Posted by linkadge on January 5, 2008, at 20:07:24
In reply to Re: correlation does not imply causality, posted by SLS on January 5, 2008, at 19:40:16
>If you happen to be stricken with depression, >how long have you been trying unsuccessfully to >treat it? More than 56 weeks? If so, you might >want to try following a protocol similar to that >used in the STAR*D study.
That doesn't make sense since we have no idea of the length of time the average STAR*D patient has been depressed. Even so, to make such an assumption one would need to perform a subgroup analysis to determine that such a proposed efficacy extends to those who have been treating depression for such a lenght of time.
>So, you see how the STAR*D investigation thus >indicates that there were quite a few members of >the study population who were chronic and >difficult to treat, otherwise their depressions >would have remitted spontaneously during what >you acknowledge is an extended period of time. >Where is the selection bias here?
Where has the study indicated the percentage of participants who have been depressed for more than a year or so? Depression can be diagnosed after two weeks. What Jamal said is very important. If a good portion of the participants had the type of depression that generally remits within a year, then the length of the study is a major factor to consider. In my opinion, extended trials are only good in so much as they can work to show continued efficacy of treatments that initially work. That a patient gets better in the remaining minuates of a trial doesn't say much.
Linkadge
poster:linkadge
thread:804126
URL: http://www.dr-bob.org/babble/20080105/msgs/804529.html