White Coat Underground

As I’ve written before, the placebo effect is a rather messy phenomenon. It usually refers to the difference in outcomes in a study that are not due to the intervention but to multiple other variables associated with being in a study. More colloquially, “placebo” often means a positive effect seen from the administration of a biologically inert substance. There’s a bit of a buzz brewing about a recent brief communication in Science. The report used fMRI to look for physiologic correlates to pain responses that were attenuated by an inert substance. (For the purposes of this discussion, I’ll assume that the fMRI technique used is valid.) According to the authors:

Placebo analgesia is a prime example of how psychological factors can influence pain perception (1). It refers to a situation where the administration of an inactive treatment has a pain-relieving effect, presumably because of the participant’s belief in the analgesic effectiveness of the treatment.

This study did not involve an “inactive treatment”. If you administer a presumably inert substance with the intent of causing an effect, you are no longer studying placebo, but a specific set of behaviors to manipulate the patient’s pain response. In other words, when you set up the theater to encourage a patient to feel better, you can no longer separate out the “placebo” response from the, er, placebo response. ”Placebo” is what is left to explain an effect when your intervention is expected to be inert, and setting strong expectations in patients that a cream will reduce pain is hardly an “inactive treatment”, as the fMRI findings showed.

Measurements with real lidocaine were never taken. Subjects were subjected only to an inert cream, one presented as lidocaine, and one presented as control, so we don’t know what would have happened had the patients been exposed to a “real” analgesic.

What these researchers measured was not “how placebo works” but how a certain type of medical theater affects pain behaviors. Medical theater is probably a significant component of placebo effects measured in clinical studies, but it is not the only component, and should not be confounded with “placebo”. So while this study is interesting in showing a possible physiologic and anatomic correlate to pain modulation, I’m not convinced it illuminates our understanding of placebo.

References

Eippert, F., Finsterbusch, J., Bingel, U., & Buchel, C. (2009). Direct Evidence for Spinal Cord Involvement in Placebo Analgesia Science, 326 (5951), 404-404 DOI: 10.1126/science.1180142 (Methods available here.)

Comments

#1 bob koepp
November 4, 2009

Pal – I think the charitable way to interpret the authors’ reference to an ‘inactive treatment’ is to take them as referring strictly to the “inert substance” they administered, not to the whole clinical context in which that administration occurred. They probably aren’t quite as stupid as you credit them with being.
#2 PalMD
November 4, 2009

I don’t think they are stupid at all…this is a quasi-ideological debate on the meaning of “placebo”.
#3 JohnV
November 4, 2009

This debate, specifically “placebo” response vs placebo response is going to make my head asplode.
#4 Diane G.
November 4, 2009

OT, but maybe you can point me to some links…
Are there any studies on a sort of “anti-placebo” effect? I.e., cases in which uber-skeptics like myself realize less benefit from medications than less suspicious types?
#5 James Sweet
November 4, 2009

heh, Diane G., anecdotally (the best kind of evidence there is!) I kinda feel like I’ve experienced that myself…

This study is sort of interesting, but for me the question of whether “placebo” (in the sense of “medical theater”/”magic feather”) can reduce reported pain levels is sort of a silly question. I mean… if you think you are experiencing less pain, then you think you are experiencing less pain — this is a tautology. I guess this study was more looking into the “how”, not the “if”, so that’s interesting I suppose.

What I find more interesting is the overwhelming public perception that “magic feather” “placebo”s can improve objective outcomes as well as subjective ones (like reported pain level), and whether this has any basis in reality — and if so, to what extent, and (the million dollar question) how?
#6 Dianne
November 4, 2009

It refers to a situation where the administration of an inactive treatment has a pain-relieving effect, presumably because of the participant’s belief in the analgesic effectiveness of the treatment.

At risk of making this discussion strange or maybe that should be even stranger, I’m not sure that that’s how a placebo works.

This is just an anecdote and may be meaningless, but…Once, about 15 years ago, before non-sedating anti-histamines were OTC, I had a nasty allergic reaction (hay fever, not drug). Worse, I didn’t have any antihistamines around and no prescription to get them with. (Actually, I don’t think I had any money then either, but that’s a different problem.) Anyway, I took an NSAID instead, hoping that the pill taking ritual itself would lead to some relief. It did.

Possible explanations include:
1. The placebo effect works (for some value of “works”) even when you know it’s a placebo. Therefore, it’s not your belief at all but something else which causes the effect.
2. The NSAID had some effect at reducing inflammation which might have eased the symptoms, even though it wasn’t the ideal medication.
3. The pollen count fell causing the allergic reaction to cease and it had nothing whatsoever to do with the medication I took or didn’t take.
4. I’m just strange and this anecdote is applicable to no other person living.

Could be any of those…
#7 Alex
November 4, 2009

It’s not quite what you’re looking for, not being anything to do with “uber-skepticism”, but there is something called the nocebo effect.

http://en.wikipedia.org/wiki/Nocebo
#8 Diane G.
November 4, 2009

Alex, that was most interesting! Thanks.

Obviously, it would be ethically difficult to devise a test of the nocebo hypothesis!

But no, that is not exactly what I was thinking of…my 1st reaction is that both the “placebo” & the “nocebo” response might be associated with a less skeptical orientation…
#9 Alex
November 4, 2009

Yes, exactly.

It would also be hard to devise an experiment to test your anti-placebo skepticism hypothesis. How would we measure skepticism?
#10 Alex
November 4, 2009

Again, not exactly what you were looking for, but this is interesting:

http://www2.med.umich.edu/prmc/media/newsroom/details.cfm?ID=1359

Another good reason for doctors not to abuse the placebo effect by telling white lies.
#11 Neil B ♠
November 4, 2009

“Placebo” is what is left to explain an effect when your intervention is expected to be inert,, and then to apply the expectation of being inert to the psychological effect itself is a silly and confused hall of mirrors. The placebo effect is what is left to explain, when the agent by itself (ie, not known to the recipient) is expected to be inert in terms of “pharmacological mechanisms.” That original definition was the correct one after all. Don’t you think that some substances just don’t affect our bodies by themselves, but our thinking about them does? And “theater” or more elaborate manipulations are part of the psychological effect, not something new.

Once you find out that thinking about something has it’s own effects, that means your new expectation has to be separated from the original basis in order to make the definition continue to imply the same sort of disparity between “direct physical” and “psychological” effects.
#12 hibob
November 5, 2009

Question for PalMD and everyone else:
Are there many good (statistically powerful, well designed, non-woo, etc) studies out there that found a strong placebo response in a symptom/condition that is measured by physicians/lab tests instead of being reported by the patient? I’d like to rule out blood pressure, cortisol, and other things that respond quickly to mood. I’m thinking more along the lines of hemoglobin A1C, CBC, tumor size …
#13 catgirl
November 5, 2009

It would also be hard to devise an experiment to test your anti-placebo skepticism hypothesis. How would we measure skepticism?

I think this would be fairly easy to study by giving everyone a placebo, but telling one group that it’s definitely a placebo (the “skeptical” group), and telling a control group that they might get a placebo or they might get something active. It’s probably unethical, but it would be interesting to give the placebo to a third group and tell them they are definitely getting an active drug.
#14 Denice Walter
November 5, 2009

@ Alex; @ catgirl: lots of questionaires to measure skepticism, especially in *consumers*.However, if woo is involved, it’d probably be easier to just have them read some articles from naturalnews and have an unbiased observer counting the number of chokes, guffaws, and expletives emitted /per a set time unit.
#15 SkeptVet
November 5, 2009

I’m a bit puzzled by the distinction you’re making between a placebo and “administering a presumably inert substance with the intent of causing an effect.” I think a reasonable understanding of placebo is that it is the effect of the context and process of a treatment or trial on the patient’s perceptions of their condition. It is pointless to argue whether it is “real” or not since anything that afffects perception must be affecting the brain in some way, unless you subscribe to some variety of mind/body dualism. So the real questions are what value does the effect have, can it be harnessed practically and ethically, how do we separate it from treatment effects not mediated solely by perception (which is where we get into distinguishing woo from real medicine), and so on.

If all this study claims is that administering a treatment with no measurable effects not mediated by patient perception has a measurable effect that is mediated by perception, I don’t see what’s objectionable about it. It doesn’t support woo to investigate how our beliefs influence our perception of our condition. In fact, I would think the better we understand how these effects work, the better we will be able to isolate them from more substantive, non-perception mediated effects and the harder it wil be for woo to generate superficially supportive clinical trial evidence that turns out just to be about manipulating how patients perceive their disease rather than affecting the disease itself directly.