Stephen J.LurieMD, PhD, Senior EditorIndividualAuthorJody W.ZylkeMD, Contributing EditorIndividualAuthor
In Reply: Although several writers take issue
with the lack of a drug treatment arm in our study, the simple placebo controlled
trial is a widely accepted design. A power analysis indicated that a 2-cell
study of this size would yield a power greater than 0.85 to detect a 2-point
difference at end point on the HAM-D. A 3-group study would have required
a prohibitively large sample size.
Dr Cott and colleagues are concerned about the low placebo response
rate in our study. In trials of therapy for depression, use of sensitive measures
of outcome lead to higher placebo response rates.1,2
Our design was intended to minimize placebo response to detect any therapeutic
advantage with St John's wort. We did note the possibility of a sampling bias
in our article. A total of 33 participants dropped out of our study (after
final data accounting) not the 43 originally reported.
Shelton RC, Crits-Christoph P. St John's Wort and Major Depression—Reply. JAMA. 2001;286(1):42-45. doi:10.1001/jama.286.1.42