used from an early age, on breast cancer risk. As current formulations are likely to be the highest doses used over the period, the assessment of their effects cannot be cut short. To date, the cohort studies performed in the United Kingdom have been among the best and have yielded the most information. European and Scandinavian populations may continue to be important resources for surveillance: some are particularly stable, and several have prescription records of high quality and long duration. Also, some oral contraceptive formulations bound for the United States are already in use in Europe. In the United States, members of large, stable health maintenance organizations may be similarly important resources for surveillance.
Plans are already being formulated to establish sound Medicare data bases and cohorts to permit studies of the effectiveness of medical interventions (Heithoff and Lohr, 1990). Consideration should be given to ways in which linkages with the new Medicare data bases could be used to explore the influence of earlier oral contraceptive use on breast cancer incidence among older, postmenopausal women.
The Surveillance, Epidemiology, and End Results (SEER) Program monitors community-wide cancer incidence among women of all races. These registries are the nation's resource for epidemiological studies. They must maintain and, if necessary, expand their capacity to support the necessary epidemiological studies of the influence of oral contraceptive use on breast cancer incidence among U.S. women of different races and cultures.
The surveillance that needs to continue to cover oral contraceptive use and all reproductive cancers will be particularly demanding. The Cancer and Steroid Hormones (CASH) Study conducted by the Centers for Disease Control cost approximately $13.7 million. Even this sizable study, however, was not large enough to answer questions concerning the risk of breast cancer in subgroups of users. Future studies must (1) involve many thousands of cases and controls to permit statistically valid analyses of subgroups of users (e.g., individuals who used the pill prior to their first pregnancy), and (2) follow sufficient numbers of women for many years to measure how long any possible effect persists after discontinuing oral contraceptive use. These are studies of daunting size that cannot be conducted on an ad hoc basis. Nevertheless, they must be undertaken—and the necessary resources must be set aside to conduct them—to protect the health of American women.