APH 2002, 60, 173-185:

The use of job exposure matrices for cancer epidemiology research and surveillance.

M. Goldberg and E. Imbernon

Keywords: occupational epidemiology, cancer, job exposure matrix

The epidemiologic investigation of occupational cancers is intended to identify new pathogenic agents or to analyse the effects of occupational hazards.  The workplace today is characterised by lower levels of exposure than in the past, and these exposures tend to induce common cancers for which there are other occupational and non occupational causal agents.

Confronted with the difficulties of studying low risks, epidemiologic research is currently developing along several pathways.  One of these involves the development of methods of retrospective assessement of occupational exposures that might have taken place several decades before the onset of the disease.  These methods can be applied to very large samples of subjects and make it possible to assess the multiple concomitant exposures that are common in occupational settings.

The principle of job-exposure matrices is to associate exposure data with occupations, jobs, or workstations, making it possible to attribute exposure to subjects "automatically" by linking the exposure data with individual work histories.  Despite some limitations, its automatic nature presents decisive advantages in very large surveys.

Multi-hazard matrices have been developed, as well as matrices specific for a given agent.  Both may be applicable in population- based or in industry-based studies.  They are used for etiological research, to describe exposures in a population, or to help ensure individual follow-up of workers.  They will surely make important progress possible in the years to come; they should also become common industrial hygiene tools for prevention purposes.