Posted by SLS on August 6, 2005, at 11:59:58
In reply to Re: We need placebo, posted by Kon on August 6, 2005, at 10:58:10
Hi Kon.
You bring up good points that I would have been disappointed had you missed them. :-)
> > My conjecture is not that people with mild depression are less responsive to antidepressants. It is that they are far more apt to respond to placebo.
> If you read over the Kirsch review they claim that this is not true. In fact, they argue that the opposite may be the case. Interestingly in the Khan study you provided in your most recent post, the placebo response was as great in patients with severe depression as those with milder depression (with the lowest placebo response actually found in the milder depressives-though not statistically significant). The authors of the study you mention write:
> "However, our results also suggest that the magnitude of placebo response was similar in the four subgroups (low moderate, high moderate, moderately severe and severe). This is in contrast to earlier findings that suggest that placebo response is higher in depressed patients with lower pre-treatment depressive symptoms."Yeah. I don't know how they got the wrong results... :-)
I was surprised to read this as it contradicts previous studies. Actually, I think there might be an "over reporting" of improvements made by placebo "responders" simply because they feel that they are getting help. They probably aren't really feeling that much better. There might even be a doctor-pleasing tendency with these people. Regardless of what the cause is of the high percentage of placebo responders (reporters), I feel that there should be some way to compensate for this in the design or interpretation of studies. Perhaps we should throw out our current primitive rating scales and move in the direction of more complete psychometric testing.
> So what does this mean? Greater response in severely depressed is due to actual drug effect and not differences in placebo response or are other confounding factors involved some of which the Khan mention:
> "Finally, specific items on the HAM-D may induce greater antidepressant/placebo differences. For example, the 17-item HAM-D contains three questions pertaining to sleep. If a patient has difficulty with sleep or has few sleep-related problems, results of the HAM-D could be skewed depending on the amount of sleep problems the patient experiences.”
> For instance if severely depressed have greater difficulty sleeping than less severely depressed and SSRIs improve this aspect then their better drug response may amount to this. I don't know if this is true but it's possible.I think you are right. In my opinion, the current self-reporting rating system is ridiculous. I don't know why there has not been as much effort applied to designing more effective rating systems for depression as has been devoted to detailing its physiology. It is quite possible that psychometric testing will discriminate between placebo "reporters" who haven't truly improved and placebo "responders" who, for whatever reason, experience a true global improvement of the illness. Perhaps the rate of the latter will more closely approximate that of rate of spontaneous remissions seen in the general population. Perhaps the placebo response rate is more of a sum of spontaneous remitters plus those who would be otherwise responsive to psychotherapy, and thus responsive to the support given by the investigating personel in the clinical study and the suggestion of a cure that the placebo brings.
I don't know. I'm just thinking out loud.
Be well.
- Scott
poster:SLS
thread:534296
URL: http://www.dr-bob.org/babble/20050803/msgs/538281.html