In epidemiologic studies of birth defects, occupational titles have frequently been used as surrogates for exposure. To avoid the error associated with such proxy exposure measures, we have designed a process in which an industrial hygienist systematically imputes exposures derived from maternal interviews. In response to a structured questionnaire, mothers of cases and controls recalled occupational and nonoccupational tasks performed or products used around the time of conception. Maternal exposures were then assigned to several a priori defined categories by an industrial hygienist. The central exposure category consists of 74 chemical families, e.g., alcohols, lead compounds. Other exposure categories are individual chemical compounds; nonchemical agents, e.g., ionizing radiation, infectious diseases; and product end-use categories, e.g., insecticides, combustion products. A detailed description of this approach and its exposure assessment potential is presented using exposure data from 220 maternal interviews. (C) 1994 Wiley-Liss, Inc.