Posted by SLS on July 31, 2005, at 7:44:25
In reply to Re: We need placebo » SLS, posted by ed_uk on July 31, 2005, at 6:56:17
Another aspect of the study of individual compounds as antidepressants is that any one antidepressant may work for only a fraction of an ultimately responsive population. We know this to be true. Yet, we place demands upon new drugs that they be able to produce response rates tantamount to being silver bullets. We don't ask this much of other drugs in most other illnesses. If each of 5 antidepressant drugs work in only 35% (compared to the generally observed 70%) of cases in different sets of people, depending on the heterogeneity of the population and the degree of overlap, you could conceivably cull a 95% rate of remission. An example of this shortsightedness is the rejection by the FDA of gepirone, a drug that did not show as robust a statistical advantage over placebo as was desired by that agency. Yet, this drug does seem to be effective enough to get some people well. From what I understand, new data on gepirone is to be included in a resubmission to the FDA by the manufacturer.
By the way, my silly little idea did not call for the abolition of placebos, just a different way of designing placebo-controlled studies. It doesn't look like it would be a viable solution, but I think it helps bring into focus the difficulty in evaluating antidepressants against placebo when a rate of placebo response of 1/3 has yet to be explained and adjustments made to the design of studies, their statistical evaluation, or perhaps only our unreasonable expectations of a single drug.
- Scott
poster:SLS
thread:534296
URL: http://www.dr-bob.org/babble/20050728/msgs/535939.html