Many countries have registries of certain diseases and most developed countries have high-quality cancer registries. Some of these registries contain information regarding the occupation of the registrant or an ability to link to other databases that contain this information. Such registries seem a potentially valuable resource for examining the relationship between occupation and cancer but there are several methodological challenges to their appropriate use.

The study by Vlaanderen and co-workers1 provides an interesting example of the use of a disease database to provide insight into a practical occupational medical problem. The paper describes an analysis of the Nordic Occupational Cancer Study (NOCCA) database2 undertaken to examine the hypothesis that there is a relationship between employment in printing occupations and risk of cholangiocarcinoma. The analysis was of interest because of a recently published report of a very high incidence of cholangiocarcinoma in printers at a printing premises in Japan and suggestion that either or both of dichloromethane or 1,2-propylene dichloride might have been responsible.3

The analysis reported in the Vlaanderen paper supports the …