The estimation and use of absolute risk for weighing the risks and benefits of selective estrogen receptor modulators for preventing breast cancer.

Abstract

In order to weigh the risks and benefits of intervention with selective estrogen response modifiers for preventing breast cancer, one needs to consider the effects of intervention on several health outcomes. For example, tamoxifen was shown to reduce the risks of breast cancer and hip fracture while increasing the risks of endometrial cancer and cardiovascular end points, including stroke. One approach to weighing risks and benefits is to estimate the net effect of the intervention on the absolute risk of each of the relevant health outcomes. To estimate this net effect, one needs to know not only the relative risk from the intervention, but also the absolute risk of the health outcome in the absence of intervention. Intervention trials yield unbiased estimates of intervention relative risks, but data are usually too limited to estimate these relative risks precisely for subgroups or for rare health outcomes. Moreover, intervention trials are usually too small to provide data for developing a model for estimating the individualized absolute risk of various health outcomes in the absence of intervention. The model of Gail et al. for projecting the individualized risk of breast cancer, as modified for use in the Breast Cancer Prevention Trial, has been validated. To weigh various risks and benefits of interventions, there is a need for research to develop such models for a range of health outcomes.