Some methodologic problems encountered in occupational health research.

The sources of error encountered in occupational health epidemiology present unique and difficult problems. Knowledge of these possible sources of error is essential for the proper design of studies interpretation of study results. This article describes important sources of bias in occupational health research including healthy worker selection, information bias and the differential distribution of workplace health services. The paper emphasizes that reference groups must be chosen with extreme care in order to avoid potential problems of healthy worker selection. In addition, knowledge of the process by which diseases are diagnosed is important in planning appropriate study designs.