AGRICOH is an international consortium of agricultural cohort studies co-coordinated by the International Agency for Research on Cancer in France, and the National Cancer Institutes in the USA, to encourage and support data sharing to study disease-exposure associations for which data pooling represents a significant gain in statistical power compared to analysis of individual cohorts. Cohorts in the consortium study a variety of health outcomes, including cancer, respiratory, neurologic, reproductive, allergic disorders, injuries, and overall mortality in association with a wide array of agricultural exposures, including potential health hazards and protective factors of relevance to agricultural populations. AGRICOH, then, offers a rich framework to conduct pooling projects on determinants of health and disease in farmers, farm workers and their families. As of May 2017, 29 cohorts from 5 continents are participating in AGRICOH. The studies, varying in size to address different objectives, are from Australia (2), Canada (2), Chile (1), Costa Rica (1), Denmark (1), France (6), New Zealand (2), Norway (3), Republic of Korea (1), South Africa (2), the UK (1), and the USA (7) (http://agricoh.iarc.fr/). Conceived independently, these cohorts adopted varying definitions of the populations and farm types they study, the approaches to select cohort members from the target populations and the extent of documenting farming related activities and exposures. In the first pooling project in AGRICOH, on application of pesticides and risk of myeloid and lymphoid malignancies, harmonisation of pesticide exposure and cancer data has made it possible to generate combined estimates of cancer-type and individual pesticide associations based on over 3 00 000 farmers and farm workers enrolled in three cohorts from France, Norway and the USA. This project illustrates challenges encountered and solutions adopted to address the study of a ubiquitous exposure suspected of being associated with excess cancer risk in this setting.