Design issues for a breast cancer chemoprevention trial

A design to evaluate the efficacy of tamoxifen as a chemopreventive agent for breast cancer in healthy, high-risk women is proposed. Factors to be considered include basic incidence rates in the general population, definition of high risk, screening efforts, sample size, and compliance to protocol. Assuming a breast cancer incidence of 3/1,000/year in women 50-59 years old with a first degree relative with breast cancer, we estimate that approximately 16,000 subjects would need to be randomized into a tamoxifen-placebo controlled trial to detect a 50% reduction in incidence. This assumes a 5% (two-sided) significance level and 90% power. Such a trial would need to be multicentered and might cost $64 million over a 5-year period.