Cohort study of oestrogen treatment and the risk of endometrial cancer: evaluation of method and its applicability.

A Swedish study of the risk of endometrial cancer associated with oestrogen treatment was started in 1977. A prospective design was chosen and a cohort of 23,233 women receiving oestrogen medication was collected over a three-year period. Cohort members were recruited on the basis of prescription forms for oestrogens produced by all 120 pharmacies in a defined geographical region, resulting in comprehensive coverage of that region. A questionnaire study of a random sample (1/30) of the cohort permitted mapping of confounding factors, estimation of drug compliance and detailed characterization of oestrogen and progestagen exposures. Efficient follow-up of the cohort members was achieved by linkage of identity numbers of cohort members and those of all incident cases of endometrial cancer in the region. The expected outcome was calculated on the basis of accumulated person-years of observation in the entire cohort - and in subgroups - and age-specific incidence rates of endometrial cancer in the reference population. The present design may be generally useful for post-marketing studies of the association between drug use and side-effects.