My colleagues and I have shown how to compare the risks and benefits of tamoxifen by combining 3 ingredients:1 the absolute risks of breast cancer and other endpoints, such as stroke, in the absence of tamoxifen; the effects of tamoxifen on these background risks (from data in Fisher and associates2); and weights for comparing the various outcomes. We used weights of 1.0 for life-threatening outcomes (invasive breast cancer, stroke, pulmonary embolism, hip fracture and endometrial cancer), 0.5 for severe outcomes (in situ breast cancer, deep vein thrombosis) and 0 for other events. We pointed out, however, that a woman's own preferred weights could be used. Tables 10 to 12 in Gail and colleagues1 indicate that the risks of tamoxifen outweigh the benefits in many women, especially older women in whom the risks from stroke and endometrial cancer are appreciable. Indeed, Rockhill and collaborators3 estimated that only 2.3% of women in the Nurses' Health Study would experience a net benefit, according to Tables 10 and 11 in our study.1 These observations reinforce the warnings outlined by Eric Wooltorton.4
Greiver suggests that the findings of Gail and colleagues1 be incorporated into a computer-based tool. Until such a program, properly validated, is available, Tables 10 to 12 in that article provide useful indications of net risk or benefit.1