Deconstructing the Placebo

By Neuroskeptic | October 17, 2009 10:30 am

Last month Wired, announced that Placebos Are Getting More Effective. Drugmakers Are Desperate to Know Why.

The article’s a good read, and the basic story is true, at least in the case of psychiatric drugs. In clinical trials, people taking placebos do seem to get better more often now than in the past (paper). This is a big problem for Big Pharma, because it means that experimental new drugs often fail to perform better than placebo, i.e. they don’t work. Wired have just noticed this, but it’s been being discussed in the academic literature for several years.

Why is this? No-one knows. There have been many suggestions – maybe people “believe in” the benefits of drugs more nowadays, so the placebo effect is greater; maybe clinical trials are recruiting people with milder illnesses that respond better to placebo, or just get better on their own. But we really don’t have any clear idea.

What if the confusion is because of the very concept of the “placebo”? Earlier this year, the BMJ ran a short opinion piece called It’s time to put the placebo out of our misery. Robin Nunn wants us to “stop thinking in terms of placebo…The placebo construct conceals more than it clarifies.”

His central argument is an analogy. If we knew nothing about humour and observed a comedian telling jokes to an audience, we might decide there was a mysterious “audience effect” at work, and busy ourselves studying it…

Imagine that you are a visitor from another world. You observe a human audience for the first time. You notice a man making vocal sounds. He is watched by an audience. Suddenly they burst into smiles and laughter. Then they’re quiet. This cycle of quietness then laughter then quietness happens several times.

What is this strange audience effect? Not all of the man’s sounds generate an audience effect, and not every audience member reacts. You deem some members of the audience to be “audience responders,” those who are particularly influenced by the audience effect. What makes them react? A theory of the audience effect could be spun into an entire literature analogous to the literature on the placebo effect.

But what we should be doing is examining the details of jokes and of laughter –

We could learn more about what makes audiences laugh by returning to fundamentals. What is laughter? Why is “fart” funnier than “flatulence”? Why are some people just not funny no matter how many jokes they try?

And this is what we should be doing with the “placebo effect” as well –

Suppose there is no such unicorn as a placebo. Then what? Just replace the thought of placebo with something more fundamental. For those who use placebo as treatment, ask what is going on. Are you using the trappings of expertise, the white coat and diploma? Are you making your patients believe because they believe in you?

Nunn’s piece is a polemic and he seems to be conclude by calling for a “post-placebo era” in which there will be no more placebo-controlled trials (although it’s not clear what he means by this). This is going too far. But his analogy with humour is an important one because it forces us to analyse the placebo in detail.

“The placebo effect” has become a vague catch-all term for anything that seems to happen to people when you give them a sugar pill. Of course, lots of things could happen. They could feel better just because of the passage of time. Or they could realize that they’re supposed to feel better and say they feel better, even if they don’t.

The “true” placebo effect refers to improvement (or worsening) of symptoms driven purely by the psychological expectation of such. But even this is something of a catch-all term. Many things could drive this improvement. Suppose you give someone a placebo pill that you claim will make them more intelligent, and they believe it.

Believing themselves to be smarter, they start doing smart things like crosswords, math puzzles, reading hard books (or even reading Neuroskeptic), etc. But the placebo itself was just a nudge in the right direction. Anything which provided that nudge would also have worked – and the nudge itself can’t take all the credit.

The strongest meaning of the “placebo effect” is a direct effect of belief upon symptoms. You give someone a sugar pill or injection, and they immediately feel less pain, or whatever. But even this effect encompasses two kinds of things. It’s one thing if the original symptoms have a “real” medical cause, like a broken leg. But it’s another thing if the original symptoms are themselves partially or wholly driven by psychological factors, i.e. if they are “psychosomatic”.

If a placebo treats a “psychosomatic” disease, then that’s not because the placebo has some mysterious, mind-over-matter “placebo effect”. All the mystery, rather, lies with the psychosomatic disease. But this is a crucial distinction.

People seem more willing to accept the mind-over-matter powers of “the placebo” than they are to accept the existence of psychosomatic illness. As if only doctors with sugar pills possess the power of suggestion. If a simple pill can convince someone that they are cured, surely the modern world in all its complexity could convince people that they’re ill.


ResearchBlogging.orgNunn, R. (2009). It’s time to put the placebo out of our misery BMJ, 338 (apr20 2) DOI: 10.1136/bmj.b1568

  • Anonymous


    Very good, what I think about this is that placebos are by their own nature well adjusted to the patient's conditions whereas truly effective drugs cannot be a perfect match for the diversity of individual physiologies, therefore the bad cases “statistically” ( 😉 ) undermine the good results in favorables cases.
    An interesting case of misguided “one size fits all” egalitarianism.

  • Todd I. Stark

    I'm honestly confused by most critiques of placebo responding in terms of the empirical data. There's no reasonable doubt that individual variation in response to many treatments is in part related to what we expect, and that the effect is large enough to matter clinically and statistically very significant.

    The questions around “deconstruction” of the concept are the best way to explain this sort of finding and what its implications are for medical testing and medical treatment.

    Yes critics rightly point out that placebo is commonly used to mean an effect so general that it is virtually meaningless in explanations. But many alternative explanations are offer even weaker, like response to “psychosomatic” illness.

    Placebo is just responding “to please,” or perhaps more technically in the most plausible theory, in accord with a response expectancy. Well at least that's a testable theory for the mechanism.

    “Psychosomatic illness” is to me just an archaic way of framing the symptoms in terms of the presumed psychological mechanisms that maintain them. It superficially seems at first to make more sense of placebo responding, but looked at more deeply I think it quickly turns out to be a step backwards because it explains even less and isn't testable except by exclusion (there's no medical cause for your symptom, so it's “psychosomatic.”)

    Also, what if the response expectancy has nothing to do with symptoms? Then the “psychosomatic” dimension fails completely. For example, some theorists explain hypnotic suggestion response in terms of response expectancy. Classic suggestibility scales have little to nothing to do with psychosomatic symptoms, they have to do with responding (theoretically in response to an expectancy) without experiencing a sense of direct volition for the action.

    Placebo responding seems to me to be better described in these terms as a subset of response expectancies: particular medically relevant response expectancies under particular kinds of situations found in medical treatment.

    Response expectancy can usefully be considered part of any treatment. We need to be able to factor it out in order to see which of our interventions are working independently of expectancy, but it also makes sense to make best use of it to the extent that we can.

  • Anonymous

    Quote: “….surely the modern world in all its complexity could convince people that they're ill.” Unquote.

    Yes, Sir. You see it going on all the time! That's how marketing people build up sales of all unnecessary goods/foods and create a “need” in the minds of people for their products to fill their own pockets!!

  • taliesin

    I've sometimes wondered whether part of the reason the American Psychiatric Organization insists its task force adhere to NDAs when planning revisions of the Diagnostic and Statistical Manual of Mental Disorders is that they are worried they will scare away the placebo effect that underlies a good fraction of psychiatry's blockbuster drugs if they were too open about how important placebo really is.

    And bizarrely, I agree. For placebo to work you must instill confidence, and simple drugs, as long as they are relatively harmless, present quite a convenient “distribution channel” for that strange intangible quality we call confidence. I think Western medicine has known about placebo for much longer than any particular practitioner cares to think, and that has always been and will always be an important part of treatment.

    Incidentally, this also explains the intensity of Western medicine's vilification of alternative medicine – it senses the threat that placebo makes to its monopoly claim to scientific validity, and it resents the implication that for many mental disorders, any old culturally sanctioned placebo-dispensing mechanism would in a large fraction of cases be just as good.

    My own take is that it is far more tainted by historical forces and intellectual fashions than it would like to believe. But that doesn't for a moment mean that it has to be scrapped, merely that its institutional proclamations must be taken with a pinch of salt.

  • Baal

    I don't remember exactly but some Scandinavian researchers were studying magical thinking in people and they did propose an interesting idea.
    They thought that because for most of our history humans had to make do with shamans and medicine that was basically 'magic', those people who could believe that the rituals performed for them would actually cure them might have a slightly greater survival than those who didn't.
    So, they would live to pass on their genes which would include a greater tendency to magical thinking; that magical thinking in the past could have had positive benefits.

    From another perspective it seems to me that magical thinking is at the root of our sense of self. Isn't investing oneself with a potency, which we call willpower, really a form of magical thinking?
    Couldn't individuals who strongly believe themselves to have the ability to direct thir own destiny actually have more agency than someone who is fatalistic, who believes themselves at the mercy of the forces outside themselves?

  • taliesin

    Baal: I'm not sure that your evolutionary just-so story even makes sense. What are you saying? Do you have a link to the original paper or article?

  • Baal

    taliesin:I'm sorry but I don't have a link. I just vaguely recall it was a Scandinavian team. I hesitate to give too much as it would probably be unreliable. I think I remember that they were doing studies on the relationship between individuals who had little rituals they did to help when doing exams and their success. It seemed to show that some magical thinking wasn't always detrimental and could help them cope with the stresses of exams.
    I think it contrasted with students who were very rational, without any such beliefs.

    Of course they were't suggesting that magical beliefs could cure cancer or any woo like that, merely that when it came to handling anxiety and boosting confidence, some magical thinking seemed to help.
    Then again, it was one team and I don't remember how big the study was or anything like that.
    Sorry for that lack of specificity.

  • Ron C. de Weijze

    Placebo's convince the experimenter that non-experimental influences on the experiment are controlled for, between the experimental and the control group. I cannot see how that means to an end is invalidated or deconstructed. If medicine turns out to be (more) ineffective, then that is exactly what the experiment shows and not the placebo but the medicine is to blame.

  • Anna Nachesa

    There have been experiments (conducted 50 or so years ago, I believe) when a person in hypnotic dream has been told that the coin pressed against her arm was glowing from heat, and consequently, her skin developed all signs of burn. I can't provide a reference, because I have read about this experiment long time before internet has become a tap-water-like facility 😉 but if that was true, then it seems that placebo is an indirect method of studying psychosomatic effects, which are the most interesting piece in this puzzle.

  • mrG

    what is a 'pharma'? Information. It is information which is received by the systems of the body where it is interpreted and triggers a cascade that gets, in addition to usually long lists of side effects, the desired result. But it is just information; only a very few drugs are direct-actors, most are X or Y inhibitor/exciters, ie information that communicates an increase or decrease in natural body function.

    So if the chemical pharma can impart information, is it so far-fetched to accept that other information pathways might be possible? We already use optogenetics to trigger gene programs, again, information that informs natural body function.

    Thus I think it is wrong to say drugs 'do not work” simply because they perform no better than placebo. They don't work if they perform no better than chance.

    Our medical science would do well to accept the obvious: Placebo is a perfectly acceptable 'information path' and need not be tacitly counted as mystic voodoo bunkum; placebo is mystic voodoo medicine!

  • Todd I. Stark

    MrG said:

    “Thus I think it is wrong to say drugs 'do not work” simply because they perform no better than placebo. They don't work if they perform no better than chance … Placebo … 'information path'”

    This reminds me of a viewpoint that had a lot of potential but never quite got off the ground “InfoMedicine.”

    The basic idea is just that patients and healing resources a complex interacting systems and that information exchange is an important aspect of medical care.

    It never got very far, perhaps because it was difficult to make the concept precise enough to be useful. That's a problem that dogged “systems” approaches since their inception. Some have survived, some have not.

    It still makes sense I think, if we can pin it down and operationalize it better.

  • Anonymous

    The fact that placebo relief of pain has been shown to be blocked by naloxone suggests endogenous opiate production as a mechanism in at least a subset of what we call placebo effects.

  • Neuroskeptic

    Anonymous – That's certainly true, I think the evidence that placebos provide pain relief is very strong & there's been some interesting biological work there.

    But going with Nunn's idea of investigating what exactly it is that drives the pain relief… is it, specifically, the belief that the pain will stop? Or could it a more general effect of the “pleasant” idea that you're going to feel better? Pleasant stimuli are known to cause opioid release.

    A nice way of testing this would be to give someone two painful stimuli and give them a placebo “treatment” for just one of them, and see if that reduces both pains, or just one of them. Does anyone know of such a study?

  • Joergo

    Hi Neuroskeptic,
    not sure if I got the point, but
    Fabrizio Benedetti and Antonella Pollo have done extended research on pain and the placebo effect.

  • Anonymous

    I think that people think they know more than they actually do about phenomena. If a comedian tells a joke and the audience laughs, the audience laughing is an element in the funniness of the joke. Jokes are funnier, I think, if they are told to big audiences, because the crowd has an effect. Human mirror neurons are firing because other people are laughing. I think that with drugs, there's a tendency to think that the drug does all the work, when in fact it's the bodies interaction with the drug that causes the effect.



No brain. No gain.

About Neuroskeptic

Neuroskeptic is a British neuroscientist who takes a skeptical look at his own field, and beyond. His blog offers a look at the latest developments in neuroscience, psychiatry and psychology through a critical lens.


See More

@Neuro_Skeptic on Twitter


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Collapse bottom bar