about deployment exposures; consequently none of the surveys can examine deployment exposure–health outcome relationships. Therefore, national surveys used for population health surveillance are not representative of the entire service member and veteran population or the population eligible for the AH&OBP Registry, and as such were not considered to be suitable alternatives to the registry for this purpose.

Finally, the committee considered the Individual Longitudinal Exposure Record (ILER) as a population health surveillance mechanism. The committee was not charged with assessing the quality and function of ILER, but it was tasked with examining whether ILER might provide comparable or better exposure or health information

___________________

⁴ On March 25, 2022, VHA issued Directive 1308 (VA, 2022c), which rescinded VHA Directive 1307 (VA, 2019) and moved details of the eligibility criteria and evaluation protocols for six VA registry programs, including the AH&OBP Registry, to the VA Health Outcomes Military Exposures intranet so they are no longer publicly available.

than the AH&OBP Registry. In the following section, the committee focuses on whether ILER could be used to conduct population health surveillance.

ILER

In Chapter 5: Use for Etiologic Research, the features of ILER that make it unsuitable for conducting etiologic research on the health effects of burn pits and other airborne hazards exposures were described in detail. Several of those factors are also relevant when considering ILER’s use for population health surveillance. These factors include:

• it is not currently fully operational;
• it does not have exposure information on the entire service member or veteran population who are eligible for the AH&OBP Registry (most exposure information is from 2006 forward);
• it cannot accurately or completely assign exposure at the individual (versus unit or area) level;
• its location information is only available from 2001 forward, thus excluding those who are eligible for the AH&OBP Registry but who separated before this time;
• it will require numerous and rigorous validation studies for its data compilations; and
• linkages with summary health data are only available from 2012 forward and will be limited to DoD records for active-duty service members and those veterans who are eligible for and receive their health care from VA.

ILER was not developed to meet a stated goal of conducting population heath surveillance (the first criterion); however, VA has stated that ILER, when fully operational, may be used to monitor the health of service members and veterans (VA, 2022a). ILER will not have complete surveillance information (i.e., exposure and health outcomes) for the entire population who are eligible for the AH&OBP Registry. In the future, however, it may have the potential to conduct monitoring by compiling and linking DoD exposure information with VA and DoD health outcome data, albeit with the caveats noted above and discussed in chapters 4 and 5.

ILER partially meets the criterion of regular data collection, analysis, and interpretation for population health surveillance (criterion 2). It compiles deployment dates, locations, military occupational specialty, environmental exposures, and industrial hygiene information as available (see full description of ILER in Chapter 4) to assign and characterize selected exposure information for the entire service member population since 2000. Depending on the specific exposures of interest, it can also create appropriate comparison groups of similar but unexposed service members and veterans.

In contrast to the AH&OBP Registry, which relies on self-reported exposures, exposures in ILER would be inferred at the unit level and based on deployment history (i.e., locations and dates) from DoD’s Defense Manpower and Data Center records, but exposures could not be determined at the individual level. Thus ILER exposure assessments would not be subject to recall bias or differential misclassification of exposure based on health status. ILER linkages of DoD and VA health information with DoD exposure information may provide more representative, if not comprehensive, coverage of the population of interest, as they are passive systems and thus do not incur the risk of selective voluntary participation or place a burden on participants. This link also allows for on-demand, prospective monitoring of objectively identified exposure and health information. However, as shown in Chapter 2, many registry participants have to manually add deployment segments to supplement the information in ILER, indicating that the ILER deployment records have inaccuracies and are incomplete for some service members (e.g., those with classified deployments).

Conducting health surveillance on all service members and veterans whose health record information is available for linkage (DoD for active-duty service members since 2012 and VA for veterans who obtain care through VHA) would increase the representativeness and generalizability of the surveillance findings. This health surveillance would not be restricted to the approximately 10% of the entire eligible population who have chosen to enroll in the AH&OBP Registry. However, many veterans are not eligible for VA health care, and therefore, have no VA health record that would be captured in ILER. Using VA electronic health record data as the only source

of health data compiled by ILER for individuals who have separated from the military would result in incomplete population health surveillance and may not be generalizable to the larger eligible veteran population. DoD and VA electronic health records may capture International Classification of Diseases (ICD) diagnostic codes but the records may not include more subjective symptoms, such as cough or wheeze, that are available from the registry self-report questionnaire.

ILER compilations for individuals (i.e., exposures and health outcomes) are available for trained clinicians within VA and DoD, and, once the capability is in place, access to ILER may be extended to researchers who apply for it. Thus, ILER partially meets the third criterion of information dissemination. However, it is unclear how VA and DoD might disseminate results from the assembled cohorts of individuals to the broader audience of service members and veterans or other key stakeholders.

Finally, whether ILER might meet the fourth criterion of a population health surveillance system—health improvement resulting from the information—is unknown. Information on an individual’s exposures may assist clinicians in diagnosing and treating them, but this does not translate to improving health on a population level. Without meeting the fourth criterion, ILER does not meet the criteria for population health surveillance. Instead ILER may have the ability to potentially conduct monitoring by compiling and linking DoD exposure information with VA and DoD health outcome data, albeit with the caveats noted in this chapter and discussed in chapters 4 and 5.

In principle, ILER could become a tool or component of a low-cost, rapid mechanism to surveil or monitor the health of service members and veterans with burn pit and airborne hazard exposures by conducting queries on an ongoing basis, including detecting health conditions with long latency periods. ILER also has the potential to track and assess the health experiences of those with other military exposures that are captured within the ILER system. To determine whether ILER could be used to perform or augment population health surveillance in practice, the development of an in-depth assessment of data linkages, use, and analysis plans would be required, which was beyond the committee’s charge. Additionally, any use of ILER to perform this function would be dependent on its actual completion, the quality of its source data, its capacity to analyze data, and its ability to reliably define groups of exposed service members and veterans for follow-up.

ILER has potential value for capturing some deployment-related exposures in the future and for those who served in the armed forces after 2012 as a component of a population health surveillance system, depending on its continued development and validation of the information it compiles. It also has the potential to include a large number of both exposed and unexposed service members and veterans for comparisons. Nevertheless, at present it is unable to fulfill any of the four criteria for a population health surveillance system, and even when fully operational it will have at best a limited capability to do so. Therefore, it is not and should not be looked to or promoted as a population health surveillance system for burn pits and airborne hazards.

SYNOPSIS

This chapter examined the ability of the AH&OBP Registry to meet each of the four criteria for a population health surveillance system (CDC, 2012; German et al., 2001). The four criteria are:

1. Decide on the overarching goals of the system to address the prevalence or incidence of a potentially harmful exposure or occurrence of a disease in a defined population;
2. Regularly collect, analyze, and interpret the data that are acquired to look for predefined “signals” or unexpected occurrences of outcomes or association patterns that would motivate some type of action;
3. Disseminate data, results, and information to key audiences; and
4. Seek to improve health as a result of the collection and dissemination of that information.

the ability to conduct monitoring by compiling and linking DoD exposure information with VA and DoD health outcome data, albeit with the caveats noted in this chapter and discussed in chapters 4 and 5.

The extent to which these criteria continue to be met (or not) will ultimately depend on ILER’s completion and capabilities. However, even with the most optimistic assumptions about the completion of ILER, there will be inherent limitations concerning the exposures that the AH&OBP Registry was developed to address. The committee cautions that ILER’s potential capability to perform population health surveillance will not include the same population that is eligible to participate in the AH&OBP Registry nor will it cover the same timeframes of deployment segments, exposure, or health outcomes as the registry. Although the availability of exposure data may improve as additional reports are digitized or as information from other sources is made available for compilation, complete information on deployment exposures, particularly at the individual level will still not be available. Moreover, ILER will access only selected health information drawn from ICD-9 and ICD-10 codes of DoD and VA health record databases that are available from 2012 onward. The impact of these information gaps will depend on the particular exposure of interest. For example, it may have a large effect on analyses of burn pit exposures since most of the DoD burn pits were in heavy operation in 2001–2009 and the incomplete exposure data before 2006 will miss many of these exposures. Not only are verified exposure locations needed, but a validated algorithm for differentiating exposure levels or gradations based on location and time period would be necessary for ILER to support a population health surveillance system.

ILER has the potential to include a large number of service members and veterans and may include both exposed and unexposed individuals for comparisons, but it is not suitable to population health surveillance related to burn pits and airborne hazards exposures. Thus, ILER has potential value for capturing some deployment-related exposures in the future and for those who served in the armed forces after 2012. It may serve as a tool or component of a population health surveillance system depending on its continued development and validation of the information it compiles, but it is not—and should not be looked to or promoted as—a population health surveillance system for burn pits and airborne hazards.

REFERENCES

This page intentionally left blank.