Skip to Content

The Central Nervous System

Expensive Placebos Work Better

The placebo effect weirds everyone out. I think we can say that without much fear of contradiction. It seems like magic, the fact that just thinking that you’re getting some treatment of benefit can have actual benefits.
Magic, though, it is not. This new study in the journal Neurology (press release here) is, at first, one of the weirder placebo effects yet documented. But if you look closer, it provides a scientific way to think about what’s going on. What the authors did was a double-blinded study in Parkinson’s patients, testing a new injectable drug affecting dopamine receptors. The patients got two shots of the same drug, one in a formulation which they were told cost $100 per dose, and the other in a formulation that they were told cost $1500 per dose. Half the group would take them in one order, and half in the other, with the second shot to be given after the effects of the first had worn off, four hours later. The entire study group was told that the purpose of study was to prove equivalence between the two formulations, and that the two were, in fact, believed to be of similar efficacy.
What everyone got, though, was saline. There was no drug. Both placebo doses improved motor function, which was expected: the placebo effect has been documented (PDF) in Parkinson’s patients before. But when patients were told that the first dose was the expensive one, the effect was greater than when they were told that the first dose was the cheap one. The belief is that that placebo effect is so noticeable in Parkinson’s (and in pain relief and in depression) because the reward and expectation system in the brain has a large dopamine signaling component, which matches well with these conditions. And this study shows another way to maximize that in turn.
The placebo effect itself is surely driven by neurotransmitters and hormonal signaling, as is the flip-side nocebo effect. Reward and expectations versus stress and anxiety – these are emotional states, to be sure, but they work through physical mechanisms that can alleviate or exacerbate other conditions. Some of those are going to have a higher signal-to-noise than others, and the effects will vary in different people according to their own emotional makeup. That showed up in this study as well:

After the study, the participants were told about the true nature of the study. “Eight of the participants said they did have greater expectations of the “expensive” drug and were amazed at the extent of the difference brought about by their expectations,” Espay said. “Interestingly, the other four participants said they had no expectation of greater benefits of the more expensive drug, and they also showed little overall changes.”

People probably feel more effect from higher-priced homeopathic preparations, too, although they’re the same distilled water as all the rest of them. (Probably a good reason to turn around and raise the prices, or launch a more expensive line from the same production runs). If you tell people that they’re drinking expensive wine, they report that it tastes better than the cheap stuff, even though they both came from the same bottle. That link raises some interesting philosophical points – when you’re reporting a sensation like taste, there’s no way to distinguish between what’s “objectively” in the substance being tasted versus what being “added” by the mind. Some parts of medicine are closer to that than we like to think.

23 comments on “Expensive Placebos Work Better”

  1. Anonymous says:

    Interesting study and this also applies to Generics Vs Brand. The brand is more expensive so it should be better. In this case the placebo is not only for the taker but also for the prescriber! 🙂

  2. Jeff says:

    Maybe Dr. Oz will get on the bandwagon and start touting homeopathic free-range gluten free vegan placebos.
    “Be ready when the moment is right. Ask your doctor if Placebo is right for you.”

  3. luysii says:

    Neuron vol. 84 pp. 623 – 637 ’14 is a terrific review of placebos and nocebos and the considerable subtleties of experiments designed to study them., Be prepared for logical intricacy worthy of Talmudic disquisitions. Any study of drug efficacy will have to deal with these issues so it’s definitely worth reading. Better to know about them before starting to design a study.
    Here are a few things to get you started. One issue in all clinical trials is the conditioning effect of past experience with treatment for the problem. This can never be avoided. People with chronic pain (almost by definition) are negatively conditioned to respond given their experience. If previous treatments had helped, they wouldn’t be in the study. Most studies have ignored the role of gender. Also the gender of the experimentalist is important in the placebo response.
    It goes on.

  4. drug_hunter says:

    Isn’t this like the MRI experiment where they gave people the same wine but group A was told it was cheap and group B was told it was expensive, and their brains reacted differently?

  5. anon the II says:

    I think Descartes said it best. “I think, therefore I am” better. They always leave off that last part.

  6. anon says:

    So… if it works so well do you still give the patients saline after?

  7. dr z says:

    Although I don’t doubt the conclusions of the paper, 12 total participants (6 in each group) seems a bit small. This is especially because this article has picked up steam on internet news sites…
    Also, the article is behind a paywall, so I can’t actually see the quantitative results. I would curious to see that statistics as to the actual magnitude of the effect.

  8. Larry says:

    This only makes me question the objectivity of the motor response evaluation. The same is true for pain (to the degree measured for these patients) and depression. If you gave someone suffering from severe burns saline or morphine, I don’ think the placebo effect would come into play. I do accept that for symptom outcomes that are not so objective, expectations can be manipulated to affect results. Is this surprising?

  9. Chrispy says:

    I’d be pretty annoyed if I enrolled in a study only to find out that it was testing the placebo effect. I wonder if there is a reverse placebo effect (“nocebo”) for these patients now: they’ll be less likely to observe a real effect from an efficacious drug in a subsequent study. Fool me once…

  10. Larry says:

    This only makes me question the objectivity of the motor response evaluation. The same is true for pain (to the degree measured for these patients) and depression. If you gave someone suffering from severe burns saline or morphine, I don’ think the placebo effect would come into play. I do accept that for symptom outcomes that are not so objective, expectations can be manipulated to affect results. Is this surprising?

  11. placebos says:

    Important to note that placebos are a part of medical practice, doctors routinely perscribe relatively harmless creams and drugs for vague symptoms. Would have been interesting to see if it worked after revealing deceptdeception.

  12. bob says:

    A placebo that is effective because it is expensive is a typical SSRI.

  13. NJBiologist says:

    @8, 10 Larry: I haven’t read this article yet, but I know I’ve seen placebo effects reported for exercise tolerance (angina patients on a treadmill) and joint circumference (patients given sham ultrasound treatment for inflammation). Patient-reported outcomes are in no way essential for a placebo effect.

  14. Anon2 says:

    I don’t have access to the pub, but I am surely curious…a informed consent for must have the study design and study drugs. If people knew there were no active pharmaceutical products in either arm that would certainly alter the outcome.
    Did these people sign a proper ICF form?

  15. annon says:

    Maybe Chris Westphall would be interested in this…oh yeah, he just did it with Flex Pharma.

  16. Verzor says:

    @14: Perhaps they were informed that the study drug was Dihydrogen Monoxide in a saline solution. I presume they use consent similar to psychological studies that use deception.

  17. Anonymous says:

    This study shows that if you’re going to go with snake oil, you had better go with the expensive snake oil! 🙂

  18. Morten G says:

    Double-blind? No way the person instructing the patients and giving the injections didn’t guess/know it was placebo.

  19. Anonymous says:

    Doesn’t this just prove that Dr. Oz should charge more for his BS???

  20. DCRogers says:

    The placebo effect also differs between countries and indications. Germans, for example, have a stronger-than-average placebo effect on antidepressants.
    I read somewhere that, if the clinical trials for any available SSRI had been performed in Germany, that the results wouldn’t have been statistically significant because of the high German placebo response rate.

  21. Allchemistry says:

    #7
    Marginally different UPDRS-III scores:
    Overall
    baseline: 27.7 (+/-5.6)
    cheap: 23.8 (+/- 7.5)
    expensive 22.8 (+/-9.7)
    p-value (ANOVA): 0.029
    Cheap placebo first:
    baseline: 26.0 (3.9)
    cheap: 21.9 (7.4)
    expensive:24.0 (12.1)
    p:0.06
    Expensive placebo first:
    baseline: 29.4 (6.8)
    cheap: 25.6 (7.7)
    expensive:21.6 (7.4)
    p:0.01; expensive vs baseline: p=0.007

  22. Larry says:

    @13 NJBiologist–thank you. From your post, I presume the exercise tolerance was some objective measure (i.e. not patient reported)? Do you have a reference for the joint circumference study? Thanks.

  23. NJBiologist says:

    @22 Larry–I can’t seem to find my copy of the mammary artery ligation paper, and my academic connection doesn’t give me access to that paper. I think it was treadmill time, but my memory isn’t perfect. In case your access is better than mine, here is the reference (plus the inflammation study you asked for):
    Cardiac endpoints after real/sham internal mammary ligation–Cobb LA et al 1959 An Evaluation of Internal-Mammary-Artery Ligation by a Double-Blind Technic. N Engl J Med 1959; 260:1115-1118
    Swelling reduced after real/sham ultrasound–Hashish I et al 1988 Reduction of postoperative pain and swelling by ultrasound treatment: a placebo effect. Pain. 1988 Jun;33(3):303-11.
    The NEJM article continues to amaze me. With our current understanding of medical ethics, I can’t imagine any IRB anywhere approving that study–and yet it (and another paper appearing around the same time) led to a real improvement in cardiac care by convincing surgeons to stop doing a surgery that did not have any demonstrable benefit relative to placebo.

Comments are closed.