2
The Herbicide Exposure Assessment Model

In 2003 Jeanne Mager Stellman, Ph.D., and colleagues completed their project to develop methods to retrospectively characterize the exposure of Vietnam veterans to herbicides used by the military in Vietnam. The project resulted in the compilation of databases on herbicide spraying, the identification and location of certain military units, and the development of a software tool—Herbicide Exposure Assessment–Vietnam (HEA-V)—that facilitates use of an exposure assessment model to calculate two proximity-based herbicide exposure metrics: a “hits” measure and an exposure opportunity index (EOI) (Stellman and Stellman, 2003). The model and the HEA-V software tool use a geographic information system (GIS) that makes it possible to link an extensive database providing timing and location of herbicide application with the location histories of individuals or military units. The work was described in detail in a final report submitted to an Institute of Medicine committee overseeing the project and to the Department of Veterans Affairs (VA) (Stellman and Stellman, 2003). In addition, Stellman and colleagues published three peer-reviewed papers on various aspects of their work (Stellman et al., 2003a,b; Stellman and Stellman, 2004).

This chapter presents an overview of the exposure assessment concepts underlying measures of exposure opportunity and provides a brief summary of the Stellman team’s exposure assessment model and the related software and databases as described in the final report to VA, the journal articles, and presentations by Dr. Stellman to the committee.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 25
2 The Herbicide Exposure Assessment Model I n 2003 Jeanne Mager Stellman, Ph.D., and colleagues completed their project to develop methods to retrospectively characterize the exposure of Vietnam veterans to herbicides used by the military in Vietnam. The project resulted in the compilation of databases on herbicide spraying, the identification and location of certain military units, and the development of a software tool—Herbicide Exposure Assessment–Vietnam (HEA-V)—that facilitates use of an exposure assessment model to calculate two proximity- based herbicide exposure metrics: a “hits” measure and an exposure oppor- tunity index (EOI) (Stellman and Stellman, 2003). The model and the HEA-V software tool use a geographic information system (GIS) that makes it possible to link an extensive database providing timing and location of herbicide application with the location histories of individuals or military units. The work was described in detail in a final report submitted to an Institute of Medicine committee overseeing the project and to the Depart- ment of Veterans Affairs (VA) (Stellman and Stellman, 2003). In addition, Stellman and colleagues published three peer-reviewed papers on various aspects of their work (Stellman et al., 2003a,b; Stellman and Stellman, 2004). This chapter presents an overview of the exposure assessment concepts underlying measures of exposure opportunity and provides a brief summary of the Stellman team’s exposure assessment model and the related software and databases as described in the final report to VA, the journal articles, and presentations by Dr. Stellman to the committee. 

OCR for page 25
 PROXIMITY-BASED HERBICIDE EXPOSURE ASSESSMENT THE EXPOSURE OPPORTUNITY CONCEPT Quantifying “exposure opportunity” refers to assessing the potential for exposure rather than assessing the exposure itself and thus is, at best, a crude approximation of toxicologic dose. Measures like the EOI are particularly useful in situations where historical data on body burden or environmental exposure concentrations are not available, and where measurements based on newly collected samples are not useful for estimat- ing past exposures. The Stellman team’s two measures of exposure opportunity are based on presence at a location that records indicate was within a specified dis- tance of herbicide spraying and can be considered surrogates of exposure. By contrast, many past studies of Vietnam veterans and their health status have used service in Vietnam—a marker even further removed from dose than the proximity-based measures—as an exposure surrogate (e.g., CDC, 1987; Breslin et al., 1988; Dalager and Kang, 1997; also see Appendix B). Because measurements of body burden, dose, or exposure concentrations are often not feasible in retrospective studies of occupational or environ- mental hazards, surrogate measures of potential for exposure (such as the opportunity index) are commonly used in studies of occupational and environmental hazards. Exposure–Disease Pathway It is often helpful to think of health effects that result from exposure to a toxic material in terms of a pathway from the introduction of the toxic material into the environment to the production of the health effects (Figure 2-1). Following the introduction of a toxic material, processes that affect the fate and transport of the material lead to its presence in varying concentrations across the environment. Contact between people and the toxic substance—modified by individual behaviors—can lead to exposure via three principal routes: inhalation, dermal exposure, or inges- tion. Pharmacokinetic processes then govern the concentration of the toxic substance and its metabolites in the body, including at the target organ. The factors determining the biologically effective dose can depend on the health effect of interest. For example, for some diseases cumulative dose might be important, while for others the peak dose may be the most relevant consid- eration. Timing of the dose can also be critical. For cancer, relevant doses may occur decades prior to diagnosis; for developmental effects, the rel- evant exposure of the developing organism can occur during specific time windows (e.g., early in pregnancy).

OCR for page 25
 THE HERBICIDE EXPOSURE ASSESSMENT MODEL Potential Transport Human Accumulation Source dose and contact: in emissions to the body transformation exposure environment Early Biologically Health Internal expression effective effect dose of disease dose Elimination Bioavailability Accumulation Transformation FIGURE 2-1 Pathway from emission of a contaminant to a health effect. SOURCE: Reprinted with permission from Lioy, 1990. Copyright 1990 American Chemical Society. fig 2-1 Exposure Assessment Hierarchy The pathway notion provides a useful conceptual structure for thinking about the relationship between exposure and outcome, but the practicalities of actually doing a study must also be considered. For example, while the pathway diagram suggests that biomarkers (which reflect internal dose in Figure 2-1) can provide an exposure measure that is close to ideal, such measures may not be practical if no biomarker has been identified for the compound under study or if, in a retrospective study, too much time has elapsed between exposure and measurement for a biomarker to be detected. Thus, when designing or evaluating studies, it is useful to keep in mind the types of data associated with the exposure assessment hierarchy (Table 2-1). Although it was developed in the context of community exposures to haz- ardous waste sites, it has wider applicability (Nieuwenhuijsen, 2003). The idea behind the exposure assessment hierarchy is that for any given study there exists a range of possible exposure metrics that vary in accu- racy. Quantified personal measurements—such as exposure biomarkers or personal air monitoring—generally provide the best data if they are biologi- cally relevant. Beyond that, various increasingly indirect exposure metrics are possible. For example, such measures may be based on combinations of ambient environmental measurements and individual-level information on behavior patterns that affect exposure; estimates of ambient environmental concentrations derived from fate and transport modeling; or distance and

OCR for page 25
 PROXIMITY-BASED HERBICIDE EXPOSURE ASSESSMENT TABLE 2-1 Data for a Hierarchy of Exposure Assessment Approximation to Types of Data Actual Exposure 1. Residence or employment in defined geographical area Poorest (e.g., a county) of the site 2. Residence or employment in geographic area in reasonable proximity to site where exposure can be assumed 3. Distance from site or duration of residence in area 4. Distance from site and duration of residence 5. Quantified surrogates of exposure (e.g., estimates of drinking water use) 6. Quantified area or ambient measurements in the vicinity of the residence or other sites of activity Best 7. Quantified personal measurements SOURCE: Adapted from NRC, 1991. duration, in combination or separately, as surrogates for fate and transport modeling. The poorest exposure approximation in the hierarchy is residence or employment in the same area (e.g., a county) as the exposure source. Despite their limitations, such relatively crude exposure surrogates as location have been used effectively in many environmental and occupa- tional health studies. Indeed, geographic approaches to exposure assess- ment can be especially useful when pollutants vary significantly over space in a systematic manner (Briggs, 2003; Colvile et al., 2003). Several considerations are important when thinking about the hierarchy. If there is a relationship between exposure and outcome, for instance, more accurate measures of exposure often tend to provide a better ability to detect effects. This occurs because errors in exposure measurement tend to dimin- ish estimated effects if the errors are non-differential, that is, if they do not depend on the outcome. Although there are exceptions, this holds true in many important cases, such as the regression analysis of a continuous expo- sure measure with random measurement error. On the other hand, while it is nearly always possible to imagine better exposure measures, additional details do not always dramatically improve results. For example, if personal behaviors relevant for exposure do not differ much between individuals, then adding information on those behaviors will not change the rank order of exposure very much (e.g., Vieira et al., 2005).

OCR for page 25
 THE HERBICIDE EXPOSURE ASSESSMENT MODEL COMPONENTS OF THE HERBICIDE EXPOSURE ASSESSMENT MODEL The exposure assessment model developed by the Stellman team brings together several components. The GIS provides a method for link- ing detailed data on the location of herbicide spraying with data on the location of military personnel or military units or facilities and can also be applied to Vietnamese communities or residents. The GIS relies on a detailed base map of Vietnam, a database for herbicide spraying missions, and software that can calculate exposure opportunity measures (both direct hits and cumulative exposure opportunity) based on entered locations. Also available is a database that assigns a soil type to each grid cell in the GIS. In the GIS, Vietnam is portioned into a grid with the smallest unit being a rectangle of 0.01° of latitude by 0.01° of longitude, which is approxi- mately 1.2 square kilometers. The center point of each of these cells is the reference location for all of the positions within that grid cell. As a result, the grid cell size defines the ultimate spatial resolution of the GIS. The Stellman team has also done extensive work to compile databases on the identification of military units that served in Vietnam and on the location of many units during the period herbicide spraying took place. Data on the location of military units or on the personnel who served in them are essential inputs for generating unit- or person-level exposure opportunity scores, but unlike the geographic parameters or the data on herbicide spraying, these databases are not contained in the herbicide exposure assessment software. Each of these components is discussed in the sections that follow. Measures of Exposure Opportunity The exposure opportunity assessment model described here and dis- cussed throughout this report is an extension and refinement of earlier approaches developed by the Stellman team (Stellman and Stellman, 1986; Stellman et al., 1988) to produce quantitative surrogates of exposure that take into account proximity in time and space to herbicide spraying over a specific period during the war in Vietnam. Two types of measures of expo- sure opportunity can be generated with the current GIS and software. One measure provides the number of “hits,” or instances of a herbicide spray application, that occurred within a certain distance (0.5 km, 1 km, 2 km, or 5 km) of a given grid location (Figure 2-2). The second measure, EOI, takes into account the potential exposure from past as well as same-day spraying. The Stellman team refers to the

OCR for page 25
0 PROXIMITY-BASED HERBICIDE EXPOSURE ASSESSMENT FIGURE 2-2 Region of Vietnam approximately 80 kilometers east of Saigon (see inset), showing the flight path of Ranch Hand mission no. 3087, flown on May 11, 1967, (straight orange line) along Highway 1 (jagged red line). The GIS grid system fig 2­2 is shown as horizontal and vertical lines spaced 0.01° apart. Contained within a 5-kilometer buffer drawn around the flight path (heavy black line) are the centroids of the 210 cells that fell within 5 kilometers of the flight path, with size and color This is a jpeg, which is like a photograph. indicating proximity to the defoliation flight path: large black circles = 0.5 kilometer We could "tone down" all colors together. or less; large open circles = 0.5–1 kilometer; small shaded squares = 1–2 kilometers; small black circles = 2–5 kilometers. SOURCE: Stellman and Stellman, 2004. Reprinted by permission from Macmillan Actually our color printer tends to print dark. Publishers Ltd: Journal of Exposure Analysis and Environmental Epidemiology If you look at the figure on screen it is quite a bit 14:354–362, Copyright 2004. paler than the printout. The final printing probably will look more like the image on the computer screen. possibility of exposure to residual herbicide from sprays on prior days as “indirect exposure.” To calculate the combined direct and indirect EOI, their model takes into account for each spraying mission the quantity of herbicide or concentration factor; the reciprocal distance of the exposure location from the herbicide spray; a time factor, chosen to be a first-order environmental decay of the sprayed herbicide over time; and the concept of continuous dispersion of herbicide all along the path of the flight. The EOI developed and refined by the Stellman team is reported to be quantitative on a ratio scale, meaning that an EOI of 1,000 indicates twice as much exposure opportunity as an EOI of 500 (Stellman et al., 2003b).

OCR for page 25
 THE HERBICIDE EXPOSURE ASSESSMENT MODEL However, given that the EOI is a measure of exposure opportunity, not actual exposure, the EOI is not necessarily proportional to the dose of herbicide that an exposed individual would have received. Software System Automating Exposure Opportunity Calculations The HEA-V software (HEA-V, 2003) was developed to facilitate and automate the calculations involved in assessing proximity to the herbicide spraying that is known to have occurred during the Vietnam War. The software, which is accompanied by a manual to guide users, prompts users to provide—either in a file that can be imported or by directly entering the data—the information needed to calculate hits or an EOI for individuals or military units of interest. The necessary inputs include the longitude and latitude for each of the individuals’ or units’ locations during a period of interest and the start and end dates for their presence at each of these locations. After the necessary inputs have been provided, the software allows users to make a series of decisions regarding the nature of the analysis sought. For example, a user may specify that hits or EOI scores be calculated for only one kind of herbicide or one type of spraying (e.g., fixed-wing aircraft versus helicopter). Users can also designate the half-life for environmental decay of the agent of interest, or limit the locations, dates, or gallonage (the quantity of herbicide for a spray mission) to be considered. The software also aids in the selection of preferred outputs and in the documentation of analyses. The Herbicide Spray Database The herbicide exposure assessment model incorporates a database with detailed information on the timing and locations of herbicide applications. The file contains information for 9,141 spray missions, 65 percent of which were conducted by the U.S. Air Force using C-123 fixed-wing air- craft as part of Operation Ranch Hand. A review of the database showed that these Ranch Hand missions delivered 95 percent of the total amount of herbicide recorded as dispersed by the U.S. military in Vietnam. The remaining 35 percent of missions recorded in the database were U.S. Army helicopter and ground spraying activities or missions with an unspecified delivery method. Records of helicopter missions were kept with the fixed wing records beginning in 1968, but ground spraying was not tracked as part of a permanent record system (Stanton, 1989). To assemble these data, Dr. Stellman’s team has described drawing on Department of Defense (DoD) files known as HERBS and Services HERBS; extensive research into military records held by the National Archives

OCR for page 25
 PROXIMITY-BASED HERBICIDE EXPOSURE ASSESSMENT and Records Administration; and assistance from the Army’s Center for Research of Unit Records (CRUR; now the Joint Services Records Research Center, JSRRC) (Stellman and Stellman, 2003).1 This work was the basis for the elimination of duplicate records, the addition of records for some previously undocumented missions, and the correction of or addition of missing information to some other records. The HERBS and Services HERBS data use the DoD’s version of the UTM coordinate system to represent the location of herbicide spraying missions. UTM refers to the Universal Transverse Mercator map projection, for which latitude and longitude are transformed into meter distances on a projection of the earth’s surface onto thin strips running from pole to pole. DoD’s version of UTM consists of a two-letter code identifying a particu- lar 100,000-by-100,000-meter grid cell within the map, followed by two three-digit sequences indicating the distance in hundreds of meters east and north, respectively, from the southwest corner of the grid cell. Although the UTM coordinates have a nominal precision of 100 meters, it was difficult for aircrews and ground troops to determine locations with that level of accuracy (Young et al., 2004). Spraying mission flight paths are denoted by a sequence of linear legs between UTM coordinates. The Stellman team converted these UTM coor- dinates into longitude and latitude using software from what is now the National Geospatial-Intelligence Agency (Stellman, 2007) and placed flight paths onto their GIS’s grid defined by the 0.01°-by-0.01° cells mentioned above (approximately 1.2 km2). UNIT LOCATION INFORMATION To use the GIS and exposure assessment model to estimate veterans’ potential exposure to herbicides, coordinates indicating location are needed for each day that a study subject was present in Vietnam (or for a specific period of Vietnam service). Locations of individual service members in Vietnam were not recorded, so the recorded locations of the military units in which they served are used to impute the locations of individuals. The Stellman team has drawn upon a variety of sources to compile a comprehensive list of units that served in Vietnam. This list exists as a 1The Army unit currently designated as the U.S. Army and Joint Services Records Research Center (JSRRC) has operated under various other designations in the past, including the Environmental Support Group (ESG, 1980–1996), the Center for Research of Unit Records (CRUR, 1996–1998), and the Center for Unit Records Research (CURR, 1998–2006) (Hakenson, 2007). The principal responsibility of JSRRC and its predecessor organizations is to research military records to provide information related to veterans’ claims filed with VA. JSRRC responds to requests from individual veterans, veterans service organizations, and VA. This report generally refers to the unit by its current designation of JSRRC.

OCR for page 25
 THE HERBICIDE EXPOSURE ASSESSMENT MODEL database called UICLIST, which is independent of the herbicide exposure assessment software. Based upon units’ missions, the Stellman team catego- rized them as stable (units whose missions did not require field maneuvers), mobile (units whose missions required leaving main base camps), or con- taining mobile elements within an otherwise stable unit. They have gathered location information on many stable units, including nearly all the combat arms support and combat support units assigned to Vietnam. They estimate that these units constitute about 80 percent of the Army troops who served in Vietnam (Stellman and Stellman, 2003). In addition to the data on these stable units, the Stellmans also have created a database capturing location information collected by JSRRC on Army combat battalions serving in the III Corps Tactical Zone during the years 1966–1969, a place and time of intense aerial herbicide spraying (Stellman et al., 2003a). Although the Stellman team has collected location information for many of the stable units in Vietnam and some of the combat units, it is likely that additional information on unit locations will have to be assembled for at least some of the participants in an epidemiologic study. The considerations involved in such efforts are discussed in Chapter 4. It is important to note that the model is not applicable to Air Force Ranch Hand personnel and Army Chemical Corps personnel, whose expo- sures to herbicides are likely to be the highest of all veterans. Their herbicide exposures were primarily a direct result of duties that required handling or applying herbicides. By contrast, the model is designed to assess the expo- sure opportunity that would result from unintended proximity to herbicide spraying. REFERENCES Breslin, P., H. K. Kang, Y. Lee, V. Burt, and B. Shepard. 1988. Proportionate mortality study of US Army and US Marine Corps veterans of the Vietnam War. Journal of Occupational Medicine 30(5):412–419. Briggs, D. 2003. Environmental measurement and modeling: Geographical information sys- tems. In Exposure assessment in occupational and environmental epidemiology, edited by M. J. Nieuwenhuijsen. New York: Oxford University Press. CDC (Centers for Disease Control and Prevention). 1987. Postservice mortality among Viet- nam veterans. Journal of the American Medical Association 257(6):790–795. Colvile, R., D. Briggs, and M. J. Nieuwenhuijsen. 2003. Environmental measurement and modeling: Introduction and source dispersion modeling. In Exposure assessment in occu- pational and environmental epidemiology, edited by M. J. Nieuwenhuijsen. New York: Oxford University Press. Dalager, N. A., and H. K. Kang. 1997. Mortality among Army Chemical Corps Vietnam veterans. American Journal of Industrial Medicine 31:719–726. Hakenson, D. 2007. Help with some names and dates? E-mail to L. Joellenbeck, Institute of Medicine, October 16. HEA-V (Herbicide Exposure Assessment–Vietnam). 2003. CD-ROM, version .0.. Software and accompanying electronic documentation. New York: Columbia University.

OCR for page 25
 PROXIMITY-BASED HERBICIDE EXPOSURE ASSESSMENT Lioy, P. J. 1990. Assessing total human exposure to contaminants: A multidisciplinary approach. Environmental Science and Technology 24(7):938–945. Nieuwenhuijsen, M. J. 2003. Introduction to exposure assessment. In Exposure assessment in occupational and environmental epidemiology, edited by M. J. Nieuwenhuijsen. New York: Oxford University Press. NRC (National Research Council). 1991. Environmental epidemiology, Volume : Public health and hazardous wastes. Washington, DC: National Academy Press. Stanton, S. L. 1989. Area-scoring methodology for estimating Agent Orange exposure status of U.S. Army personnel in the Republic of Vietnam. In Comparison of serum levels of ,,,-tetrachlorodibenzo-p-dioxin with indirect estimates of Agent Orange exposure among Vietnam veterans, by Centers for Disease Control and Prevention. Atlanta, GA: Centers for Disease Control and Prevention. Stellman, J. M. 2007. Responses to IOM 00. Unpublished document submitted to the IOM Committee on Making the Best Use of the Agent Orange Reconstruction Model, September 14. Stellman, S. D., and J. M. Stellman. 1986. Estimation of exposure to Agent Orange and other defoliants among American troops in Vietnam: A methodological approach. American Journal of Industrial Medicine 9(4):305–321. Stellman, J. M., and S. D. Stellman. 2003. Contractor’s final report: Characterizing exposure of veterans to Agent Orange and other herbicides in Vietnam. Submitted to the National Academy of Sciences, Institute of Medicine, in fulfillment of Subcontract VA-5124-98- 0019, June 30, 2003. Stellman, S. D., and J. M. Stellman. 2004. Exposure opportunity models for Agent Orange, dioxin, and other military herbicides used in Vietnam, 1961–1971. Journal of Exposure Analysis and Environmental Epidemiology 14:354–362. Stellman, S. D., J. M. Stellman, and J. F. Sommer, Jr. 1988. Combat and herbicide expo- sures in Vietnam among a sample of American Legionnaires. Environmental Research 47:112–128. Stellman, J. M., S. D. Stellman, R. Christian, T. Weber, and C. Tomasallo. 2003a. The extent and patterns of usage of Agent Orange and other herbicides in Vietnam. Nature 422:681–687. Stellman, J. M., S. D. Stellman, T. Weber, C. Tomasallo, A. B. Stellman, and R. Christian, Jr. 2003b. A geographic information system for characterizing exposure to Agent Orange and other herbicides in Vietnam. Environmental Health Perspectives 111:321–328. Vieira, V., A. Aschengrau, and D. Ozonoff. 2005. Impact of tetrachloroethylene-contaminated drinking water on the risk of breast cancer: Using a dose model to assess exposure in a case-control study. Environmental Health 4(1):3. http://www.ehjournal.net/content/4/1/3 (accessed August 7, 2007). Young, A. L., P. F. Cecil, and J. F. Guilmartin, Jr. 2004. Assessing possible exposure of ground troops to Agent Orange during the Vietnam War: The use of contemporary military records. Environmental Science and Pollution Research 11(6):349–358.