A. INTRODUCTION

The Directorate for Deployment Health Support (formerly the Office of the Special Assistant to the Deputy Secretary of Defense for Gulf War Illnesses, Medical Readiness, and Military Deployments) conducted a health risk assessment (HRA) for pesticides and related products used by or in proximity to ground troops deployed to the Kuwait theater of operations (KTO) during the Gulf War. Risk assessment, as defined by the US Environmental Protection Agency (EPA), Office of Pesticide Programs (OPP), is the process used to estimate the likelihood and magnitude of health effects from environmental exposures.^[148] "Exposure" is defined here as the contact of a chemical with the outer boundary of a human or test animal.^[149] The outer boundary means the skin, the lining of the alimentary canal (stomach, intestines, etc.), and the lining of the airways. Most exposures to pesticides are harmless; however, exposures of sufficient concentration and duration may elicit adverse physical symptoms, indicating overexposure.

The purpose of the HRA is to provide an estimate of the likelihood and magnitude of various specific effects from pesticide exposures during deployment. Such effects would have been limited to the time of deployment, and may or may not have implications for long-term health effects. While the HRA alone cannot prove either that overexposures occurred during deployment, or that any connection exists between pesticide exposures and chronic health effects months or years after exposure, the goal of the analysis is to answer the question, "Is it possible that some US personnel were exposed to concentrations of pesticides which exceeded reasonable risk-based levels?" The answer is yes; however, significant uncertainty is inherent in the overall analysis and limits the extent to which the analysis can be used.

This HRA was conducted in a manner that would allow it to be as comprehensive as possible, taking into account exposures due to supervised and unsupervised pesticide applications. Supervised applications were pesticide applications conducted under the supervision of DoD-certified pesticide applicators. Unsupervised applications were those DoD-certified pesticide applicators did not supervise. The HRA also addresses the cumulative health effects of various combinations of pesticide active ingredients.

The HRA is based on information from a number of sources. Foremost among these were the veterans of the Gulf War. RAND personnel contacted veterans in the course of a scientific survey. Investigators interviewed preventive medicine personnel, personnel with specific knowledge of fly bait use, and personnel who participated in or witnessed delousing of enemy prisoners of war. Investigators also conducted searches of military records. Additionally, personnel from the Health Effects Division (HED), a division of the EPA Office of Pesticide Programs, provided much helpful assistance during the preparation of the HRA. HED personnel met with staff from this office on numerous occasions, and otherwise communicated on an ongoing basis from 1997 to 2000. HED personnel reviewed preliminary draft portions of the HRA in depth, and provided many helpful comments, suggestions, and recommendations.

While the risk assessment is quantitative, it is associated with significant levels of uncertainty. No real-time monitoring data exist for pesticide usage during the Gulf War, consequently the HRA must rely heavily on limited written records and on veteran recall of events which occurred 7 to 9 years prior to discussions with investigators. Thus, the descriptions of the HRA findings in Section IV.D, "Health Risk Assessment Methodology and Exposure Scenarios," are qualitative. However, investigators employed standard quantitative risk assessment methodology to arrive at the conclusions.

The HRA follows four steps commonly employed in risk assessments:

• data collection and evaluation;
  
  
• exposure assessment;
  
  
• toxicity assessment; and
  
  
• risk characterization.

The HRA's data collection and evaluation section summarizes and evaluates the data used, such as from the survey, interviews, and records searches. The HRA presents and screens pesticides and related products against specific criteria to determine which will be evaluated in detail. In the exposure assessment, investigators estimated possible human doses for the pesticide products retained for detailed evaluation. The toxicity assessment comprises hazard identification and dose-response evaluation. Hazard identification is the process of determining whether exposure to a pesticide can cause an adverse health effect relevant to humans. Dose-response evaluation is the process of quantitatively characterizing the relationship between dose and one or more defined effects, which may or may not be accurately characterized as truly "adverse." Finally, risk characterization combines the information from the exposure assessment and the toxicity assessment to estimate noncancer hazards and cancer risks. However, it must be emphasized that the HRA relies on many assumptions, associated with varying levels of uncertainty, to arrive at the estimates of hazard and risk. The hazards and risks so determined may be considered "hypothetical" in that while they reflect known use and application practices as determined from the veteran interviews, we do not have the benefit of data that characterize application rates and the actual levels of the pesticides to which the applicators were exposed. Therefore, while the risk estimates may indicate areas of potential concern for large groups, investigators cannot know with confidence the full extent to which the estimates reflect the actual hazards and risks to any particular individual.

A toxicity assessment and risk characterization can be accomplished using a "toxicological approach," and/or an "epidemiological approach." This report employed both. The toxicological approach combines the doses estimated in the exposure assessment with the "standard" toxicity values identified in the toxicity assessment. These standard toxicity values are reference doses (RfDs) derived from controlled laboratory studies of animals or humans. Most of the values listed were derived by EPA, or investigators extrapolated them from EPA-derived values. The epidemiological approach compares the doses estimated in the exposure assessment to the available benchmarks from human studies (see Tab J). The relevant studies include epidemiological studies of exposed individuals. The latter benchmarks have not been verified by an appropriate agency, such as EPA, for purposes of risk assessment, but provide another means of comparison. It should be noted that the epidemiological approach does not constitute an epidemiological study of Gulf War veterans.

Investigators used the toxicological approach to determine whether or not specific biological effects would be hypothetically likely at the possible exposure levels encountered. The epidemiological approach was used to relate doses to specific human health effects. The appropriateness in the present case of the epidemiological approach is self-evident; that is, troops were exposed to pesticides and comparison of estimated exposure levels to published levels believed to be associated with specific health outcomes is a relevant analysis. The toxicological approach is called for, given the lack of hard data on exposure, the number and variety of chemicals involved, and the need to assess potential cumulative effects. Additionally, the toxicological approach addresses, to a limited extent, the potential for future health problems. Each approach has its own strengths and weaknesses, as summarized in Table 6.

Throughout the HRA various assumptions, inputs, and the analysis in general are frequently referred to as "conservative." In the HRA, consistent with conventional risk assessment methodology, conservative means "marked by moderation or caution." A conservative assumption is one that is used due to the uncertainty present, and is more likely to overestimate than it is to underestimate exposure and/or risk. A conservative analysis, containing one or more conservative assumptions, is conducted in order to estimate an upper limit on exposure and/or risk. In a regulatory environment, where risk assessment is used to determine (future) pesticide tolerances, or drive cleanup at contaminated sites, conservative means health protective. To the extent that the HRA is an evaluation of past exposures, "health protective" is not an appropriate description. However, given that some recommendations for future pesticide use by DoD may result from the HRA, there is potentially a health-protective component.

Table 6. Toxicological and epidemiological approaches

B. DATA COLLECTION AND EVALUATION

Investigators collected and evaluated data from the following sources for use in the HRA:

• Survey. During 1999 RAND conducted a scientific survey of 2,005 Gulf War veterans randomly selected to statistically represent the entire Gulf War population on the ground in the Kuwait theater of operations (KTO).^[150]
  
  
• Preventive Medicine (PM) Interviews. Investigators interviewed 252 preventive medicine personnel, deployed on the ground in the KTO during the Gulf War, who provided specific information on exposures to pesticides and related products. Investigators conducted interviews from February 1998 through November 1999. Tab F contains additional information.
  
  
• Fly Bait Interviews. Investigators conducted interviews from December 1997 through February 1998 of 35 veterans who provided specific information on fly bait usage on the ground in the KTO during the Gulf War. Tab G contains additional information.
  
  
• Delousing Interviews. Investigators conducted interviews from November 1998 through December 1999 of 60 personnel who participated in or witnessed delousing of enemy prisoners of war (EPWs). Investigators conducted 35 interviews expressly in connection with delousing and 25 interviews as part of the PM interviews. Tab H contains additional information.
  
  
• Records Search. Investigators conducted a search of available contemporaneous deployment-derived information from sources such as the Defense Logistics Agency (DLA) and Armed Forces Pest Management Board (AFPMB). Tab I contains additional information.

The main result of the data collection and evaluation step was the identification of pesticide formulations retained for evaluation in the HRA.

Much of the data used in the HRA, which was drawn from the survey, interviews, and published sources, are in the form of percentiles. A percentile is a value on a scale of >0 to 100 that indicates the percent of the distribution that is equal to or below it. For example, if the 5^th percentile value from the survey for exposure duration (ED) is 2 days/month, then 5% of servicemembers were exposed 2 days or fewer per month. Likewise, if the 50^th percentile for ED is 20 days/month, then 50% of servicemembers were exposed 20 or fewer days per month.

In 1999, RAND scientifically surveyed 2,005 Gulf War veterans randomly selected to statistically represent all Gulf War KTO ground troops. RAND conducted the telephone survey from May to October 1999, drawing the sample from Army, Air Force, Marine Corps, and Navy servicemembers who served on the ground in the KTO between August 1, 1990 and July 31, 1991. RAND estimates that 469,047 servicemembers were actually on the ground in-theater during the period of interest. The report, "Pesticide Use During the Gulf War: A Survey of Gulf War Veterans," describes the survey in detail.^[151]

Table 7 briefly summarizes pesticide use by service according to the RAND findings. The two major categories of pesticides for survey purposes were "personal-use" pesticides and "field-use" pesticides. Personal-use pesticides are those deliberately applied to skin and/or clothing, while field-use pesticides cover all other types addressed by the survey. The most useful information for the HRA pertains to these two categories. Other types of pesticide products were those applied by certified applicators, and the lindane used for delousing EPWs. The most relevant information on these came from the preventive medicine (PM), fly bait, and delousing interviews, which are all covered following this subsection.

Table 7. Survey, percentage of servicemembers estimated to have used pesticides from a given category, by service^[152]

Among other things, servicemembers were asked if they used various personal-use pesticides and if they used or saw others use various field-use pesticides. As can be seen in Table 7, 62% of servicemembers surveyed used personal-use pesticides, while 49% saw and/or used field-use pesticides.

As determined by the survey, Table 8 presents the most common pesticide products used and/or observed by servicemembers. The products are listed in descending order of prevalence within each category. Of the products listed, DEET was associated with the highest usage rate, while benzocaine and flea/tick collars were associated with the lowest. The Army used the highest levels of the pesticides listed, while the Air Force tended to use the lowest levels, although there are three exceptions. Six of the pesticide products listed in Table 8 are quantitatively evaluated later in the HRA, while benzocaine and flea/tick collars are not. The basis for excluding the latter is explained under "Screening of Pesticides for Risk Assessment."

Table 8. Survey, percentage of servicemembers estimated to have used the most common pesticides products, by service^[153]

Significant interpretation is necessary to extract the most accurate understanding of conditions in the Gulf from the information in the survey. For example, Table 8 lists total permethrin usage at 44%. This is based on the survey usage rate for "personal-use spray" rather than the survey usage rate listed for "permethrin." The usage rate of 44% agrees fairly well with the usage rate of 40% for permethrin aerosol determined by the PM interviews. PM personnel are the most knowledgeable servicemembers regarding pesticides products. Permethrin would have been the most common personal-use aerosol spray. Although some DEET may have been present as a "personal-use spray," no DEET was issued in spray form from the military supply system. While investigators conducted the interpretation of data carefully, it is possible that some bias was introduced in extrapolating survey and interview data to the entire servicemember population deployed to the Gulf.

Table 9 and Table 10 present exposure factors of potential use in the HRA. Some of these data are more reliable than others, and investigators handled them accordingly, as the exposure assessment explains. Table 11 presents application rates for resin strips. These application rates are reliable data for risk assessment purposes since servicemembers would be likely to remember seeing a common product with which they were probably familiar. Table 12 provides selected exposure location data by service.

Table 9. Survey, exposure factors for personal-use products^[154]

Table 10. Survey, exposure factors for field-use products^[155]

Table 11. Survey, application rates for resin strips^[156]

Table 12. Survey, selected exposure location/facility data (percent)^[157]

b. Preventive Medicine Interviews

Investigators conducted preventive medicine (PM) interviews from February 1998 through November 1999 (Tab F). PM personnel are the individuals most likely to have knowledge of pesticide products used in the Army, Navy, and Marines. The PM personnel interviewed included physicians, entomologists, environmental science officers, preventive medicine specialists, and field sanitation team members. In the Air Force, pest controllers are assigned to civil engineering units. Although not technically "preventive medicine" personnel, investigators covered interviews with Air Force pest controllers as part of the PM interviews, and as used below, "PM personnel" generally includes Air Force pest controllers. The PM personnel interviewed included those who were trained in pesticide application, conducted pesticide application, supervised, or witnessed pesticide application during the Gulf War.

Of the approximately 500 PM interviews conducted, investigators identified 252 of these as "exposure interviews." To be classified as an exposure interview, the interviewee had to demonstrate at least some knowledge of pesticide usage on the ground during the Gulf War. Interviews which did not demonstrate this knowledge were excluded. All information gleaned from the interviews is subjective, as it was based on the recollections of the veterans as interpreted and compiled by investigators.

During initial reviews of the PM interviews, it became obvious that, in addition to the mass of complete, correct, and unequivocal information, there was also some incomplete, incorrect, and ambiguous information. Thus, investigators made certain assumptions to interpret such problematic data, where possible, in order to make it useful for exposure assessment. These assumptions were based on the accumulated knowledge of pesticide product availability and usage during the Gulf War. Investigators developed the following set of directions to use when processing the interviews:

While we acknowledge that employment of the directions above potentially introduced some bias, we believe that it was necessary in order to use some data for exposure assessment which would otherwise have been incorrect or useless.

The discussion in this subsection centers on Table 13 through Table 15, which present a summary of exposure data from the exposure interviews. The data presented in Table 13 and Table 15 are primarily those which have a direct bearing on the HRA. In order to facilitate discussion of the data presented in Table 13, covering 14 of the 15 pesticides of potential concern (POPCs), the HRA will refer throughout to the "block" numbers listed in the left-hand column. (See Tab D, Section B.2, "Screening Pesticide Products for Risk Assessment," below for a description of how the POPCs were identified.) Lindane (a POPC) is not covered in this section; rather, it is addressed under "Delousing Interviews." Each block is readily identifiable by the thick black line at the top, and the HRA refers to it below as "B#" (e.g., B1 for Block 1). The information presented in the summary tables is limited. The original responses contain much additional information which the tables do not reflect.

Table 13 summarizes formulation-specific exposure data. The table's approach is conservative because (1) all data and statistics are based exclusively on the 252 exposure interviews, and (2) investigators did not count negative responses or responses equal to zero. Although reliance on the 252 exposure interviews may bias the results high, it is necessary because very few of the other servicemembers were in a position to say that pesticide products were not used by their units or that they had not witnessed pesticide use; if anything, they were more likely to say that they just did not remember.

Table 13 presents various statistics, including the number of servicemembers providing a response, and the fraction of the total this represents. "Total" means the 252 exposure interviews, or all responses for a given formulation, as indicated in the table. Also listed, where appropriate, are the 10^th percentile, average, and 90^th percentile values calculated using Microsoft Excel^� based on the specific numerical responses provided by the interviewees. The percentile algorithm will calculate percentile values based on as few as two data points; however, the fewer the data points, the higher the uncertainty. In many cases, the 90^th percentile values are equal to the maximum values.

The levels of uncertainty associated with the data presented in Table 13 vary from low to high, depending upon the specific datum in question, but probably tend to be high, in part, because the data are all based on the memories of events that occurred 7-9 years prior to the interviews. Some interviewees stated that they had been deployed to many different locations around the globe over the course of many years, and were unsure sometimes about which things had occurred in which locations. Higher quality data would be possible, for example, if more detailed records had been kept and real-time monitoring, with appropriate quality controls, had been conducted during operations. The level of certainty is also related to the question asked. For example, the interviewees were probably more accurate about the names of the pesticide products used than they were about estimating how many hours per day people were exposed. A detailed explanation of each section of Table 13 follows the table.

Table 13. PM interviews, summary of formulation-specific exposure data^a

Block Number/Descriptionb B1. Pesticide Products Cited                             Total Responses 78 46 102 19 68 108 16 62 39 42 14 14 27 33 Fraction of Exposure Interviews 31% 18% 40% 8% 27% 43% 6% 25% 15% 17% 6% 6% 11% 13% B2. Final Concentration                             Number of Responses               10 12 9 2 0 0 7 Fraction of Total Responses               16% 31% 21% 14% - - 21% 10th Percentile (%)               1 0.3 2 2 - - 18 Average (%)               13 14 15 8 - - 70 90th Percentile (%)               45 44 35 13 - - 92 B3. Individuals Exposed                             Number of Responses 31 18 43 7 20 50 3 30 17 24 6 6 15 13 Fraction of Total Responses 40% 39% 42% 37% 29% 46% 19% 48% 44% 57% 43% 43% 56% 39% 10th Percentile (%) 25 75 13 1 2 1 91 1 1 1 8 1 1 3 Average (%) 76 91 76 24 75 62 95 67 48 58 69 36 55 57 90th Percentile (%) 100 100 100 55 100 100 99 100 100 100 100 90 100 100 B4. Individuals Wearing PPEc                             Number of Responses     14 5 15 37 2 27 21 27 8 12 21 20 Fraction of Total Responses     14% 26% 22% 34% 13% 44% 54% 64% 57% 86% 78% 61% 10th Percentile (%)     58 30 100 97 15 73 100 95 85 100 100 75 Average (%)     91 74 94 97 53 92 97 98 94 100 100 95 90th Percentile (%)     100 100 100 100 91 100 100 100 100 100 100 100 B5. Exposure Time                             Responses 30 18 39 8 27 50 8 33 19 24 8 6 17 17 Fraction of Total Responses 38% 39% 38% 42% 40% 46% 50% 53% 49% 57% 57% 43% 63% 52% 10th Percentile (h/d) 2 2 1 1 1 1 3 1 1 1 1 1 1 1 Average (h/d) 14 15 15 3 7 6 11 5 4 3 2 1 2 3 90th Percentile (h/d) 24 24 24 9 20 16 22 12 11 6 3 3 4 7 B6. Exposure Frequency                             Number of Responses 29 19 39 8 27 53 9 32 18 20 8 7 18 15 Fraction of Total Responses 37% 41% 38% 42% 40% 49% 56% 52% 46% 48% 57% 50% 67% 45% 10th Percentile (d/mo) 5 1 1 1 4 1 23 1 1 1 1 1 1 1 Average (d/MO) 22 23 20 11 22 18 27 13 11 10 4 6 8 10 90th Percentile (d/MO) 30 30 30 30 30 30 30 30 30 30 10 12 21 30 B7. Exposure Duration                             Number of Responses 28 16 36 9 25 50 7 31 17 22 8 6 16 12 Fraction of Total Responses 36% 35% 35% 47% 36% 46% 44% 50% 44% 52% 57% 43% 59% 36% 10th Percentile (MO) 2 2 1 2 3 1 2 1 1 1 1 1 1 1 Average (MO) 4 5 4 4 5 4 4 4 4 5 3 4 4 5 90th Percentile (MO) 8 8 8 6 9 8 6 8 9 8 7 7 8 8 B8. Applied to Skin                             Number of Responses 43 28 8 -                     Fraction of Total Responses 55% 61% 8% -                     B9. Applied to Clothing                             Number of Responses 2 3 60 2                     Fraction of Total Responses 3% 7% 59% 11%                     B10. Used Mainly Outdoors                             Number of Responses     27 1 30 62 - 36 25 29 7 5 18 22 Fraction of Total Responses     26% 5% 44% 57% - 58% 64% 69% 50% 36% 67% 67% B11. Used Mainly Indoors                             Number of Responses     9 15 13 14 9 5 5 7 5 7 3 1 Fraction of Total Responses     9% 79% 19% 13% 56% 8% 13% 17% 36% 50% 11% 3% B12. Exposed Mainly Outdrs.                             Number of Responses     19 1 28 61 - 32 23 28 6 6 16 23 Fraction of Total Responses     19% 5% 41% 56% - 52% 59% 67% 43% 43% 59% 70% B13. Exposed Mainly Indoors                             Number of Responses     10 14 15 16 9 8 5 9 5 6 5 2 Fraction of Total Responses     10% 74% 22% 15% 56% 13% 13% 21% 36% 43% 19% 6% B14. Body Parts Exposed                             Hands, arms, face, headd 38 25 19 8 18 35 2 19 14 14 5 7 10 11 Entire body 5 3 21 2 - - - 4 1 4 - - 1 1 Other parts - - 2 2 1 2 - 4 2 3 1 - 1 2

a) Based on 252 preventive medicine (PM) interviews yielding at least some apparently reliable exposure data. See "Exposure Data from Preventive Medicine Interviews" in text for detailed information. Shaded areas are either unnecessary or not relevant. For example, no one would wear PPE to apply DEET to skin. A dash (-) indicates that there were no responses in this category. b) "Total responses" in B1 means the total number of exposure interviews citing the formulation. "Fraction of Total Responses" in subsequent blocks means the number of responses/total responses. c) PPE = personal protective equipment. Each interviewee was asked to estimate the percent of personnel who wore adequate PPE. Some of those who did not respond to the specific PPE question stated elsewhere that PPE was inadequate. The latter case is not reflected in these data. d) Exposure was reported for one or more of the parts listed. "Head" includes head, face, and neck.

B1. Pesticide Products Cited presents the number of "hits" for each of the 14 POPCs listed (the 15^th POPC, lindane, is addressed under "Delousing Interviews," later). A single hit means that the pesticide product was mentioned one or more times in the course of an interview. Thus, 68 hits for 1% azamethiphos means that 1% azamethiphos was mentioned at least once in each of 68 interviews. The fraction, represented as a percent, equals the number of hits divided by the number of exposure interviews (i.e., 252). The level of certainty associated with the data presented for B1 is high.

B2. Final Concentration presents data on the percent concentration of active ingredient (a.i.) reportedly applied; that is, the final concentration after dilution, if any. For example, the 90^th percentile final concentration for chlorpyrifos, 45% liquid (EC) is 45%. What this means is that at least 10% of personnel may have applied undiluted chlorpyrifos EC. This would be a violation of the label under any circumstances.

Investigators consider these values unreliable for the HRA. The actual concentrations applied were difficult for most people to recall, as partly evidenced by the low rates of response. The values listed are generally out of line with how trained applicators (pesticide workers) would most likely have used these formulations. However, there were a small number of apparently reliable reports, based on the level of detail provided, of extremely high application rates for some formulations.

B3. Individuals Exposed presents data on the number of individuals reportedly exposed to each formulation. There is moderate to high uncertainty associated with these data. Some respondents were referring only to applicators, while some were referring to all servicemembers within their units. Also, based on a review of the actual interview narratives, it appears that some interviewees excluded servicemembers wearing personal protective equipment (PPE), while others included those wearing PPE.

B4. Individuals Wearing PPE presents data on applicators wearing PPE. These data are associated with moderate uncertainty. One factor which lowers uncertainty is that respondents were referring only to applicators. However, there are several factors that tend to increase uncertainty. First, since "adequate PPE" was not defined in the interview, some personnel may be referring to applicators who were in fact inadequately protected (e.g., they had gloves, but not respirators when respirators were required). Second, there is a low response rate for most formulations. Finally, the data do not reflect many of the qualitative statements of veterans who did not answer this specific question that they frequently had little or no PPE.

B5 through B7. Data present the exposure times (hours/day), frequencies (days), and durations (months) for each pesticide product. These data are associated with moderate to high uncertainty. The response rates are low for some formulations. Some respondents were referring to applicators only, while some were referring to all members of their units. Recall was clearly a problem.

B8. Applied to Skin presents the numbers of respondents confirming that skin application occurred for DEET and permethrin. One can safely assume that most servicemembers who used 33% DEET deliberately applied it to their skin, while 61% of respondents identified skin application of 75% DEET. Eight percent stated that servicemembers applied permethrin to skin, which constitutes a misuse of the product.

B9. Applied to Clothing presents the numbers of respondents stating that pesticide products were applied to clothing. Permethrin was by far the most widely applied, as expected. It is noteworthy interviewees did not identify 75% DEET liquid as being used very much for clothing application, as military guidance issued following the Gulf War specified clothing application as the only appropriate use for 75% DEET.^[158] Chances are, troops applied much more DEET to clothing than is reflected here, since clothing application was an intended purpose of DEET, and such application would have enhanced protection against pests. It seems unlikely that much d-phenothrin was deliberately applied to clothing; only two respondents identified this use of d-phenothrin, and they may have been referring to incidental and/or infrequent application rather than intentional and/or frequent application. No further information on this issue was obtained.

B10 and B11. Data describe the predominant location of pesticide product use as either outdoors or indoors; that is, outside or inside any enclosed structure, including a tent. Low response rates for some of the formulations limit data usability for the HRA, and the actual differences for some formulations are highly uncertain. For example, 26% said permethrin was used mainly outdoors, while 9% said it was used mainly indoors. It would be useful to know where the other 65% recalled seeing it used. On the other hand, some of the responses confirm what would be expected. For example, 91% malathion was reportedly used outdoors the majority of the time. Despite the shortcomings of the data, it is safe to conclude that most pesticide products were used more outdoors than indoors, with the exception of dichlorvos, d-phenothrin, and bendiocarb. Consideration of outdoor versus indoor use is irrelevant for DEET, since an individual is in continuous contact with it after it is applied.

B12 and B13. Data describe the predominant exposure point to the pesticide product as either outdoors or indoors. There is a high level of agreement between B10 and B12, and B11 and B13, indicating respondents answered conscientiously. This fact may be viewed as increasing the reliability of these answers, although the low response rate for some formulations has the opposite effect.

B14. Body Parts Exposed presents the numbers of veterans reporting that various body parts were predominantly exposed to pesticide products. As expected, the greatest dermal exposure for most of the pesticide products used was to hands, arms, face, head, and/or neck. The only exception is permethrin, where about half of those providing an answer to this question indicated predominant exposure was whole body.

Table 15 indicates a few other points worth noting. At least 12% of interviewees believed specific documentation may have existed regarding pesticide product use and/or exposures. At least 17% of interviewees provided information on pesticide product disposal practices. Seven percent of interviewees could describe specific incidents of overexposure to pesticide products that caused an immediate reaction or symptoms. Fifteen percent of interviewees witnessed host nation preparation and application of pesticide products.

Table 15. PM interviews, summary of other exposure data

The survey and the PM interview results are not directly comparable, as investigators undertook each component for a different purpose, employing radically different methods. Despite this, there is reasonable agreement regarding reported usage rates for DEET, permethrin, and dichlorvos. There appear to be significant discrepancies for fly baits and d-phenothrin. The survey indicates that there was a 12% usage rate for fly baits, while the PM interviews indicate a rate of at least 43%. For d-phenothrin, apparent usage rates were 28% versus 8% from the survey and PM interviews, respectively. The survey usage rates for fly baits and d-phenothrin are likely overestimates.

c. Fly Bait Interviews

Early in the pesticides exposure investigation it became apparent that troops used fly baits widely during the Gulf War. Consequently, investigators completed 35 interviews, addressing mainly fly baits, from December 1997 through February 1998 (Tab G). The RAND survey confirmed the prevalence of fly baits, as did the PM interviews. Table 16 summarizes the qualitative and quantitative exposure information from the fly bait interviews.

Table 16. Fly bait interviews, summary of exposure data^a