Acupuncture appears to show results.

[Sherman Bay]

I can't tell what you are trying to test with this proposed design. If all you want to test is acupuncture, then all you need are two groups - an acupuncture group and a sham (as in non-acupoint needle insertion) acupuncture group. Previous studies have shown no difference between these groups.

Some studies have shown improvements in both groups that have received treatment (true acu) and groups that have not received treatment (false acu), if the non-treatment group thinks they are being treated. That effect is most likely because of the placebo effect.

The placebo effect is the measurable, observable, or felt improvement in health or behavior not attributable to a medication or treatment that has been administered.

If a non-treatment group shows improvement, what's going on here? That's why you need to control for not only no treatment (the group doesn't know it is being used) and no treatment but with expectation of treatment (the placebo effect).

The whole point is to isolate the effect of the procedure to what is being tested only. If the possible cause of a favorable effect is multiple factors, you haven't learned much.

I stick needles in, the subject gets better. It was either the needles or the placebo effect. Not a valid study for needles or acupuncture.

Well then, tell me specifically what is wrong with using just the following two groups:
1) True Acupuncture.
2) Sham Acupuncture.
The groups are identical except that, in the Sham Acupuncture group, the wrong acupuncture points are used.

If you believe any other groups should be used, please explain why?

I thought I had, and perhaps the above part of this post will expand on what I already said.

With your 1)2) setup, you did not mention if either groups were told what to expect. It is impossible to stick needles in someone without their knowledge, which makes tests of this sort harder to control for than drug tests. It's hard to make this double-blind, too.

In your example, what would you say if both groups showed an improvement? Would you conclude that putting needles in random points was a good thing? How do you know -- you have no controls for NON-needled subjects? Or could the cause be the various effects talked about here: Experimenter, Hawthorne, Placebo? You simply don't know. You have too many things going at once.

Please, please, read Bausell's book. Although I was well aware of the placebo effect before, I think he explains and expounds on the topic better than anyone. He also includes a list of specific CAM studies, both pro and con, with a brief analysis of each.

 

[fls]

Some studies have shown improvements in both groups that have received treatment (true acu) and groups that have not received treatment (false acu), if the non-treatment group thinks they are being treated. That effect is most likely because of the placebo effect.

Yes. And it supports the idea that acupuncture has no specific effect.

If a non-treatment group shows improvement, what's going on here?

Expectation.

That's why you need to control for not only no treatment (the group doesn't know it is being used) and no treatment but with expectation of treatment (the placebo effect).

The whole point is to isolate the effect of the procedure to what is being tested only. If the possible cause of a favorable effect is multiple factors, you haven't learned much.

You have yet to make it clear to me the purpose of your test. It seems obvious to me that people who have needles stuck in them will have different expectations from those who do not.

I stick needles in, the subject gets better. It was either the needles or the placebo effect. Not a valid study for needles or acupuncture.I thought I had, and perhaps the above part of this post will expand on what I already said.

I don't see how your set-up is any better. It still simply shows that different expectations lead to different outcomes. But we knew that already.

Are you trying to show that sticking needles in to people has an effect that is different from mere expectation? Then you don't need an acupuncture group at all. What you need is two groups that think they have had needles stuck in them, but in one group the needles don't penetrate the skin.

Linda

 

[Sherman Bay]

fls, I don't see how I can make it any clearer than I have. Maybe someone else would like to try.

 

[JJM]

{snip} Are you trying to show that sticking needles in to people has an effect that is different from mere expectation? Then you don't need an acupuncture group at all. What you need is two groups that think they have had needles stuck in them, but in one group the needles don't penetrate the skin.

Linda

Sorry I don't have the citation handy; but sham needles have been developed. The "needle" sits in a handle into which it retracts, rather than penetrates the skin, when pushed against the skin. Subjects don't notice the difference.

The problem remains that the person administering sham acupuncture still knows it, and can give the subject subtle clues (a.k.a. "sensory leakage").

 

[fls]

fls, I don't see how I can make it any clearer than I have.

The simplest thing would be for you to state what hypothesis you are testing with your study design (I have asked twice now ).

Maybe someone else would like to try.

You suspect this site is populated by mind-readers?

Linda

 

[fls]

Sorry I don't have the citation handy; but sham needles have been developed. The "needle" sits in a handle into which it retracts, rather than penetrates the skin, when pushed against the skin. Subjects don't notice the difference.

The problem remains that the person administering sham acupuncture still knows it, and can give the subject subtle clues (a.k.a. "sensory leakage").

I think you are referring to the Streitberger needle.

http://www.ncbi.nlm.nih.gov/pubmed/10534595

Linda

 

[JJM]

{snip} You suspect this site is populated by mind-readers?

Linda

I think you are referring to the Streitberger needle.

http://www.ncbi.nlm.nih.gov/pubmed/10534595

Linda

Thanks, you read my mind ...

 

[Sherman Bay]

The simplest thing would be for you to state what hypothesis you are testing with your study design (I have asked twice now ).

And I have answered at least thrice. Once more, from the top...

The hypothesis is that traditional acupuncture produces beneficial results in humans.

The placebo effect does exist. People do get better, or at least report feeling better, when there is no actual treatment or substance administered.

The placebo effect requires the subject to think that they are being treated. This means that a control group, which has no expectation of treatment, is not the same as a placebo control group, which does.

Whether the mind actually can facilitate a cure, or the patient merely feels better psychologically, the jury is still out on this one.

If you design a study to test the effectiveness of a particular treatment, you must be sure that any improvement you observe is not due to the placebo effect. Otherwise, what have you learned?

So the minimum number of groups for most tests would be 3.

1. Treated, with expectation of treatment
2. Not treated, but with expectation of treatment (placebo group) and
3. Not treated, with no expectation of treatment (control group)

Of course, subjects should be assigned to each group randomly and neither subject nor experimenter should know which group each is assigned to until the test is done.

To translate the above general principle to an acupuncture one,

1. Treated with traditional acupuncture points and told they are being treated,
2. Treated with sham acupuncture points but told they are being treated,
3. Not treated and not told they are being treated

You suspect this site is populated by mind-readers?

No, but it is populated by many smart people, some of which may be better able to explain this concept than I have been. It's not a new or unique idea.

Here's an example. Effectiveness of Leech Therapy in Osteoarthritis of the Knee

This study fails on several levels -- inadequate size per group for one. It's not surprising that the results were positive. "After seven days, patients receiving leech therapy suffered less pain than those receiving the traditional therapy (topical diclofenac)."

What important impediment to making causal inferences does this trial not control for? Answer: the placebo effect. Bausell says:

People volunteer for trials that they believe might be able to help them. Few if any people volunteered for this trial because they had a burning desire to see if topical diclofenac would help them...Thus while the diclofenac might result in a very small pacebo effect, the leeches would engender far and away the larger placebo response for this strange little sample of people.

So how should the test have been done? Bausell has an answer:

If the investigators had consulted me, I would have suggested that they spend several years developing placebo leeches to adequately blind their patients. Or, if they were in a hurry, I might have suggested they try common garden-variety slugs attached with a drop of superglue. (Which perhaps explains why I am not in more demand as a CAM consultant.)

Penn & Teller did exactly that in one BS eposide, using live garden snails as a facial treatment. People said it worked.

 

[fls]

And I have answered at least thrice. Once more, from the top...

The hypothesis is that traditional acupuncture produces beneficial results in humans.

The placebo effect does exist. People do get better, or at least report feeling better, when there is no actual treatment or substance administered.

The placebo effect requires the subject to think that they are being treated. This means that a control group, which has no expectation of treatment, is not the same as a placebo control group, which does.

Whether the mind actually can facilitate a cure, or the patient merely feels better psychologically, the jury is still out on this one.

If you design a study to test the effectiveness of a particular treatment, you must be sure that any improvement you observe is not due to the placebo effect. Otherwise, what have you learned?

Exactly.

If acupuncture performs better than placebo, it also performs better than no treatment (the information is redundant). If acupuncture does not perform better than placebo then we don't care whether it performs better than no treatment (the information is irrelevant).

I think you are saying that you do care whether it performs better than no treatment. And, of course, if one were trying to gather more information on the placebo effect, that may be useful information to have. However, it does not provide any additional information on acupuncture. And acupuncture is what Billy Joe and you claim to be interested in.

No, but it is populated by many smart people, some of which may be better able to explain this concept than I have been. It's not a new or unique idea.

But the problem is not with my understanding of the concept of placebo, it is with trying to understand why you are insisting on adding a control group (no treatment) that does not give you any additional information about the benefit of acupuncture (only about the benefit of placebo), but reduces the quality of the trial by making the groups smaller.

Here's an example. Effectiveness of Leech Therapy in Osteoarthritis of the Knee

This study fails on several levels -- inadequate size per group for one. It's not surprising that the results were positive. "After seven days, patients receiving leech therapy suffered less pain than those receiving the traditional therapy (topical diclofenac)."

What important impediment to making causal inferences does this trial not control for? Answer: the placebo effect. Bausell says:So how should the test have been done? Bausell has an answerenn & Teller did exactly that in one BS eposide, using live garden snails as a facial treatment. People said it worked.

Yes. Which is the same thing that Billy Joe suggested - compare leeches (acupuncture) with snails (sham acupuncture).

Linda

 

[trvlr2]

It was not a personal attack on Rolfe but on her ideas. My impression was that she attacked me claiming that I'm unfamiliar with the qualifications of members of this Forum.

When I set up a Saanen goat-farm after making the business plan I asked the expertise of an experienced Vet and then a second opinion of a second Vet. I just asked questions because they were the experts and I acted accordingly to their advice because their knoledge in this field is much better than mine and I am (was) almost a quack in this field. I didn't say on a general term that Rolfe is uneducated but only in this particular field. Rolfe is certainly much more educated in Vet Med than I am and I'm am aware of it. I will never pretend to know more than Rolfe in Vet Med. On the other hand, I'm an MD a a Pharmacology specialist, working and researching in the field of Phytotherapy for 40 years, and then someone unqualified (ignorant) in this field pretends to know more than I doand sentences that Phytotherapy doesn't work. This is unacceptable for me.

I repeat my earlier statement: I think Prof Facco isby far much better qualified in his field of competence than anyone in this Forum (me included). If Rolfe or someone else can prove to me that she/he more qualified than Prof Facco than I'm willing to pay for his summer holidays on the Adriatic coast (where I live). By the way, why did't she write the names of Ivor, The Atheist, Dr. Imago, Robinson etc.?

I agree that personal attacks are in bad taste, so stop making them.

I can't say which company because I have still a working relationship and a contractual agreement not to use the company's name outside the imposed limits.

It is true that the Italian political situation was bad but with yesterdays elections things will, probably, change (for better or for worse). The third richest man in Italy (Berluscony) is the winner and communist finally disappeared from our Parlamient. I wonder if this happend in the US (with Bill Gates). Would it be a better or a worse Country?

I live and work in Italy which is not the ideal place but I prefer this than to live in a country where nurses are allowed to prescribe, where anyone has the right to open a practice and practise almost anything (and I'll stop here because I have a great respect for that Monarchy).

Does anyone other than myself, find this guy howlingly funny!??





Bet I could guess what he may be doing with the goats...

 

[Sherman Bay]

Yes. Which is the same thing that Billy Joe suggested - compare leeches (acupuncture) with snails (sham acupuncture).

Then you are missing the third group: the one with no treatment and no expectation of it. Bausell didn't suggest replacing one group, but adding a third.

With the first two, if both show similar, favorable results, what have you learned? That doing something, when the expectation is there, produces positive results. But add a control group, if those results are negative, and it shows that the "doing something" is most likely the force at work, not the actual treatment itself, whether valid or sham.

I urge you to read some of Robert Carroll's essays (long):

Evaluating Acupuncture Studies and Carroll's book review of Snake Oil Science

 

[BillyJoe]

I thought I had, and perhaps the above part of this post will expand on what I already said.

I think you are having trouble stating clearly exactly what groups you want and exactly why you want them. Perhaps, as Linda has already hinted, you need to start off by defining exactly what it is that you want to test.

With your 1)2) setup, you did not mention if either groups were told what to expect.

I thought it was clear from my original post, though not expressly stated, that they would be told that they will have their migraines treated with either True Acupuncture or Sham Acupuncture (it is probably unethical to tell them that they will all be treated with true acupuncture, though that would be the ideal). They would, of course, not be told which of the two they will be receiving.

It is impossible to stick needles in someone without their knowledge, which makes tests of this sort harder to control for than drug tests.

It's not a problem for this setup, because they all have needles stuck into them. Or, in other words, we are not pretending to stick needles in anyone. One group gets the true acupuncture points, the other group gets the false acupuncture points.

It's hard to make this double-blind, too.

It is easy to make it single-blind, but hard to make it double-blind, I agree, and I said so in my original post. My suggestion was that the false acupuncture points in the Sham Acupuncture group should actually be true acupuncture points for another condition. The acupuncturists would then be told they are treating that condition in this group, not migraine. For this protocol to be successful, however, there would need to be no communication between the the acupuncturinst and the patient, otherwise the acupuncturist would soon find out that the patient is really suffering from migraine and the double-blinding would be broken.

In your example, what would you say if both groups showed an improvement? Would you conclude that putting needles in random points was a good thing?

I would conclude that there is no difference between True Acupuncture and Sham Acupuncture in the treatment of migraine. Read that again. We were trying to show whether acupuncture is an effective treatment for migraine and we conclude that true acupuncture is no more effective than sham acupuncture.

How do you know -- you have no controls for NON-needled subjects? Or could the cause be the various effects talked about here: Experimenter, Hawthorne, Placebo? You simply don't know. You have too many things going at once.

I don't care what the various components of the placebo effects are. What I am trying to do is make "the placebo effect" the same in both groups and therefore any differences found in the trial will be between True Acupuncture and Sham Acupuncture. Therefore, if there is no difference, then True Acupuncture is no more effective than Sham Acupuncture.

 

[BillyJoe]

Then you are missing the third group: the one with no treatment and no expectation of it. Bausell didn't suggest replacing one group, but adding a third.

And you have yet to show us why that third group is necessary when what we want to know is whether acupuncture works.

With the first two, if both show similar, favorable results, what have you learned?

That true acupuncture is no more effective than sham acupuncture.

That doing something, when the expectation is there, produces positive results.

That as well, but don't we already know that.

But add a control group, if those results are negative, and it shows that the "doing something" is most likely the force at work, not the actual treatment itself, whether valid or sham.

Add that control and you have decreased the strength of your trial by decreasing the number in the true and sham acupuncture groups. These groups will actually show you what you set out to show. The control group is an unecessary add on that tells you nothing more than you already know and adds nothing to the question you are actually interested in.

 

[fls]

Then you are missing the third group: the one with no treatment and no expectation of it. Bausell didn't suggest replacing one group, but adding a third.

Bausell did not suggest anything other than providing an appropriate control group, as the topical diclofenac was not. In addition, Bausell's own acupuncture studies did not include a non-treatment group, but rather one or two sham acupuncture groups.

With the first two, if both show similar, favorable results, what have you learned? That doing something, when the expectation is there, produces positive results. But add a control group, if those results are negative, and it shows that the "doing something" is most likely the force at work, not the actual treatment itself, whether valid or sham.

We already know that doing something produces positive results. There really is no point in wasting the power of your study on providing proof for something as well-established as the placebo effect. Also, you specifically stated that you were interested in discovering whether acupuncture was effective, not in discovering whether placebo was effective.

I urge you to read some of Robert Carroll's essays (long):

Evaluating Acupuncture Studies and Carroll's book review of Snake Oil Science

Robert Carroll says the same thing that Billy Joe and I have been saying, and contradicts your claim.

From your link:

A properly controlled acupuncture study is one that is double-blind, randomized, and uses a control group. The only proper control group for an acupuncture study is a group that receives sham acupuncture.

A proper acupuncture study would randomly assign at least 50 patients to either a true or sham acupuncture group and not reveal to the patients which group they have been assigned to. Comparing an acupuncture group to a group that does not get acupuncture is an invalid study, unless one is trying to measure the different degrees of effectiveness of placebo treatments.

Linda

 

[BillyJoe]

Sherman Bay,

I have not yet had time to read your referenced book, but I did have time tonight to read your linked article:
Evaluating Acupuncture Studies
It's a very good article and I thank you for sharing it.

However, as Linda demonstrated above, that article does not support your position.

You have been repeatedly urging us to read certain books and articles so that we may properly understand the placebo effect, but I am wondering about your own understanding of the placebo effect even after you have read those books and articles. I will add a quote to the two provided by Linda above and ask you to answer a question about what you think it means, because I get the distinct impression that you are somehow misunderstanding what you are reading and this may be a good example of that misunderstanding:

Some researchers obviously do not understand the placebo effect. When they compare true and sham acupuncture groups to a third group getting some other kind of therapy and find that the two acupuncture groups show a significant positive effect over the third group, they don't realize such a result supports the position that acupuncture works by the placebo effect.

My question is:
Do you take this to mean that the third group is necessary in order to demonstrate that acupuncture is a placebo?

Sorry if this question sounds stupid, but I'm just trying to figure out where you are coming from with the need for that third group in a trial of acupuncture, especially when your linked article specifically says that it should not be used (the second quote in Linda's post above).

regards,
BillyJoe

 

[Sherman Bay]

My question is:
Do you take this to mean that the third group is necessary in order to demonstrate that acupuncture is a placebo?

Yes. Or may strongly suggest it. If, of three groups, the two that used needles showed improvement over one that did not, one conclusion would be that needles, not particular acupuncture points, was the operative factor.

Or, if one placebo group was told they were receiving acupuncture, but did not have actual needles inserted into their skin (the special needles mentioned above), where we could be sure that there was no treatment of any kind but there was an expectation of it, then the placebo effect is for sure the only factor.

Without a true control, if other groups show improvement, what was it? Was everyone feeling better that day? Did the stock market go up that week? Where's your baseline?

And you have yet to show us why that third group is necessary when what we want to know is whether acupuncture works.

"Works" is such an abused term. Do you mean "causes a measurable effect of some kind" or "cures an ailment"? If all groups show a significant change, does that mean it "works"?

 

[BillyJoe]

I've now also read two reviews of Bausell's book

Carroll's book review of Snake Oil Science

At the end of this review there is a link to Dr. Harriet Hall's review of the same book (you'll need to scroll down a bit to find it).
She quotes from the book as follows:

Bausell offers some simpler criteria that can rule out the worst offenders:

• Subjects are randomly assigned to a CAM therapy or a credible placebo
• At least 50 subjects per group
• Less than 25% dropout rate
• Publication in a high-quality, prestigious, peer-reviewed journal

In one of the studies on acupuncture by Bausell's own group....

...they compared “true” acupuncture to the most credible “sham” acupuncture they could devise.

 

[BillyJoe]

My question is:
Do you take this to mean that the third group is necessary in order to demonstrate that acupuncture is a placebo?

Yes. Or may strongly suggest it. If, of three groups, the two that used needles showed improvement over one that did not, one conclusion would be that needles, not particular acupuncture points, was the operative factor.

$#!+, let me rephrase that:
Do you take this to mean that the third group is necessary in order to demonstrate whether acupuncture is more than just a placebo?

All treatments have a placebo effect, so what we want to know is not whether acupuncture has a placebo effect (because we know that it does), but whether acupuncture has more than just a placebo effect? Is the third group necessary to determine this?

And you have yet to show us why that third group is necessary when what we want to know is whether acupuncture works.

"Works" is such an abused term. Do you mean "causes a measurable effect of some kind" or "cures an ailment"? If all groups show a significant change, does that mean it "works"?

I would much rather you just answer the question.

Again, I will rephrase it:
And you have yet to show us why that third group (the first two groups being the true acupuncture and sham acupuncture groups) is necessary when all we want to know is whether acupuncture is more than just a placebo.

 

[fls]

Yes. Or may strongly suggest it. If, of three groups, the two that used needles showed improvement over one that did not, one conclusion would be that needles, not particular acupuncture points, was the operative factor.

Or, if one placebo group was told they were receiving acupuncture, but did not have actual needles inserted into their skin (the special needles mentioned above), where we could be sure that there was no treatment of any kind but there was an expectation of it, then the placebo effect is for sure the only factor.

Yes, if you have additional hypotheses about the effectiveness of needling, then additional controls groups would provide additional information about the details. However, an untreated control group provides no additional information about the effect of needling, only about the placebo effect.

For example. Let us say that you performed a study with 4 groups. One group received acupuncture, one group received acupuncture but at non-acupoints (needling), one group received acupuncture with the Streitberger needles (placebo) and one group received no acupuncture. Group one showed 21% improvement, group 2 19%, group 3 15% and group 4 10%. You could then say that there was a 2% difference between acupuncture and needling, a 4% difference between needling and placebo control, and a 5% difference based on expectation between placebo and untreated controls. The untreated control group provides you with information about expectation, but it does not provide you with information about needling or acupuncture.

Without a true control, if other groups show improvement, what was it? Was everyone feeling better that day? Did the stock market go up that week? Where's your baseline?

Are you now saying that Bausell, Carroll and hundreds of thousands of controlled clinical trials are wrong?

Linda

 

[fls]

I've written several posts on this before, but I think part of the problem is that there isn't really a placebo effect, per se. The improvements that we see in a placebo group are due to a hodge podge of different things, the bulk of which is simply regression to the mean and bias. The only differences we find between groups who think they've been given an effective treatment and those that haven't, are in the strength of their subjective perceptions (http://content.nejm.org/cgi/content/short/344/21/1594). We could, of course, study the various bits and pieces of what it is that leads to the improvement seen in a placebo group. But when we are using a placebo group in order to study a particular treatment, all we are really interested in is whether there are any effects specific to that particular treatment. Both the active treatment group and the placebo group are subject to all those things that lead to improvement - the weather was nice, no disasters were on the news, the body's immune system stifled the germs that were causing the disease, they had a chance to sleep in on the weekend, etc. We don't really need to know the specifics. All we really want to know is whether a particular treatment adds something more to the mix - whether there is an additional effect specific from that treatment. We want to know that the treatment effect was 10%. We don't care that the placebo effect was 20% expectation, 30% regression, 40% other effective treatments, and 10% desire to please, because none of that is generalizable when it comes to making a recommendation for a patient. Only the 10% treatment effect is.

Linda

Note: For simplicity, I am excluding some specific effects that can be found from placebo on the basis of conditioning, as most of the time these effects are not present in the placebo groups in clinical trials.