1. ^aDepartment of Epidemiology, Mailman School of Public Health, Columbia University, New York, NY 10032, USA
2. ^bInstitute for Cancer Prevention, Valhalla, NY 10595, USA
3. ^cDepartment of Health Policy and Management, Mailman School of Public Health, Columbia University, New York, NY 10032, USA

Correspondence: Dr. Steven D. Stellman, Department of Epidemiology, Mailman School of Public Health, Columbia University, 722 West 168th Street, New York, NY 10032, USA. Tel.: +1-212-305-4911; Fax: +1-212-305-9413; E-mail: sds91@columbia.edu

Received 10 January 2003; Accepted 28 October 2003.

Abstract

Nearly 19.5 million gallons of herbicides were sprayed on the Republic of Vietnam between 1961 and 1971 for military purposes. Amounts of spray and patterns of applications are available in an electronic file called HERBS that contains records of 9141 defoliation missions, including detailed coordinates of US Air Force Ranch Hand aircraft flight paths, along with chemical agent and gallonage sprayed. Two classes of models for use in epidemiological and environmental studies that utilize the HERBS data for estimating relative exposure opportunity indices are presented: a discrete "hits" model that counts instances of proximity in time and space to known herbicide applications, and a continuous exposure opportunity index, E4, that takes into account type and amount of herbicide sprayed, distance from spray application, and time interval when exposure may have occurred. Both direct spraying and indirect exposure to herbicide (or dioxin) that may have remained in the local environment are considered, using a conservative first-order model for environmental disappearance. A correction factor for dermal versus respiratory routes of entry has been incorporated. E4 has a log-normal distribution that spans six orders of magnitude, thus providing a substantial amount of discrimination between sprayed and unsprayed areas. The models improve on earlier ones by making full use of the geometry of the HERBS spray flight paths of Ranch Hand aircraft. To the extent possible so many decades after the War, the models have been qualitatively validated by comparison with recent dioxin soil and biota samples from heavily contaminated areas of Vietnam, and quantitatively validated against adipose dioxin obtained in epidemiological studies of Vietnamese. These models are incorporated within a geographic information system (GIS) that may be used, as one would expect, to identify locations such as hamlets, villages, and military installations sprayed by herbicide. In a novel application, the GIS also facilitates quantitative risk assessment in epidemiological and ecological studies by applying the models within a framework of historical reconstruction of exposure history of individuals based upon their location histories.