Does the placebo effect inflate the effectiveness of psychotherapy?

[Bob]

Does the placebo effect inflate the effectiveness of psychotherapy?
Sarah Knowles.
The Mental Elf.
23rd March 2016
http://www.nationalelfservice.net/t...t-inflate-the-effectiveness-of-psychotherapy/

A review of this 2015 paper...

Cuijpers P, Karyotaki E, Andersson G, Li J, Mergl R, Hegerl U.
The effects of blinding on the outcomes of psychotherapy and pharmacotherapy for adult depression: A meta-analysis.
Eur Psychiatry. 2015; 30:685-93.
http://www.europsy-journal.com/article/S0924-9338(15)00131-5/abstract
http://www.sciencedirect.com/science/article/pii/S0924933815001315

Background
Randomized trials with antidepressants are often run under double blind placebo-controlled conditions, whereas those with psychotherapies are mostly unblinded. This can introduce bias in favor of psychotherapy when the treatments are directly compared. In this meta-analysis, we examine this potential source of bias.

Methods
We searched Pubmed, PsycInfo, Embase and the Cochrane database (1966 to January 2014) by combining terms indicative of psychological treatment and depression, and limited to randomized trials. We included 35 trials (with 3721 patients) in which psychotherapy and pharmacotherapy for adult depression were directly compared with each other. We calculated effect sizes for each study indicating the difference between psychotherapy and pharmacotherapy at post-test. Then, we examined the difference between studies with a placebo condition and those without in moderator analyses.

Results
We did not find a significant difference between the studies with and those without a placebo condition. The studies in which a placebo condition was included indicated no significant difference between psychotherapy and pharmacotherapy (g = −0.07; NNT = 25). Studies in which no placebo condition was included (and patients and clinicians in both conditions were not blinded), resulted in a small, but significant difference between psychotherapy and pharmacotherapy in favor of pharmacotherapy (g = −0.13; NNT = 14).

Conclusions
Studies comparing psychotherapy and pharmacotherapy in which both groups of patients (and therapists) are not blinded (no placebo condition is included) result in a very small, but significantly higher effect for pharmacotherapy.

 

[A.B.]

Is the sky blue? Is water wet?

The article is still unwilling to face the truth however, as it subscribes to the view that improvement in a placebo group is due to some mental healing effect. Perhaps the author doesn't know any better.

PS: I mean the mental elf article.

 

[Bob]

The Mental Elf review is a handy intro to the study, but I didn't find their analysis very helpful. I think the conclusion in the study's abstract is probably more helpful.

 

[Jonathan Edwards]

My impression is that Sarah Knowles is honest enough to admit she does not really understand this study. The problem is that the authors of the study do not seem to understand their study. As Knowles says to begin with, it is upside down - or maybe worse than that. It is a bit like picking up a sock and a glove and asking which is right and which is left. I cannot see why the inclusion of a placebo has anything to do with the apparent difference between two treatments you can obviously tell apart in all cases. This seems to be the sort of dross we have met before - but in this case the bid is in No Trumps.

 

[jimells]

NNT = 25

Doesn't this mean that they had to treat 25 patients to find one that responded to treatment? Or is there some other meaning for "NNT" besides "Number Needed to Treat"?

Instead of worrying about placebo effect, perhaps they should investigate why doctors are using treatments that work so poorly.

 

[Esther12]

I don't understand what they were trying to do from the abstract. Seems pretty unlikely that the placebo effect would not inflate claims about the effectiveness of psychotherapy.

 

[A.B.]

I don't understand what they were trying to do from the abstract. Seems pretty unlikely that the placebo effect would not inflate claims about the effectiveness of psychotherapy.

Indeed. We know that the placebo effect is the sum of various factors that can give the false appearance of treatment related improvement. The largest of these factors is regression to the mean. Patients and doctors having their own biases when interpreting the situation is also a big one. Arguing that the effectiveness of psychotherapy is not inflated by this is akin to arguing that the respective conditionsdon't get better without treatment, and that in a psychotherapy setting both patients and therapists are mostly free of bias.

 

[alex3619]

I am at the point where you can throw as much data at me as you like involving significance and effect size that without also seeing a scatter plot I think its potentially nonsense. A few outliers can massively distort the results. In all potentially heterogeneous conditions I think scatter plots are mandatory or the study is very hard to interpret.

I also want to see far more use of objective measures.

NNT of 25 tells me: don't bother using this treatment.

 

[Sean]

NNT of 25 tells me: don't bother using this treatment.

Yep. A big fat meh.

 

[Bob]

Doesn't this mean that they had to treat 25 patients to find one that responded to treatment? Or is there some other meaning for "NNT" besides "Number Needed to Treat"?

It definitely means the "number needed to treat". "NNT=25" or "NNT=1 in 25" means that for 25 treated there is 1 positive outcome (i.e. a 4% positive outcome rate). But in this case I think it might be referring to the number of occasions in which there was a difference in treatment effect when comparing psychotherapy and pharmacotherapy studies. (i.e. a difference between psychotherapy and pharmacotherapy studies was detected for 1 in 25 comparisons.) It seems too low to indicate how many patients respond to treatment. But i haven't read the full paper so I may be way off the mark.

 

[A.B.]

The NNT offers a measurement of the impact of a medicine or therapy by estimating the number of patients that need to be treated in order to have an impact on one person.

Imagine for instance a fictional heart attack treatment called ‘StopAttack’. Imagine that 75 percent of heart attack victims who take StopAttack survive, but only 25 percent survive if they are not given StopAttack, as shown in the graph below. This is certainly a major reduction in deaths. But note that 25 percent of people will die whether they get StopAttack or not (the bottom of the graph) and 25 percent will survive whether they get StopAttack or not (the top). These two groups are therefore unaffected by StopAttack, and the treatment neither helped nor hurt them. For the 50 percent in the middle, however, StopAttack was life-saving.

Here’s how that estimation works: If we calculate how many people we need to treat with StopAttack in order for one person to be positively affected, the number is 2. This is because StopAttack positively affected (saved the lives of) 50 percent, but did not help the 25 percent who would have died, nor the 25 percent who would have survived either way.

http://www.thennt.com/thennt-explained/

 

[Snow Leopard]

Given that there is no placebo control for a psychotherapy study, I don't really understand any conclusion that claims there is a higher or lower effect of placebo on studies of psychotherapy interventions.

 

[jimells]

http://www.thennt.com/thennt-explained/

Over the years I have found this website to be very interesting and helpful, especially when doctors are trying to sell me statins, baby aspirin, etc.