Beware Active Placebos

In my health econ class I review a classic ’98 paper by Irving Kirsch and others, suggesting that the apparent benefit of antidepressants is actually an “active placebo effect,” a stronger placebo effect than from sugar pills; patients who see they get side effects, like sleepiness or dry-mouth, feel assured they have a “real” drug:

Mean effect sizes for changes in depression were calculated for 2,318 patients who had been randomly assigned to either antidepressant medication or placebo in 19 double-blind clinical trials. As a proportion of the drug response, the placebo response was constant across different types of medication (75%), and the correlation between placebo effect and drug effect was .90. … These data raise the possibility that the apparent drug effect (25% of the drug response) is actually an active placebo effect. (more)

Kirsch is still working on the topic. His new paper avoids publication selection bias by looking at all 47 trials ever done. He still gets similar results:

We requested from the FDA all publicly releasable information about the clinical trials for efficacy conducted for marketing approval of fluoxetine, venlafaxine, nefazodone, paroxetine, sertraline, and citalopram, the six most widely prescribed antidepressants approved between 1987 and 1999. … Forty-seven clinical trials were identified in the data obtained from the FDA. … Unlike prior studies, we restricted our analysis to complete datasets that included all trials conducted, whether published or not. Thus, simple publication bias cannot underlie the results. … We found no linear relation between severity and response to medication. … The differences between drug and placebo were not clinically significant in clinical trials involving either moderately or very severely depressed patients, but did reach the criterion for trials involving patients whose mean initial depression scores were at the upper end of the very severe depression category. … The response to placebo in these trials was exceptionally large, duplicating more than 80% of the improvement observed in the drug groups. In contrast, the effect of placebo on pain is estimated to be about 50% of the response to pain medication. (more)

GD Star Rating
Tagged as:
Trackback URL:
  • The Editor’s disclaimer in the first paper says a lot about how rarely the status quo is challenged in scientific journals:

    The article that follows is a controversial one. It reaches a controversial conclusion—that much of the therapeutic benefit of antidepressant medications actually derives from placebo responding. The article reaches this conclusion by utilizing a controversial statistical approach—meta-analysis. And it employs meta-analysis controversially—by meta-analyzing studies that are very heterogeneous in subject selection criteria, treatments employed, and statistical methods used. Nonetheless, we have chosen to publish the article.

    That is one nervous editor!

    • I’m not sure we should be conflating medical and scientific journals. True, some medical research is scientific in quality and style, but more is at best clinical, and a large number of articles in medical journals are purely (if often covertly) political. For examples, the Hot-Dog fear-mongering mentioned yesterday by Bruce Schneier, the AMA and NEJM anti-gun frauds in the 1990s, and the more recent Lancet frauds about the war in Iraq.

  • Proper Dave

    So why is the active placebo effect not immediate? Most anti-depressants work only after a few weeks. Did the researchers control for that?
    Interesting finding that may explain the delayed effect:

    • The active placebo effect relies on the side-effects of the drug; if the side-effects are immediate, then the placebo effect should be, also.

  • dWj

    If it works better than any other option — to a cost-effective extent — does it matter whether it’s a placebo effect? (Perhaps as a guide to research: look into better placebos!)

    I do a lot of self-medication with placebos. If it doesn’t work for someone else because they don’t believe in it, that doesn’t affect me.

    • Well if a drug that gives dry mouth reduces depression, we should be looking for the safest drugs that give dry mouth as part of a depression reduction strategy. Rather than more dangerous “anti-deprresant” drugs that happen to reduce dry mouth.

  • One of the reasons to be nervous about this and the reason I have never mentioned Kirsch work and still won’t on my on blog is that publizing this information could cause the placebo effect to collapse.

    Kirsch’s work suggests not that anti-depressants “don’t work” but that anything you can convince someone is an anti-depressant will work.

    Well given that there are people struggling mightly with depression and given that we have already convinced them that SSRIs work this makes one reluctant to rock the boat.

    What it does suggest, however, that we should pursue “anti-depressants” with a more pleasant side effect profile. For example Buproprin. We can say “take this, but be forewarned it could really increase the intensity of your orgasms”

    • big other effect

      oh christ

      so, this is your reasoning

      “Millions of people across the nation are taking a drug that does not work. Instead of telling them not to waste their money,as I would with homeopathy, I am going to have them keep taking the drug.

      Therapy has the same results as antidepressants, so it also does not do anything. But hey. Fuck them. You know what, who cares if life is a lie, if pharmaticutical companies make billions of dollars off of a scam, and the ‘therapists” make 50,000 bucks a year for not really doing a damn thing. ”

      Yeah. Fuck the dumb people who think this might make them better. Just fuck them.Why not.

      I love your attitude.

  • Someone from the other side

    An anti depressant that does not make fat, does not cause erectile dysfunction/anorgasmia but instead goes the other way? Maybe we should put it into the drinking water…

  • Bock

    A few years ago there was some noise questioning whether the placebo effect was real. The claim was that the initial study which claimed to discover the placebo effect was flawed. Whatever happened with that?

  • Pingback: Tweets that mention Overcoming Bias : Beware Active Placebos --

  • Metacognition

    Will the active placebo effect the animal models I wonder..

  • it is important to note that even if the minor effect observed in very severly depressed patients isn’t the result of unblinding it doesn’t mean there is a ‘true’ anti-depressant effect. I mean imagine we just gave a random substance screwing with neurotransmitters to a bunch of extremely depressed people then purely by regression to the mean we should probably expect it to have some tiny positive effect. I mean if you are sufficently depressed it’s probably much harder to shift your brain state into even deeper depression than to move it further towards the normal range.

    Also I strongly object to the idea that we should cover up the failure of SSRIs to preserve the placebo effect. That might be a reasonable response if we didn’t think we could do any better but we certainly can and indeed already have agents that are far more efficacious. The reason we don’t use them is an unwillingness to give up the disease model of depression, an unwillingness bolstered by the fairy tail that SSRIs are effective treatments for depression.

    As a society we need to face up to the fact that depression may likely be better understood as the extremely unhappy end of natural variation in happiness. In other words we shouldn’t expect to find drugs that make depressed people better but don’t also make happy people feel better. Rather we should treat depression the way we treat people whose extremely slow rate of growth threatens to cause medical problems, give them drugs that make everyone feel happier/grow faster. We already know that adderall has significant anti-depressive effects and pretending that SSRIs are effective lets doctors avoid weighing the risks and benefits of these agents. Indeed, other anti-depressive agents have actually been denied by the FDA merely because non-depressed people liked to take them as well.

    • Doug S.

      Incidentally, Adderall is basically made of amphetamines…

  • If the tested intervention is noticably “active”, then there is a case for making sure the placebo is too. Otherwise the tested intervention gets an “unfair” advantage – by more effectively activating the placebo effect – making it seem more effective than it actually is.

    • That is, for example, part of why “sham surgery” is used as a control in surgical testing – rather than a pill. It is important for the validity of the test that the patient gets the signal that they have been surgically operated on by a qualified medical practitioner.

  • David Olsen

    Anyone with an interest in this comment should read Crazy Like Us by Ethan Watters, which (potentially) claims that the SSRI-depression linkage is both a myth and a marketing scheme.

    You’ll have to read the book to draw your own opinion, but I think the author did a great job in illustrating that the drug companies which sell SSRI’s created the disease out of thin air, and then marketed the pills to cure it.

  • Paul Strait

    Interesting, but I don’t think this quite explains why two antidepressants with equivalent side effect profiles can have dramatically different effects. For example, imipramine is considered the gold standard tricyclic for melancholic depression — which is consistent with your hypothesis, since imipramine has a larger side effect profile than the newer SSRIs. But, for atypical depression, imipramine has no effect, while the SSRIs are often helpful, and even more helpful are the MAOIs (which again have large side effect profiles so this is somewhat consistent with your hypothesis).

    In any event, I think it is overly simplistic to reduce the effects of these drugs to an especially strong placebo effect.

    There is one more possibility. Certain drugs increase the placebo effect. For example, when you secretly inject proglumide, there is no effect. But, when you visibly inject proglumide, it has a much larger analgesic effect than saline. As proglumide does not have any significant side effects, and as this effect exists even when no side effects are reported, I think it provides evidence that certain drugs interact biologically (rather than psychologically) with the placebo effect.

  • Pingback: Overcoming Bias : Beware Mind Drugs()