As part of the surveillance of work-related cancers, it would be useful to have cancer incidence rates broken down by occupation, an indicator that does not currently exist in France. To achieve this, it is necessary to know the occupational characteristics of each new cancer case. Cancer registries in some departments already track these cases, but their data does not include occupational information. The objective of these two studies was to test different methods for collecting basic information (most recent occupation and longest-held occupation) from patients prior to registration in the registries. The studies focused on urological cancers in the Hérault department and on bronchopulmonary cancers in the Vendée department. The overall response rate (validated questionnaires/total cases in the registries) was 14% and 16% in these two departments, respectively. However, in both departments, 90% of the available information on occupation and 70% on industry sectors could be coded for the last job held. However, the low overall response rate does not support extending these procedures to all registries. One of the challenges for monitoring work-related cancers will be to utilize professional career data held by pension funds and link them to individual health data. (R.A.)