Reproductive variables as possible confounders in occupational studies of breast and ovarian cancer in females.

In a study of occupational mortality among all females dying at age 20 years or over in British Columbia during the period 1950 through 1978, significantly elevated proportional mortality ratios (PMRs) were seen for breast and/or ovarian cancer among teachers, nurses, office clerks, and sales clerks. Approximately 91% of the female deaths were recorded as occurring among "homemakers." When the PMRs were recalculated for "working women" only (ie., excluding homemakers), most of the excesses in risks for breast and ovarian cancer observed among the four occupational groups disappeared. The authors suggest that the differences in PMR values between all women and working women are due to protective higher parity and lower maternal age at first birth among homemakers as opposed to working women in general. In occupational studies of hormone-related tumors among women, every effort should be made to control for parity level and maternal age at first birth.