The papers today are full of the news of the meta-analysis of clinical trials of anti-depressant drugs (including Prozac) that shows such drugs only have a clinically significant effect for a very small subset of the most depressed patients. The authors also concluded that the larger effect compared to placebo for the most depressed patients was caused by those patients responding less well to placebo, not to the SSRI drugs being more effective for those patients. [I should add that if you are prescribed such drugs, you shouldn't stop taking them unless you are advised to by your GP].
Obviously, this is not good news for the companies who make these drugs. GlaxoSmithKline, who make Seroxat, said that the conclusions were at odds "with the very positive benefits seen in actual clinical practice" and that "This one study should not be used to cause unnecessary alarm". What the study is suggesting is that the benefits of the drugs seen in clinical practice may be partly related to the placebo effect, and that those benefits not related to the placebo effect are small and usually not clinically significant. That's the point that GSK need to address. But they don't, and not only that but they disingenuously describe a meta-analysis of 47 trials as 'one study'.
Eli Lilly (makers of Prozac) said that "Extensive scientific and medical experience has demonstrated that [Prozac] is an effective anti-depressant". This just ignores the evidence from the meta-analysis that says the opposite.
When I read these responses, I couldn't help but be reminded of the responses of homeopaths to the evidence from double-blinded, randomised, placebo-controlled trials that homeopathy doesn't work. They simply ignore that evidence and say that they know it works, through clinical experience or anecdotes.
It's disappointing, but perhaps not surprising, to see drug companies falling back on these unscientific statements. After all, homeopathy transparently doesn't work, but plenty of people still use it. Perhaps the companies are hoping for a similar outcome for their anti-depressants. The point here is not that these anti-depressant drugs are useless: clearly they do 'work', especially for the most depressed patients. The point is that for most people taking them, they don't work particularly well.
Subscribe to:
Post Comments (Atom)
12 comments:
I was also struck by their appeals to anecdote.
But, I'd like to take issue with your line:
"the larger effect compared to placebo for the most depressed patients was caused by those patients responding less well to placebo, not to the SSRI drugs being more effective for those patients"
As I'm not sure what that is actually supposed to mean - it leapt out at me from the paper as well.
The intimation seems to be that rather than antidepressants having a greater effect in severe depression this is 'really' the result of a reduced placebo effect and thus not 'real'.
Which is utterly meaningless! If the placebo effect is supposed to underlie most of the antidepressant effect in less severe depression then a reduced placebo effect n severe depression means greater antidepressant effect over and above the placebo effect in this group.
you say homeopathy 'clearly does not work'....well it works at least as well as placebo (perhaps better?) but without the toxic effects of drugs such as anti-depressants!! Visit my blog for more on placebo power, and on this anti-depressant issue: http://chaitowschat-leon.blogspot.com/
email me at leonchaitow1@mac.com
You know, they DON'T want to hear our "anecdotal" evidence of the terrible side effects of SSRI withdrawal, yet they'll gladly wave anecdotal evidence in our face today.
Funny.
Hmmm. I shall be reserving judgment on this one just for now, Paul. While the new analysis does suggest that antidepressants are marginal at best for a lot of the people that get given them, the medics I know are adamant that they are useful in severely depressed patients. Anecdote, I know, but the opinions on this are very strong so I think I'd like to wait a bit and let the dust settle. If I had a penny for every time I'd seen someone come along and argue about the analysis in a meta-analysis...
Of course, the major confounder in all studies of depression - except perhaps really severe depression - is whether the patients were really depressed, or whether they were just sad. For mild to moderate depression, even just diagnosing depression is a minefield - hence the whole ongoing controversy about "medicating sadness" and pushing SSRIs to people who would probably be best served with a talk about reassessing their position in the global scheme of things, drink less, exercise more, count your blessings and here's a kick up the backside.
What else.... I am no great friend of the PharmaCombines, but it is unfair to blame them exclusively or even primarily for this - it is the "pills for ills" (real or imagined) culture that Ben Goldacre often blogs about. The Pharmas take advantage of this culture of expectation to make $$££ money, that is much is clear, but they couldn't do it unless the patients wanted to be given something.
There is also the element of the harrassed doctor with the clearly unhappy recurrent patient sat in his/her waiting room. I would bet my mortgage that the mere existence of "friendly not-too-many-side-effects not-lethal-in-overdose well-marketed mood brightener" - as the SSRIs were first sold - makes doctors per se significantly more likely to diagnose "mild to moderate depression". The point is that the doc and the patient both leave the consultation feeling that something has been done. Even if psychotherapy is now more in vogue, if the waiting list is 6-12 months (which is not untypical) then I wouldn't count on SSRI prescriptions dropping off just yet. If doctors believe SSRIs are mostly placebo, will they stop prescribing them? I doubt it.
Anyway... while one can see this as just a question of trials, numbers and data analysis, drugs for things like depression are such a broad social issue that I wonder whether it is really possible to arrive at a "clear-cut answer". It's going to be a Hell of a next few months for NICE, that much is clear.
PS
An excellent blog written by a psychiatrist which covers aspects of depression (and many other things) is Lake Cocytus.
See e.g. this post.
Bastard Bastard Bastard.
I had thought in the car this morning that this was an excellent, witty, intelligent idea for my blog and (only because I have gone and got drunk instead of writing my blog) you have STOLEN my idea AND done a better job than I would have done.
I think it was listening to Jeremy Vine on Radio2 that really brought the message home.
Good blog entry :)
Re: 'For mild to moderate depression, even just diagnosing depression is a minefield - hence the whole ongoing controversy about "medicating sadness"...'
This is a big problem, not just in terms of treatment but for these types of studies that attempt to judge how well various treatments work. There is no blood test for depression. How is a GP equipped to decide when someone they barely know is "depressed", just "sad", or "having a bad day"? Does the diagnosis of "clinical depression" really have any meaning?
Cheers all for your comments.
Regarding the line that "the larger effect compared to placebo for the most depressed patients was caused by those patients responding less well to placebo, not to the SSRI drugs being more effective for those patients". This is slippery territory, isn't it? I took the authors to mean that the absolute effect of the drugs was the same for more severely depressed patients. But, as you say, that doesn't negate the fact that the drugs show a considerable benefit over placebo for that subset of patients. So the drugs are not 'more effective': they are 'more effective in comparison to placebo', which is similar to saying they were 'clinically more effective'...I think.
That doesn't change the fact that the drugs are probably acting more or less as placebo for large numbers of people who are being prescribed them. Of course, you can argue about whether that's a good thing or not, and you can argue about whether the NICE guidelines for cost effectiveness are appropriate.
I also agree with Dr Aust that over-medicalisation of what Jolan might call 'having a bad day' isn't entirely the fault of evil Big Pharma. I also agree that it would be very difficult to come up with a 'clear-cut answer'. But I did think that the responses of the companies to the research was somewhat reminiscent of the way homeopaths dismiss evidence against homeopathy.
And apologies to dr* t for 'stealing' your idea...
I still don't think it is particularly meaningful to imply that decreased placebo response causes increasing efficacy, but it is interesting to note that it doesn't in fact seem to be the case anyway when you look at the raw data.
PJ, I should belatedly thank you for these links. I've been away in Sinai, yet again. This is very interesting stuff.
Hello, not being a scientist I am being more and more confused: we are told that our drugs ( in this case antidepressant are scientifically tested, and so we cannot really argue withe the doctor who prescribe this to us because he has the almighty power of scientific evidence to back it up, and then some other scientist as righteous as the others come up with an other meta-analysis that prove that the previous evidence was crap.
since you brought up a comparison with homeopathy, I see on various websites that the anti-alternative medicine group always come up with irrefutable evidence that all alternative medicine is rubbish, whil the pro alternative medicine camp brings in various scientific papers and meta-analysis that prove that alternative medicine work.
Who am I suppose to believe?
It seems to me that Scientists keep on contradicting each other all the time, throwing at each others scientific papers and proofs that cannot be taken seriously.
Post a Comment