Posted by JLM on November 6, 2002, at 3:23:31
In reply to Re: Lexapro 'failed' trial: I get confused, posted by Anyuser on November 5, 2002, at 19:57:03
> I get confused by all the dissatisfaction with the FDA. My impression is that it's harder than hell to get a drug approved. There are interest groups that are very unhappy that drugs aren't approved more readily. Think of AIDs patients, or cancer patients. Think of all the biotech companies that go bust because they can't afford recurrent phase IIIs. Think of the posters on this board who advocate approval of a drug. On the other hand, there are people that blame the FDA because an approved drug isn't efficacious. Many posters on this board seem to be mad at the FDA for two contradictory reasons at once: approving drugs they don't like and refusing to approve drugs they want. I think by and large being steamed at the FDA is beside the point. If a drug doesn't sicken people, and there's a chance it works as intended, it seems to me the default mode should be that the FDA approves it, regulates advertising, and lets practitioners and the market figure out if it works.
Here's an interesting read for you:http://www.journals.apa.org/prevention/volume5/pre0050025c.html
a snippet:
"In spite of the research design flaws that may favor the drug condition, there is a huge advantage to the FDA database from a scientific perspective. The database includes all of the data from initial trials, published or not, and therefore it is not subject to the usual "file drawer" and publication biases. It would be interesting to know how many of these studies actually were published, to get an estimate of the publication bias in this literature. An indirect estimate of publication bias is possible by looking at the percentage of studies that resulted in significant differences. Antidepressants are significantly more effective than inert placebos in about two thirds of published studies (Thase, 1999). In the FDA database, Kirsch et al. (2002) found the study medication had a significant advantage over inert placebo less than half the time (in 20 of 46 trials), similar to findings from other independent analyses of the same database (Khan, Khan, & Brown, in press; Laughren, 2001). Such a pattern would be consistent with a failure to publish about 23% of antidepressant trials, presumably those showing no advantage to the antidepressant, slightly more than has been estimated in the fluoxetine literature (Gram, 1994). This is a serious problem, because clinical practice guidelines based on the published literature fail to take this distortion of the literature into account (Gilbody & Song, 2000)."
poster:JLM
thread:109458
URL: http://www.dr-bob.org/babble/20021101/msgs/126637.html