Estrogens and Breast Cancer-Reply

In Reply. —  Randomized clinical trials provide stronger evidence than do comparable observational studies because they are less susceptible to the biases that can afflict the most rigorously conducted observational investigation. There are, however, many important medical questions that cannot be addressed by clinical trials for either practical or ethical reasons. One of these concerns the presence or absence of long-term side effects of treatments such as estrogen replacement therapy. It is thus no accident that this important question has been addressed by a large number of observational studies. How then should we evaluate this literature? The traditional approach is an informal review article by a highly regarded clinician. These reviews suffer from the fact that they are often not comprehensive and emphasize those studies whose findings confirm the biases of the reviewer.¹ In contrast, meta-analyses can provide a comprehensive evaluation and synthesis of all articles that meet specified