military data is not known, but Young (2009) has criticized methods used to adjust the records.
Nevertheless, the committee cannot dismiss EOI model findings solely on the basis that command directives prohibited spraying, given that the EOI model is based on actual military data. Inasmuch as the AgDRIFT and EOI models focus on different outcomes, however, the committee does not recommend that one model be used instead of the other for the purposes of epidemiologic studies, nor does it advocate or discourage use of either the AgDRIFT or the EOI model in epidemiologic studies. If either model is used in epidemiologic studies to predict exposure, results should be interpreted in light of the model limitations noted.
The controversy surrounding the use of the EOI and AgDRIFT models points to the difficulties inherent in assessing Agent Orange exposures of Vietnam veterans. For Operation Ranch Hand and ACC cohorts, exposure assessment is the most straightforward of all assessments of Vietnam veterans in that their exposures to Agent Orange originate predominantly from one source and one exposure route. Nevertheless, attempts to quantify their exposures, even at the level of serum biomarkers of exposure, have been less than satisfactory. For ground troops, who make up the largest group of concern, exposure assessment is considerably more complex; multiple, dispersed sources of Agent Orange exposure over multiple possible routes occurred over an extended period long ago. As a result, few studies have characterized exposure beyond “in-theater” vs “not-in-theater” comparisons. Considerable work, however, has been done by National Academies committees and others to develop exposure assessments for ground troops based on numbers, patterns, and timing of aerial spray missions combined with troop-location information. Aerial spraying has been the focus of much of the committee’s efforts given that 95% of Agent Orange used during the Vietnam War was applied by aerial spraying (Stellman et al., 2003a,b).
Regardless, it is important to note that sole emphasis on aerial spraying as an exposure source should be reconsidered. To ascribe a health effect to an exposure in an epidemiologic study accurately, one must account for all sources and routes of exposure—a concept now popularly termed total exposure assessment. In the Vietnam theater, there were undoubtedly multiple sources and routes of TCDD exposure of ground troops other than being directly under an aerial-spray mission. The relative magnitudes of those sources and whether the aerial spray route predominated are unknown and now probably unknowable. For instance, troops in the field commonly collected drinking water from streams. Some of those streams are still highly polluted with TCDD. Although the ultimate source of the TCDD in the streams may have been aerial spraying, the concentration of TCDD in the water would not necessarily correlate with spray-mission exposure estimates and could conceivably far exceed the “direct exposure” estimates, depending on the terrain, rainfall, timing of water collection, and other unknown factors. The dynamic nature of TCDD released into the environment is largely unknown quantitatively, so an exposure assessment that accounts for all sources