Problems encountered in the design of epidemiologic studies of cancer in pesticide users.

Epidemiologic studies of cancer in pesticide users suffer from three main difficulties: lack of appropriate controls, exposure ascertainment, and exposure to multiple pesticides. Principles of control selection are discussed in the context of cohort and case-control studies. Residence on a farm is associated with unique dietary and lifestyle factors which are important to account for in case-control studies of pesticides users on farms. Exposure ascertainment methods include information collected from job descriptions, interviews, pesticide suppliers, expert panels, application records, passive dosimetry, and biological monitoring. Sensitive methods based on cell receptor-pesticide complexes have been proposed, but have not yet been validated and field tested as methods of quantitative exposure assessment. Examination of studies which successfully helped demonstrate cancer causation is suggested as a means of determining which epidemiologic methods are most fruitful and efficient at identifying human carcinogens.