The Placebo Effect

[Darat]

"public at large" Id say many in the medical professions also have vastly exaggerated ideas of what the placebo effect is capable of.

 

[arthwollipot]

"public at large" Id say many in the medical professions also have vastly exaggerated ideas of what the placebo effect is capable of.

It's near-universal, as we have seen in this thread.

 

[Roboramma]

It's near-universal, as we have seen in this thread.

Have we? I just reread the thread. There's a near consensus that the strong placebo effect doesn't exist. There's some discussion of whether or not the weak placebo effect is caused by a real effect or just confounding factors.

I don't think that we have seen in this thread "vastly exaggerated ideas of what the placebo effect is capable of", certainly not near-universally.

 

[dann]

See what I mean? Missing the point.

The article is about media interpretation of scientific literature. You will notice that Mike has no problem with what the study concludes. He is not disputing it. He is not questioning it. He is reporting it accurately. But read this again:

Overly simplistic headlines muddy the water around placebo effects and mislead the public

The headline says what the article is about. The article is about overly simplistic headlines, and how they muddy the water around placebo effects and mislead the public.

Arthwollipot, you really need to read that article. It's not enough to read the headline!
As I already told you, Mike Hall writes a too-long introduction to his article about a particular sturdy. The introduction is about headlines. The article isn't. It doesn't make it any better that he then gives the article a misleading headline. His reason for doing so only appears after the introduction:

What happened in the study This particular headline was in reference to a recent study published in Nature titled ‘Neural Circuit Basis of Placebo Pain Relief’.
Overly simplistic headlines muddy the water around placebo effects and mislead the public (TheSkeptic, Aug 27, 2024)

See?! He actually says so explicitly even though he himself has a peculiar and confusing headline for his own article! It would have made it clearer for you if he hadn't come up with the misleading introduction + misleading headline, obviously. It is particularly dumb to come up with a misleading headline for an article where he complains about .... headlines!

The article is explicitly and directly about how the media misrepresents scientific papers for sensational headlines. That's what it's about. The mouse study happens to be a convenient recent example of that.

Mike's entire thesis regarding the placebo effect (that he has developed through years of reporting on the podcast, and which you can follow through all of his articles that I have linked to here on the forum) is that due to poor science communication and sensationalist headlines, the public at large has a vastly exaggerated idea of what the placebo effect is capable of. This article is another piece of that thesis.

And yet, that is not what he writes about in the article! He is upset because the authors of the new study claim to have found out that there is an actual placebo effect that is independent of confounding factors, actual pain relief. It contradicts his own idea of placebo, so he wants to diminish the new study with his introduction about headlines:

The problem is that we use the same term, ‘the placebo effect’ to describe a wide array of phenomena, when the reality is that there is no singular effect. What we have instead is a morass of non-specific, inconsistent effects that confound clinical trials so perniciously they are often mistaken for clinical effects in their own right. They vary enormously depending on the context, and can include regression to the mean, experimenter bias, parallel interventions, classical conditioning, and more.
(...)
it is just one part of a much larger, more intricate picture that includes other effects like regression to the mean, reporting biases, parallel interventions, and so on.

regression to the mean ≠ pain relief
reporting biases ≠ pain relief
parallel interventions ≠ pain relief

All three things are aspects of the science of the placebo effect that, but they aren't aspects of any real pain relief. And he doesn't point to any relevance of any of these confounding effects in the new study!!!

Why won't you tell us why you and Mike Hall think that the new study is fascinating?

 

[arthwollipot]

Arthwollipot, you really need to read that article. It's not enough to read the headline!

Some of the article is simply reporting the methods and results of the study, which anybody can read by just looking at the study itself. I did not feel the need to quote that part of Mike's article, because that's not the point of the article.

 

[Roboramma]

I haven't watched it yet, but I'm willing to bet that it's not controlling the placebo effect, it's controlling the subjective perception of pain.

I'm also willing to bet that Mike Hall is going to talk about this on the show, because I'm going to send it to him after I have watched it.

I still haven't watched it. Does he mention any studies out of Harvard? Ted Kapchuk perhaps?

If so, that casts serious doubt on any conclusions.

My initial look at the paper has not thrown up a lot of red flags, though I find it interesting that a lot of people who seem to be doing research on placebo effects appear to be Chinese.

One co-author is from Harvard, but I can't see any obvious links to the thoroughly discredited acupuncturist Ted Kapchuk, so no red flags there either.

Has your view changed since you wrote the above posts? If so, how?

 

[arthwollipot]

Has your view changed since you wrote the above posts? If so, how?

Yes. It's clear that the paper is not about controlling the subjective perception of pain, it's about classical conditioning being able to stimulate the release of endorphins. My initial suspicion was incorrect.

And it also appears to have little to do with discredited acupuncturist Ted Kapchuk, so my suspicion there was also incorect.

 

[Roboramma]

Yes. It's clear that the paper is not about controlling the subjective perception of pain, it's about classical conditioning being able to stimulate the release of endorphins. My initial suspicion was incorrect.

And it also appears to have little to do with discredited acupuncturist Ted Kapchuk, so my suspicion there was also incorect.

Thanks. I very much appreciate this response.

I'm not really clear on what the paper says, but do you think it should shift our credence in the idea that at least sometimes placebo effects related to pain are at least partially due to real physiological effects (such as increased release of endorphins)? Or is that just completely orthogonal to what was being studied?

 

[Roger Ramjets]

Look at what the study says. Now look at the headline of, say, the New Scientist article, or Petrov's video. Heck, just search YouTube for the term "placebo effect" and you will find plenty more examples.

The study does not say that we have found out "how the placebo effect works", yet that is exactly how the articles and the YouTubes are presenting it...

Here's what the study actually says,

Neural circuit basis of placebo pain relief
Placebo effects are notable demonstrations of mind–body interactions. During pain perception, in the absence of any treatment, an expectation of pain relief can reduce the experience of pain—a phenomenon known as placebo analgesia. However, despite the strength of placebo effects and their impact on everyday human experience and the failure of clinical trials for new therapeutics, the neural circuit basis of placebo effects has remained unclear. Here we show that analgesia from the expectation of pain relief is mediated by rostral anterior cingulate cortex... These findings open the possibility of targeting this prefrontal cortico-ponto-cerebellar pathway with drugs or neurostimulation to treat pain.

You describe articles saying eg. ‘We may finally know how the placebo effect relieves pain’ (New Scientist, 24 July 2024) as irresponsible journalism. Yet this is exactly what the study itself is claiming.

Of course the proper headline would be 'Experiment that hasn't been reproduced makes claim that you should ignore until it is (multiple times)", because you can never be too skeptical - and no responsible journalist would presume to have the skills to interpret something they themselves are not an expert on (which is why if I was in charge of New Scientist it would just consist of the editorial and page upon page of ads, with no articles).

As for YouTube, you know any idiot can create videos and put them up there, right? I'd say any that manage to get it this right are a welcome exception to the usual dross.

But hey, let's excoriate the ones who are trying to get it right, and treat them just like the nutcases and trolls spreading misinformation to boost their egos. Because we skeptics have our own egos to boost...

 

[arthwollipot]

I'm not really clear on what the paper says, but do you think it should shift our credence in the idea that at least sometimes placebo effects related to pain are at least partially due to real physiological effects (such as increased release of endorphins)? Or is that just completely orthogonal to what was being studied?

No, as Roger points out below, this is pretty much what the paper concludes. Classical conditioning can produce an endorphin response in the absence of actual pain. The fallacy that is being reported is that this is the entire definition of the placebo effect. It isn't.

You describe articles saying eg. ‘We may finally know how the placebo effect relieves pain’ (New Scientist, 24 July 2024) as irresponsible journalism. Yet this is exactly what the study itself is claiming.

The paper concludes that they have documented a mechanism for an effect that has previously been assumed to be placebo. But by documenting it, it is no longer placebo. There is an actual mechanism involved beyond "mind over matter" which is what the public at large believes the placebo effect is, thanks to people like Ted Kapchuk and others who promote this belief.

 

[dann]

Some of the article is simply reporting the methods and results of the study, which anybody can read by just looking at the study itself. I did not feel the need to quote that part of Mike's article, because that's not the point of the article.

So when Mike Hall finally, after a misleading headline and introduction, gets to the actual point of the article, "that's not the point of the article," in your opinion.
That's a very convoluted argument! It's also interesting that it took Mike Hall so long after you allegedly sent it to him:

I'm also willing to bet that Mike Hall is going to talk about this on the show, because I'm going to send it to him after I have watched it.

I wonder why it took him so long ...

No, as Roger points out below, this is pretty much what the paper concludes. Classical conditioning can produce an endorphin response in the absence of actual pain. The fallacy that is being reported is that this is the entire definition of the placebo effect. It isn't.

It only isn't because you (and Mike Hall) insist on defining the placebo effect as everything that may give the impression that pain is being relieved in the brain-body of an individual, including confounding factors, which is nonsense.

The paper concludes that they have documented a mechanism for an effect that has previously been assumed to be placebo. But by documenting it, it is no longer placebo. There is an actual mechanism involved beyond "mind over matter" which is what the public at large believes the placebo effect is, thanks to people like Ted Kapchuk and others who promote this belief.

So by documenting a pain-relieving effect in the brain-body of the individual mouse that cannot possibly be dismissed as a confounding factor, they have documented something that isn't placebo?!
That is nonsense!
You confuse it with the argument that when documenting an actual beneficial effect of traditional medicine, it stops being alt.med. and becomes an integral part of science-based medicine instead.
It's not at all the same thing!

In order to avoid playing around with definitions and words:
The study appears to have discovered a pain-relieving effect in mice based on their previous experiences with a behavior that relieved pain, i.e. moving to a place that didn't induce pain. And it turns out that the mice seem to continue to feel a relief from pain when moving to the same place even when that place objectively exposes them to the same level of heat.

I would call that the real placebo effect, i.e. what is left when the confounding factors have been removed. You, on the other hand, would like to call it something else, and you are obviously free to do so.

 

[Lithrael]

I'm with Arth on this one. The analogy would be, a paper comes out where they can successfully prevent basal cell carcinoma from progressing, and the reporting says 'scientists finally find the trick to preventing skin cancer!'

That Arth was overly skeptical of the paper before looking at it is orthagonal to his point about pop science reporting.

His other bit about 'if there's a demonstrable mechanism/effect it's not placebo' is a linguistic argument like '(...) then it's not alternative medicine, it's just medicine' and I have no dog in that race.

 

[Roboramma]

I'm with Arth on this one. The analogy would be, a paper comes out where they can successfully prevent basal cell carcinoma from progressing, and the reporting says 'scientists finally find the trick to preventing skin cancer!'

Sure, I think we all agree about that point.

That Arth was overly skeptical of the paper before looking at it is orthagonal to his point about pop science reporting.

Yes, also agreed.

His other bit about 'if there's a demonstrable mechanism/effect it's not placebo' is a linguistic argument like '(...) then it's not alternative medicine, it's just medicine' and I have no dog in that race.

That seems to be pretty important to the subject of this thread, though, don't you think? Is there some effect where the expectation that a treatment can cause physiological changes similar to what was expected of that treatment? Or is there no such effect but there are only confounding factors, such as regression to the mean, social expectations, etc? The claim at the start of this thread is the latter: that there is no actual physiological effect.
The linguistic claim that Arth is making is that to the extent that such an effect exists, it's not the placebo effect.

This is just defining the placebo effect out of existence. The real scientific question is whether or not such an effect exists. To dismiss evidence that such an effect exists on the basis that having discovered it, it's no longer placebo, seems ridiculous to me.

When doing placebo controlled trials, the reason to control for placebo is to control for all of these things. That includes any real effect and all possible confounding factors. In that case we don't really care whether what we're controlling for is real (because insomuch as the placebo controls for it, it exists for the actual treatment as well as the sham treatment), we just care about detecting the signal of the actual treatment.

But if you want to ask the question "is there a real physiological mechanism underlying the placebo effect?" then you have to accept the possibility that the answer is yes. If you define the term such that any such mechanism can't be placebo, then what are you even asking?

 

[arthwollipot]

I wonder why it took him so long ...

Well, because he has a job, and a family, and always talks about things on the podcast before cleaning them up for the magazine, and that's on a two-weekly schedule.

In order to avoid playing around with definitions and words:
The study appears to have discovered a pain-relieving effect in mice based on their previous experiences with a behavior that relieved pain, i.e. moving to a place that didn't induce pain. And it turns out that the mice seem to continue to feel a relief from pain when moving to the same place even when that place objectively exposes them to the same level of heat.

And if they'd reported it as such without referring to the placebo effect at all, there would have been no problem. But no, this is THE placebo effect according to the reporting.

But if you want to ask the question "is there a real physiological mechanism underlying the placebo effect?" then you have to accept the possibility that the answer is yes. If you define the term such that any such mechanism can't be placebo, then what are you even asking?

The point is there is no one placebo effect. The placebo effect is a blanket term covering a range of uncontrolled variables. The referenced article controls and isolates one variable. That doesn't mean that "the placebo effect" is explained.

 

[Roboramma]

The point is there is no one placebo effect. The placebo effect is a blanket term covering a range of uncontrolled variables. The referenced article controls and isolates one variable. That doesn't mean that "the placebo effect" is explained.

I certainly don't think that everything that falls under the blanket term "placebo effect" is an actual physiological mechanism. I do think that there's a real question as to whether or not such mechanisms exist. And if they do, they are part of the placebo effect.

I've already said I agree with you about the issue of hyperbolic articles saying that the placebo effect has been explained. It's clear that in many cases there probably isn't any such physiological mechanism.

That doesn't mean that in cases when such a mechanism exists, such as in the mouse study, it's not a placebo effect.

Aren't you interested in the question of whether or not such mechanisms exist?

 

[arthwollipot]

I certainly don't think that everything that falls under the blanket term "placebo effect" is an actual physiological mechanism. I do think that there's a real question as to whether or not such mechanisms exist. And if they do, they are part of the placebo effect.

I've already said I agree with you about the issue of hyperbolic articles saying that the placebo effect has been explained. It's clear that in many cases there probably isn't any such physiological mechanism.

That doesn't mean that in cases when such a mechanism exists, such as in the mouse study, it's not a placebo effect.

Aren't you interested in the question of whether or not such mechanisms exist?

I'm a bit confused as to why when something is known and controlled, it somehow is still part of a placebo effect. The placebo effect is defined as all of the responses that are not controlled for.

Yes. They have documented a method by which classical conditioning can cause the release of endorphins in the absence of a pain stimulus. That means that this particular mechanism is no longer part of the placebo effect, by definition.

It also, more importantly, goes towards dispelling the idea that such effects are magical, somehow "mind over matter" effects, as Kapchuk and others claim.

 

[Roboramma]

I'm a bit confused as to why when something is known and controlled, it somehow is still part of a placebo effect. The placebo effect is defined as all of the responses that are not controlled for.

Yes. They have documented a method by which classical conditioning can cause the release of endorphins in the absence of a pain stimulus. That means that this particular mechanism is no longer part of the placebo effect, by definition.

So when you say that the placebo effect is a sham, you are explicitly not saying that the idea that the expectation that, for instance, a pain medication will relieve pain, will itself cause pain relief (perhaps by the release of endorphins), even if the patient is given a sugar pill, is a sham?

You're saying, such effects, if they exist, just aren't the placebo effect.

Let's say we find, through clever controlled experiments, that a mechanism exists in humans similar to the mechanism that we discovered in mice. The next time someone studies a pain medication, based on these experiments, they expect that mechanism to come into play. However, they don't have the time or expertise to tease out this effect from other potential confounders. Instead, they just give half of their experimental subjects a sugar pill. Part of the thing they are controlling for with this placebo will be the actual effect of the release of endorphins associated with the expectation of pain relief in the placebo group. If they didn't give placebo, and just had a separate control group that didn't get any treatment, this effect would exist in those given the medication but not the control group and would confound their results, leading them to report a higher efficacy of their medication than is actually justified. It seems to me that if this effect exists, we should be controlling for it when studying new medicines, and placebo controlled studies do control for it. So to not call it placebo is completely unjustified.

The fact that we can potentially track down the different aspects of placebo doesn't make them not placebo.

 

[arthwollipot]

So when you say that the placebo effect is a sham, you are explicitly not saying that the idea that the expectation that, for instance, a pain medication will relieve pain, will itself cause pain relief (perhaps by the release of endorphins), even if the patient is given a sugar pill, is a sham?

You're saying, such effects, if they exist, just aren't the placebo effect.

I'm saying that such effects, if they exist, are an effect.

I don't quite know why but I'm having trouble parsing your next paragraph, so let's cut to the chase.

The fact that we can potentially track down the different aspects of placebo doesn't make them not placebo.

Why would they be? If they are known and can be controlled for, why are they still placebo? That doesn't make sense.

 

[Roboramma]

Why would they be? If they are known and can be controlled for, why are they still placebo? That doesn't make sense.

Because when we use placebo controls, they are one of the things that's being controlled for.

We also know that the effect "people experience social pressure to say the medicine makes them feel better" is a real effect that can be controlled for (we control for it using placebos), that doesn't mean it's not part of the placebo effect.

Every aspect of the placebo effect can at least potentially be known and controlled for. Why would that make them not placebo?

 

[Roboramma]

I'm saying that such effects, if they exist, are an effect.

But to be clear, when you say that the placebo effect is a sham, you aren't saying that such effects don't exist?