Stephen J.LurieMD, PhD, Senior Editor
To the Editor: While the WHI authors1 present both nominal and adjusted confidence intervals (CIs) for clinical outcomes in their Table 2, only the nominal results are discussed in the text. This may be misleading. For example, using the nominal CI, the results suggest a significant increase in stroke; however, using the adjusted CI, this result is no longer significant since the reported range includes 1. In the "nominal" column, differences in only nonfatal strokes are statistically significant (95% CI, 1.08-2.08). Results are similar for invasive breast cancer. The authors should justify why they used only the "nominal" column in describing their results.
Stokes HH. Risks of Postmenopausal Hormone Replacement. JAMA. 2002;288(22):2822. doi:10.1001/jama.288.22.2822-JLT1211-1-9