Posted by linkadge on September 1, 2010, at 13:34:07 [reposted on September 3, 2010, at 18:58:53 | original URL]
In reply to Re: Trk-b sigma-1r » linkadge, posted by ed_uk2010 on September 1, 2010, at 6:56:40
I resent the notion that my opinions change with the wind. I have always contended that effects from antidepressants are slim to nonexistent. Don't get me wrong, I will still try them, and I like reading about their purported mechanisms, but ultimately I think they are much overreated.
>In any particular individual, it's very >difficult to know. On the other hand, we can see >from placebo controlled trials that the overall >response rate is usually higher with >antidepressants.........so someone must be >responding, even if others obtain no benefit.
I would argue with the choice of wording i.e. "usually" higher with antidepressants. The problem is publication bias. Sure, many of the *published* trials show that antidepressants are marginally more effective than placebo, but when the unpublished data is included via mata-analysis, the drug effect becomes more similar to that of placebo.
Some of the largest meta-analysis done to date by Kirsh, have taken into account drug data that up to this date, was not available for inclusion. When looking at all the data, antidepressants don't really differentiate from placebo.
Linkadge
poster:linkadge
thread:960610
URL: http://www.dr-bob.org/babble/neuro/20100607/msgs/961184.html