This is from an open-access article in the Canadian Journal of
Psychiatry: an article featuring a debate about the relevance of
randomized, controlled trials to clinical practice. It is mostly
about research on psychotherapy, but with some treatment of
psychopharmacology.
href="http://publications.cpa-apc.org/browse/documents/468&xwm=true">Are
Randomized Controlled Trials Relevant to Clinical Practice?
Can J Psychiatry. 2009;54(9):637-643.
Steven D Hollon, Bruce E Wampold
There is no abstract. Click on the title to go to the journal
page, then click on the title there to download the PDF.
I'm not going to summarize the article. It is short and concise enough
that there is little point in that. Rather, I will present a
couple of points that I consider to be interesting.
From page 4:
In a pharmacotherapy trial, it was found that an
antidepressant was more effective than a placebo, accounting for 3% of
the variability in outcome; however, the psychiatrist, providing weekly
clinical management, accounted for 9% of the variability in outcomes--
indeed, the more effective psychiatrists had better outcomes
administering placebos than the poorer psychiatrists had administering
the antidepressant. [emphasis added]
Hmmm. Now, as it turns out, the study cited was a study on
imipramine. That is one of those drugs that does require a fair
amount of experience to administer properly. Often, the rate of
dropouts -- persons who quit the study due to adverse effects -- will
run about 30%, in a study of any tricyclic antidepressant.
Therefore, managing the adverse effects is a critical part of using the
drug effectively. This requires a good understanding of the
pharmacology, and a good relationship with the patient: If they
trust you and think you know what you are doing, they are more likely
to put up with adverse effects; if you understand the pharmacology, you
will titrate the dosage more adroitly, thus minimizing the adverse
effect burden.
I am not sure that you would see the same magnitude of variability
among practitioners in a study of a drug that is easier to use, such as
an SSRI. Some, sure, but not as much.
Having said that, it does not explain why some psychiatrists get better
results with placebos. Perhaps it also takes expertise to manage
the adverse effects from placebo. (I'm serious about that, by the
way.) My leading hypothesis, however, would be that it is the
relationship that makes the difference.
Here's how I look at it. Often, in clinical trials of
antidepressants, symptoms will be reduced by ~50%. That's great
if you started out with a HAM-D score of 18. Get it down to 9,
the patient feels OK. But if they start at 40, getting it down to
20 still leaves the patient with a high degree of misery and
dysfunction. So you look to multimodal interventions to get the
score down dome more. Get the patient to exercise, which might
get you another 10% reduction. Improve nutrition, another
10%. So there is list of things to do, in order to incrementally
drive the Ham-D score down.
One of the things on that list: spend time with trusted, supportive
people. Learn what that feels like, so you can: 1) seek it out
more effectively, 2) provide the same thing to other people, and 3)
model the behavior to others, so you are more likely to get it
back from others.
Another way that the relationship (therapeutic alliance) is
important has to do with some of the other items on the
list of treatment adjuncts. We know it helps if people exercise
and eat well. We know it is hard to get people to do those
things, especially when they are depressed. Perhaps a good
treatment relationship helps with motivation for such things.
The other thing about the effect of the clinician in the outcome of
clinical trials, is this: Remember how some people improve when they
are on a waiting list for therapy? Why is that? Does it
mean that the therapy is not really doing anything? After all,
intuitively, it might seem that the wait-list condition is inert.
No, it is not. How does one get on a waiting list? By
making the decision to do something, and by being willing to make
changes. Taking the step of getting on a waiting list is a
symbolic marker of willingness to change. What is the first
critical ingredient of change? ...
----------
Addendum: this has got to be the strangest spurious spell-checker
suggestion of all time: Gewürztraminer, for imipramine.
- Log in to post comments
Another interesting quote, from page two:
I think the author just pushed traditional psychotherapy into the category of woo without meaning to.
Perhaps we need a double-blind, placebo-controlled trial of Gewürztraminer.
Well, like homeopathy, psychotherapies are all in the mind. But, unlike homeopathy, they're meant to be.
The usual medical trial methodology does break down the closer you get to psychotherapy. Experienced psychotherapists are going to know what they're administering, and patients aren't stupid. Double-blind methodology is pretty difficult, you can't control what people know about a psychotherapy in the same way you can with a pill. And psychotherapies actively and explicitly attempt to change patients' view of things - this is what these treatments are designed to do - whereas the usual medical trial methodology is about minimising the effect of patients' view of things so we can study physiological stuff.
Some treatments are harder to test than others, and it seems a bit unfair that interventions such as medication treatment and the cognitive and behavioural therapies lend themselves so nicely to empirical evaluation."
A lot of the problem stems from the use of an inappropriate definition of âplaceboâ. The definition I use is âa treatment not based on administering a physiologically active drug or a physically active treatment (i.e. surgery)â. By that definition psychotherapy is a placebo, a placebo that happens to be effective.
SBM treatments are treatments that are more effective than placebo.
âWooâ is treatments that are not more effective than placebo.
The effects of placebos are mediated through communication, and the triggering of some physiological effects via the effects of that communication on the CNS. How physiology mediates the placebo effect is a large part of my research into the physiology of basal nitric oxide. If a placebo âdoes somethingâ, there has to be physiology that mediates those placebo-derived effects.
We know that stress reduction âdoes somethingâ, and that placebos can be pretty good stress reducers. The opposite of a stress reducer, a stress increaser is the opposite of a placebo; it makes people heal more slowly.
Healing is a very complicated process and is under very complicated control. For a placebo to improve a very complicated process by accelerating healing, that placebo has to be tying into the normal physiology of healing (which is very complicated). I see a lot of this âcomplicationâ as being due to the normal regulation of the allocation of resources by normal physiology.
When under stress (such as when running from a bear), resources are allocated away from long term needs (such as healing) and held ready for short term needs (an extra burst of speed to escape from the bear). If an organism is caught by a predator, it is certain death. Any injury short of certain death is infinitely better than being caught by the bear.
http://daedalus2u.blogspot.com/2007/04/placebo-and-nocebo-effects.html
Turning off healing so as to escape from a bear is an excellent short-term trade-off. The problems of stress are long term, and derive from not turning off the âfight or flightâ state. Placebos are memes that help people to turn off the âfight or flightâ state.
It is not a surprise that a placebo from a good psychotherapist would be more effective than a placebo from a bad one, or even better than an effective antidepressant.