In order to assess the occupational health risk to workers using pesticides, accurate data on exposure (including knowledge of the primary route of exposure) and on absorption are needed. In addition, a well-defined no-effect level (NOEL) derived from suitable animal data must be available. Biological monitoring, urinary metabolite excretion in particular, frequently is used to indicate whether a worker has been exposed. Interpretation of the toxicological significance of the observed urinary metabolite levels is often difficult because the relationship between these levels and toxic dose are generally unknown. Another complication is the apparent lack of correlation between patch data and urinary metabolite data. The usefulness of a metabolite to predict exposure depends on many things, including detailed knowledge of absorption and excretion characteristics of the parent compound and identification of the metabolites. These data, when combined with appropriate toxicology data, permit an analysis of the potential health risks associated with an occupational exposure to toxic chemicals. This paper will correlate data from a number of studies in which the dermal penetration of azinphosmethyl (AM) was measured in rats, rabbits, monkeys and man; and urinary alkyl phosphate metabolites were measured in orchardists exposed to AM. The feasibility of utilizing metabolite excretion to estimate exposure and ultimate risk will be discussed.