Placebo Effect - No proof that it exists

Puppycow said:
I think we need a strict definition of the term "placebo effect" before we can determine whether it exists or not.
From the Wikipedia article in post #2:

So if it is defined as "a therapeutic effect, causing the patient's condition to improve," rather than simply "an effect causing the patient to report an improvement," (or some similar lower hurdle) perhaps it does not exist? I think there's little doubt that there is "an effect causing the patient to report an improvement," but I don't know if there is "a therapeutic effect, (actually) causing the patient's condition to (actually) improve." For one thing, we know that correlation does not necessarily mean causation.

Does everyone agree that the correct definition of "placebo effect" is "a therapeutic effect, (actually) causing the patient's condition to (actually) improve"?
Click to expand...

I agree that it is the effect of interest. Other than pain, people aren't really interested in whether they can be induced to alter their subjective feelings about an illness while the actual disease process marches on unabated.

Linda
 
sol invictus said:
Again: what's the difference? How can you distinguish between an "actual improvement" in pain and mere "motivation...to behave differently"?

The only thing I can think of is some kind of brain scan. Since the patients were behaving differently, however, that might reveal a difference for the same reason this data did. It's going to take something extraordinarily sophisticated to do produce the evidence you seem to be demanding, and in fact it's not clear to me that the distinction you're drawing exists even in principle.
Click to expand...

I think the distinction may be a lot less clear for something like pain management, but even there I think there are certainly ethical implications. Yes, I agree that it would be extremely difficult, if not completely unethical, to do the tests to determine whether there is a real effect at all, or under which circumstances it might arise. But I'm not so demanding, really. I have no problem at all with just dismissing it as an unknown and assuming, for practical purposes, that it doesn't exist. After all, this course does not in any way negate the value of placebo controlled trials, it just makes life more difficult for quacks.
 
Prometheus said:
I think the distinction may be a lot less clear for something like pain management, but even there I think there are certainly ethical implications. Yes, I agree that it would be extremely difficult, if not completely unethical, to do the tests to determine whether there is a real effect at all, or under which circumstances it might arise.
Click to expand...

What would those tests be?

I have no problem at all with just dismissing it as an unknown and assuming, for practical purposes, that it doesn't exist. After all, this course does not in any way negate the value of placebo controlled trials, it just makes life more difficult for quacks.
Click to expand...

Then you "have no problem" dismissing an effect which has been clearly demonstrated to be a statistically significant effect in multiple experiments, merely because you don't like it. There's a word for that: woo.
 
If you look at the study I referenced earlier, you can see that differences in reporting cannot account for the results, there must also be an actual improvement in the pain experienced. Other studies show that requests for additional analgesia correlate with ratings of pain if you confine yourself to the range of moderate pain. Analgesics are rarely requested for changes in pain ratings if the pain is mild and are usually requested regardless of change if the pain is severe. But with moderate pain, improvements (of 10mm or more on the VAS) are associated with fewer analgesia requests while worsening pain of a similar change are associated with more analgesia requests.

Linda
 
My question was what do others think about this, as I've always believed that placebo was a scientifically proven phenomena.
Click to expand...

One wonders how they did a scientific test to check that -- one group received a placebo, the other, control group, received a placebo...
 
Skeptic said:
One wonders how they did a scientific test to check that -- one group received a placebo, the other, control group, received a placebo...
Click to expand...

The other group received no treatment (i.e. no expectation).

Linda
 
sol invictus said:
What would those tests be?
Click to expand...

I don't know.

sol invictus said:
Then you "have no problem" dismissing an effect which has been clearly demonstrated to be a statistically significant effect in multiple experiments, merely because you don't like it. There's a word for that: woo.
Click to expand...

I think that you misunderstand my position. I'm not dismissing anything merely because I don't like it. As best I'm able to analyze my own feelings, I harbor neither like nor dislike for the Placebo Effect. And I'm not suggesting that there is no statistically significant effect. I'm quibbling over just what that known effect really is.

Popular belief about PE seems to be that humans actually have some sort of "mind over matter" ability that may sometimes be triggered by beliefs and expectations. To me this seems like an extraordinary claim, far less plausible than the idea that PE is a conflation of a variety confounding factors that cannot realistically always be controlled for in all research. I'm provisionally dismissing this notion of PE until I see convincing evidence in its favor.

I'm reminded of part of Feynman's famous "Cargo Cult Science" lecture in which he describes some research done on rats running mazes, and just what sorts of herculean efforts it really takes to actually control for all the possible confounding variables in such studies.

I'm also not arguing one side of a hard dichotomy. It seems entirely possible to me, for instance, that there is a real, causative physiological placebo effect with some conditions, such as pain, but that confounding factors may--in whole or in part--actually explain reported effects with other conditions.

Of course, I'm also open to the possibility that I'm in error. As I mentioned before, I'm not well versed in the relevant literature--but the studies I have looked at have not been convincing. To characterize my stance as "woo" is uncharitable, at best.
 
Prometheus said:
Popular belief about PE seems to be that humans actually have some sort of "mind over matter" ability that may sometimes be triggered by beliefs and expectations.
Click to expand...

Pain is in the mind. So are many symptoms of illness. What's the matter? :)

To me this seems like an extraordinary claim, far less plausible than the idea that PE is a conflation of a variety confounding factors that cannot realistically always be controlled for in all research.
Click to expand...

Really? You find it extraordinary that sensations, which are in your mind, can be affected by your state of mind?

I don't know much about the medical side of this, but I do know that nearly everything people believe and feel emotionally, and many aspects of their responses to stimuli that have been studied by psychologists, are very strongly influenced by expectations.

I think the claim that feeling well or feeling pain is not affected by expectation is the extraordinary one.

As I mentioned before, I'm not well versed in the relevant literature--but the studies I have looked at have not been convincing.
Click to expand...

What's wrong with the one I linked to? It was very carefully controlled.

To characterize my stance as "woo" is uncharitable, at best.
Click to expand...

You haven't presented any argument as to why you don't believe that later study, or why you don't accept the results of the meta-study posted earlier (which are in accord with it).
 
sol invictus said:
Pain is in the mind. So are many symptoms of illness. What's the matter? :)


Really? You find it extraordinary that sensations, which are in your mind, can be affected by your state of mind?
Click to expand...

If we're limiting the discussion only to pain and other subjective sensations, sure. I'm addressing the popular belief that treatment by placebo sometimes results in healing of less subjective conditions. For example, I remember once seeing a study on the effects of minoxidil, in which some small fraction of the control group experienced dense hair regrowth despite being treated only with placebo. I did a cursory search and was unable to find the original study, so I'm sorry I can't link it. Something like this, though, seems to me like a prime candidate for the sort of thing I'm talking about.

sol invictus said:
I don't know much about the medical side of this, but I do know that nearly everything people believe and feel emotionally, and many aspects of their responses to stimuli that have been studied by psychologists, are very strongly influenced by expectations.

I think the claim that feeling well or feeling pain is not affected by expectation is the extraordinary one.
Click to expand...

Maybe I'm just splitting hairs, but to me there's an important ethical difference between altering the physiological basis for feelings, and altering people's subjective reports of those experiences. For instance, suppose it turned out that telling subjects they're receiving a strong painkiller really just guilt-trips some of them in to not requesting more drugs. They still feel all the pain, but, for whatever reason, don't think they should ask for more drugs, so they are just resigned to it. Isn't this possibility ethically distinct from a situation where, say, some people's actual experience really is improved by placebo? Suppose we were talking about a drug with seriously risky, known, potential side-effects. If there's even a small chance that a mere placebo could really help some people and allow them to avoid taking on the risks of the real drug then isn't it more ethical to try a sugar pill first?

sol invictus said:
What's wrong with the one I linked to? It was very carefully controlled.



You haven't presented any argument as to why you don't believe that later study, or why you don't accept the results of the meta-study posted earlier (which are in accord with it).
Click to expand...

I'm in general agreement with uncayimmy's post #25, above (which I'd missed on my first read through the thread).
 
I have to apologize because I have been dragging my feet on a placebo effect presentation.

Basically, sol invictus and fls have identified the problem: there's some misunderstanding about what is being referred to by the phrase 'placebo effect,' and this is understsandable because the term has different colloquial and clinical uses, depending on context.

As a consequence, many conversations about whether the placebo effect is 'real' get conflated in miscommunication and participants end up talking past each other a bit.

Beecher was referring to the overall improvement in the placebo group, which we know could have many origins, not the least of which is the natural progress of the illness since the trial selection. ie: baseline recovery regardless of whether the patient is being treated. I believe he concluded that this could be around 30% of positive response. Some people call this the placebo response.

Since then, we've come up with explanations for a good portion of this positive response that are present even when the patient is untreated.

The slice of this postitive response that remains unexplained is also sometimes called "the placebo response". It's not 30%. It's transient (meaning it shows up in one study, but doesn't get replicated). It's not present in binary outcomes ("are you better yes or no?") It's very rarely present in independently verifiable outcomes, and appears limited to things that have to be reported by or dependent on the opinion of the patient.

Interpreting the original post's use of the word "proof" as really meaning "evidence," I do think that in light of the body of research there is evidence of a placebo effect from suggestions made by the practitioner. I call this a placebo response.

The more recent line of investigation has been to determine if it's simply a reporting bias (eg: is the patient convincing themselves that no change in pain is actually a change in pain - poor evaluation of perception and memory is not a controversial skeptical assumption), or if there could be a genuine improvement reported accurately.

We are all familiar with an unrelated but relevant study that showed that swearing reduced the perception of pain. This is just a natural consequence of what we already know: our reporting of pain has a component that depends on how much we are paying attention to the signal.

Nevertheless, I feel that this line of research has produced a body of literature that suggests one type of placebo response to practitioner suggestion may have a legitemate and relevant underlying mechanism: placebo analgesia.

On the other hand, I believe that the specific types of experiment that generate placebo analgesia through conditioning are a dead end.

What is in debate are more vague questions:

  • does the evidence suggest that the effect could be exploited to obtain meaningful clinical outcomes (ie: can placebo analgesia be used to treat pain?)
  • does the evidence suggest that the effect could be exploited to obtain meaningful clinical outcomes for complaints that the body of literature currently shows are unaffected? (ie: can placebo response be used to treat cancer?)

Dr. Hall did produce an article in the recent Skeptic magazine, but I think she has made an error in the way she interprets the body of literature on the subject. Specifically, she dismisses a comprehensive literature review with negative findings because there are a handful of unreplicated studies with positive findings. (When homeopaths do this, we consider it bad science.) Dr. Hall is quite bright and I'm not a doctor, so I'm sure I'm missing something. I was hoping she could explain her reasoning at some point.
 
Prometheus said:
Maybe I'm just splitting hairs, but to me there's an important ethical difference between altering the physiological basis for feelings, and altering people's subjective reports of those experiences. For instance, suppose it turned out that telling subjects they're receiving a strong painkiller really just guilt-trips some of them in to not requesting more drugs. They still feel all the pain, but, for whatever reason, don't think they should ask for more drugs, so they are just resigned to it. Isn't this possibility ethically distinct from a situation where, say, some people's actual experience really is improved by placebo? Suppose we were talking about a drug with seriously risky, known, potential side-effects. If there's even a small chance that a mere placebo could really help some people and allow them to avoid taking on the risks of the real drug then isn't it more ethical to try a sugar pill first?
Click to expand...

This is a very important distinction, and the question is quite valid.

It does rely on evidence of efficacy, and also a comparison with other approaches. eg: if swearing also reduces pain, then is it ethical to charge them for a sugar pill misrepresented as an analgesic?
 
You must log in or register to reply here.

Back
Top Bottom