We do not agree with Mr. Krueger that a within-participant design would have been preferable. A valid crossover study requires sufficient time for the effects of 1 drug to "wash out" before treatment with a second drug can begin. The time needed for the effects of herbal treatments to wash out is unknown, so a crossover study would be subject to the criticism that the effects of 1 drug contaminated the effects of the others. Exposing each woman in our study to all treatments would have required 14 months of participation, introducing unnecessary logistical and analytic complexity. This approach would have negated our opportunity to evaluate bone mineral density and to examine adverse events associated with longer duration of therapy. Menopausal symptoms change through time, further complicating the interpretation of crossover studies.

Dr. Ge raises a principal limitation of many studies of alternative therapies for menopause symptoms: the use of generalized scales to summarize the menopause experience. These scales use a single question or small sets of questions for each outcome (for example, hot flashes and mood). As Dr. Ge observed, changes in scale scores are usually driven by vasomotor symptoms. We believe it is preferable to use validated, outcome-specific instruments (such as hot flash diaries and sternal skin conductance for vasomotor symptoms and the Patient Health Questionnaire-9 for depression). Our study was powered to rule out differences between treatment groups greater than 1.5 vasomotor symptoms daily—a change we consider small.

One cannot conclude that a large placebo effect occurred in the HALT (Herbal Alternatives for Menopause Symptoms) study simply because symptoms decreased from baseline to month 3 in the placebo group. We excluded participants if they did not have at least 2 symptoms per day. We expected symptoms to decline from baseline to month 3 because of regression to the mean. Also, change in the placebo group was minimal from months 3 to 12, so it is incorrect to claim that there was a placebo effect after month 3.

Black cohosh was standardized to 2.5% triterpene glycosides. The doses we used were those most frequently used by naturopathic physicians in Seattle, Washington, at the time; were consistent with manufacturer recommendations; and are considerably higher than doses currently recommended by many manufacturers (40 to 80 mg). We are unaware of recommendations for doses as high as 1000 to 1500 mg and consider such a practice inadvisable. Dr. Byron may be confusing the dose of dried root with that for standardized extract.