Despite growing public and scientific concerns about the possible hazards to the fetus from workplace exposures, there has been very little research focused on this important subject. A major limitation to such research has been the difficulty in identifying specific occupational exposures. In recent years, NIOSH has developed a matrix (the Job Exposure Matrix, or JEM) that allows one to impute specific occupational exposures based on an individual's job title and industry. Unlike previous matrices, the JEM is based on-site assessments of exposure, and provides estimates of the probability of exposure to given agents among persons in specific job/industry combinations. While potentially of great value, the JEM has not yet been applied to a birth defects data base. Over the last decade, we have interviewed parents of approximately 10,000 malformed infants as part of our Birth Defects Study (BDS). We propose to take advantage of these two resources in order to develop a large-scale, epidemiologic data base for the study of specific birth defects in relation to occupational exposures. Over a two-year study period, we will apply the JEM to the current BDS data base in order to develop joint distributions of specific occupational exposures and specific birth defects; these distributions will then be used as a resource for testing and generating hypotheses concerning occupation and birth defects. The project will be undertaken in three phases. In phase 1, we will use the JEM to assign specific workplace exposures to BDS parents; in phase II, we will develop tables that compare exposure frequencies in each birth defect category against all other defect categories; and in phase Ill, we will conduct preliminary explorations of hypotheses for which there is sufficient information within this new data set. By the end of the study, we expect to have developed a large and uniquely valuable epidemiologic data resource for studying the relationship of occupational exposures to birth defects. In addition, our experience in assigning specific occupational exposures may serve as a model for application to other data bases, thereby advancing in a more general way the study of occupational hazards.