Why Does Pharma Study Placebos?

As head of Lilly’s early-stage psychiatric drug development in the late ’90s, Potter saw that … the company’s next-generation antidepressants were faring badly, too, doing no better than placebo in seven out of 10 trials. … Potter discovered, however, that geographic location alone could determine whether a drug bested placebo or crossed the futility boundary. By the late ’90s, for example, the classic antianxiety drug diazepam (also known as Valium) was still beating placebo in France and Belgium. But when the drug was tested in the US, it was likely to fail. Conversely, Prozac performed better in America than it did in western Europe and South Africa. It was an unsettling prospect: FDA approval could hinge on where the company chose to conduct a trial. …

AsPotter and his colleagues [also] discovered that ratings by trial observers varied significantly from one testing site to another. It was like finding out that the judges in a tight race each had a different idea about the placement of the finish line. … The placebo response is highly sensitive to cultural differences. Anthropologist Daniel Moerman found that Germans are high placebo reactors in trials of ulcer drugs but low in trials of drugs for hypertension—an undertreated condition in Germany, where many people pop pills for herzinsuffizienz, or low blood pressure. Moreover, a pill’s shape, size, branding, and price all influence its effects on the body. Soothing blue capsules make more effective tranquilizers than angry red ones, except among Italian men, for whom the color blue is associated with their national soccer team—Forza Azzurri! …

AsIn the spring, Potter, who is now a VP at Merck, helped rev up a massive data-gathering effort called the Placebo Response Drug Trials Survey.  Under the auspices of the NIH, Potter and his colleagues are acquiring decades of trial data—including blood and DNA samples—to determine which variables are responsible for the apparent rise in the placebo effect. Merck, Lilly, Pfizer, AstraZeneca, GlaxoSmithKline, Sanofi-Aventis, Johnson & Johnson, and other major firms are funding the study, and the process of scrubbing volunteers’ names and other personal information from the database is about to begin.  In typically secretive industry fashion, the existence of the project itself is being kept under wraps. NIH staffers are willing to talk about it only anonymously, concerned about offending the companies paying for it.

More here.  Gee, do you think drug companies will use their better understanding of placebo effects to help us all better distinguish effective from useless drugs, or do you think they will instead use it to game the FDA approval process, to make more of their drugs look better than placebos?  What do these two theories predict about how secretive they would be about such research?

GD Star Rating
Tagged as: ,
Trackback URL:
  • bellisaurius

    Maybe I’m just a sneaky bastard, but my real question is why aren’t doctors taking advantage of placebo effects (although, maybe they already do, like when they prescribe an antibiotic for a probable viral infection, so the patient feels “they’re doing something.”).

    • JohnnyZoom

      Short answer: ethics

      Long answer: ethics

      The whole so-called doctor-patient relationship is built on trust and honesty. Under this paradigm, one doesn’t prescribe a treatment claiming its utility while knowing its utility has not been demonstrated, let alone even expected to have utility at all. Of course, if one does study placebos rigorously in their own right, then one can say this Buffon-blue sugar pill really works, with a straight face.

      One can argue that there is merit to changing that paradigm to allow for placebo exceptions, but currently that is not a popular view.

    • gwern

      I would say the reason is obvious: they would have to have insane amounts of faith in placebo to risk it.

      Can you imagine the testimony in the malpractice lawsuits? ‘Yes, I did realize that she was suffering from .’ ‘And did you prescribe any antibiotics?’ ‘No.’ ‘What *did* you prescribe?’ ‘A sugar pill.’ ‘I rest my case, your honor.’

      • The trouble with your argument is that placebos are known to be effective. There is actual medical value to prescribing a placebo … particularly if the person is non-responsive to existing antibiotics.

    • I agree with the others that, for ethical (and probably practical!) reasons, doctors shouldn’t exploit the placebo effect by prescribing known placebos.

      However, it seems that the medical community’s near-top priority right now should be to study why the placebo effect exists at all, so that it can be used in a more controlled, predictable, and ethical way. I mean, seriously, you’ve got patients who can be tricked into reacting to saline solution as if it were morphine, and you don’t want to learn what neurological process is going on so it can be tapped some other way?

      The question should be, “how can there be a placebo effect at all?” not “how can we better conduct placebo trials?” (although of course the latter is important too)

    • rcriii

      Actually if you read the article they do try to use the placebo effect, prescribing drugs they know are ineffective, or in ineffective doses.

  • Pingback: EconTech » Links for 2009.08.26()

  • Jordan Amdahl

    Would this be a problem if the placebos were made to be identical in appearance to the drug being tested? That way, drug companies could reap no advantage from altering the appearance of the drug (and therefore, its consumers’ expectations) because the placebo it would be tested against would be given the exact same advantages.

    This seems like too easy of an answer, so I wouldn’t be surprised if they do this already. But if so, where is the problem?

    • Do you think pharma is studying this to reduce or increase the advantage their drugs get via manipulating appearances?

      • JohnnyZoom

        False dichotomy. From the article:

        “Studies like this open the door to hybrid treatment strategies that exploit the placebo effect to make real drugs safer and more effective.”

        Of course, “this” is not the big NIH effort. But the same goal would apply to that research as well.

        “But why would the placebo effect seem to be getting stronger worldwide? Part of the answer may be found in the drug industry’s own success in marketing its products.”

        All the more reason to (hopefully voluntarily) stop DTC advertising. Any marketing strategy which requires the tacit presumption that one’s physician is incompetent, is insane.

        >> Gee […]
        >> game the FDA approval process

        It always boggles my mind that in the public eye, big pharma plays nearly the same part as big tobacco played 25 years ago. And yet when business schools in the 22nd century look back at these two topics, the disparities between the actual products/business will make the current one look like McCartyism in hindsight.

      • Jordan Amdahl

        Obviously, pharma wants their drugs to look as effective as possible in comparison to the placebo, but this is not the only thing in play. I have no knowledge of the regulations concerning drug trials.

        What standards, if any, are there concerning the appearance of placebos?

    • Robert Koslover

      Jordan, I like your thinking, but I’m not sure they can do that. If I understand this situation correctly (and I may not), almost all prescription pills are deliberately shaped, colored, and coded with various numbers and letters to ensure their accurate identification. Making placebos look identical to active medicines might possibly represent a hazard, or perhaps a violation of statutes? That said, I suspect they could still make them look similar.

      • Jordan Amdahl

        It certainly wouldn’t be the first time that regulatory standards got in the way of good practices, but I don’t know anything about this either.

        It is at least possible that the regulations concerning the shape, color and markings of the drugs are different before the drug is approved, but that is just speculation on my part.

    • tim

      Shouldn’t they already be made to look the same? This seems less about the ineffectiveness of pharmaceutical drugs than the effectiveness of placebos. If a blue placebo doesn’t calm Italian men but a blue drug does, while a red placebo performs the same as a red drug, the trial is saying that blue drugs work… but no better than red placebos. Then the question is whether red drugs perform better than blue ones, etc. If these variables aren’t already controlled then the trials are worthless.

  • Eric Johnson

    > the classic antianxiety drug diazepam (also known as Valium) was still beating placebo in France and Belgium. But when the drug was tested in the US, it was likely to fail

    This smells a little off, because benzodiazapines are the sort of drug that is thoroughly effective in the vast majority of patients – at least at sufficient doses. Not like anti-depressants. Reading this is a little like reading that vodka’s effects couldn’t be distinguished from those of placebo.

    It’s possible the doses were limited; that might explain such findings. The chief “toxicity” of benzos is addictiveness, so maybe some limit was imposed on doses for that reason. I have heard that they were prescribed liberally in America in the 1960s or so, the way anti-depressants are today, and that a backlash ensued when their ability to produce hard core addiction in some people became more appreciated.

  • Alan

    The observation about the cultural variation to placebo effects is striking. Such phenomena suggest a significant role for social construction in denominating what constitutes an illness or disease process or status. One can look to the disparate frequency in the U.S. of certain surgical interventions based on geography for evidence. What role does a logical positivist approach play in determining what constitutes illness, and what role social or cultural iatrogenesis? Might one hypothesize that placebo response would be more pronounced in cases mostly characterized by the latter than the former?

  • diogenese

    I’ll just second Eric — benzo’s (i.e. valium) are a very effective drug class — any model suggesting a geographic bias for their effectiveness is probably a very wrong model.

    Benzo’s are often used with opiates as a light form of anaesthesia for many minor medical / surgical procedures. (this is not a “placebo” effect).

  • George

    While I totally agree with your implication, Robin – they’re clearly trying to game the system – I think it can be argued that drug approval procedures are too strict and cumbersome. That is, the system is too risk-averse; it takes way too long for important drugs to be approved and so many people needlessly suffer from being deprived the best treatments.

    • the system is too risk-averse; it takes way too long for important drugs to be approved

      Perhaps so, but that is no reason for a drug to get approved if its sponsor is willing to defraud the FDA and denied it if its sponsor is unwilling.

  • tut

    About the cultural variation in placebo effects when testing antidepressants: Might it be that the placebo response is bigger in places where people believe in antidepressants? And the effect of prozac is bigger in places where a bigger portion of those diagnosed with depression have something that prozac is effective against?

  • Ben A

    As a former biotech guy, I would note that:

    a) Having a drug that obviously works (e.g., diazepam) is not the same as having a drug that is easy to show working in a clinical setting. Placebo creep can be a real problem, and can destroy your power calculation. We don’t (usually) have the benefit of saying to FDA “surely p = .06 on the primary is good enough”
    b) JohnnyZoom is correct — it’s both. There’s no doubt large pharma will game if they can. There’s also little doubt that they also want to know what drugs work. You’d hate to advance a crummy drug to Phase III because you had an unusually low placebo effect in Phase II. Because that crummy drug will fail in Phase III and cost you $100M.
    c) usually there is an attempt to color and size match placebos perfectly.

  • Robert Speirs

    Since depression is a “disease” with no physical cause or effect, treatment by placebos, which have no physical effect, would appear to be entirely appropriate. To say that tranquilizers “cure” depression is like saying that knockout drops “cure” hyperactivity, another physiologically nonexistent disorder. The only ethical outrage here is the use of medical language for behavioral problems.

    • tim

      Suggesting that a brain state has no physical cause or effect is patently absurd. You can’t “cure” depression but you can treat it as an undesirable brain condition that can be changed into a different condition by the administration of various drugs. It’s no different than, say, someone who doesn’t want to be sober who consumes alcohol in order to modify his state of mind, just like I will be doing later tonight.

  • TGGP

    It isn’t about pharma’s motivations, but Mind Hacks had an interesting response:

  • jonathan

    1. Placebo effects must be studied. If something has a real effect, then it needs to be analyzed.
    2. Misuse is in the eye of the beholder. If a drug is made to look better than a placebo, then how is it not?
    3. An example of why placebo effects need to be studied: if we have more “lifestyle” drugs, then we need to determine how much of the improvement is chemical and how much is lifestyle improvement related to the constant reminder of taking the pill. People seem to eat less saturated fat if they’re on Lipitor and less salt if they’re on an ACE inhibitor. Pull those apart to see which effect is chemical.

  • Floccina

    do you think drug companies will use their better understanding of placebo effects to help us all better distinguish effective from useless drugs, or do you think they will instead use it to game the FDA approval process, to make more of their drugs look better than placebos?

    Both, more of the latter.

  • Eric Johnson

    > Since depression is a “disease” with no physical cause or effect, treatment by placebos, which have no physical effect, would appear to be entirely appropriate.

    Robert Speirs, it’s exhilarating that you have plumbed the depths of the brain, and penciled in the neuron-level details of how each one of the several different human mind-states is produced. How singular that you were able to get so far ahead of the scientific community at large. Just ten years ago, diseases like narcolepsy were having a physical basis in the brain discovered all of a sudden, after generations of mystery (hypocretin) – now we will experience no more such shocks. I eagerly await the full report from your lab, and for the nonce can only bow very low indeed before one of the greatest men of all time.

  • Eric Johnson

    > Placebo creep can be a real problem

    I agree with Ben A’s points. Anxiety, depression, and chronic pain are notable for their degree of placebo responsiveness, as I recall. And that probably goes a long way towards explaining why a diazepam trial could fail to beat placebo.

    Variable sensitivity is probably an issue here too, and leads to a lower trial dose. If you hit a set of anxious people with an “elephant dose” of diazepam, they would probably all respond very strongly indeed – but such a dose wouldn’t be appropriate for everyone (would put some to sleep), and wouldn’t be used in the trial.

    Perhaps chronic effectiveness is also a problem. The acute efficacy is certainly high but if someone did a trial of chronic use (over a month) it might be hazier due to tolerance.

    • The placebo responsiveness of pain makes me wonder what the function of pain is. Analogising the human body to a warship I’ve always thought of pain as like the warning lights on the damage control panel. When Alpha Turret is on fire the Alpha-Turret-Fire light is on (and boy does that hurt). When the damage control crew extinguish the fire, the light goes out and the pain stops.

      But maybe pain is higher up the chain of command. Damage control reports are fused with an appreciation of the tactical situation to produce pain that modulates behaviour in the obvious way without being impossible to over-ride if the strategic situation demands it. So after a hit on Alpha Turret the Alpha-Turret-Fire light comes on. Another hit starts a fire close to the magazine. Then the Alpha-Turret-Fire light goes out; Alpha Turret is still on fire but the tactical situation demands that damage control crews go to fight the fire threatening the magazine. Perhaps the captain will over-rule this, unless the forward turret can be brought back into action the approaching enemy torpedo boat will sink the ship anyway.

      Scuttling the nautical metaphor and returning to the body, pain has an obvious role of modulating behaviour so as to rest an injury and then to gradually bring the affected limb back into use as it heals. Suppose that pain is integrating both the degree of damage and an instinctive estimate of how much rest is a good trade-off between healing and being out of action. Perhaps the instinctive estimate is sophisticated enough to pick up on the idea that the magic medical pill will take care of the healing, so rest is no longer required and activity need not be discouraged. Result: pain relief.

      Perhaps placebo effects are sensitive the fine distinction of whether the patient thinks of a pill as paliative or curative. Since it is probably the patients subconscious beliefs that matter to the bodies pain system, this is probably hard to research.

      • Cyan

        Pain and the resulting suffering are actually distinct neurological phenomena, although they are typically strongly coupled in neurologically intact people. Pain is a pure informational damage-report sent to the brain. This signal is processed by the insular cortex, a brain structure that handles many functions related to homeostasis and emotion. People with pain asymbolia perceive and recognize pain but do not suffer from it.

  • Pingback: Moonlit Minds - »sketches in the wind« » Blog Archive » Daily Digest for August 27th()

  • Pingback: false symmetry()

  • A different Diogenes


    Doctor’s cannot prescribe placebos as it is considered unethical to prescribe an inactive compound. As you say, we often will give innocuous medications or change to another drug in the same class if the first one doesn’t work which is getting close to placebo.

    And anyone who thinks benzos only work as well as placebo has never met a patient with anxiety disorder. Short-term use is very effective, long-term is not.


    There was a great article on the Placebo Effect at Skeptic Magazine here which answers your questions.


  • Did you know that:The proportion of Americans using antidepressants in a given year nearly doubled from 5.8% in 1996 to 10.1% in 2005. APalling, big pharma wom.
    Maybe it is due to the placebo efect, cause they are afraid that is is shown to the public that other things work.
    Some think it is the process of administering it. It is thought that the touching, the caring, the attention, and other interpersonal communication that is part of the controlled study process (or the therapeutic setting), along with the hopefulness and encouragement provided by the experimenter/healer, affect the mood, expectations, and beliefs of the subject, which in turn triggers physical changes such as release of endorphins, catecholamines, cortisol, or adrenaline. The process reduces stress by providing hope or reducing uncertainty about what treatment to take or what the outcome will be. The reduction in stress prevents or slows down further harmful physical changes from occurring. The healing situation provokes a conditioned response. The patient’s been healed before by the doctor (or thinks she’s been healed before by the doctor) and expects to be healed again.

  • Pingback: Health care reform « N=1()