How is sex considered in recent epidemiological publications on occupational risks?

Although women account for almost half the working population in industrialised countries, a sex bias persists in publications on medical research in general and occupational health in particular. The objective was to review recent publications on how sex is considered in epidemiological studies of occupational health, and to answer the following questions: are men and women studied equally, what are the respective characteristics of the studies which comprise only men, only women, and both, and what strategy of data analysis is chosen by the authors to take account of the sex factor in mixed studies.

This review was based on publications in six journals during the year 1997, and included all the original articles reporting an epidemiological study of occupational health.

In all, 348 articles were reviewed. In 40 articles (11%), the sex of the study population was not specified. In 177 articles (51%), the study population was mixed. In 108 (31%), the population consisted exclusively of men, and in only 23 (7%), exclusively of women. Even when study populations were mixed, they included fewer women than men. The sex composition of the population was related to the occupational risk factor considered, and also to health outcome. Industrial sector workers, and exposure to chemicals were more likely to be studied in samples of men. Mortality and health outcomes such as neoplasms and cardiovascular diseases were also more often studied among men. Surprisingly, study design differed significantly according to the sex of the population, and prospective studies, cohort studies, and exposed versus non-exposed studies were more often carried out in samples of men. Among the 177 mixed studies, sex was not investigated in over a quarter (27%). In 26 articles (15%), sex was not taken into account, but the authors attempted to justify this decision. In 46 mixed studies (26%), the results were adjusted for sex, and in 46 (26%), the authors gave separate results for men and women. In 11 studies (6%), more complete strategies of data analysis were chosen, including research for interactions or adjustment, followed by stratification.

This review of recent publications in occupational health epidemiology showed that women are still less often studied than men, and that the sex factor is not investigated in many mixed studies. The results therefore underline the need for further research on occupational hazards among women, and on sex differences.