Acupuncture & Science

[gmalivuk]

Immediately after the treatment, the real acupuncture group had lower NPY levels than the sham acupuncture group. After ten days, they were both the same, and both were lower than the no-treatment group.

Interesting, though eSOANEM did point out some mechanisms that could explain that without acupuncture points making the difference.

[ahammel]

Immediately after the treatment, the real acupuncture group had lower NPY levels than the sham acupuncture group. After ten days, they were both the same, and both were lower than the no-treatment group.

Interesting, though eSOANEM did point out some mechanisms that could explain that without acupuncture points making the difference.

They only used a single acupuncture point in the treatment group (if I'm reading that right), so it's really a difference between "stabbed in the hindleg" and "stabbed near the tail".

[qetzal]

There's another very big possible source of bias that I didn't see explicitly mentioned here so far. The researcher is an acupuncturist, who presumably believes acupuncture works, right? How do we know she didn't handle the rats differently in ways that affected their NPY levels? Not to say she purposely skewed anything; it could easily be entirely subconscious.

Remember the Benveniste affair? They really thought they were seeing significant effects from applying homeopathically diluted materials to cultured cells. No chance that a cell in a dish could exhibit a placebo effect, right? But in fact, the researchers knew which groups of cells were treated or not, and that was subconsciously biasing their measurements. When the researchers were blinded to how the cells were treated, the apparently significant effect disappeared.

That may not be the case for this study, but it's something to keep in mind.

I'd also echo the previous comments about using electrical stimulation. Calling that acupuncture is a bit bogus, IMO. It's really just a form of electrical stimulation using needle electrodes. Perhaps it had real effects in this study, but that say virtually nothing about efficacy of standard acupuncture. Yet proponents seem all too willing to conflate the two procedures in an effort to prove that 'acupuncture works.'

I have to say, drunken, that you seem a bit guilty of that here. Not that you've claimed acupuncture really does work. But by bringing up the electro study in rats, you seem to be implying that if it's real, it suggests acupuncture in general may be real. I submit that that's wrong.

[gmalivuk]

There's another very big possible source of bias that I didn't see explicitly mentioned here so far. The researcher is an acupuncturist, who presumably believes acupuncture works, right? How do we know she didn't handle the rats differently in ways that affected their NPY levels? Not to say she purposely skewed anything; it could easily be entirely subconscious.

Yeah, hence why double-blinding is so important.

[BlackSails]

Whether that is enough to have an effect on the nervous/circulatory/lymphatic system to change your health I can't say, similarly it seems odd that (eg) a point on your foot relates to your spleen.

When I ruptured my spleen it was actually my shoulder that was in a ton of pain. The doctors said that was because there's a nerve that runs up the side of your body from the spleen to the shoulder.

This isnt quite true. Pain from the phrenic nerve is referred to the shoulder. This is because the phrenic nerve is formed by the spinal nerve roots C3, C4 and C5, which also innervate the shoulder via the superclavicular nerve.

[ahammel]

There's another very big possible source of bias that I didn't see explicitly mentioned here so far. The researcher is an acupuncturist, who presumably believes acupuncture works, right? How do we know she didn't handle the rats differently in ways that affected their NPY levels? Not to say she purposely skewed anything; it could easily be entirely subconscious.

The paper does not mention a double-blinding technique.

[qetzal]

The paper does not mention a double-blinding technique.

Admittedly, that would be hard to do. You have to handle the rats while you treat them, and since the controls were needled at a different point, you can't really be blinded to treatment.

But that just means it's extremely difficult to ensure that the apparent differences in NPY levels are due to the actual treatment, as opposed to some other uncontrolled difference between the groups. Which makes any conclusions suspect.

[ahammel]

Admittedly, that would be hard to do. You have to handle the rats while you treat them, and since the controls were needled at a different point, you can't really be blinded to treatment.

You would have to train one student (or group of students) to perform acupuncture at real TCM acupuncture point, and another at the sham acupuncture site without telling them which was which.

Not trivial, but not impossible.

[drunken]

I will refrain from quoting the posts I am replying to, I hope it is clear.

Firstly thanks to ahammel for the details of the sham treatment results, that removes a lot of the speculation from the argument. I assume you have paywall access?

In my own words (I tried looking it up to see if I was wrong first but the subject appears rather complicated) the placebo effect is the change in physiological symptoms caused by a psychosomatic mechanism. So the purely psychological effect of the treatment as opposed to any chemical or physical effects. This means that if I have a certain bacterial infection and the doctor gives me a sugar pill, and the infection goes away faster than statistically normal, this could be attributed to the placebo effect. If on the other hand it is later found that this particular strain of bacterium is actually killed by sugar (unrealistic I know, but it is just an example), this would mean that it was in fact not the placebo effect, and that sugar is not the correct placebo for testing treatments for this form of bacteria. I think the endorphin example falls into this category, although it is admittedly a borderline case. If it can be shown that the endorphins from pain cause a reduction in stress in rats, this would be a clinical effect which, psychological in the sense that the mechanism is based in brain chemistry, is not a true placebo effect as it works through a specific known chemical pathway and is not dependent on the psychological state of the test subject. To say that this is the placebo effect would mean that drugs such as anti-depressants work via the placebo effect. This may indeed be the case, and my understanding may be flawed. Please correct me if I am wrong. Is the placebo effect simply a name for anything that happens within the bounds of neurology and the brain?

With regard to scientifically minded people, the phrase 'believe nothing without evidence' is key. Whilst this might be a an idealistic goal of some idealised scientific mind, it is in reality neither true for any individual, nor psychologically possible. There are two reasons for this. Firstly all systems of logic and evidence evaluation are based on axioms, these axioms are accepted without evidence and therefore to believe nothing without evidence would mean not believing any axiom based system and therefore believing nothing at all. The second issue is, as I mentioned already, the fundamental subjectiveness of the human mind. For example I believe that I like milk chocolate better than I like dark chocolate. I have however not seen any objective evidence of this whatsoever, and I do not see any way of obtaining any such evidence. My life, and those of all other humans, are largely made possible by such assumptions based on purely subjective evidence and to do away with all such assumptions would render one incapable of functioning.

I have similar issues with Occam's razor as I do with with 'extraordinary claims'. While I am mentioning these I will also add 'the burden of proof is on one side but not the other'. These are all tools of argumentation which neither facilitate scientific discovery nor people's understanding of it except in cases where they are common sense anyway. Yes if I claim that I have a pet dragon this is clearly extraordinary, the burden of proof is largely on me, and the simpler explanation that I am lying is probably correct. But this kind of argument has no place in reasoned scientific debate between educated and intelligent individuals. Anyone that has to be reminded that it is unlikely I have a dragon and they should not simply believe my claim, should really be moved to another forum, perhaps one discussing where to get a highschool education. Occam's razor specifically, while a good rule of thumb to use while evaluating potential explanations for something, is based on abstract efficiency of thought, and not on reality. The universe does not seem to be simple, the explanation 'god created everything' is simpler than modern cosmology, and the idea of nihilism is the simplest of them all. It is true Occam's razor does not favour these due to their lack of explanatory power, but the fact remains that simplicity is merely more convenient and workable, not more correct. The argument 'my explanation is simpler than yours and therefore more correct' is fallacious. Similarly, unless someone is making obviously outrageous claims, the claim 'x is not true' has its own burden of proof and while it may not be equal to the burden on 'x is true', it is still there. If you hear someone claim x and say 'I do not accept the evidence for this' that is fine, but as soon as you claim 'not x' to be the case, you are making a claim of your own and have assumed your own burden of proof. This may seem like common sense but I see this happen everywhere, including on this forum. It may also seem like a trivial distinction but it matters a lot, at least to me.

It also sometimes goes even further. I have been in a situation where I was interested in learning more about a particular set of experiments. My claim was 'I want to know more about this'. I was basically shut down by people saying 'the results of this experiment are impossible, therefore the methodology was flawed, and as the person claiming otherwise it is solely your burden to prove me wrong about this fact.'. Not only was I not making any such claim, but the claim that the results are impossible and the methodology was flawed was totally unsupported by anything other than the phrases 'burden of proof' 'Occam's razor' and 'extraordinary claims'. I hope you can see how this may have frustrated me to the point where I lost all respect for those disagreeing with me, and left the discussion in disgust. This is the case that prompted me to change to the sig I am currently using on this forum. With that disclaimer I can claim 'acupuncture is the only true medicine' and that is not a statement of pseudoscience, as I have established that it is my own opinion and not a scientific claim. Without such a disclaimer I believe it is fair to assume that any statement presented as fact in a forum dedicated to scientific discussion is a scientific statement, and if not supported by empirical facts, a statement of pseudo-scientific conclusions. This includes statements such as 'this is caused by the placebo effect', and 'your hypothesis is wrong'. If the evidence strongly supports the statement it is fine, and if not, then it is pseudoscience, not merely an unsupported claim. I expect a higher standard of debate in a forum with the word 'science' in the name, otherwise the name should be changed to 'opinion'.

[ahammel]

I think the endorphin example falls into this category, although it is admittedly a borderline case. If it can be shown that the endorphins from pain cause a reduction in stress in rats, this would be a clinical effect which, psychological in the sense that the mechanism is based in brain chemistry, is not a true placebo effect as it works through a specific known chemical pathway and is not dependent on the psychological state of the test subject.

I wouldn't get caught up in whether it should be called a placebo or not. The conclusion would be that sticking rats with needles reduces stress via such-and-such a mechanism. Call it what you like.

Irrelevant fillibuster ignored.

[gmalivuk]

the placebo effect is the change in physiological symptoms caused by a psychosomatic mechanism.

No, that's far too narrow. Why worry about the specific cause at all?

Better to think of placebo as *every* effect of a therapeutic interaction apart from the specific thing being tested. So it includes psychosomatic mechanisms as well as any other purely physiological mechanisms that aren't responding to the treatment under investigation. Any response to being prodded with an electrically charged needle that doesn't result from placing the needle at a specific acupuncture point, to pick one example.
---
The rest of your anti-science rant doesn't really warrant a response, as ahammel said.

[BlackSails]

the placebo effect is the change in physiological symptoms caused by a psychosomatic mechanism.

No, that's far too narrow. Why worry about the specific cause at all?

Better to think of placebo as *every* effect of a therapeutic interaction apart from the specific thing being tested.

Thats not right either. If I am testing drug X to see if it reduces blood pressure by blockade of a1-receptors, and it turns out to also block serotonin reuptake, that doesnt mean its a placebo antidepressant - it just has two actions.

[eSOANEM]

With regard to scientifically minded people, the phrase 'believe nothing without evidence' is key. Whilst this might be a an idealistic goal of some idealised scientific mind, it is in reality neither true for any individual, nor psychologically possible. There are two reasons for this. Firstly all systems of logic and evidence evaluation are based on axioms, these axioms are accepted without evidence and therefore to believe nothing without evidence would mean not believing any axiom based system and therefore believing nothing at all. The second issue is, as I mentioned already, the fundamental subjectiveness of the human mind. For example I believe that I like milk chocolate better than I like dark chocolate. I have however not seen any objective evidence of this whatsoever, and I do not see any way of obtaining any such evidence. My life, and those of all other humans, are largely made possible by such assumptions based on purely subjective evidence and to do away with all such assumptions would render one incapable of functioning.

Yeah, good luck finding many fields of science which have been properly axiomatised. Most fields are a collection of statements which seem to be true (having been experimentally verified). It's only once these statements have been found that people try to find axioms which can produce those statements. The axioms are not believed without evidence but because they explain the evidence or are themselves directly falsifiable (such as the conservation of momentum in physics for example).

As for you liking milk chocolate better than dark chocolate, you do have evidence for that. Your evidence may not be especially significant statistically (because, for obvious reasons, getting significant results would be tricky) but you have the evidence that, when you eat milk chocolate you experience nicer sensations than when you eat dark chocolate. Not all evidence needs to be quantitative (although if you want other people's money, it'd pretty douchey for it not to be).

I have similar issues with Occam's razor as I do with with 'extraordinary claims'. While I am mentioning these I will also add 'the burden of proof is on one side but not the other'. These are all tools of argumentation which neither facilitate scientific discovery nor people's understanding of it except in cases where they are common sense anyway. Yes if I claim that I have a pet dragon this is clearly extraordinary, the burden of proof is largely on me, and the simpler explanation that I am lying is probably correct. But this kind of argument has no place in reasoned scientific debate between educated and intelligent individuals. Anyone that has to be reminded that it is unlikely I have a dragon and they should not simply believe my claim, should really be moved to another forum, perhaps one discussing where to get a highschool education.

Occam's razor specifically, while a good rule of thumb to use while evaluating potential explanations for something, is based on abstract efficiency of thought, and not on reality. The universe does not seem to be simple, the explanation 'god created everything' is simpler than modern cosmology, and the idea of nihilism is the simplest of them all. It is true Occam's razor does not favour these due to their lack of explanatory power, but the fact remains that simplicity is merely more convenient and workable, not more correct. The argument 'my explanation is simpler than yours and therefore more correct' is fallacious. Similarly, unless someone is making obviously outrageous claims, the claim 'x is not true' has its own burden of proof and while it may not be equal to the burden on 'x is true', it is still there. If you hear someone claim x and say 'I do not accept the evidence for this' that is fine, but as soon as you claim 'not x' to be the case, you are making a claim of your own and have assumed your own burden of proof. This may seem like common sense but I see this happen everywhere, including on this forum. It may also seem like a trivial distinction but it matters a lot, at least to me.

[paragraph break mine]

If you reject any idea of the burden of proof then I'm afraid you fail scientific logic.

Batting the burden of proof about is pointless yes and this is why the principle is that, because science is fundamentally sceptical, the burden of proof always lies with the positive claimant (the one claiming something is the case).

Furthermore, both Occam's razor is a razor not a law. It does not say which theory is correct, it does not even say which is more likely to be correct. In fact, all it says is which explanation is to be preferred. Furthermore, in all formal contexts it requires that the predictions of the explanations (as far as reasonably verifiable (this is clearly context dependent)) be the same and therefore that the two explanations are (to your reasonable degree of accuracy) both as correct.

If you claim to have a pet dragon, this is a not a situation where Occam's razor should be applied because the predictions are different between the two theories (and the two can be easily distinguished by you showing me your dragon). It is however, a case for the burden of proof. You claim to have a pet dragon, that makes you the positive claimant and so it is on you to demonstrate this to me. Until then, the scientific thing to do is assume that you are either a) lying or b) crazy.

Furthermore, you cite "god created everything" vs. modern cosmology as a situation where Occam's razor seems to give the wrong answer. This is only the case when you allow the predictions to be different. If however, you require that god also created the CMBR, galactic redshift etc. you end up having to add so many additional causes as to why god decided to add galactic redshift not blueshift, why hubble's constant takes the value it does etc. that you have at least as many as in standard comsology plus the additional one that there is a god.

As for where the burden of proof lies in you X vs. not X argument, it depends on whether X is a positive or negative statement. If the claim is "A exists" or "B is impossible" (because B being impossible requires a restriction against B exists) then the claim is positive and the burden of proof does lie on the person claiming "X"; if the claim is "A does not exist" or "B is possible" then the burden of proof lies on the person claiming "not X".

Now you're probably thinking "but if someone claims cats don't exist, surely the burden of proof lies on them". This is true, but this is because you have evidence that cats do exist and, as such, they are not presenting a claim that "cats don't exist" but rather that "your evidence that cats exist is incorrect" which is a positive statement.

Furthermore, you say that it is ok to say "I do not accept the evidence for this" but not to say "X is false", whilst this is true in colloquial language, in a scientific context "X is false" can mean "there is no/insufficient evidence of X". If you asked a physicist in the mid-early 1800s whether length depended on reference frame, they'd say no but now that SR is well-demonstrated, a physicist would say yes.

Science is fundamentally unlike maths because the truthiness of a statement is never certain and so can change in the light of new evidence.

It also sometimes goes even further. I have been in a situation where I was interested in learning more about a particular set of experiments. My claim was 'I want to know more about this'. I was basically shut down by people saying 'the results of this experiment are impossible, therefore the methodology was flawed, and as the person claiming otherwise it is solely your burden to prove me wrong about this fact.'. Not only was I not making any such claim, but the claim that the results are impossible and the methodology was flawed was totally unsupported by anything other than the phrases 'burden of proof' 'Occam's razor' and 'extraordinary claims'. I hope you can see how this may have frustrated me to the point where I lost all respect for those disagreeing with me, and left the discussion in disgust. This is the case that prompted me to change to the sig I am currently using on this forum. With that disclaimer I can claim 'acupuncture is the only true medicine' and that is not a statement of pseudoscience, as I have established that it is my own opinion and not a scientific claim. Without such a disclaimer I believe it is fair to assume that any statement presented as fact in a forum dedicated to scientific discussion is a scientific statement, and if not supported by empirical facts, a statement of pseudo-scientific conclusions. This includes statements such as 'this is caused by the placebo effect', and 'your hypothesis is wrong'. If the evidence strongly supports the statement it is fine, and if not, then it is pseudoscience, not merely an unsupported claim. I expect a higher standard of debate in a forum with the word 'science' in the name, otherwise the name should be changed to 'opinion'.

I would be interested to see this experiment but, in general, a single experiment with unusual results (particularly if not published in a proper peer-reviewed journal) is insufficient to overturn a large body of other experiments. Science will always have to make judgements based on the weight of evidence and, if the results of one experiment seem to be impossible given a large body of previous data, it is very probably a flaw in the odd experiment/experimental noise which is causing the impossible results.

As a general rule, to overturn a body of evidence you need more and better evidence. I'm sorry to go back to physics here, but as it's my main area of interest in science, most of my examples must come from there by necessity; the recent faster-than-light neutrinos result was not sufficient to overturn relativity because it was a smaller and less good body of evidence than that supporting relativity. As such, the scientific community universally thought it more likely that something was wrong with the experiment (although many wished it wasn't because how cool would that be) and, as it happens, they were right to do so; one of the cables wasn't connected properly.

Again, science is not like maths, a single counter-example cannot overturn a whole string of verifications.

[gmalivuk]

Better to think of placebo as *every* effect of a therapeutic interaction apart from the specific thing being tested.

Thats not right either. If I am testing drug X to see if it reduces blood pressure by blockade of a1-receptors, and it turns out to also block serotonin reuptake, that doesnt mean its a placebo antidepressant - it just has two actions.

Sorry, by "thing being tested" I meant the specific intervention being tested, not the specific hypothesis. Probably could have worded that better.

a single counter-example cannot overturn a whole string of verifications.

Especially because, in science unlike math, it's never quite so easy to see that an observation is a real counterexample in the first place, or if it is, what exactly it's a counterexample *to*. Neptune moving oddly indeed meant our previous explanation wasn't entirely correct, but the fix involved adding one more planet to the mix, rather than throwing out Newtonian gravity. So it was only a counterexample to the "theory" that there weren't any other significant bodies out in that area of space.

[ahammel]

I would be interested to see this experiment but, in general, a single experiment with unusual results (particularly if not published in a proper peer-reviewed journal) is insufficient to overturn a large body of other experiments.

Just to belabour this point a bit: the commonly accepted false-positive rate for publication is 5%. In other words: scientists are totally ok with being wrong one time in twenty. To return to the accupuncture example: one experiment which shows that it's effective and 19 others that show that it's not is exactly what we would expect if acupuncture does nothing. Because of this.

[qetzal]

Right. But most of the 19 studies where acupuncture failed wouldn't get published (the file drawer effect), whereas the one positive study has a good chance to be published because it seems novel and exciting. So when you look at the literature, you won't see anything like a 19:1 ratio, even if in fact positive acupuncture studies are all just statistical flukes.

More reason not to be persuaded by one seemingly positive rat study.