I think this essay is valuable in helping me clarify just what is meant by placebo effects, and their putative biological mechanisms and processes.
The arguments should not be summarily dismissed (a lot of thinking and research went into this essay). The arguments should be taken apart and critiqued individually.
In particular I agree with the key idea that it is important to critically evsluate the strength of evidence and the strength of experimental designs that favor long-lasting functional placebo effects.
If a design is a perfectly implemented triple-blinded study (https://www.scribbr.com/methodology/double-blind-study/) in which a reliable outcome measure is significantly affected with a reasonably large effect size—-for example, a Cohen’s d score above 0.2) then the evidence in favor would be pretty compelling (https://www.statology.org/interpret-cohens-d/). And very compelling us independently replicated in a second case/control study.
But how many studies reach this level? I doubt that there are any in animal model research with which I am more familiar.
If “faith and prayer” are regarded as a placebo effect then how much does this reduce the incidence and impact of illness when controlling well for socioeconomic factors? Was COVID severity modulated by said factors (or ivermectin ;-) ?
Interesting I had not heard about triple blinding. Is it much more difficult to design a triple blinded study over a double blind? My colloquial understanding is that designing a double blinded study is a significant hurdle over a single blinded study, but maybe the gap between a double and triple blind is not so high.
Keeping the analysts and statisticians in the dark should not be too hard. I can even imagine writing a study in which two or more conditions are just labeled A, B, and C. This would not be much fun though.
Might be true, but I don't think the case presented here is very compelling.
The general problem is it interprets the evidence to fit the conclusion: studies showing a placebo effect are interpreted as having bad methodology while ones showing no effect are interpreted as having correct methodology.
Of course, study design can exaggerate the effect. Does that mean it's doesn't exist, though?
Probably the more general problem is that contrarian views get clicks (and all that goes with that) which creates an unfortunate incentive.
I'm not sure what is supposed to be true and what we are supposed to all incorrectly believe.
We clearly double blind instead of single blind because we are all as concerned about researcher bias as the writer of the article. Their beliefs about our beliefs on where the errors are introduced in single blind studies seem at best presumptuous.
I have a friend who potentially contributed to the placebo effect for a study she was in.
She was in a study to see if a vitamin supplement would have a beneficial effect to a condition she suffered. She didn’t know if she was getting placebo or the supplement so after two weeks of not feeling any positive effect, she decided to start taking additional supplements of the vitamin in question on her own next to the study tablets to “be sure”. I remember choking on my coffee when she casually told me this later. I’m a scientist and it never occurred to me that anyone would act like this in a study.
But this would "just" reduce the power of the study and not increase bias, right? It's effectively turning it into an intention-to-treat design.
I know about one case of dieting old lady that adhered to perscribed meal plan and avoided sweets. But she was happily snacking on sausages between the meals.
> [...] after two weeks of not feeling any positive effect, she decided to start taking additional supplements of the vitamin in question on her own next to the study tablets to “be sure”.
If unrecorded, that would be fatal to the study's design and intent. I think you know what people are going to ask:
* Did she report this change to the research program?
* Was she dropped from the study?
* Did the study's authors consider writing/enforcing better, clearer subject instructions after this action was (hypothetically) brought to light?
* Does she understand how science works?every scientist who tries to do dietary interventions knows this happens all the time and it's basically impossible to prevent (which is also a big piece of why the 'science' in this area is not so great)
nothing will change as long as researchers get to keep publishing as they are
source: did some of this myself
Some trials are run in an isolated environment in which the participants are sequestered for the duration of the trial, including making sure subjects are chaperoned when going outside. For example, this study reported on by The NY Times [1] seems to try very hard to prevent cheating.
[1] https://www.nytimes.com/2024/07/30/well/eat/ultraprocessed-f...
We don't need to prevent it, it's enough to convince (almost all of) the test subjects to honestly admit if they have done something like this.
Is that largely impossible as well?
Why should she care how science works? She's just getting free vitamins (maybe).
People in these studies often don’t care about the science. They have selfish goals and are hoping they won’t get the placebo, because they don’t want to waste their time.
The article claims that the placebo effect has never been shown as a physical, objectively measurable effect, but this is wrong. For example, I look at autoimmune studies a lot. In those studies, when placebo-controlled, a fair number (often 20-30%) of patients on the placebo arm see significant improvement in symptoms across many objective, physical metrics. Measurement errors exist, of course, but I would expect physical measurements to not be prone to so much psychological bias, especially if double-blinded, that this would explain away the difference in results.
The article is also wrong about there not being a physiological mechanism. Dopamine and endorphins can be powerful mediators of pain relief, for example.
This could be placebo, but it could also be "self-healing" effects or just natural variation over time (which would cause a similar amount of patients with worse results) or a combination of this or something completely else.
Maybe you can rule out some of this, but I'm asking the question: Is the placebo effect the only plausible explanation for this numbers? I don't think so.
Did anyone manage to read the whole essay? I often notice with text like these, that they try to imitate scientific writing, but never quite manage (they always end in some kind of uncanny valley). The author states his premise quite nicely in the abstract, but the article does not really provide a consise supporting argument and instead just makes semi related points. This is another pattern I noticed, layman authors often seem to think that dumping lots of loosely related "evidence" somehow strengthens the argument.
I wonder if there are good resources which would help lay scientists to write good articles (maybe an opportunity for an LLM proofreader?)
I notice a lot of articles end abruptly when they just start to get going. Maybe people get tired of writing and publish whatever they have?
The "role playing" aspect is treated as fake and inconsequential here, but for a doctor, doesn't it translate to "this patient is all set and will not call later saying the problem got worse"?
In contrast, hearing from your doctor that "your labs are normal, you just have anxiety" is a common and horrible experience. It's likely the end of the doctor-patient relationship and if anything is being missed, maybe even a lawsuit.
I get that there are some interesting Objective Science issues here but for studies aimed at improving real world clinical practice, subjective outcomes matter.
Here’s an interesting study we could do.
Is there a placebo effect for car repair. And how does it compare to human “repair”?
You could pretend to fix peoples car issues and see if they report it improving.
Assuming the mind can’t fix cars mentally it would let you isolate what part of the placebo effect is perception.
Discuss..
There are devices that plug in to the OBD2 or cigarette lighter charging port & claim to boost power or MPG, but in reality only light up an LED. I think some people are fooled into thinking they work by wishful thinking, expected result bias, etc., so I'd expect we could find evidence for particular kinds of placebo effect among some people WRT cars
> Although placebo and “mind-cure” beliefs are widespread, the most parsimonious interpretation of the evidence is that the “placebo effect” is not a real healing effect, but a product of response bias and questionable research practices
This article is built upon a questionable definition of the placebo effect. Debatably, it’s a straw man argument that sidesteps the complexities of placebo response by using an oversimplification that is easier to dismiss.
The placebo effect varies by treatment and illness. Placebo isn’t fixing broken legs or making cancer disappear. Placebo effects are strongest where measures involve perception and are related to more complex mental processes.
Pain is a common example. Pain is a sensation, but sharp pain is associated with additional fear and panic response. Placebo addresses some of the fear and panic response, thereby addressing part of the pain. I agree that some of the studies showing placebo performing as well as fentanyl and morphine are likely flawed. We have a lot of studies showing dose-response curves for pain management using OTC analgesics, so it’s laughable to see a study showing a dose-response curve that is basically flat or inverted (placebo == 0mg dose).
The placebo effect gets much more complicated in the context of mental health. Anti-depressant studies are the classic example, with placebo often performing close to (but not as well) as powerful SSRIs. This has created internet backlash that “SSRIs are almost placebo” but the truth is that placebo groups improve dramatically during the study, too. There is something about the hope of receiving treatment, the way patients are told that they’re possibly receiving a treatment, the act of self-measuring and reporting mental health scores, and the interactions with clinicians that causes many (though not all) people to turn a corner and start improving. It’s almost as if they’ve received a signal that the end of their suffering is near so they start reactivating themselves, thereby improving the depression. The SSRI groups usually perform better by a statistically significant amount, but the challenge is that the placebo group improves so much that there is little numeric resolution left in the depression inventory to see much signal!
It’s a strange phenomenon, but I don’t think this article is a good overview of the effect let alone a rebuttal to it.
I wonder how much of placebo is things like lessening stress effects of the illness which decreases inflammation?
Thanks, I did find this article to to bend the definition of to fit the point.
There’s this whole genre of scientific questions it seems we’re just going to go back and forth about and mostly your opinion is based on your priors/personal beliefs. Kinda interesting but doesn’t seem like there will ever be a real definitive answer on the placebo effect. I wonder if there’s some kind of proof you could do on certain questions that would categorize them in this way.
Did anyone actually manage to read the whole essay? I noticed that often texts like these try to imitate
They could do with defining what "the placebo effect" they are refuting actually is.
I thought placebo proponents were usually advocating it for things like pain relief, tinnitus, fatigue, etc. Things where what you are consciously thinking can make a difference to your experience even if it doesn't actually fix the underlying issue.
I think that's quite a bit different from thinking that placebos actually heal wounds.
As I understand it, the placebo effect is seen in controlled trials that use a placebo, where both the drug under study, and the placebo, produce effects. To argue that a drug is effective, it has to be more effective than the placebo.
A completely separate issue is whether placebos should be used clinically.
It's hard to tell if the author is using either of those definitions.
This is pretty controversial, but the placebo effect does appear to heal physical problems sometimes too.
https://movementdisorders.onlinelibrary.wiley.com/doi/10.100...
Isn't stress known to reduce body's healing processes. Stress, lack of sleep and proper rest. If you believe something is taking care of you, you will have reduced stress, sleep and rest better. Wouldn't it be logical to expect faster healing. Also human bodies can heal naturally as well.
Stress increases cortisol and cortisol for a long time wrecks the immune system. There are a lot of other things that effect cortisol levels, but it's at least on very plausible pathway for some getting sick more easily.
Some placebo proponents would advocate for things like pain, nausea etc.
However, others advocate it for cancer, and run studies where people imagine they don’t have cancer and then measure their tumours to see if they shrink.
It's placebo effect all the way down
> The current scientific consensus is that the placebo effect is a real healing effect operating through belief and suggestion.
Wait what? I’ve never heard this being the scientific consensus. It’s more like the lay person’s view. I think that many scientists agree that the placebo effect is primarily a combination of regression to the mean and participant’s desire to please the experimenter (ie the literal definition of “placebo”).
There is scientific consensus that placebos have a subjective effect on symptoms, argued to be some unknown internal/mental effect. This is very widely held in the medical sciences community, and has more broad acceptance than most things held as a so-called consensus. It is not and has never been a lay person thing.
e.g. An article like https://www.health.harvard.edu/newsletter_article/the-power-... is not kooky or controversial, or the creation of a lay person. It's the accepted understanding of what's happening -- at least what is seemingly happening -- as viewed by medical experts.
As an aside, while placebo does come from Latin for "to please", you have it backwards. Placebos are to please either the patient or the participant, giving them a feeling that they're being treated/studied.
For some psychological conditions there is also an effect from merely participating in a study. For example, if you have depression then the human interaction and feeling that someone cares about you has some therapeutic effect.
What about a "reverse" effect: https://en.wikipedia.org/wiki/Voodoo_death
The "reverse" effect would be a Nocebo. Which also exists. https://en.wikipedia.org/wiki/Nocebo
Far from being convinced. It sounds like superstition.
Isn't it more like a superstition to assume the placebo effect is real (and not just a measurement artifact or regression to the mean)? It is hard to make a scientific experiment that could show the placebo effect to be real:
> t’s impossible to achieve a blind when comparing placebo to no treatment, since the placebo itself is the main method for blinding in the first place.
"Reality is that which, when you stop believing in it, doesn't go away", as Philip K. Dick said.
I don't find it unprobable that a psychological bias of a patient can change the result of an medical observation, since medical problems aren't usually 100% only of physical nature either — when your kid is ill in the morning and they can avoid school they will suddenly feel very unwell, when it is the day you wanted to go to Disneyworld they will suddenly feel very well. This ofc could just be a transparent lie, but often it is not and it truly changes how people feel about their own bodies. Similar that example with the hurt kid being kissed, now the post gives some rational explaination why the pain is soothed, but that doesn't make the effect go away: the psychological presentation of a thing makes an actual difference. A sandwich in a barren break room on a Monday tastes worse than that exact sandwich on a mountain summit after four hours of hiking on a Saturday.
Psychological context won't unbreak a broken leg or anything, so there are clear limits, but I am not convinced that the placebo effect is something that only works when researchers believe in the placebo effect themselves. As a Film student I don't find it unlikely that the acting performance of active anti-placebo critics could be unconvincing, but that just means they didn't do a good study then. If you want to study the effect where people are convinced of a thing (placebo) and you half-ass the convincing-part, then you are not researching the effect. But as someone who worked with actors I can assure everybody that you don't necessarily need to believe in a thing to create a convincing performance.
Now the actual question is how far that psychological effect goes when it comes to actual measurable numbers.
Psychological context won't unbreak a broken leg or anything, so there are clear limits
---
I am going to argue against that, proper rest and sleep helps your leg heal faster and you get better sleep when yoi are less stressed. You are less stressed when you believe something is taking care of you.
[delayed]
I may be off, but are you serious about asking that question ?
https://pubmed.ncbi.nlm.nih.gov/1422109/#:~:text=About%2050%....
That's not what I hinted to, but to this:
>The picture that emerges is that a placebo pill has almost no effect when administered by researchers who do not care about the placebo effect, but the exact same pill has an enormous effect larger than all existing treatments when administered by a researcher who really wants the placebo effect to be real.
No
The case against cases.
Even the most compelling and logically consistent case stands upon "mere" observation, convention and authority.
Couldn't you have replied to this post instead of modding it into invisibility?
-- quote --
sieabahlpark 2 minutes ago [dead] | parent | next [–]
Don't you know, the appeal to authority is the new "I'm smart enough to be credentialed, don't question me".
College really has made a whole group of people think they're elite when all they did was acquire debt and some watered down education.
-- unquote --
Not that it’s a strictly uninteresting topic but it 1) was presented in a way that will cause many people to be defensive and 2) has an unclear association with the topic of these comments. Both of those are common reasons for downvotes on a comment here.
Some of the people here really like to censor.
> The picture that emerges is that a placebo pill has almost no effect when administered by researchers who do not care about the placebo effect, but the exact same pill has an enormous effect larger than all existing treatments when administered by a researcher who really wants the placebo effect to be real.
Stopped reading right here because it is nonsense. Antidepressant researchers would love nothing more than the 30-50% placebo response rate to go away and yet there it is, competing with whatever tiny effect their pill does.
Isn't most research also double-blind? Meaning that the "treatments when administered by a researcher who really wants the placebo effect" are also probably badly designed which might be the reason for this effect?
The patient needs to have no doubts that have received medicine and the placebo effect will work 30-50% according to the previous comment, which is really close to what I thought.
If the patient doubts the effect won't take at all.
So that matters but not all for the reasons OP thinks.
It matters bc you can heal yourself, ppl literally do it everyday and it is measured by science as the placebo effect but if you don't think you can, you can't.
It has been widely reported that placebos can work even if you know it's a placebo:
https://www.health.harvard.edu/blog/placebo-can-work-even-kn...
Although there is an obvious reason why the placebo effect would be stronger for psychological issues.
Also the entire point the post makes is that these studies are flawed. Where is this person getting the idea that antidepressant companies are competing with placebos?
To bring a drug to market it has to be demonstrated as being more effective than a placebo
Most doctors don't prescribe placebos, and may not even be allowed to. Even though a placebo might very well work.
Placebos are ideal drugs in some cases. Effective, no side-effects...
not necessarily. It can be against another existing drug.
That were already proven to be stronger than placebo?
Usually we test against best current treatment and not just always against placebo. If something clearly improves outcomes for a patient, it's unethical to not give them a treatment so that you can test your new intervention against a placebo.
Right, like this person clearly doesn't understand the placebo effect and wrote a whole article about it
They summarize their positions clearly in the introduction which is basically on objective observation the placebo effect doesn’t emerge in the physiology but in the self reported effects. While they seem to hold psychology as not relevant because it’s not objectively observable, the less hard ass way to read this is the placebo effect doesn’t produce objectively measurable effects on physiology but it does appear to have a psychological effect.
They are positing the psychological effects are confirmation bias and conforming to expectations, but what they don’t address is whether the psychological outcome is real. Depression is greatly impacted by our psychology and whether the effect is due to some social conformity, beliefs, or other non physiological basis, is irrelevant to the efficacy.
Ultimately the authors argument is based on the notion that psychology is not real science, because science deals with things you can directly observe, and anything that isn’t real science is false. This is obviously fallacious. Even assuming psychology isn’t a real science, and psychology researchers have done themselves no favors here, the second statement is where the fallacy lies. Well done science is generally not false. But not everything that is true has been explained by science. Something can be true without us knowing the facts of how it works. Psychology requires self reporting by its very nature and as such is subject to all sorts of bias and difficult to control aspects - some impossible to control - and it’s likely some aspects of psychology are cultural and contextual and change with time. This doesn’t make it not real and not true. It just makes it difficult.
I think there's a placebo effect that is related strictly to the human interaction effect - you go to see a doctor, the doctor pays attention to you, you feel like people care about what happens to you, and so your mood improves. I have no idea how to quantify this effect (send pills in the mail with no human interaction?) but it appears to be significant in some cases of depression.
Just randomly pump antidepressants into peoples water supplies and check the suicide rate
Note that placebos aren't just limited to areas like pain and depression, but were also used in clincal trials that involved infectious disease treatments, including antibiotics and vaccines. In these cases, the issue is not the effect of the mind/brain/nervous system so much as the body's immune and biochemical processing systems (liver/kidneys).
Thus, in some kinds of infectious disease, the body is entirely capable of clearing the infection on its own in most cases as long as the subject isn't suffering some kind of immunodeficiency (e.g. all the strange diseases that HIV patients succumbed to before antiviral therapy was introduced). Studies like this led to the movement to reduce antibiotic treatment for minor infections that are very unlikely to lead to issues like sepsis etc, given the deleterious side effects on the human microbiome and so on. A placebo study might also reveal significant problems with toxicity, such as liver-kidney damage to the subject caused by a problem with the antibiotic under study.
Likewise with vaccines, careful study with placebos could demonstrate that the vaccine under study didn't prevent transmissibility of the infectious virus any better than the placebo did. Notably this was a problem with all the Sars-Cov2 vaccines, none seem to have prevented transmision from an infected person to a vaccinated person. Why wasn't this seen in the clincal trials, one wonders?
As far as depression/mental health studies, placebos are often readily detectable by the patient, which is a big problem in today's research into the 5-HT2a receptor drug research (eg psilocybin, LSD, DMT, mescaline). They've tried to get around this by giving subjects drugs like methyphenidate and methamphetamine as the 'placebo' (because the subject clearly can tell when they've been given a powerful psychoactive substance), but this is not a usual placebo, it's an alternative treatment.
> Likewise with vaccines, careful study with placebos could demonstrate that the vaccine under study didn't prevent transmissibility of the infectious virus any better than the placebo did. Notably this was a problem with all the Sars-Cov2 vaccines, none seem to have prevented transmision from an infected person to a vaccinated person. Why wasn't this seen in the clincal trials, one wonders?
The obvious answer is because vaccines are not designed to prevent viruses from ever becoming transmissible through the air according to some impossible standard devised by vaccine sceptics, they were designed to ensure that the body on average killed the virus off more quickly with less strain to their health[1]. So they tested the thing the vaccines were intended to do (against placebo), and not the thing they were not intended to do.
[1]a corollary of this is that people who are infected with viruses for shorter periods are less likely to transmit it, which is in fact what happened....
Well, what you want to see reported in a clinical vaccine trial is the breakthrough percentage, that is, the percentage of vaccinated subjects who then became infected. In the case of the smallpox vaccine, this percentage was extremely low, which is why 'vaccination ring' strategies in India were effective in wiping out the disease, so that it now only exists in a few freezers globally (though the genetic sequence is widely available and thus it could be synthesized from scratch using modern molecular biology techniques).
Since this wasn't reported to the public, but many official government sources claimed it was a sterilizing vaccine (0% breakthrough), this resulted in great public distrust of the health authorities. It's a lesson for the future: be honest with people.
>Likewise with vaccines, careful study with placebos could demonstrate that the vaccine under study didn't prevent transmissibility of the infectious virus any better than the placebo did. Notably this was a problem with all the Sars-Cov2 vaccines, none seem to have prevented transmision from an infected person to a vaccinated person. Why wasn't this seen in the clincal trials, one wonders?
I think this is from a misunderstanding of what transmissibility people were talking about. People getting the vaccine did not transmit the disease to others simply by virtue of being vaccinated. What that means is, you can get the shot and not worry that, by getting the shot, you are going to transmit the disease to others.
What that does not mean is that, once vaccinated, you cannot ever transmit the disease to others. If you get sick, the vaccine will likely make you get over it faster and have it less impactful, reducing transmission as an added bonus of those two things happening (you're coughing less so less is put in the air, you are sick for less time so the period you will be heavily contagious for and also symptomatic is shorter) but it will not entirely prevent transmission.
> Notably this was a problem with all the Sars-Cov2 vaccines, none seem to have prevented transmision from an infected person to a vaccinated person. Why wasn't this seen in the clincal trials, one wonders?
It was seen, then ignored by the media so the average person didn't realize what was going on. Remember that 95% effective? It wasn't 100% because vaccinated people not only got infected, they actually got sick (they were only testing once symptoms showed, not everyone in the study). The study it came from also only lasted 3 months, if it had gone on the planned 6 months it would have almost certainly gone much lower. I think they did end up having a follow-up at the 6-month mark (don't remember the results), but they got the EUA on the 3-month results.
A difficulty with the COVID vaccine is that it received emergency approval, which I think is great. But it means traditional clinical study now has to be conducted in an environment where the population, the behavioral environment (isolation, masks, etc) -- and possibly the virus itself -- are drastically changed by widespread use of the vaccine.