Periodic Occupational and Environmental Monitoring Summary

The overall conclusion of this analysis, however, is 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. The information currently publicly available in POEMS is insufficient to adequately determine the degree of risk individuals may face based on the dates spent at a given location.

• Over half the reviewed POEMS files recorded insufficient data to assess short- and long-term health risks from exposure to particulate matter.
• Of summaries that did assess health risks, most were not confident in their assessments.
• POEMS files generally contain insufficient information to assess variations in short- or long-term health risks over time.
• Frequency of sample collection downrange was inconsistent.
• Risk assessments were inconsistent and subjective.

To demonstrate what can be gleaned from the available data, this project presents the visualizations of risk from those sites with minimally adequate information. This project’s focus was on exposure to airborne particulate matter, a mixture of extremely small solid particles and liquid droplets emitted into the air by sources including power plants, vehicles, construction, fires, and dust. Exposure to excessive airborne particulate matter has been linked to a variety of potential negative health effects. Graphics in this pilot project show the health risks of two types of particulate matter at U.S. military sites in Iraq/Kuwait and Afghanistan over time. Both show the POEMS assessment of the risk itself (from low to high) by color saturation, as well as assessment of confidence in that risk assessment (low, moderate, or high) shown by one, two, or three bars respectively.

Heat maps for Iraq/Kuwait and Afghanistan are presented separately; the zoom function allows viewers to distinguish between sites that are geographically close together. These maps are optimized for desktop viewing. Within each map, users can toggle between graphics displaying information about the potential health risks of two types of particulate matter, PM₁₀ and PM_2.5:

• PM₁₀ – Short-Term Health Risks: PM₁₀ includes coarse particles with a diameter of 10 micrometers or less. Since there are no available health evaluations for its long-term risk, the assessments of short-term health risks are presented here.
• PM_2.5 – Long-Term Health Risks: PM_2.5 consists of finer particles of 2.5 micrometers or less that can reach deeper in the lungs; these maps present the POEMS assessments of its long-term health risks.

Iraq and Kuwait

Afghanistan

Additional details contained in the source documents, which are available at the Army Public Health Center.

The data provided from POEMS makes it impossible to visually represent changes in risk over time for most sites due to inadequacies in the available underlying data, as described in greater detail. These heat maps present the information by year to visually indicate known risk by site by years, as well as—in those instances where such information was available—changes in risk within sites over time. The limited number of sites visualized is a stark demonstration of how limited the data is: the many sites that listed “no data” or “insufficient data” do not appear on the map at all.

The following sections provide additional information on background and methodology, as well as summary observations and additional questions raised.

Background

Early in the conflicts in Iraq and Afghanistan, the U.S. Department of Defense’s forward bases began burning the trash generated in open air “burn pits.” Refuse of all kind, including plastic, medical waste, lithium batteries, paper, and human feces, were burned; jet fuel was used as an accelerant. Journalist Kelly Kennedy reported in 2008 on service member concerns about possible negative long-term health effects related to breathing in the smoke from these fires; she also cited an internal memo, dated 2006, from the former bioenvironmental flight commander for Joint Base Balad raising concerns.¹

Since then, additional news stories have regularly been published about the burn pits themselves, some of which were later operated by contractors such as KBR. Claims of health problems related to exposure by service members, veterans, and family members also increased. In response to these concerns, the Institute of Medicine (now the National Academy of Medicine) issued a report “Long-Term Health Consequences of Exposure to Burn Pits in Iraq and Afghanistan,” conducted at the request of the Department of Veterans Affairs (VA).² The committee conducting the study reported insufficient evidence to develop firm conclusions about long-term health risks and recommended multi-year study, since many chronic diseases such as cancer take decades to develop. In addition, the VA in 2014 created the Airborne Hazards and Open Burn Pit Registry (AHOBPR), which allows veterans to self-report their exposures and health concerns; by the end of fiscal year 2019, over 186,000 veterans had completed the registry questionnaire.³

Today, numerous veterans service organizations and other nonprofits dedicated to supporting service members, veterans, military families, and survivors continue to advocate for greater attention to the potential long-term health effects of burn pits and other military toxic exposures. Advocates argue that some health conditions among those who served near burn pits should be considered “presumptive diseases,” which the VA assumes are related to military service, in the same way that certain health problems linked to Agent Orange exposure now are.⁴ This designation would ease the burden on veterans to prove the nexus between a current condition and military service, which is important when applying for a service-connected disability rating and potentially also when seeking VA health care for that condition. Ongoing pressure and media coverage have led members of Congress to introduce multiple pieces of legislation related to toxic exposures.⁵ In response to the acknowledged risks of open air burn pits, the Department of Defense (DoD) also reports having switched to “conventional solid waste management practices” at established locations; nine burn pits remained operational by U.S. Central Command in July 2019 at short-term contingency operations where “no feasible alternative exists” despite DoD’s investment in research for deployable incinerators.⁶

To date, comprehensive data on individual or population exposures to environmental toxins among deployed service members and the potential for long-term health effects remains incomplete and difficult to access. As a result of a 2013 White House Presidential Review Directive, the DoD and VA are in the process of launching a long-term exposure record for each service member, called the Individual Longitudinal Exposure Record (ILER). The ILER, which draws from a variety of databases across the service branches, includes occupational and environmental health, matching veterans to place, time, location, and events to consolidate data on exposures. This allows improved tracking and data analysis at the individual level.⁷

In addition, the APHC maintains the POEMS website, which allows service members and health care providers to view and/or download PDFs containing summary information about DoD’s assessment of potential environmental exposures at locations to which troops have been deployed worldwide. The documents “describe overall general population exposure estimates and do not describe any specific person’s unique exposures."⁸ POEMS was designed to standardize recordkeeping across the services and to meet DoD requirements on occupational and environmental health as well as deployment health surveillance; their format was concurred by the Joint Environmental Surveillance Work Group.⁹ POEMS files provide estimates on population-based health risk assessments; despite this limitation, for those who deployed before the time frame that will be covered by the ILER, they provide the most comprehensive available public source of information.

At the time of this review (May 2019), a total of 146 records were available for sites in 16 countries in the Middle East, Central Asia, and Africa; there were multiple files for some sites, sometimes covering overlapping time frames, and only one for others. Each summary follows a template. Included at the beginning are the purpose of the document and a site description, which generally includes a brief description of the operation, details on the base camp and vicinity, major tenant organizations, and information on the climate. The first table summarizes occupational and environmental conditions with moderate or greater short- and long-term health risks, such as dust inhalation, heat injury, and leishmaniasis, to be entered by a medical provider. The second table contains population-based health risk estimates for an array of sources, including those in the air, soil, and water; endemic disease; venomous animals/insects; heat/cold stress; and noise. The tables are followed by lengthier discussions of the health risks, sampling frequency, and confidence in the risk assessment provided.

While the amount of information is significant, there has been little outreach on its availability or how to interpret it, and understanding what components are relevant may be challenging for the lay individual. In addition, while this is the most comprehensive source publicly available for overseas exposures, it is not easy to navigate.¹⁰ Accordingly, with support from Wounded Warrior Project, CNAS undertook a pilot project to present information visually about key potential risks. By reviewing “heat maps” of various exposures and risks through time, service members, veterans, and their families may be better able to identify potential concerns that they should share with their health care providers, who can then seek additional information from the source documents or other available data.

Methodology

As noted, the POEMS files list a wide array of potential short- and long-term risks. For this limited demonstration project, CNAS decided to focus on airborne particulate matter:

"Particulate matter (PM) is a complex mixture of extremely small particles suspended in the air. The PM includes solid particles and liquid droplets emitted directly into the air by sources such as: power plants, motor vehicles, aircraft, generators, construction activities, fires, and natural windblown dust. The PM can include sand, soil, metals, volatile organic compounds (VOC), allergens, and other compounds such as nitrates or sulfates that are formed by condensation or transformation of combustion exhaust. The PM composition and particle size vary considerably depending on the source. Generally, PM of health concern is divided into two fractions: PM₁₀, which includes coarse particles with a diameter of 10 micrometers or less, and fine particles less than 2.5 micrometers (PM_2.5), which can reach the deepest regions of the lungs when inhaled. Exposure to excessive PM is linked to a variety of potential health effects."¹¹

Additionally, CNAS chose to limit the focus of this pilot project to sites in Iraq, Kuwait, and Afghanistan.

Researchers selected and captured the following information from the POEMS files: country, location, presence of burn pits, whether environmental samples were collected near burn pits, study period beginning and end dates, sampling beginning and end dates, number of samples, average concentration, and short- and long-term risks for PM_2.5 and PM₁₀, along with confidence in those risk assessments. Sites were named using the naming conventions in the POEMS dataset; since several locations have multiple POEMS summaries, each location was also assigned a number so all summaries for that location could be easily sorted; there were 82 total sites. To facilitate mapping, specific grid coordinates were entered when possible; for POEMS summaries that cover broader regions (e.g., “East Baghdad”) the coordinates for the largest base listed in the summary were entered when possible. If specific coordinates were unavailable, the coordinates of the neighborhood, town, or city that best corresponds to the location were entered. The full dataset is available upon request.

Several important caveats should be noted. The POEMS template calls for recording the total number of air samples collected, the time frame during which they were collected, and the range of and average concentrations. Environmental samples are listed with only the date range in which they were collected. Accordingly, for the majority of sites, it is not possible based on how the data is provided to determine whether concentration levels varied by year or within years—there is typically no record of whether a particular year had higher or lower concentrations, whether the average listed is significantly affected by single outlier samples, whether samples were evenly collected across the date range, or if they were clustered at specific periods in time. It is also impossible to tell whether varying weather conditions may have reduced or enhanced risk, overall or during specific time frames. Without this information, it cannot be assessed whether specific years or months saw higher risk to health than others. This also makes it difficult to determine if the health risk assessed can honestly be attributed to the whole date range provided. This is a significant limitation in the utility of POEMS.

Long-term health risks are coded as low, moderate, high, or extremely high for PM_2.5 exposure, using specific definitions laid out in the template, along with confidence levels in the risk assessment. However, it is unclear what threshold needs to be passed for a POEMS file to have enough data to make a health assessment. For example, while one dataset (al Qaim and vicinity) makes health risk assessments based off nine samples over five years for PM_2.5, another (al Asad Airbase) does not make a health risk assessment based off 25 samples over seven months for PM_2.5. No clear or transparent standards are provided to explain what amount of data was considered sufficient by those who drafted the reports to enter a health risk assessment. Similarly, it is not clear what the standards are for establishing confidence in the risk assessment. For example, Kandahar Airfield from 2009 to 2014 had a total of 446 samples taken and had “low” confidence in its risk assessment, while Mosul Airfield from 2003 to 2010 had a total of 46 samples and “moderate” confidence in its risk assessment. That said, more samples do seem to loosely correlate with higher levels of confidence. Another limitation is that long-term health risks were not evaluated for PM₁₀, as there are “no available health guidelines”: the U.S. Environmental Protection Agency does not have standards “due to an inability to clearly link chronic health effects with chronic PM₁₀ exposure levels."¹²

For visual presentation, all sites with no data or insufficient data were deleted, and only one set of data was used for visualization when sites had multiple entries in the same time frame.¹³ Two types of data are presented: PM₁₀ short-term health risks and PM_2.5 long-term health risks. After excluding sites with no or insufficient data, from a total of 82 sites coded, the following were visualized:

• Iraq and Kuwait PM₁₀: 35
• Iraq and Kuwait PM_2.5: 22
• Afghanistan PM₁₀: 19
• Afghanistan PM_2.5: 16

Both maps show the POEMS assessment of the risk itself by color (from low to high) and by confidence in risk assessment (low, moderate, or high, shown by one, two, or three bars respectively). The data provided from POEMS makes it impossible to visually represent the data in a temporal manner to show changes in risk over time for most sites. However, these heat maps do present the overall information temporally by year to visually indicate what information is available for what date ranges for most sites, as well as changes within sites in the rare instances when that information was provided.

Three POEMS reports did provide additional data; however, none of it was comprehensive enough to serve as an exemplar:

• Bagram Airfield, as well as some others, provided information regarding what percent of time the daily average risk levels were considered to represent no, low, moderate, or high hazard. Unfortunately, this data did not provide additional specificity on when hazard severity peaked.
• Al Asad airbase provided the number of samples collected each quarter. It was the only location to provide this information. However, it did not indicate what particulate concentration levels were associated with these samples, which would have provided a better picture of health risk over time.
• The Victory Base Complex provided a year-by-year assessment of health risk from both peak and average exposures. It was the only location to do so. However, several years are recorded as “insufficient data,” and there is not adequate explanation of how to interpret the difference between peak and average risks when available.

Summary, Observations, and Questions

Over half of the POEMS summaries coded did not have enough data to assess short- or long-term health risks from exposure to particulate matter:

• 51 percent did not have enough data to assess PM₁₀ short-term health risk.
• 61 percent did not have enough data to assess PM_2.5 short-term health risk.
• 63 percent did not have enough data to assess PM_2.5 long-term health risk.

Of those that did assess health risks, the majority were not confident in their assessments:

• 64 percent expressed low confidence in the health risk assessed.
• 11 percent expressed low to moderate confidence in the health risk assessed.
• 22 percent expressed moderate confidence in the health risk assessed.
• 3 percent expressed high confidence in the health risk assessed.

There is a general trend that the more environmental samples were collected, the more likely a given location is to be assessed as a higher health risk, raising several questions: Would other locations with low sample counts have a higher health risk assessment if more samples had been collected? Or is the higher health risk because the locations with high sample counts are large bases that may have more activities contributing to air pollution?

The amount of information currently publicly available in POEMS is inadequate for individuals to determine the degree of risk they may face based on the dates they spent at a given location. A more useful dataset would lay out consistent standards for assessing health risks and confidence in those risks as well as a clear and distinct methodology for the sampling. Additionally, the dataset would provide month-by-month, quarter-by-quarter, or at least year-by-year details to give a better sense of any fluctuations in the environmental samples collected and their related health risk assessments. Access to the underlying environmental sample information could make this dataset more useful for service members, veterans, health care providers, and researchers. Should that information be available, DoD should release it. To facilitate future analyses, an upgraded system for monitoring such exposures should record and publicize the details of each sample taken in database form rather than requiring users to pore through multiple documents or create their own database.