Posted by Larry Hoover on April 22, 2008, at 21:34:34
In reply to Re: OOOOOOOOOOOOooooooooo, posted by linkadge on April 20, 2008, at 9:04:23
> >Well, the question of good science is one not >clearly ascribed to him, but Kirsch found that >antidepressants were more effective than >placebo, p <.001. Kirsch himself established >efficacy.
>
> I'm sorry, I don't have time to review the data right now. I don't exactly agree with the conclusions you are reaching, but I can't say more until I review certain data myself. From what I understand this study reveals more when disected.I'm arguing that it reveals much less when dissected. I await your further input.
> >Same goes for those who claim they didn't work. >Unfortunately, the plural of anecdote is not >data. That's why I rely on the science.
>
> Statistical significance can mean different things in different contexts.Statistical significance has no meaning in and of itself, other than the likelihood of Type 1 error is less than some arbitrary level. Any other meaning depends entirely on context.
> >The same dataset has been studied and analyzed >ad nauseum.
>
> Thats why the conscensus is that antidepressants are only marginally better than placebo.No, that is anything but the case. That is not the concensus opinion, and basing categorical conclusions about clinical efficacy on this limited dataset is irresponsible.
The unpublished FDA data were first analyzed by Arif Khan, over a decade ago. This is what he had to say about his findings:
"The less-than-impressive results in these and other studies also calls to mind the fact that patients assigned to placebo treatment in clinical trials are not "getting nothing." The capsule they receive is pharmacologically inert but hardly inert with respect to its symbolic value and its power as a conditioned stimulus. In addition, placebo-treated patients receive all of the commonly employed treatment techniques: a thorough evaluation; an explanation for their distress; an expert healer; a plausible treatment; expectation of improvement; a healer's commitment, enthusiasm and positive regard; and an opportunity to verbalize their distress. Jerome Frank, Ph.D., in his book Persuasion and Healing: A Comparative Study of Psychotherapy made a compelling case that these parts of treatment are the active ingredients of all the psychotherapies (1993).
We also attempted to address the question about the benefits of including placebo in antidepressant clinical trials. In other words, would excluding placebo support the null hypothesis assumption in comparisons of antidepressants? As we had access to mean weekly HAM-D scores in some of the trials, we compared the magnitude of change with various antidepressants in this subsample (n=2,159). When placebo was included in a hierarchical linear regression, all of the antidepressants were shown to be statistically significantly superior to placebo. However, when placebo was excluded from such an analysis, it was found that amitriptyline was superior to nefazodone (p=0.03) and imipramine (p=0.07). Although not considered previously, it is interesting to note that antidepressants such as nefazodone and imipramine may not have seen daylight if only compared against an antidepressant such as amitriptyline, excluding placebo. In other words, placebo use may also prevent exclusion of potential antidepressants because of false negative clinical trial data.
A cautionary note is indicated about the generalization of these data to the clinical management of depressed patients. The less-than-impressive difference between drug and placebo in this and other studies of clinical trials does not speak directly to the effectiveness of antidepressants in clinical practice. Participants in antidepressant clinical trials are a highly select group and are not representative of the general population of depressed patients. They are not actively suicidal, they are almost always outpatients who are moderately rather than severely or mildly depressed, and they are free of comorbid physical or psychiatric illness. They are likely to have a higher placebo response rate than more severely ill depressed patients.
Furthermore, the primary aim of these studies is not to assess the optimal effect of antidepressants, but rather to rapidly assess efficacy of new drugs so they can be brought to the market. Therefore, dose, duration and diagnosis in clinical trials are not necessarily ideally suited to identify the optimal effects of antidepressants. Accordingly, clinical trials may identify the lower bound of the effect size compared to placebo. "
NICE published the following findings. "Strong evidence of clinical significance" is their absolute top rating. "Some evidence" is just lesser, and is used when the 95% confidence interval for relative risk includes any value greater than 0.80. Still very significant, just not quite as robust. Here we go:
There is strong evidence suggesting that there is a clinically significant difference favouring SSRIs over placebo on increasing the likelihood of patients achieving a 50% reduction in depression symptoms as measured by the HRSD (N = 1719; n = 3143; RR = 0.73; 95% CI, 0.69 to 0.78).
In moderate depression there is some evidence suggesting that there is a clinically significant difference favouring SSRIs over placebo on increasing the likelihood of patients achieving a 50% reduction in depression symptoms as measured by the HRSD (N = 320; n = 729; RR = 0.75; 95% CI, 0.65 to 0.87).
In severe depression there is strong evidence suggesting that there is a clinically significant difference favouring SSRIs over placebo on increasing the likelihood of patients achieving a 50% reduction in depression symptoms as measured by the HRSD (N = 5; n = 619; RR = 0.63; 95% CI, 0.54 to 0.73).
In very severe depression there is strong evidence suggesting that there is a clinically significant difference favouring SSRIs over placebo on increasing the likelihood of patients achieving a 50% reduction in depression symptoms as measured by the HRSD (N = 6; n = 866; RR = 0.72; 95% CI, 0.65 to 0.8).
In trials lasting eight weeks or longer, there is strong evidence suggesting that there is a clinically significant difference favouring SSRIs over placebo on increasing the likelihood of achieving a 50% reduction in depression symptoms as measured by the HRSD (N = 8; n = 1764; RR = 0.72; 95% CI, 0.66 to 0.79).
In moderate depression in trials lasting eight weeks or longer, there is some evidence suggesting that there is a clinically significant difference favouring SSRIs over placebo on increasing the likelihood of achieving a 50% reduction in depression symptoms as measured by the HRSD (N = 3; n = 729; RR = 0.75; 95% CI, 0.65 to 0.87).
In severe depression in trials lasting eight weeks or longer, there is strong evidence suggesting that there is a clinically significant difference favouring SSRIs over placebo on increasing the likelihood of achieving a 50% reduction in depression symptoms as measured by the HRSD (N = 3; n = 535; RR = 0.63; 95% CI, 0.53 to 0.74).
In very severe depression in trials lasting eight weeks or longer, there is some evidence suggesting that there is a clinically significant difference favouring SSRIs over placebo on increasing the likelihood of achieving a 50% reduction in depression symptoms as measured by the HRSD (N= 1; n= 299; RR= 0.72; 95% CI, 0.59 to 0.88).
> >No, I'm saying this Kirsch paper really sucks. >It is meaningless, but for its propaganda value. >NICE did a far better job of it, and before he >published.
>
> NICE already suggests that for mild/moderate depression the benifit/risk ratio for antidepressants is poor.No. No they do not.
They say that only about mild depression, but if and only if it is a first presentation. If a person has a history of moderate or severe depression, antidepressants are indicated for mild depression, and if simple interventions fail. In moderate depression, antidepressants are recommended *before* psychological interventions, and SSRIs are the recommended treatment.
4.4.3.1 Antidepressants are not recommended for the initial treatment of mild depression, because the riskbenefit ratio is poor.
4.4.3.2 The use of antidepressants should be considered for patients with mild depression that is persisting after other interventions, and those whose depression is associated with psychosocial and medical problems.
4.4.3.3 The use of antidepressants should be considered when patients with a past history of moderate or severe depression present with mild depression.
4.5.2.1 In moderate depression, antidepressant medication should be routinely offered to all patients before psychological interventions.
4.5.2.13 When an antidepressant is to be prescribed in routine care, it should be a
selective serotonin reuptake inhibitor (SSRI), because SSRIs are as effective as tricyclic antidepressants and are less likely to be discontinued because of side effects.> >Antidepressants superior to placebo, p <.001.
>
> From what I understand, in most trials analyzed, the difference between drug and placebo was not enough to exceed an arbitrary threshold established by NICE.Again, no. No that is not true. Even Kirsch said: "A visual inspection of Figure 4 suggests that studies' effects are fairly evenly distributed above and below the NICE criterion...", although that didn't stop him from publishing statistics that fly in the face of what even his eye can clearly see.
> >And has always failed when put to critical >thinking tests. Only people like Kirsch, who can >ignore his own antidepressants significantly >better than placebo result, one chance in 1000
> >(or less) that it is not a "real difference", >support the equivalence theory.
>
> You really think Kirsh was the first one to suggest that the difference between AD's and placebos is small.That's not at all what I said. It takes a certain sort of closed mind to conclude something that your own statistics have shown to be otherwise.
> Kirsh's applicaton of a 'clinical significance' threshold is not arbitrary. Didn't he borrow it from NICE?
It is inappropriate to apply it in post hoc analysis, link. The studies were not conducted with any relevance to such a test, and it is inappropriate to apply it retroactively. Moreover, he purposely limits his analysis to a very small sample of all available data, includes data for a drug no longer even on the market, and generalizes to "antidepressants" when nearly 50% of the data are from paroxetine trials. And just because I can speak to it, having read the complete clinical trial data for more than one of the paroxetine trials, I can attest that the early trials (all included in Kirsch's analysis) were of very poor quality.
When new questions are posed, new research needs to be done. Looking at existing datasets can guide hypothesis construction, and perhaps influence trial methodology, but new research needs to be conducted. The subject trials in Kirsch's analysis were never meant to demonstrate clinical efficacy. They were a regulatory hurdle for licensure, and you can make out of that whatever you wish. Efficacy trials are very different......but you didn't like STAR*D because it was different.....
> >Nobody ever argued that drugs work for all >depressives, but the clear evidence is that they >robustly work for some of them.
>
> And placebos work for others.Placebo treatment is not a standard clinical practise. It is an artefact of a methodology. It has no external validity.
> >And, as the combination of e.g. psychotherapy >and antidepressant works better than either one >alone
>
> That is certainly not a repeated finding. It is, however, logical. If you were in a study that gave half the patients two placebos (two supposed AD's) and the other have one placebo. Who do you think would fair better? Its all relative.I've never heard of anyone testing your hypothesis. I was trying to incorporate something you said in earlier debates, that you thought the drug effect was an active placebo effect.
> >that convincingly shows (IMHO) that there is >more to antidepressant response than simply >obtaining placebo response
>
> Placebo + CBT also works better than placebo alone.> >....otherwise, therapy plus antidepressant would >be the same as therapy alone.
>
> Well logically therapy plus placebo should be the same as therapy alone, but it isn't.Therapy plus antidepressant is the most effective treatment for depression, is it not?
> >Plural of anecdote not data, again.
>
> But I am not claiming its data. When you make a case statement that is what is meant to be.It is uncontrolled data. It is not meaningless, but it is not the basis for any conclusions. It certainly can help shape hypotheses and future experimentation.
> >Dismissing them outright is dangerous, IMHO. >Look at what happened to child suicide rates as >a result of the warnings. Better management is >the answer, not drug avoidance.
>
> Thats a separate issue. Don't get me started.I was referring to a recent Canadian paper, which showed that adolescent suicide rates increase from 40/100,000 to 150/100,000 after the antidepressant warnings came out in 2004.
> Its called relative deprivation. SSRI's have no proven antisuicide effect.
I wasn't referring to SSRIs. Medical interventions declined, and antidepressant prescriptions declined.
> Take a look at say, overall US data on child suicide rates from 1950-2003 what do you see?
I can't find my normal data source, but this one, covering 1970 onwards, shows a steady decline from the early 90's onwards. http://www.medscape.com/viewarticle/545555_Tables#T1
One American study of adolescent antidepressant use did a county by county comparison between ADs and suicide rates, and the correlation was so robust that they developed a linear regression equation from it......for each X prescriptions, one suicide was (probably) prevented. Of course, the AD prescription rate could be a proxy for interaction with supportive medical care, but the correlation was of net benefit.
Lar
poster:Larry Hoover
thread:823248
URL: http://www.dr-bob.org/babble/20080420/msgs/824891.html